Coronary and Systemic Hemodynamic Effects of Sustained Inhibition of Nitric Oxide Synthesis in Conscious Dogs
Evidence for Cross Talk Between Nitric Oxide and Cyclooxygenase in Coronary Vessels
Sustained inhibition of NO synthesis (Nω-nitro-l-arginine [L-NNA], 20 mg·kg−1·d−1, 7 days) was investigated at rest and during exercise in conscious dogs. At rest, L-NNA did not alter mean arterial blood pressure but markedly increased total peripheral resistance (+ 73±14%, P<.01). Exaggerated hypertension was observed during exercise (+132±5 mm Hg after L-NNA versus +113±5 mm Hg before L-NNA, P<.01). L-NNA decreased the resting coronary artery diameter by 6±1% and suppressed its exercise-induced dilation but had no effect on coronary blood flow and resistance. L-NNA decreased flow repayment volumes during reactive hyperemia, but corresponding flow debt volumes remained unchanged. The cyclooxygenase inhibitor diclofenac (10 mg/kg) had no effect on reactive hyperemia parameters before L-NNA but reduced flow repayment volumes, durations, and corresponding debt-to-repayment ratios in L-NNA–treated dogs (all P<.05). In vitro, indomethacin blunted the residual relaxation to bradykinin of large coronary arteries taken from L-NNA–treated, but not from control, dogs. Bradykinin-induced increase in 6-ketoprostaglandin F1α production was greater in coronary arteries taken from L-NNA–treated dogs (+179±41 pg/mm2) than from control dogs (+66±18 pg/mm2) (P<.05). These results indicate that (1) NO is of major importance in the control of systemic but not coronary resistance vessels at rest and during exercise, and (2) after L-NNA, the cyclooxygenase pathway is involved in myocardial reactive hyperemia and in the residual relaxation to bradykinin of isolated coronary arteries. Thus, in conscious dogs, the cyclooxygenase pathway might act as a protective mechanism of the coronary circulation when endothelial nitric oxide synthesis is altered.
Nitric oxide has been identified as one of the predominant EDRFs involved in the maintenance of a basal vasodilator tone1 and in the control of tissue oxygenation.2 Most of this knowledge issued from experiments conducted with l-arginine analogues substituted at the guanidino group such as L-NMMA or L-NNA, which inhibit NO synthase in a concentration-dependent and stereospecific manner both in vitro and in vivo.3 4 In rats and rabbits, acute and chronic administration of these inhibitors results in vivo in marked increases in arterial blood pressure and systemic peripheral resistance.1 5
At the level of coronary arteries, L-NMMA increases vascular resistance and attenuates the acetylcholine-induced vasodilation in the Langendorff-perfused rabbit heart preparation.6 However, controversial results have been obtained regarding the effects of NO synthase inhibitors in this vascular bed in dogs. On the one hand, Chu et al7 demonstrated in resting conscious dogs that systemic infusion of L-NMMA, in doses sufficient to increase blood pressure, caused a decrease in epicardial coronary artery diameter together with a small reduction in coronary blood flow, indicating a rise in coronary vascular resistance. On the other hand, Woodman and Dusting8 showed that in the anesthetized dog, inhibition of NO synthase with L-NNA caused a decrease in epicardial coronary artery diameter but did not affect coronary blood flow and resistance despite a significant rise in arterial blood pressure. Similar results were reported after systemic or intracoronary administration of L-NNA in resting conscious dogs.9 10 11 12 Duncker and Bache12 even concluded that NO did not play an obligatory role in the regulation of coronary vasomotor tone at rest and during exercise in conscious dogs, suggesting that NO synthase inhibition might have led to activation of other compensatory vasodilator mechanisms in the coronary resistance vessels. In all these studies, administration of NO synthase inhibitors resulted in a significant decrease of coronary vasodilation secondary to hypoxia or ischemia13 and in a reduction of both reactive hyperemic blood flow and duration.9 10 11
The first goal of the present study was thus to compare the effects of a sustained administration of L-NNA on systemic and coronary hemodynamic parameters in resting and exercising conscious dogs. Subchronic administration of L-NNA was chosen in order to assess whether the sustained hypertension observed in rats after repeated administration of this powerful inhibitor of NO synthesis also occurs in dogs. Moreover, since important cross talk between NO synthase and COX has recently been described in vascular endothelial cells,14 15 we hypothesized that endothelium-derived prostanoids produced by activation of local COX might play a different role in the control of vasomotor tone when investigated in the same animal under physiological conditions, ie, when NO synthase is normally activated, and during sustained inhibition of NO synthesis. Therefore, the second goal of the present study was to investigate the potential role of the COX metabolic pathway at rest and during myocardial reactive hyperemia, before and after subchronic administration of L-NNA. Finally, responses to the endothelium-dependent vasodilators bradykinin and acetylcholine were compared in isolated coronary and femoral arteries taken from L-NNA–treated and sham-operated control dogs, in the absence and presence of a COX inhibitor.
Materials and Methods
Studies were performed in 41 male mongrel dogs weighing 16 to 30 kg. Surgical procedures and all subsequent experiments were performed in accordance with the official regulations of the French Ministry of Agriculture (approval No. 94148).
Dogs were anesthetized with sodium pentobarbital (30 mg/kg IV), intubated, and ventilated with room air (Harvard Apparatus). A left thoracotomy was performed either in the fifth or the fourth intercostal space under sterile surgical conditions. All dogs were first instrumented with a catheter in the thoracic descending aorta and in the main pulmonary artery. In 6 dogs (group A), an electromagnetic flow probe (Skalar Instruments) was positioned around the ascending aorta. In 15 other dogs (groups B and C), the left circumflex coronary artery was dissected free, and a Doppler flow probe (10 MHz, Crystal Biotech) and a hydraulic occluder (Jones Instruments) were positioned around the artery. In 6 of these 15 dogs (group B), a pair of ultrasonic dimension transducers (5 MHz piezoelectric crystals, VD 5S, Triton Technology, Inc) attached to a Dacron backing were also sutured to opposing surfaces of the left circumflex coronary artery 2 to 4 cm from its origin.16 A solid state micromanometer (model P7A, Konigsberg Instruments) was introduced in all dogs into the left ventricle through the apical dimple and secured with purse-string sutures. Finally, 9 sham-operated dogs (group D) were not further instrumented and served as controls for ex vivo studies. The pericardium was loosely closed, and all catheters and electrical wires were tunneled subcutaneously to exit at the base of the neck. The chest was closed in successive layers, and the pneumothorax was evacuated. Gentamicin (40 mg) and cefazolin (1 g) were administered before and after the surgical procedure.
Studies were performed 2 to 3 weeks after surgery, when animals were accustomed to resting quietly on their right sides on the experiment table. Some of them were also trained to exercise on a motor-driven treadmill. Aortic blood pressure was measured with a P23 XL pressure gauge transducer (Spectramed Instruments, Inc) positioned at the midchest level, and mean aortic blood pressure was calculated as diastolic pressure +1/3 pulsed pressure. Left ventricular pressure was measured from the micromanometer, and LV dP/dt was obtained via electrical differentiation of the left ventricular signal. Cardiac output was measured from the aortic electromagnetic flow transducer using a square-wave electromagnetic flowmeter (Skalar Instruments). Zero blood flow was set in each experiment using the late diastolic blood flow as the zero reference. Total peripheral resistance (dyne·s·cm−5) was calculated as mean arterial pressure/cardiac output×80. The external diameter of the left circumflex coronary artery was measured instantaneously and continuously with an ultrasonic transit-time dimension system (System 6, Triton Technology, Inc). Coronary blood flow velocity was measured using a Doppler flowmeter (System 6, Triton Technology, Inc). The Doppler shift measured with the coronary flow velocity probe was converted to blood flow using the following equation: q=2.5×d2×f, where q is blood flow (mL/min), d is the internal diameter of the coronary artery, and f is the Doppler shift in kHz.17 External coronary diameter was taken as the internal diameter of the flow probe, which becomes adherent to the coronary artery during the healing process, and according to previous postmortem evaluation, wall thickness was considered to be 20% of the external diameter. Circumflex coronary vascular resistance was calculated as the ratio of mean arterial pressure to mean coronary blood flow.
In 6 dogs from groups B and C, the day-to-day variability in measurements of resting heart rate, mean arterial pressure, coronary blood flow, and epicardial coronary artery diameter during 8 consecutive days was as follows (mean±SEM value of the standard deviation): 5±1 bpm, 4±1 mm Hg, 6±1 mL/min, and 40±8 μm, respectively.
All parameters were measured from the strip-chart recordings (Gould ES 2000). During exercise, heart rate, aortic blood pressure, LV dP/dt max, cardiac output, coronary artery diameter, and coronary blood flow velocity were measured from the strip-chart recordings and averaged over the last 30 seconds of the resting period and each of the 3-minute exercise stages. When acetylcholine, bradykinin, and nitroglycerin were administered at increasing doses, data were analyzed under baseline conditions and (1) at the peak decrease in systolic aortic blood pressure for the calculation of changes in mean arterial pressure and (2) at the peak increase in coronary blood flow for the calculation of changes in coronary blood flow and resistance. For each reactive hyperemia, the area under the curve (representing the volume of the coronary blood flow deficit during coronary occlusion, ie, the flow debt) and the excess of coronary blood flow that followed the release of occlusion (ie, the flow-repayment) were calculated by using a data analysis software (HEM, Notocord Systems) after all parameters were digitized. The duration of reactive hyperemia was taken as the time from the release of occlusion to the point at which flow returned to within 5% of baseline. The peak increase in coronary blood flow and peak decrease in coronary resistance were also measured.
Sustained Inhibition of NO Synthesis
The NO synthase inhibitor L-NNA was dissolved daily in 100 mL of saline to which hydrochloric acid (0.01N) was added to obtain complete dissolution. The final pH was adjusted to 7.0 just before infusion by adding sodium bicarbonate. L-NNA was slowly infused over 30 minutes at the initial dose of 30 mg/kg on the first day and subsequently at the dose of 20 mg/kg on the 7 following days. Sham-operated dogs (group D, n=9) received 100 mL saline per day during the same period.
Treadmill exercises were performed in dogs instrumented for the continuous measurement of cardiac output (group A, n=6) or circumflex coronary artery diameter and blood flow (group B, n=6), before and 7 days after L-NNA administration. Briefly, after a resting period, with the dog standing quietly on the treadmill, three successive 3-minute exercise stages at 5, 10, and 12 km/h with 5% slope (Ex5, Ex10, and Ex12) were implemented. Dogs were trained to run every 2 days on the motor treadmill during the 1-week L-NNA administration. In addition, 4 dogs of group A underwent an additional treadmill exercise 10 minutes after the injection of methylatropine bromide (0.1 mg/kg) before and 4 days after L-NNA administration.
Effects of Acetylcholine, Bradykinin, and Nitroglycerin on the Systemic and Coronary Circulations
The effects of COX inhibition on reactivity of small coronary arteries to acetylcholine (0.1, 0.3, and 1 μg/kg) and bradykinin (0.1, 0.3, and 1 μg/kg) were studied before and 7 days after L-NNA administration in dogs instrumented solely with a coronary Doppler flow probe and a hydraulic occluder (group C, n=9). Reactivity of small coronary arteries to nitroglycerin (0.1, 1, and 10 μg/kg) was also assessed in the same conditions. Inhibition of COX was performed in vivo using diclofenac infused over 10 minutes at a dose18 of 10 mg/kg. This dose was shown to abolish the hypotensive effect of arachidonic acid (500 μg/kg IV bolus). A delay of 20 minutes after completion of diclofenac infusion was respected before conducting other experiments in order to allow a steady state in hemodynamic and coronary parameters to be reached. Experiments on the effects of diclofenac during sustained NO synthase inhibition were performed 5 to 7 days after the onset of L-NNA administration.
In dogs from group C, the effects of COX inhibition (diclofenac, 10 mg/kg) on reactive hyperemia were studied before and on the last day of the 1-week L-NNA administration. A week of simultaneous administration of diclofenac and L-NNA proved in a pilot study to be impossible, because within 3 to 4 days dogs (n=3) developed severe gut hemorrhage (arguing for mesenteric ischemia and marked hypertension); therefore, only the effects of acute diclofenac administration could be investigated in chronically L-NNA–treated dogs. Reactive hyperemic responses were induced by using the hydraulic occluder by abrupt coronary occlusions of 10, 20, and 30 seconds. Reactive hyperemia and reactivity to acetylcholine and bradykinin were studied on the same experimental day, with the sequence of coronary arterial occlusions and drug administration being randomly selected. At least 5 minutes was allowed between two interventions for the return of coronary blood flow to the steady state baseline value. Reactivity to nitroglycerin was studied on a separate day. A washout period of 48 hours was implemented after diclofenac administration.
In 3 dogs from group C, the effects of enalaprilic acid (1 mg/kg IV bolus) on systemic and coronary hemodynamics were assessed before and at the last day of L-NNA administration. At this dose of enalaprilic acid, inhibition of angiotensin I–converting enzyme was almost complete, as shown by the reduction (−90±5%) in the maximal systolic pressor response to angiotensin I (300 ng IV bolus).
In Vitro Relaxation of Coronary and Femoral Arteries
In vitro studies were performed on coronary (group A), femoral (groups A and B), and carotid (groups A and B) arteries taken from L-NNA–treated dogs and in the same arteries taken from sham-operated dogs (group D). No arteries were taken from dogs of group C because these dogs received diclofenac in vivo as part of the experimental protocol. Seven days after administration of L-NNA or saline, dogs were reanesthetized with sodium pentobarbital (30 mg/kg) and artificially ventilated. The left carotid artery was dissected free, excised, immediately frozen in liquid nitrogen, and stored at −80°C for subsequent measurement of cGMP content. Hearts from groups A and D dogs were rapidly excised, and proximal epicardial left circumflex coronary arteries were dissected free (up to 3 cm from the origin). A segment of each proximal femoral artery taken from dogs in groups B and D was dissected free and excised. Coronary and femoral arteries were cleaned of adherent connective tissue, cut into rings of ≈3 to 4 mm in length, and placed into a cold modified Krebs-Ringer bicarbonate solution (mmol/L): NaCl 118, KCl 4.7, CaCl2 2.5, MgSO4 1.2, KH2PO4 1.2, NaHCO3 25.0, edetate calcium disodium 0.026, and glucose 11.1 (control solution). Endothelium removal was performed by rubbing the intimal layer with a forceps in a number of rings. The rings were suspended between two stirrups in organ chambers filled with 20 mL control solution (37°C) gassed with 95% O2/5% CO2 for isometric tension recording as previously described.19 The rings were allowed to equilibrate for 45 minutes and were then contracted with the thromboxane A2 analogue U46619 (1 μmol/L) for coronary arteries19 and norepinephrine (0.3 μmol/L) for femoral arteries.20 A first relaxation to acetylcholine (1 μmol/L) was implemented to test the integrity of the endothelium. After another period of equilibration, each ring was allocated to one of the following pretreatments: control solution, L-NMMA (100 μmol/L), indomethacin (10 μmol/L), or L-NMMA plus indomethacin (same concentrations). Twenty minutes later, relaxations to increasing concentrations of acetylcholine (1 nmol/L to 10 μmol/L), bradykinin (0.1 nmol/L to 10 μmol/L), and the NO donor Sin-1 (1 nmol/L to 10 μmol/L) were performed in all rings after contraction with either U46619 (coronary artery) or norepinephrine (femoral artery) and in the presence of the above-mentioned pretreatment agent. Diclofenac (10 μmol/L) was used instead of indomethacin in a number of rings. In rings without endothelium, the same experimental setup was used in control solution only.
Contractions were expressed in absolute tension (g). Relaxations were measured at the peak decrease in tension for each concentration of the added agent using a PC software (IOSLab, EMKA Technologies). Results were expressed as percentage of the maximal contraction to the agonist and as EC50 values (concentrations of the relaxing agent in the presence of which the maximal response to the agonist was reduced by 50%). EC50 values were not determined when the relaxation was <20% or when a contraction instead of a relaxation was observed.
Effects of Bradykinin on In Vitro Production of Prostacyclin
In an additional group of 11 dogs (group E) instrumented with a pulmonary artery catheter, 6 dogs were treated with L-NNA for 7 days, and 5 dogs served as controls. After the animal was killed, the proximal left anterior descending coronary artery was isolated as described for groups A and B. The artery was cleaned of adherent tissue, cut into rings of 4 to 5 mm in length, and placed in 2 mL of Krebs-Ringer solution in a water bath at 37°C (control solution). In some of the rings, endothelium removal was performed by rubbing the intimal layer with a forceps. The rings were allowed to equilibrate for 30 minutes, after which the incubation medium was removed and replaced by 2 mL of fresh control solution. Incubation was then pursued for 30 minutes at 37°C, as preliminary experiments showed that 6-ketoprostaglandin F1α, the stable metabolite of prostacyclin, plateaued by that time of incubation following addition of bradykinin into the medium. At the end of this 30-minute incubation period, 1 mL was collected and frozen for measurement of the basal production of prostacyclin and immediately replaced by 1 mL of a bradykinin solution (1 μmol/L). Incubation was further pursued for 30 minutes, after which the medium was collected and frozen until measurement of bradykinin-stimulated production of prostacyclin. Length and width of each ring were measured, and the corresponding area was calculated. Measurement of 6-ketoprostaglandin F1α was performed by an enzyme immunoassay using a commercial kit (Amersham), and results are expressed as pg/mm2.
Tissue cGMP and Plasma Catecholamines and TNF Concentration Measurements
Measurement of cGMP content was made in carotid arteries using the method previously described by Arnal et al.5 Briefly, extraction was performed in 1 mL of ice-cold HCl 0.1N with a homogenizer at 13 500 rpm. The homogenate was then centrifuged 15 minutes at 4°C at 13 500 rpm, and the supernatant was frozen at −20°C until the assay. Radioimmunoassay of cGMP was performed with a commercial kit (Amersham) using the acetylation procedure. Protein contents of the extract were measured by Bradford's method21 with Coomassie brilliant blue (Bio-Rad). BSA was used as the standard. In 8 dogs from groups A and B, catecholamine (norepinephrine and epinephrine) plasma levels were assayed by liquid chromatography with electrochemical detection before and on the seventh day of L-NNA administration.
TNF activity was measured in plasma before and after 7 days of L-NNA administration by a specific in vitro cell cytotoxicity assay using murine fibroblasts according to the method previously described by Meager et al.22
All data were expressed as mean±SEM. The hypothesis that values of the hemodynamic parameters were different after L-NNA from before L-NNA was tested by a one-way ANOVA. For such a comparison, only values recorded at distance of L-NNA administration (ie, before the daily administration) were analyzed. The effects of exercise on systemic and coronary hemodynamic parameters in the absence or presence of L-NNA were analyzed by a two-way ANOVA for two within factors, ie, factor “L-NNA” (control, day 7) and factor “exercise stage” (rest, Ex5, Ex10, and Ex12). The effects of diclofenac on baseline and reactive hyperemia parameters in the absence or presence of L-NNA were analyzed by a three-way ANOVA for three within factors, ie, factor “diclofenac” (absence or presence), factor “L-NNA” (absence or presence), and factor “reactive hyperemia duration” (10, 20, and 30 seconds), followed by Bonferroni's t test. The analyses of the effects of diclofenac alone either before or during NO synthase inhibition and of the effects of L-NNA alone were performed by a two-way ANOVA only if the three-way ANOVA was significant for the parameter tested. The interaction between the factor “diclofenac” and the factor “L-NNA” allowed us to test the possibility that diclofenac had different effects whether or not L-NNA was administered. A similar analysis was performed for the effects of acetylcholine and bradykinin (the third within factor being the factor “dose”). Concentration-response curves for acetylcholine, bradykinin, and Sin-1 were compared between and within groups by a two-way ANOVA. Maximal contractions and relaxations were compared between groups by a two-tailed Student's t test for unpaired data and within groups by a one-way ANOVA for repeated measures, followed by a Student's paired t test with Bonferroni's correction. EC50 values were analyzed by nonparametric tests: Mann-Whitney U test between groups and Friedman test within groups followed by a Wilcoxon test. The statistical analyses were performed using Statview 4.0.2 and SuperANOVA statistical software (Abacus Concepts, Inc). The significance level was fixed at 5%.
L-NNA, acetylcholine, bradykinin, L-NMMA, l-arginine, indomethacin, and the thromboxane analogue U46619 (dideoxy-epoxy-methanoprostaglandin F2α) were obtained from Sigma Chimie, and diclofenac was from CIBA-GEIGY. Enalaprilic acid and Sin-1 were generous gifts from Laboratoires MSD-Chibret and Hoechst, respectively. All drugs were dissolved in water except for indomethacin, which was dissolved in dimethyl sulfoxide, and L-NNA (see above). Dimethyl sulfoxide had no effect per se at the final concentration reached in the bath.
Dogs exhibited anorexia and a reduction of body weight (−8±1% from 21±7 kg) during the treatment with L-NNA. Such side effects were not observed during the same postoperative period in sham-operated untreated dogs.
Effects of L-NNA on Systemic and Coronary Hemodynamic Parameters at Rest
Changes in systemic and coronary hemodynamic parameters measured at rest in lying dogs and resulting from administration of L-NNA are illustrated in Fig 1⇓. Because changes in heart rate and arterial blood pressure were similar in dogs from groups A and B, corresponding data were pooled. As shown in Fig 1A⇓, a marked increase in systolic (+30±6% from 114±3 mm Hg), diastolic (+38±9% from 77±2 mm Hg), and mean (+35±8% from 90±3 mm Hg) arterial blood pressures was observed after the first infusion of L-NNA (all P<.001). Thereafter, diastolic and mean arterial blood pressures rapidly returned to their corresponding control levels and were not further modified throughout the following week of treatment with L-NNA. A slight but significant increase in systolic blood pressure persisted during this period (+7±3%, P<.01). Heart rate decreased immediately after the first infusion of L-NNA (−24±8% from 94±6 bpm) and then plateaued between the first and the seventh day of treatment (−38±3% and −37±3%, respectively; all P<.001). Except for the first day, phasic and mean arterial blood pressures and heart rate values measured daily before and 10 minutes after completion of L-NNA administration were identical.
As illustrated in Fig 1B⇑, cardiac output decreased during the first infusion of L-NNA (−33±4% from 2.21±0.19 L/min, P<.001) and remained at this low level throughout the L-NNA treatment period (−40±5% at day 7), whereas simultaneously stroke volume and LV dP/dt remained unchanged. Total peripheral resistance increased markedly after the first infusion of L-NNA (+112±10% from 3468±378 dyne·s·cm−5, P<.001) and then decreased slightly before plateauing until the end of treatment (+73±14% at day 7).
As illustrated in Fig 1C and 1D⇑⇑, L-NNA induced a rapid and sustained constriction of large epicardial coronary arteries (−6±1% and −7±3% on the first and seventh days of treatment by L-NNA, respectively, from a control value of 3314±224 μm; all P<.01). At the level of resistance coronary arteries, slight but not significant decreases in coronary blood flow and increases in coronary vascular resistance were observed throughout the week of treatment with L-NNA.
Most frequently, dogs exhibited atrioventricular block throughout the week of L-NNA administration. The mean PR interval increased from 120±13 milliseconds at baseline to 140±25 milliseconds at the end of this period (P<.05), whereas QRS duration remained unchanged.
Effects of L-NNA on Systemic and Coronary Hemodynamic Parameters During Exercise
Treatment with L-NNA induced an increase in total peripheral resistance and decreases in heart rate and cardiac output that were of the same magnitude as those observed when measured in the same dogs in the lying position. At day 7, L-NNA did not affect baseline mean arterial blood pressure and stroke volume values (Figs 2⇓ and 3).
Changes in heart rate and mean arterial blood pressure during exercise were similar in groups A and B, and data were pooled accordingly. Before L-NNA, Fig 2⇑ shows that heart rate and mean arterial blood pressure increased at each level of exercise. After 7 days of L-NNA administration, the marked reduction in heart rate observed at rest persisted although attenuated at all stages of exercise (−32±3% at baseline, −12±4% at 5 km/h, −8±4% at 10 km/h, and −7±3% at 12 km/h). As shown in Fig 2B⇑, mean arterial blood pressure markedly increased after L-NNA (eg, +34±4 mm Hg on the seventh day of treatment versus +17±3 mm Hg during control exercise at 12 km/h, P<.01).
As shown in Fig 3⇓, cardiac output, stroke volume, and LV dP/dt increased, whereas total peripheral resistance decreased during exercise before L-NNA. Cardiac output and stroke volume remained significantly lower and total peripheral resistance remained higher throughout exercise compared with before L-NNA. Expressed as a relative value, the increase in total peripheral resistance induced by L-NNA was similar at rest (+69±12%) and at all stages of exercise (+75±15% at Ex5, +68±16% at Ex10, and +71±18% at Ex12). Finally, the increase in LV dP/dt induced by exercise remained unaffected by the L-NNA treatment.
Heart rate measured after administration of methylatropine was similar before and 4 days after the administration of L-NNA, either at rest (199±9 versus 195±17 bpm) or during exercise (+254±9 versus +250±3 bpm at 12 km/h). However, exercise-induced increases in cardiac output and stroke volume observed after methylatropine before L-NNA were significantly reduced after (eg, cardiac output, +8.4±0.1 versus +5.5±0.6 L/min at 12 km/h, P<.01; stroke volume, +33±1 versus +22±3 mL at 12 km/h; P<.05).
As shown in Fig 4C⇓, exercise-induced dilation of large epicardial coronary arteries was totally blunted after L-NNA (eg, at 12 km/h, +149±53 μm before versus −6±44 μm on the seventh day of L-NNA; P<.001). In contrast, exercise-induced increase in coronary blood flow and decrease in coronary vascular resistance were not significantly different before and after L-NNA (Fig 4A and 4B⇓⇓). Regardless of the level of exercise, the calculated double product of mean arterial blood pressure by heart rate remained unchanged by L-NNA (Fig 4D⇓).
Effects of Enalaprilic Acid
In 3 dogs from group C, enalaprilic acid had no effect on systemic and coronary hemodynamic parameters when administered before and on the seventh day of L-NNA treatment.
Effects of Diclofenac Before and After L-NNA on Coronary Reactive Hyperemia
Table 1⇓ summarizes the baseline values of the investigated parameters as measured before the reactive hyperemic responses in dogs of group C. Before administration of L-NNA, diclofenac decreased heart rate and LV dP/dt. These effects were still observed after 7 days of L-NNA administration. Heart rate was reduced after L-NNA, whereas L-NNA and diclofenac had no significant effect on resting mean arterial blood pressure, coronary blood flow, and coronary vascular resistance.
Table 2⇓ summarizes the values of peak coronary blood flow, minimal coronary vascular resistance, flow repayment duration, and repayment-to-debt ratio measured at increasing levels of reactive hyperemia under control conditions and after administration of diclofenac before and after 1 week of treatment with L-NNA (group C). The relationships between flow debt and flow repayment volumes of reactive hyperemia are illustrated in Fig 5⇓.
Before L-NNA, release of 10-, 20-, and 30-second coronary occlusions induced peak increases in coronary blood flow that averaged 362±33%, 373±35%, and 366±35%, respectively, and peak decreases in coronary vascular resistance that averaged 77±1%, 78±1%, and 77±2%, respectively. Corresponding flow debt and flow repayment volumes and flow repayment duration increased with the duration of coronary occlusion (all P<.001). Diclofenac administration had no effect on these reactive hyperemia parameters.
After 1 week of L-NNA administration, a significant decrease in flow repayment volumes was observed, although corresponding flow debt volumes remained unchanged (Fig 5⇑). Such decreases in flow repayment volumes mainly resulted from significant decreases in flow repayment durations (Table 2⇑), since simultaneously, no changes in peak increases in coronary blood flow and decreases in coronary vascular resistance were observed. These effects of L-NNA resulted in a significant decrease in the repayment-to-debt ratio regardless of the duration of coronary artery occlusion (Table 2⇑).
Diclofenac administered after L-NNA had no significant effect on the flow debt volumes of reactive hyperemia (Fig 5⇑). A significant interaction (P<.05) was observed between the “diclofenac” and the “L-NNA” factors for the flow repayment volume and duration and for the repayment-to-debt ratio; ie, diclofenac decreased these parameters after L-NNA only. The decrease in flow repayment volume was due to a reduction in flow repayment duration and was also related to the duration of coronary artery occlusion (Fig 5⇑).
Effects of L-NNA and Diclofenac on Systemic and Coronary Hemodynamic Responses to Acetylcholine and Bradykinin
As shown in Table 3⇓, L-NNA attenuated the dose-dependent decrease in mean arterial blood pressure and increase in coronary blood flow induced by acetylcholine and bradykinin (all P<.05). Diclofenac did not affect these responses whether administered before or after L-NNA.
Effects of L-NNA on Responses of Isolated Coronary and Femoral Arteries to Bradykinin, Acetylcholine, and Sin-1
Contractions of Coronary and Femoral Arteries
In coronary arteries with an intact endothelium, contractions to 1 μmol/L of U46619 were similar in vessels taken from sham-operated dogs and from L-NNA–treated dogs (9.8±1.2 and 11.3±1.2 g, respectively). Addition of L-NMMA did not alter the contractile response to U46619 (10.7±0.9 and 12.3±1.3 g, respectively), but this contraction was slightly increased after the addition of L-NMMA plus indomethacin (11.8±1.0 and 14.2±1.6 g, respectively; both P<.01). In femoral arteries with an intact endothelium, the contractile response to norepinephrine (0.3 μmol/L) was similar in control and in L-NNA–treated dogs (7.0±0.6 and 8.5±0.7 g, respectively). In vessels taken from control dogs, neither L-NMMA nor L-NMMA plus indomethacin had a significant effect on the contractile responses to norepinephrine. In contrast, in vessels taken from L-NNA–treated dogs, whereas L-NMMA did not affect norepinephrine contraction (8.6±0.8 g, P=NS), the latter was increased by L-NMMA plus indomethacin (10.0±0.5 g, P<.01).
Relaxations to Bradykinin
As shown in Fig 6⇓, bradykinin induced a concentration- and endothelium-dependent relaxation of coronary arteries taken from sham-operated dogs (controls, group D). L-NMMA significantly shifted the concentration-relaxation curve rightward and reduced the maximal relaxation to bradykinin (Table 4⇓), but indomethacin had no additional effect on these responses. In coronary arteries taken from L-NNA–treated dogs, the maximal relaxation induced by bradykinin was reduced by 45% and remained unchanged after the addition of L-NMMA. However, the residual relaxation to bradykinin was almost completely abolished when indomethacin was added to L-NMMA (Fig 6⇓). In a number of experiments, (1) indomethacin alone inhibited the relaxant effect of bradykinin to an extent similar to that achieved when indomethacin associated with L-NMMA, and (2) the effects elicited by diclofenac were similar to those observed with indomethacin (data not shown).
In femoral arteries taken from sham-operated dogs, bradykinin induced concentration- and endothelium-dependent relaxations that were abolished by L-NMMA alone or associated with indomethacin. In femoral arteries taken from all but 2 L-NNA–treated dogs (Fig 6⇑ and Table 4⇑), a concentration-dependent contraction to bradykinin was observed regardless of the presence or absence of an intact endothelium and regardless of the presence or absence of L-NMMA or indomethacin. In femoral segments from 2 dogs (data not included in Fig 6⇓ and Table 4⇑), bradykinin still exhibited a slight concentration-dependent relaxation after L-NMMA (maximal relaxation, 29±12%) that was abolished after the addition of indomethacin (5±12%) and restored to contraction after endothelium removal (−11±16%).
Relaxations to Acetylcholine
As shown in Fig 7⇑, acetylcholine induced a concentration- and endothelium-dependent relaxation of coronary arteries taken from sham-operated dogs (control, group D). L-NMMA significantly shifted the concentration-relaxation curve rightward and reduced by 48% the maximal relaxation to acetylcholine (Table 4⇑), and indomethacin had no additional effect (Table 4⇑). In contrast, in coronary arteries isolated from L-NNA–treated dogs, administration of acetylcholine was not followed by relaxation (Fig 7⇑ and Table 4⇑).
In segments of femoral arteries taken from L-NNA–treated dogs, the concentration-relaxation curve was shifted rightward compared with the corresponding one obtained in control segments, and maximal relaxation was decreased (Fig 7⇑ and Table 4⇑). L-NMMA had no additional effect on the relaxant responses to acetylcholine in these vessels, but indomethacin associated with L-NMMA further reduced the maximal relaxation to acetylcholine by 50%. Unlike bradykinin, acetylcholine never induced contractions of femoral arteries taken from L-NNA–treated dogs, even in the absence of an intact endothelium (Fig 7⇑).
Relaxations to Sin-1
Maximal relaxations of coronary and femoral arteries to the NO donor Sin-1 were not different in segments taken from control and L-NNA–treated dogs. L-NMMA decreased the EC50 of Sin-1 in coronary (63±21 versus 31±4 nmol/L, P<.05) and femoral (165±25 versus 117±16 nmol/L, P<.05) arteries taken from control dogs but did not affect that in coronary (143±59 versus 153±54 nmol/L) and femoral (147±26 versus 130±21 nmol/L) arteries taken from L-NNA–treated dogs. Indomethacin had no effect on Sin-1–induced relaxations in all conditions.
Effects of Bradykinin on In Vitro Production of Prostacyclin
In rings of the left anterior descending coronary artery isolated from control dogs, bradykinin (1 μmol/L) increased the production of 6-ketoprostaglandin F1α by 66±18 pg/mm2 from a baseline value of 50±18 pg/mm2 (P<.05, n=5). In rings taken from L-NNA–treated dogs during 7 days, the baseline value of 6-ketoprostaglandin F1α was not significantly different from that of control dogs (53±10 pg/mm2), but bradykinin induced a much greater increase in the production of 6-ketoprostaglandin F1α (+179±41 pg/mm2, n=6) than in corresponding control dogs (P<.05), thus reflecting the activation of the COX pathway in these vessels after sustained blockade of NO synthesis. This activation was located on the vascular endothelium as no production of 6-ketoprostaglandin F1α occurred after the addition of bradykinin to deendothelialized rings (−3±12 pg/mm2, from a baseline value of 39±10 pg/mm2, n=6).
Effects of L-NNA on Tissue cGMP and on Plasma Levels of Catecholamines and TNF
cGMP content in carotid arteries taken from L-NNA–treated dogs was reduced by 47% compared with its value in carotid arteries taken from sham-operated dogs (32±3 versus 60±7 pmol/μg of proteins, P<.001).
Plasma levels of TNF were not significantly modified by the L-NNA treatment (18±5 U/mL before versus 30±7 U/mL after L-NNA, n=6).
Plasma levels of norepinephrine and epinephrine were not significantly modified by the L-NNA treatment (211±32/75±28 pg/mL before versus 295±27/62±15 pg/mL after L-NNA, n=8).
The present study demonstrates that sustained inhibition of NO synthesis in dogs does not induce systemic hypertension at rest but results in an exaggerated hypertensive response during treadmill exercise. Although NO synthesis inhibition induces an increase in total peripheral resistance, both at rest and during exercise, coronary vascular resistance is not significantly modified. Finally, this study provides evidence suggesting that during sustained inhibition of NO synthesis, the COX metabolic pathway is activated or induced within coronary arteries and that this pathway (1) participates in myocardial reactive hyperemia in vivo and (2) contributes through increased production of prostacyclin to the residual relaxant effect of bradykinin in isolated coronary vessels. This phenomenon might be considered as a regulatory mechanism aimed at compensating for NO deficiency.
L-NNA was chosen for its high potency at inhibiting NO synthase23 and for its lack of muscarinic antagonist effect when administered intravenously.24 L-NNA was administered once a day because previous data showed that a single injection of the drug induces a stable and long-lasting reduction of heart rate and because the terminal half-life of L-NNA is 20.0±4.9 hours after an intravenous bolus injection of 10 mg/kg.25 Finally, the regimen of 20 mg·kg−1·d−1 of L-NNA was chosen because higher doses resulted in ventricular tachycardia in pilot experiments.
After the first administration of L-NNA, a marked systolic/diastolic hypertension was observed similar to that previously described in rats,5 rabbits,26 and dogs.2 8 9 10 11 12 27 However, as early as 24 hours later, diastolic and mean arterial blood pressures had returned to their control levels, and only a slight but significant systolic hypertension persisted throughout the following days of L-NNA treatment. This surprising observation strongly contrasts with the major systemic hypertension that develops in rats throughout chronic administration of NO synthase inhibitors.5 That hypertension was not maintained in our dogs during chronic L-NNA treatment cannot be accounted for by an insufficient degree of NO synthase blockade because simultaneously (1) in vivo, the dilating responses to the endothelium-dependent vasodilators acetylcholine and bradykinin were strongly reduced as previously reported in rats28 and rabbits,26 (2) ex vivo, the carotid artery cGMP content was decreased,5 and (3) in vitro, the acute addition of another NO synthase inhibitor, L-NMMA, did not further inhibit the relaxant responses to acetylcholine and bradykinin of isolated coronary and femoral arteries taken from L-NNA–treated dogs.
In the present study, arterial blood pressure was measured in conscious dogs lying in quiet resting conditions and after long periods of table training. An explanation for the lack of major systemic hypertension during subchronic administration of L-NNA might lie in species differences in the intrinsic ability to decrease heart rate and subsequently cardiac output through baroreceptor-mediated vagal activation. On the basis of the well-documented buffering effect of NO on sympathetic tone,29 the unstressful conditions in which blood pressure was measured in our dogs (at difference with the situation that prevails in rats or rabbits, in which this parameter is measured either by the tail-cuff method in conscious animals or invasively under general anesthesia) might also explain, in part, our results. Finally, arterial blood pressure was measured during a relatively short period of time (≈1 hour after the infusion of L-NNA), and it is possible that continuous recording of this parameter with telemetry systems might have revealed a blood pressure pattern similar to that previously reported.27
L-NNA strongly increased total peripheral resistance; hence, the simultaneous decrease in cardiac output was clearly responsible for the lack of hypertension at rest. That the substantial weight loss experienced by the animals during the week of treatment with L-NNA could contribute to the increase in peripheral resistance through activation of the sympathetic and/or of the renin-angiotensin systems is unlikely, since the plasma catecholamine levels as well as the hemodynamic responses to enalaprilic acid were similar before and after the 1-week treatment with L-NNA. However, alterations in other neurohumoral pathways such as the endothelin30 or the arginine vasopressin systems cannot be excluded. Reduction of cardiac output was itself the consequence of the decrease in heart rate, since calculated stroke volume and cardiac contractility, as assessed by LV dP/dt, remained simultaneously unchanged. The increase in PR interval, the occurrence of atrioventricular blocks, and the complete correction of L-NNA–induced bradycardia by methylatropine strongly argue for a sustained L-NNA–induced activation of the parasympathetic tone in the present experimental model. Similar correction of L-NNA–induced bradycardia by atropine has also been reported in guinea pigs31 and rats,32 but the hypothesis of a direct depressant effect of L-NNA on atrioventricular nodes cannot be ruled out.
Contrasting with its minor effect on systemic blood pressure at rest, L-NNA markedly potentiated exercise-induced increase in blood pressure. This was due to the maintenance throughout exercise of total peripheral resistance values higher after than before L-NNA, despite L-NNA–induced limitation of exercise-induced increases in cardiac output and stroke volume. Flow-dependent vasodilation has been shown to occur in muscular resistance arteries as small as 20 μm in diameter,33 and it is likely that abolition of this mechanism is responsible for the higher values of total peripheral resistance observed after L-NNA both at rest and during exercise. That L-NNA–induced increase in total peripheral resistance could be completely overcome by nitroglycerin administration (data not shown) strongly supports this view.
The increase in cardiac output observed during control exercise was clearly blunted by L-NNA. This was due to (1) the maintenance throughout exercise of heart rate values lower after than before L-NNA and (2) the suppression by L-NNA of exercise-induced increase in stroke volume. The fact that both cardiac output and stroke volume were still significantly blunted during exercise when L-NNA was combined with methylatropine is in accordance with the preceding statement. These unadapted hemodynamic responses to exercise were observed despite a likely enhancement of global oxygen consumption, since NO is known to exert an inhibitory effect on tissue metabolism.2
Exercise-induced dilation of epicardial coronary artery was totally blunted by L-NNA, as already demonstrated by Wang et al.10 In contrast and as previously reported after an acute administration,9 11 12 L-NNA administered subchronically does not significantly alter coronary blood flow and resistance either at rest or during exercise. This finding suggests either that NO is not of critical importance for metabolic autoregulation of coronary blood flow or that other vasodilatory mechanisms, acting to compensate for the loss of NO, also operate in this particular vascular bed. For the first time, we provide functional arguments demonstrating that after sustained blockade of NO synthase, the COX metabolic pathway can play such a compensatory role in the coronary circulation as a result of a complex cross talk between vascular NO synthase and COX. This statement is based on in vivo and in vitro observations made on L-NNA–treated dogs showing that after chronic NO synthase blockade, myocardial reactive hyperemic responses and endothelium-dependent relaxations to bradykinin of isolated coronary arteries were blunted by the COX inhibitors diclofenac and indomethacin, respectively. Although previous studies34 have reported that endothelium-derived prostanoids may act as mediators of receptor-dependent relaxing responses in coronary vessels, our observations indicate that vascular prostanoids are involved in myocardial reactive hyperemia and relaxant responses of isolated coronary vessels to bradykinin only after L-NNA, suggesting that sustained inhibition of NO synthesis could shift the metabolic pathway of EDRF/NO toward COX-derived products. This hypothesis is further supported by our experiments showing that bradykinin-induced production of 6-ketoprostaglandin F1α by isolated coronary arteries was increased to a much greater extent (P<.05) in vessels obtained from dogs previously treated for 1 week with L-NNA (+337%) than from sham-operated control dogs (+132%). This shift from the EDRF/NO toward the COX pathway appears (1) to partly compensate for the effects of NO synthesis inhibition on myocardial reactive hyperemia, (2) to account for the residual relaxation to bradykinin (but surprisingly not to acetylcholine) of isolated coronary vessels, (3) to involve endothelium, because it is no longer observed after endothelium removal, and (4) to selectively develop in certain vascular beds (eg, the coronary artery) but not in all (eg, the femoral artery).
The COX metabolic pathway was not involved in myocardial reactive hyperemia before L-NNA administration. This result is in agreement with previous studies by Holtz et al,35 who used a protocol similar to ours, and by Dai and Bache,18 who reported that prostaglandins were not involved in the mediation of exercise-induced coronary vasodilation in conscious dogs. As already demonstrated in several studies,9 11 inhibition of NO synthesis blunted the flow repayment duration and volume and debt-to-repayment ratios after 10-, 20-, and 30-second coronary occlusions, although flow debt volumes were the same. After 1 week of NO synthesis inhibition, diclofenac further decreased flow repayment duration and volume and debt-to-repayment ratios, suggesting that COX-derived prostanoids partly mediated the late phase of reactive hyperemia and that their synthesis was triggered by shear stress. The fact that diclofenac-induced decrease in blood flow repayment was proportional to the duration of the occlusion is another argument in favor of this hypothesis. Since bradykinin has been postulated to be involved in flow-dependent vasodilation of human coronary conductance vessels36 and since our present in vitro data suggest that endothelial COX and bradykinin could be functionally linked, we hypothesized that the in vivo participation of the COX metabolic pathway observed after sustained NO synthesis inhibition at the level of coronary resistance vessels could also depend on endogenous bradykinin. However, this hypothesis was not confirmed, since HOE 140, a selective bradykinin B2 receptor antagonist37 administered at a dose that completely inhibited the effects of 1 μg/kg of intravenous bradykinin, had no significant effect on myocardial reactive hyperemia, before as well as after NO synthesis inhibition (data not shown).
An apparent discrepancy in the present study lies in the fact that administration of a COX inhibitor significantly blunted the in vitro relaxant responses to bradykinin in isolated large coronary arteries taken from L-NNA–treated dogs, whereas the coronary vasodilator effects of bradykinin were not altered by diclofenac in vivo. This could result from the difference in the size of the coronary vessels studied, ie, conductance vessels in vitro versus resistance vessels in vivo. The fact that diclofenac did not alter the coronary vasodilator effects of acetylcholine and bradykinin but clearly blunted myocardial reactive hyperemia in dogs treated by L-NNA suggests that in vivo and at the level of coronary resistance vessels, the COX metabolic pathway is triggered by flow-dependent, but not receptor-dependent, mechanisms.
Numerous studies have previously reported that important differences exist in the mediation of the endothelium-dependent relaxant effects of bradykinin and acetylcholine at the level of canine and porcine coronary arteries.38 39 As observed in the present study, a large component of the bradykinin-induced relaxation of the coronary arteries of control dogs was independent of the EDRF/NO pathway, as L-NMMA attenuated the relaxant effect of bradykinin by only 17%, and neither indomethacin nor diclofenac displayed any additional inhibitory effect to that of L-NMMA. In contrast, L-NMMA attenuated by >50% the relaxant effect of acetylcholine in the same vessels. This “non-NO, nonprostanoid”-dependent component of bradykinin-induced relaxation has been attributed generally to the endothelial synthesis of a diffusible hyperpolarizing factor (EDHF).40 In coronary arteries taken from L-NNA–treated dogs, bradykinin-induced residual relaxation was insensitive to L-NMMA, an observation consistent with a total inhibition of NO synthase. In these vessels, the relaxation to bradykinin became mostly dependent on the COX metabolic pathway, since it was blunted by 79% after COX inhibition by indomethacin. This again is in line with the greater increase in bradykinin-induced production of 6-ketoprostaglandin F1α observed in coronary vessels obtained from L-NNA–treated dogs compared with sham-operated control dogs. These findings indicate that the COX metabolic pathway could have been either activated or induced after sustained NO synthesis inhibition. For reasons that remain to be elucidated, our results suggest that this shift toward the COX pathway is evidenced only when the relaxation observed in control arteries is predominantly mediated by NO. Accordingly, this mechanism was observed in vitro (1) on coronary arteries when bradykinin, but not acetylcholine, was used as the stimulating agent, and (2) on femoral arteries when acetylcholine, but not bradykinin, was used as the stimulating agent.
The present results suggest that the involvement of the COX pathway observed in coronary arteries taken from L-NNA–treated dogs is an endothelium-dependent phenomenon, because it is abolished after mechanical disruption of the endothelium in vitro. Indeed, recent biochemical and pharmacological data suggest the existence of both constitutive (COX-1) and inducible (COX-2) isoforms of COX, the latter being expressed in inflammation states. COX-2 has initially been described in human and murine inflammatory cells, such as monocytes or macrophages stimulated by endotoxin or cytokines.41 42 COX-2 can also be induced in smooth muscle cells in response to injury43 and in endothelial cells after exposition to cytokines.15 Nonsteroidal anti-inflammatory drugs have different selectivities for constitutive and inducible COX isoforms. Indomethacin is ≈60 times more potent at inhibiting COX-1 than COX-2 in intact cells, whereas this ratio is only 0.7 for diclofenac.44 The high concentrations of indomethacin and diclofenac used either in vivo or in vitro in the present study preclude us from drawing any conclusion on the type of COX isoform involved in the coronary vessels. The fact that plasma TNF concentrations remained unchanged after sustained inhibition of NO synthesis rules out the hypothesis of an induction of COX-2 isoform due to an inflammation state secondary to tissue hypoperfusion induced by NO synthase inhibition.28 Because of the absence of a significant hypertension during the week of treatment with L-NNA, the potential role of high vascular pressure in COX activation can also be ruled out. The hypothesis of direct cross talk between NO and COX may be evoked, but this question remains unclear to date. NO has been shown to activate COX in different cultured cell lines45 46 and in different models in vivo.47 48 Conversely, an inhibitory effect of NO on COX has also been reported in murine lipopolysaccharide-activated macrophages,49 bovine endothelial cells,14 rat Kupffer cells,50 and articular chondrocytes.51 Interestingly, NO synthase inhibitors induced an increase in the amount of COX-2 protein synthesized by lipopolysaccharide-stimulated macrophages.49 The fact that in the present study COX metabolic pathway involvement occurred in vivo only after sustained inhibition of NO by L-NNA but not in vitro after acute inhibition of NO synthesis by L-NMMA favors the hypothesis of COX induction rather than that of a COX activation. Finally, the occurrence of severe gut hemorrhage and of rapid health deterioration when diclofenac and L-NNA were simultaneously administered to 3 dogs in our pilot experiment is another argument to support that EDRF/NO and COX systems compensate for each other to maintain homeostasis.
In conclusion, the present study indicates that in conscious dogs (1) subchronic NO blockade is not associated with hypertension at rest, (2) NO does not play a major role in the control of coronary resistance vessels in both resting and exercising conditions, and (3) an involvement of the COX pathway is evidenced after L-NNA, which participates in reactive hyperemia in vivo and mediates relaxation to bradykinin of large epicardial coronary arteries in vitro. This involvement of the COX pathway might protect the coronary circulation from the deleterious effects of NO synthesis inhibition.
Selected Abbreviations and Acronyms
|EDHF||=||endothelium-derived hyperpolarizing factor|
|EDRF||=||endothelium-derived relaxing factor|
|Ex5, Ex10, Ex12||=||exercise at 5, 10, and 12 km/h|
|LV dP/dt||=||first derivative of left ventricular pressure|
|TNF||=||tissue necrosis factor|
This study was supported by a grant from La Fondation de France. Dr Puybasset was supported by fellowship grants from the Groupe de Reflexion sur la Recherche Cardiovasculaire (GRRC) and from the Conseil Scientifique de la Faculté de Médecine Paris-Sud. The authors are greatly indebted to Isabelle Dubus and Alain Bizé for their excellent technical assistance, to Eric Pussard and Régine Merval for plasma catecholamines and TNF measurements, respectively, and to Dr Eric Vicaut for his statistical advice.
This manuscript was sent to Peter Libby, Consulting Editor, for review by expert referees, editorial decision, and final disposition.
- Received October 27, 1995.
- Accepted May 13, 1996.
Rees DD, Palmer RMJ, Moncada S. Role of endothelium-derived nitric oxide in the regulation of blood pressure. Proc Natl Acad Sci U S A. 1989;86:3375-3378.
Shen W, Xu X, Ochoa M, Wolin MS, Hintze TH. Role of nitric oxide in the regulation of oxygen consumption in conscious dogs. Circ Res. 1994;75:1086-1095.
Arnal JF, Warin L, Michel JB. Determinants of aortic cyclic guanosine monophosphate in hypertension induced by chronic inhibition of nitric oxide synthase. J Clin Invest. 1992;90:647-652.
Chu A, Chambers DE, Lin CC, Kuehl WD, Palmer MJ, Moncada S, Cobb FR. Effects of inhibition of nitric oxide formation on basal vasomotion and endothelium-dependent responses of the coronary arteries in awake dogs. J Clin Invest. 1991;87:1964-1968.
Parent R, Paré R, Lavallée M. Contribution of nitric oxide to dilation of resistance coronary vessels in conscious dogs. Am J Physiol. 1992;262:H10-H16.
Wang J, Wolin MS, Hintze TH. Chronic exercise enhances endothelium-mediated dilation of epicardial coronary artery in conscious dogs. Circ Res. 1993;73:829-838.
Altman JD, Kinn J, Duncker DJ, Bache R. Effect of inhibition of nitric oxide formation on coronary blood flow during exercise in the dog. Cardiovasc Res. 1994;28:119-124.
Duncker DJ, Bache RJ. Inhibition of nitric oxide production aggravates myocardial hypoperfusion during exercise in the presence of a coronary artery stenosis. Circ Res. 1994;74:629-640.
Park KH, Rubin LE, Gross SS, Levi R. Nitric oxide is a mediator of hypoxic coronary vasodilatation: relation to adenosine and cyclooxygenase-derived metabolites. Circ Res. 1992;71:992-1001.
Habib A, Creminon C, Frobert Y, Grassi J, Pradelles P, Maclouf J. Demonstration of an inducible cyclooxygenase in human endothelial cells using antibodies raised against the carboxyl-terminal region of the cyclooxygenase-2. J Biol Chem. 1993;268:32448-32454.
Berdeaux A, Ghaleh B, Dubois-Randé JL, Vigué B, Drieu La Rochelle C, Hittinger L, Giudicelli JF. Role of vascular endothelium in exercise-induced dilation of large epicardial coronary arteries in conscious dogs. Circulation. 1994;89:2799-2808.
Ishida T, Lewis RM, Hartley CJ, Entman ML, Field JB. Comparison of hepatic extraction of insulin and glucagon in conscious and anaesthetized dogs. J Endocrinol. 1983;112:1098-1109.
Dai X, Bache RJ. Effect of indomethacin on coronary blood flow during graded treadmill exercise in the dog. Am J Physiol. 1984;247:H452-H458.
Vargas HM, Cuevas JM, Ignarro LI, Chaudhuri G. Comparison of the inhibitory potencies of NG-methyl, NG-nitro and NG-amino-L-arginine on EDRF function in the rat: evidence for continuous basal EDRF release. J Pharmacol Exp Ther. 1991;257:1208-1215.
Buxton IL, Cheek DJ, Eckman D, Westfall DP, Sanders KM, Keef KD. NG-Nitro l-arginine methyl ester and other alkyl esters of arginine are muscarinic receptor antagonists. Circ Res. 1993;72:387-395.
Manning RD, Hu L, Mizelle HL, Montani JP, Norton MW. Cardiovascular responses to long-term blockade of nitric oxide synthesis. Hypertension. 1993;22:40-48.
Zanzinger J, Czachurski J, Seller H. Inhibition of sympathetic vasoconstriction is a major principle of vasodilation by nitric oxide in vivo. Circ Res. 1994;75:1073-1077.
Richard V, Hogie M, Clozel M, Löffler BM, Thuillez C. In vivo evidence of an endothelin-induced vasopressor tone after inhibition of nitric oxide synthesis in rats. Circulation. 1995;91:771-775.
Koller A, Kaley G. Endothelial regulation of wall shear stress and blood flow in skeletal muscle microcirculation. Am J Physiol. 1991;260:H862-H868.
Pinto A, Abraham NG, Mullane KM. Arachidonic acid-induced endothelial-dependent relaxations of canine coronary arteries: contribution of cytochrome P450 dependent pathway. J Pharmacol Exp Ther. 1987;240:856-863.
Groves PH, Kurz S, Drexler H. The role of bradykinin in basal and flow-mediated endothelium-dependent vasodilation in the human coronary circulation. Circulation. 1994;90(suppl I):I-36. Abstract.
Mombouli J-V, Illiano S, Nagao T, Scott-Burden T, Vanhoutte PM. Potentiation of endothelium-dependent relaxations to bradykinin by angiotensin I converting enzyme inhibitors in canine coronary artery involves both endothelium-derived relaxing and hyperpolarizing factors. Circ Res. 1992;71:137-144.
Richard V, Tanner FC, Tschudi M, Lüscher TF. Different activation of L-arginine pathway by bradykinin, serotonin, and clonidine in coronary arteries. Am J Physiol. 1990;259:H1433-H1439.
Masferrer JL, Seibert K, Zweifel BS, Needleman P. Endogeneous glucocorticoids regulate an inducible cyclooxygenase enzyme. Proc Natl Acad Sci U S A.. 1992;89:3917-3921.
Ristimäki A, Garfinkel S, Wessendorf J, Maciag T, Hla T. Induction of cyclooxygenase-2 by interleukin 1 alpha. J Biol Chem. 1994;269:11769-11775.
Pritchard KA, O'Banion MK, Miano JM, Vlasic N, Bhatia UG, Young DA, Stemerman MB. Induction of cyclooxygenase-2 in rat vascular smooth muscle cells in vitro and in vivo. J Biol Chem. 1994;269:8504-8509.
Mitchell JA, Akarasereenont P, Thiemermann C, Flower RJ, Vane JR. Selectivity of nonsteroidal antiinflammatory drugs as inhibitors of constitutive and inducible cyclooxygenase. Proc Natl Acad Sci U S A. 1994;90:11693-11697.
Salvemini D, Misko TP, Masferrer JL, Seibert K, Currie MG, Needleman P. Nitric oxide activates cyclooxygenase enzymes. Proc Natl Acad Sci U S A. 1993;90:7240-7244.
Davidge ST, Baker PN, McLaughin MK, Roberts JM. Nitric oxide produced by endothelial cells increases production of eicosanoids through activation of prostaglandin H synthase. Circ Res. 1995;77:274-283.
Salvemini D, Seibert K, Masferrer JL, Misko TP, Currie MG, Needleman P. Endogeneous nitric oxide enhances prostaglandin production in a model of renal inflammation. J Clin Invest. 1994;93:1940-1947.
Stadler J, Harbrecht BG, Di Silvio M, Curran RD, Jordan ML, Simmons RL, Billiar TR. Endogenous nitric oxide inhibits the synthesis of cyclooxygenase products and interleukin-6 by rat Kupffer cells. J Leukoc Biol. 1993;53:165-172.
Stadler J, Stefanovic-Racic M, Billiar TR, Curran RD, McIntyre LA, Georgescu HI, Simmons RL, Evans CH. Articular chondrocytes synthesize nitric oxide in response to cytokines and lipopolysaccharide. J Immunol. 1991;147:3915-3920.