Nitric oxide (NO) and L-citrulline (L-cit) are released by endothelial NO synthase (eNOS) to induce vasodilation via guanylyl cyclase and cyclic guanosine monophosphate (cGMP). Volatile anesthetics directly reduce vascular muscle tone, but their effects on the eNOS cGMP pathway is controversial. The aim of this study was to examine the effects of anesthetics on bradykinin-induced increases in flow, NO, and L-cit in isolated hearts.


Guinea pig hearts were isolated, perfused at 55 mmHg with a crystalloid or erythrocyte perfusate at 37 degrees C, and heart rate, left ventricular pressure, coronary flow (CF), effluent pH, and oxygen tension were monitored. Effluent [NO] was measured by a Clark-type electrode (sensitivity > or = 1 nM = 3 pA) with a selectively permeable membrane. Effluent [L-cit] was measured by chromatography. Before, during, and after exposure to halothane, isoflurane, or sevoflurane, hearts were infused with as much as 100 nM bradykinin to induce increases in CF and effluent release of NO and L-cit.


In crystalloid-perfused hearts, 10 nm bradykinin produced maximal concentration-dependent increases in CF (87+/-2%), [NO] (24+/-4 nM), NO release (128+/-18 pmol x g(-1) x min(-1)), and [L-cit] (58+/-8 nM). Isoflurane slightly increased CF but not NO. Anesthetics did not alter the bradykinin-induced CF, NO slope relationship, or change [L-cit]. In erythrocyte-perfused hearts, isoflurane also did not alter the bradykinin-induced increase in CF and decrease in percentage of oxygen extracted.


This is the first study to simultaneously measure CF with bradykinin-induced changes in [NO] and [L-cit] in the presence of halothane, isoflurane, and sevoflurane in intact hearts. The study shows for the first time that volatile anesthetics do not alter the CF to NO relationship and suggests that NO production, NO release, and NO vasodilatory effects mediated by the eNOS cGMP pathway are not significantly affected by anesthetics in crystalloid or erythrocyte-perfused guinea pig hearts.

Since the discovery of endothelium-derived relaxing factor [1]and its identification as nitric oxide (NO), [2,3]it has been shown that NO gas is released from the coronary endothelium and modulates local coronary vascular tone. [4]Nitric oxide is generated by stimulation of endothelial NO synthase (eNOS) or by metabolism of nitrosyl compounds at soluble guanylyl cyclase (sGC), [5,6]which in turn promotes vasodilation by increasing smooth muscle cyclic guanosine monophosphate (cGMP). [6]Nitric oxide contributes importantly to coronary vasodilation elicited by various physiologic and pharmacologic stimuli. [7] 

It is possible that some of the vascular effects of anesthetics may be mediated through the endothelium. The role of volatile anesthetics on vascular endothelium has come under intense investigation, but most studies have been done in vitro in endothelial - vessel strip or co-culture preparations in the absence and presence of eNOS antagonism. [8–19]There are several reports that volatile anesthetics alter vascular endothelial function, or the effects of endothelial factors on vascular smooth muscle activity, yet there is no consensus that anesthetics significantly alter the NO signaling pathway. [10,13,20–25] 

We proposed to test in the intact heart if volatile anesthetics significantly alter endothelium-dependent vasodilation induced by bradykinin and if they correspondingly alter the rate of NO and L-citrulline (L-cit) production and release by eNOS. We reasoned that NO release could be monitored continuously on-line from the intact coronary endothelium in crystalloid-perfused hearts because much less luminally released NO is metabolized in the absence of hemoglobin. Coronary flow (CF) measured at constant perfusion pressure assesses global coronary vascular resistance; simultaneous monitoring of the percentage of oxygen extracted assesses vascular tone independent of cardiac metabolism. Only endothelium-dependent vasodilator agents that stimulate NO production should cause both an increase in effluent NO and L-cit concentrations and CF. [26]To assess the possible pathways involved, we measured changes in coronary effluent NO concentration ([NO]), NO release, and L-cit concentration ([L-cit]) with increasing concentrations of an eNOS stimulator, bradykinin, with and without halothane, isoflurane, or sevoflurane. We also determined if volatile agents also significantly alter endothelium-mediated increases in CF in intact, erythrocyte perfused intact hearts.

Langendorff Heart Preparation

The investigation conformed to the Laboratory Animal Care and Use Standards (National Institutes of Health publication 85–23, revised 1995) and was approved by the Medical College of Wisconsin animal studies committee. Hearts were harvested for study after 25 mg ketamine and 1,000 units heparin were injected intraperitoneally into 52 albino English short-haired guinea pigs (250 - 300 g). Description of the surgical preparation for this model has been reported in detail. [26,27]Each heart was perfused in a retrograde manner through the aorta at 55 mmHg with a 95% oxygenated, modified Krebs-Ringer's solution as described before. [26,27]Seventeen of these hearts were perfused with erythrocytes. Isovolumetric left ventricular (LV) pressure, CF, and spontaneous heart rate were measured continuously, as described previously. [26,27]Maximal CF was elicited with adenosine (0.2 ml of 200 [micro sign]M stock solution) injected directly into the aortic root cannula during the initial control period and after the last control reading.

Nitric Oxide, Oxygen, pH, and Anesthetic Measurements

Coronary sinus effluent was collected by placing a small gas-impermeable cannula into the right ventricle through the pulmonary artery after ligating the superior and inferior vena cavae. Coronary outflow (coronary sinus) oxygen tension and pH were measured continuously on-line with a miniature thermostabile Clark oxygen electrode (model 203B; Instech Laboratories, Plymouth Meeting, PA) and a temperature-compensated pH electrode (microcomputer pro-vision pH meter, model 05669–20; pH electrode PHE 2121; Cole Palmer Instruments, Vernon Hills, IL). Perfusate, bath, oxygen, pH, and NO electrode temperatures were maintained tightly at 37.2 +/- 0.1 [degree sign]C using a thermostatically controlled water circulator to jacketted glass tubing, bath, and aluminum heat exchangers. Coronary sinus and NO electrode temperature was tightly controlled because NO electrode current is affected by the artifact of temperature alone (equivalent to +24.5 nM/[degree sign]C.)[26] 

Coronary inflow and effluent pH and oxygen and carbon dioxide tensions were measured during each maneuver off-line at 37 [degree sign]C with an intermittently self-calibrating analyzer system (Radiometer ABL-2; Medtron Chicago, Des Plaines, IL). Coronary effluent carbon dioxide tension was calculated from the on-line pH signal with bicarbonate assumed to be constant at 15 mM; calculated effluent carbon dioxide tension at a given pH was verified off-line with the gas analyzer. The NO electrode was slightly sensitive to changes in carbon dioxide induced by changing the percentage of carbon dioxide flow in the gas mixing chamber in the absence of hearts. [26]The effect of changes in effluent carbon dioxide to alter the pH of the internal electrolyte solution was minimized by using a two-part buffered electrolyte solution (World Precision Instruments, Sarasota, FL, catalog #7521). The small effect of a change in effluent H+concentration (-0.22 nM/mV), or partial carbon dioxide pressure (-0.50 nM/mmHg) was adjusted mathematically at each data point based on the coronary venous pH. The maximal decrease in coronary effluent carbon dioxide tension from control CF (6.3 +/- 0.2 ml [middle dot]-1g [middle dot] min-1) to a maximal steady state increase in CF (13.3 +/- ml [middle dot]-1g [middle dot] min-1) was 4 mmHg, an apparent NO increase of about 2 nM. The concentrations of each anesthetic used did not alone alter NO current.

The NO concentration was measured as the change in redox current (pA) generated by a gas-permeable, water-impermeable NO electrode (ISO-NOP 2 mm; World Precision Instruments, Sarasota, FL). The Clark-type electrode measures dissolved NO gas concentration in aqueous solutions polarographically. [28]An electric current is generated as NO diffuses through the membrane and becomes oxidized at the platinum electrode. Generated current is proportional to the diffusion of NO through the membrane, and diffusion is based on its partial pressure, which, in turn, is proportional to [NO] at the probe tip. Current is measured with a sensitive amperometer during zero voltage suppression to expand the response range. The membrane selectivity for NO over other gases is determined by the potential applied to the electrode. The presence of dissolved NO2gas (2NO + O2= 2NO2) could generate a current, but NO2is highly unstable in aqueous solution and reacts at pH >7 with alkali (2NO2+ 2OH-= NO2-+ NO3-+ H2O) to form nitrite and nitrate ions, which are not permeable.

The NO calibration curves were generated by graded chemical production of NO, where KI and H2SO4are in excess: 2NaNO2+ 2KI + 2H2SO4= 2NO + I2+ 2H2O + K2SO4+ Na (2) SO4. Calibration was done 1 day after the gas-permeable membrane and internal filling solution were changed (they were changed weekly). There was no significant change in NO electrode sensitivity (>or= to 1 nM) during a 1-week period. Calibrations, conducted at the beginning and end of a 5-day use period, were unchanged over time. The average change in current generated for a given concentration of chemically generated NO (i.e., the slope [P <or= to 0.0001]) amounted to 2.92 pA/nM +/- 0.17 (95% CI), so that for every pA increase in current, the NO concentration was increased by 0.34 nM. The Y intercept was not significantly different from zero. Transient time for perfusate from the heart to the NO electrode via the coronary sinus and pulmonary artery tubing was between 1 to 1.5 s, as assessed by the appearance of methylene blue at a CF rate of 20 ml/min. Nitric oxide release (pmol [middle dot] g-1[middle dot] min-1) was calculated as CF (ml/min) times NO concentration (nM) divided by heart wet weight (average 1.80 +/- 0.06 g [SEM]). The basal No concentration cannot be measured directly by this method, but we have shown that perfusion of hearts with 100 [micro sign]M N (G) nitro L-arginine methyl ester (L-NAME) decreases [NO] by about 12 +/- 3 nM, which suggests that this is the basal effluent [NO] in this preparation. [26] 

Anesthetics were delivered by calibrated agent-specific vaporizes, and anesthetic concentrations were measured in the effluent by gas chromatography, as described before. [27]Effective anesthetic fractions, in volume percentage (vol %), were determined as noted previously [27]and are shown in Table 1. The percentage of oxygen extracted was calculated as the inflow and outflow tension difference multiplied by 100 and divided by inflow oxygen tension. The percentage of O2extraction assesses a direct vasodilatory response separate from that caused by an autoregulatory response resulting from altered contractility; this measurement assumes that local metabolites are produced in proportion to myocardial oxygen consumption and that local metabolites are important factors in controlling CF autoregulation. Inflow perfusate oxygen tension was kept constant by maintaining the reservoir pressure 5 mmHg above atmospheric pressure. Heart rate, inflow, and outflow oxygen (mmHg), pH (in mV), CF, systolic and diastolic isovolumetric LV pressure, and NO electrode current (pA) were displayed continuously on a fast-writing (3 kHz), high-resolution, eight-channel chart recorder (Astro-Med Inc., West Warwick, RI). The NO (pA) and pH (mV) electrode signals were zero suppressed and amplified for continuous display.

Measurement of L-Citrulline

L-Citrulline was measured by high-performance liquid chromatography in 250 coronary effluent samples from 32 hearts before and during exposure to each anesthetic in the absence and presence of 0.1, 1, 10, and 100 nM bradykinin. The high-performance liquid chromatography system consisted of a Laboratory Data Control constametric III G pump (Rivera Beach, FL), a Gilson Automatic Sampler model 231 (Middletown, WI), and an electrochemical detector (BAS LC-4B; Bioanalytical Systems, West Lafayette, IN). The column (Beckman Ultrasphere ODS 5 [micro sign], 4.6 mm x 25 cm, Fullerton, CA) was perfused at a mobile phase flow rate of 1.5 ml/min. The detector potential was set at +0.7 V. The mobile phase consisted of 800 ml 0.1 M sodium acetate, adjusted to pH 5.7, plus 260 ml acetonitrile. L-Citrulline was detected electrochemically as the o-phthaldialdehyde derivative. The o-phthaldialdehyde reagent consisted of 25 ml 0.1 M borate buffer (pH 9.5), 50 [micro sign] 2-methyl 2-propanethiol, 2.5 ml methanol, and 135 mg o-phthaldialdehyde. All chemicals were high-performance liquid chromatography grade. The chromatographic data were collected on a Hewlett Packard 3393A integrator and stored on a Hewlett Packard 9122 disc drive (Wilmington, DE). Coronary venous effluent was collected (2 ml) during drug-free control periods and during the last 30 s of bradykinin infusion and was frozen at - 15 [degree sign]C. Samples were later prepared and analyzed as follows: To each 0.5-ml sample was added 25 [micro sign]l methyl-L-arginine (2 [micro sign]g/ml) as an internal standard. Three milliliters of ethanol was added to each sample, mixed, and centrifuged. The supernatant was transferred to a clean tube and evaporated to dryness under a stream of air at 40 [degree sign]C. The dried residue was redissolved in 2 ml mobile phase and 400 [micro sign]l was mixed with 40 [micro sign]l of the o-phthaldialdehyde reagent for exactly 2 min before injection of 100 [micro sign]l into the high-performance liquid chromatograph. The L-cit concentration was calculated from the standard curve of the respective peak height ratio versus concentration. The standard curve data were derived using perfusate that did not pass through the isolated heart. The L-cit standard curves were linear over the concentration range studied. The limit of detection was 1 ng/ml perfusate. The absolute retention time for L-cit was 10.3 min.


Crystalloid-perfused hearts were randomized to three treatments, halothane (HAL), isoflurane (ISO), or sevoflurane (SEV), at several concentrations. Fourteen hearts were exposed to 0.5% HAL or to 1% HAL by vaporizing HAL in random order into Krebs-Ringer's solution. Similarly, 14 other hearts were exposed to 1% or 2% ISO, and seven hearts were exposed only to 2% SEV because effects on NO did not appear to be concentration dependent. Each heart was infused with 0, 0.1, 1, and 10 nM bradykinin given in random order. After a period of stabilization, adenosine was bolus injected to measure maximal CF. A given bradykinin concentration was then infused for 1 min. This was followed 10 min later by a 15-min exposure to a given anesthetic concentration, while the same bradykinin concentration was again infused for 1 min; then the anesthetic was discontinued, bradykinin alone was again infused for 1 min, and this was followed by a 15-min drug-free control period. This protocol was repeated twice again in the same heart for the other two concentrations of bradykinin. Adenosine was again bolus injected at the end of each experiment at the same concentration into the aortic cannula to observe any change in maximal CF reserve. For all groups (Table 1), basal CF was 6.4 +/- 0.2 ml [middle dot] g-1[middle dot] min-1. Beginning and ending CF responses to adenosine were 16.2 +/- 0.9 and 12.7 +/- 1.3 ml [middle dot] g-1[middle dot] min-1, respectively, for all groups combined (P < 0.05). Measurements and L-cit samples were obtained during the peak steady state change in CF during exposure to each concentration of an anesthetic or bradykinin. The CF and NO current returned approximately to the pre-drug level after a 15-min washout of bradykinin, anesthetic, or both. Drug-free control periods were interspersed between treatment. The basal NO levels can be determined only indirectly, [26]so NO concentration and NO release were calculated and expressed as changes from the preceding control.

Seventeen additional hearts were perfused with oxygenated Krebs-Ringer's solution containing erythrocytes. One liter of freshly obtained and heparin-processed canine whole blood was washed in saline and centrifuged three times and the buffy coat discarded. The resulting packed erythrocyte suspension was filtered (40-[micro sign]m pore size) and diluted to a hematocrit of about 9% immediately before use. Ionized Na+, K+, Ca2+, and pH were within normal limits and not different in erythrocyte and erythrocyte-free Krebs-Ringer's perfusate. Using the same randomized protocol as described before, bradykinin-induced changes in CF and oxygen extraction were measured in the absence and presence of 2% and 4% isoflurane delivered by vaporizer.


All data were expressed as mean +/- SEM. Each heart served as its own control. Individual drug responses were compared with the preceding control by paired Student's t tests. The effects of anesthetic concentrations on CF, NO release, and NO and L-cit concentrations during increasing concentrations of bradykinin were compared by Tukey's comparison of means' tests after analysis of variance for repeated measures (Super Anova 1.11 [registered sign] software for Macintosh; Abacus Concepts, Berkeley, CA). The change in CF induced by increasing concentrations of bradykinin was plotted as a function of the change in NO concentration for each coordinate, and data were fitted by linear regression analysis to determine slopes, y intercepts, correlation coefficients, significance of slopes, and 95% confidence intervals. Differences in slopes in the absence (control) and presence of an anesthetic concentration were determined by slope heterogeneity comparison tests. Differences among means were considered significant when P <or= to 0.05.

(Table 1) displays the effects of each volatile anesthetic concentration alone on LV pressure, heart rate, CF, and percent oxygen extraction. Anesthetic concentration is expressed by the inflow vaporizer setting (vol %), coronary effluent concentration (mM), and effective vapor pressure (vol %). Each concentration of each anesthetic significantly decreased LV pressure and heart rate. The decrease in LV pressure by HAL, but not by ISO, was concentration dependent. Only ISO increased CF, and both HAL and ISO decreased oxygen extraction. The increase in CF by ISO was concentration dependent.

(Figure 1) shows a typical response to 2% ISO alone and bradykinin plus 2% ISO. Bradykinin increased CF, effluent oxygen tension, LV pressure, and [NO], and it decreased the RR wave internal and pH (mV). The time course of the NO response closely matched that of the CF response. Although ISO slightly increased CF, it had no appreciable effect on other responses to bradykinin.

(Figure 2A, Figure 2B, Figure 2C) summarizes the effect of bradykinin on increasing CF in the presence and absence of HAL, ISO, and SEV in crystalloid-perfused hearts. In the absence of bradykinin, ISO and SEV, but not HAL, alone significantly increased CF. Bradykinin alone significantly increased CF in each anesthetic group in a concentration-dependent manner. The presence of any of the three volatile anesthetics did not significantly alter the bradykinin-induced increase in CF, with the single exception of a significant increase in CF by 2% ISO with 0.1 nM bradykinin. Bradykinin had no significant effect on LV pressure or heart rate (data not shown).

(Figure 3A, Figure 3B, Figure 3C) summarizes the effect of increasing concentrations of bradykinin on increasing coronary effluent [NO] in the presence and absence of HAL, ISO, and SEV in crystalloid-perfused hearts. None of the anesthetics altered [NO] from the drug-free control. Bradykinin alone significantly increased [NO] in each anesthetic group in a concentration-dependent manner. The presence of any of the three anesthetics did not significantly alter the increase in [NO].

(Figure 4A, Figure 4B, Figure 4C) summarizes the effect of increasing concentrations of bradykinin on changing coronary effluent NO release in the presence and absence of HAL, ISO, and SEV in crystalloid-perfused hearts. None of the anesthetics altered NO release from the drug-free control. Bradykinin alone significantly increased NO release in each anesthetic group in a concentration-dependent manner. The presence of any of the three anesthetics did not significantly alter the increase in NO release, with the single exception of a significant increase in NO release by 2% ISO in the absence of bradykinin, an effect likely caused by the increase in CF by ISO. None of the anesthetics significantly altered the effect of bradykinin to increase effluent NO release.

(Figure 5A and Figure 5B, Figure 6A and Figure 6B, and Figure 7) show increases in CF as a linear correlate function of the increase in effluent [NO] induced by 0.1, 1, and 10 nM bradykinin in the absence and presence of lower (A panels) and higher (B panels) concentrations of HAL and ISO and of SEV (2% only) in crystalloid-perfused hearts. All slopes are significant, which indicates that the increases in [NO] were associated with the increases in CF. There were no differences in slopes among the three anesthetic controls or between a concentration of any anesthetic and its control. Only the y intercepts for 1% and 2% ISO were significantly different from controls.

(Figure 8) displays bradykinin-induced increases in [L-cit] in the presence and absence of each anesthetic in crystalloid-perfused hearts. Significant increases in [L-cit] were induced by 1 - 100 nM bradykinin. The responses to 0.1 to 10 nM bradykinin were not altered significantly by the given concentrations of each anesthetic.

(Figure 9A and Figure 9B) summarizes data in 12 erythrocyte-perfused hearts on the effect of two concentrations of bradykinin on altering CF and the percentage of oxygen extracted in the presence and absence of 2% and 4% isoflurane. Although ISO alone did not alter CF, it decreased the percentage of oxygen extracted significantly. Bradykinin (10 nM) increased CF and decreased the percentage of oxygen extracted significantly. Whereas ISO had no added effect on the increase in CF by 10 nM bradykinin, it accentuated bradykinin's effect to decrease the percentage of oxygen extracted. In five additional erythrocyte-perfused hearts, 4% isoflurane did not alter CF (4.8 +/- 0.6 to 4.9 +/- 0.5 ml [middle dot] g-1[middle dot] min-1) but significantly increased effluent oxygen tension (200 +/- 5 to 251 +/- 6 mmHg), whereas 100 [micro sign]M L-NAME slightly decreased basal CF (3.8 +/- 0.6 ml [middle dot] g-1[middle dot] min-1) and oxygen tension (158 +/- 8 mmHg); 4% isoflurane had no added effects (3.6 +/- 0.4 ml [middle dot] g-1[middle dot] min-1; 142 +/- 9 mmHg) when combined with L-NAME.

This is the first study to measure simultaneously bradykinin-induced CF and NO release in the presence of a volatile anesthetic in intact, crystalloid-perfused hearts. The study shows, first, a significant correlation between increases in coronary effluent [NO] and CF induced by the eNOS dependent drug bradykinin; and, second, that this relationship is not altered by volatile anesthetics. Furthermore, the anesthetics did not alter bradykinin-induced increases in effluent [L-cit]. The finding of no significant influence of anesthetics on bradykinin-induced CF, NO release, [NO], and [L-cit] suggests that they do not interfere significantly in the pathway between endothelial release of NO and coronary vasodilation. A small increase in [NO] accompanied a small CF increase by ISO that is likely a flow-dependent but ISO-independent effect. In the companion article, we reported that perfusion pressure-induced changes in CF were associated with proportional changes in [NO] and that these effects were attenuated by L-NAME. [26]The lack of a significant alteration in the CF to NO relationship at the anesthetic concentrations selected for the crystalloid-perfused heart studies does not preclude that higher concentrations may have an inhibitory effect on the NO signaling pathway between coronary endothelium and vasculature. However, the erythrocyte-perfused heart studies confirm that a higher ISO concentration also does not alter the bradykinin-induced increase in CF and decrease in oxygen extraction. In fact, ISO enhanced the bradykinin-induced decrease in oxygen extraction, suggesting that after normalizing for an auto-regulatory influence on CF, vascular tone is additively reduced by bradykinin and ISO. In erythrocyte-perfused hearts, basal CF was lower and bradykinin-induced increases in CF, but not decreases in oxygen extraction, were less than those in crystalloid-perfused hearts. Differences in oxygen-carrying capacity may account for these differences.

There are several reports that CF is locally regulated by NO in the guinea pig heart. [4,29,30]A direct measure of [NO] in hemoblobin-free coronary effluent along with CF allows an assessment of the amount and rate of vascular endothelial cell NO production by NO synthase (eNOS) and the coronary vasodilation produced by NO. Volatile anesthetics could alter specific endothelial receptor responses (e.g., B2-kininreceptor subtype), [31]modulate NO production or release, or alter the effects of NO on the sGC cGMP system that promotes vascular relaxation. Until now, most studies on NO synthesis, NO release, and NO effects have been based inferentially on antagonism of eNOS by L-arginine analogs such as NG-monomethylL-arginine or L-NAME. These antagonists may not completely inhibit eNOS. [26]Our conclusions are based on a lack of effect by anesthetics to alter the correlation between coronary effluent [NO] and CF rather than on antagonism by L-arginine analogs. Although unlikely, this could be an oversimplification. For example, if an anesthetic agent were to decrease the effect of a given amount of NO released on the sGC - cGMP pathway and thus to indirectly reduce endothelium-dependent relaxation, it might also proportionally enhance relaxation by another endothelium-dependent effect.

Several methods are reported to quantitate NO release from endothelial cells. Techniques to measure NO release directly or indirectly from tissue include methemoglobin, [4]bioassay, [32,33]chemoluminescence, [34]and electron paramagnetic response spectra. [35]Compared with these techniques, an NO-sensing electrode has the distinct advantage of measuring NO directly and continuously in real time. There are a few preliminary reports on the use of an amperometric NO sensor to measure NO release from cultured endothelial cells, [36,37]but there are no other known reports of its use in isolated organs. In our experiment, the cardiac effluent reached the NO electrode within 1 to 1.5 s. The half-life of NO in aqueous solution is 3.8 to 5.6 s, [4]so most NO released luminally from the coronary bed can be detected. In the absence of hemoglobin and other scavengers of NO in an aqueous solution, it is assumed that the amount of NO released abluminally to bind to sGC in vascular smooth muscle cells is equivalent, or at least proportional, to that released luminally. Anesthetic partition coefficients on either side of the endothelial cell would be equivalent. We reported recently that NO release induced by 100 nM bradykinin amounted to 146 +/- 19 pmol [middle dot] g-1[middle dot] min-1, which is about 225 pmol/min at a mean heart weight of 1.54 g. [26]Kelm and Schrader [4]quantitated NO release indirectly as the product of methemoglobin from the reaction of NO with oxyhemoglobin and reported that 100 nM of bradykinin increased NO release from isolated hearts to about 300 pmol/min.

We cannot detect effluent NO by our method in erythrocyte-perfused hearts, so it is likely that NO is largely scavenged by hemoglobin. It could be argued that because luminal [NO] and NO release is greater in a crystalloid perfusate devoid of hemoglobin to scavenge NO, any inhibitory effect of an anesthetic on the NO pathway might be overwhelmed by excess NO. However, ISO, at a higher concentration than used in the crystalloid perfusion studies, also did not attenuate bradykinin-induced changes in erythrocyte-perfused hearts when luminal [NO] was probably very low. Crystalloid perfusion studies are helpful to determine if anesthetics alter the NO pathway because measured effluent (luminal)[NO] is more likely to be proportional to abluminal (interstitial)[NO], where an anesthetic might be expected to alter NO signaling between abluminal release of NO and NO activation of vascular smooth muscle sGC and cGMP. The inhibitory effect of anesthetics on the NO vasodilator pathway noted by some investigators may be a result of an effect on lowering basal endothelial Ca2+for Ca2+-dependenteNOS stimulation. The current investigation supports studies that found that lower levels of anesthetics have no effect on bradykinin-stimulated eNOS-dependent vasodilation.

As summarized in recent reviews, [22,23]volatile anesthetics have been reported to modify endothelium-dependent modulation of smooth muscle tone in vitro. This is important because it would indicate that at least some part of the vasoactive effects of anesthetics could be attributed to altered endothelial control of vascular tone. But most studies suggest that anesthetics neither stimulate release of endothelium-dependent relaxing factors nor have any, or little, effect on endothelium-independent vasodilation induced by nitro vasodilators. [9,10]However, anesthetics may inhibit vascular relaxation secondary to stimulated release of relaxing factors by endothelial muscarinic receptors, bradykinin receptors, thrombin, and Ca2+inophores. [22,23] 

Several indirect studies [12–16,25,38]suggest that anesthetics modify responses to endothelium-dependent vasodilators. HAL may inhibit vasodilation induced by NO and nitroglycerin, [19]but HAL did not inhibit measured sGC, [39]and Johns et al. [13]reported that >2% HAL or 1% ISO depressed cGMP formation independent of sGC activation when endothelial and vascular smooth muscle cell co-cultures were treated with bradykinin. Chelly et al., [38]using L-NAME-treated intact dogs, reported that 1.2% HAL may interfere with NOS activity in carotid, mesenteric, and renal vascular beds, but not in the coronary vascular bed. Sigmon et al. [25]also suggested that HAL impairs NO-mediated regulation of some vascular beds, but not the coronary vascular bed, because it largely eliminated systemic responses of eNOS inhibition by L-NAME in rats.

Most investigations suggest that anesthetic-induced vasodilator effects are independent of vascular smooth muscle sGC activation, but anesthetics could attenuate endothelium-dependent relaxation by inhibiting the production, transport, or release of relaxing factors, primarily NO. They could alter activation of endothelial receptor sites, modify second messengers, or affect intracellular Ca2+flux in the cascade of events leading to stimulation of eNOS. [22]In one study it was suggested that a receptor mechanism is not the primary site of anesthetic inhibition of NO signaling, because anesthetics attenuated non-receptor-mediated NO production, an effect that bypasses the receptor. [12]Much as anesthetic effects on Ca2+have been examined in cardiac and smooth muscle cells, it is possible that anesthetics decrease Ca2+availability for NOS activation. [40]Volatile anesthetics might act on endothelial cells by altering sarcolemmal Ca2+entry, by altering release of Ca2+or uptake of Ca2+from intracellular stores, or by inhibiting phospholipase C and IP3-mediatedCa2+release. [18]Also within the endothelium, it is possible that anesthetics interact with or inhibit other cofactors involved with activation of eNOS.

Volatile anesthetics might also inactivate NO or reduce NO efficacy once released from eNOS. It has been reported that HAL attenuates NO-mediated relaxation, but that it does not interfere with endothelial cell release of NO, based on a bioassay technique, or with relaxation induced by nitroprusside. [16]From this finding it was hypothesized that HAL might have interfered with NO stability once it was released from the endothelium. If anesthetics cause production of free radicals by reductive metabolic pathways (e.g., during ischemia), this could enhance inactivation of NO. [41]Nakamura et al. [15]reported that 4% SEV reduces relaxation of intact rat aortas by acetylcholine and suggested that SEV inactivates NO or the vasodilator effect of NO. Yoshida et al. [14]reported that SEV impairs isolated canine mesenteric artery relaxation induced by a high concentration of bradykinin and suggested that SEV functions as a free radical to interact with NO, also a free radical, because superoxide dismutase attenuated the decrease in bradykinin-induced relaxation caused by SEV. However, it is doubtful that any volatile anesthetic is metabolized by a free radical-producing reductive pathway in the absence of ischemia. With ischemia, their effect on NO-dependent control of vascular tone could be very different.

Just as there are many reports that anesthetics interfere in endothelial vascular muscle signaling, there are others that anesthetics do not alter NO-induced vasodilation. [9,10,13,21]Flynn et al. [9]showed that ISO relaxes isolated cerebral arteries rings independently of the endothelium. Brendel and Johns [10]reported that ISO-induced relaxation of isolated aortic rings is independent of endothelium-derived relaxing factor, or NO, and cGMP-mediated mechanisms. Crystal et al. [21]reported that coronary vasodilation induced by 1.4% ISO in open-chest dogs was not affected by L-NAME and suggested that NO does not mediate or modify this response. Johns et al. [13]furnished data that volatile anesthetics do not stimulate or inhibit cGMP production in endothelial-vascular co-cultures and suggested that they do not activate eNOS or GC. The present results concur with the conclusions made from these later studies that anesthetics have no significant effect on NO-mediated relaxation. Differences in methods, organ specificity, or type of preparation may account for different results and conclusions in various studies.

A possible limitation of our study is that only bradykinin was given to induce NO production by eNOS. Other drugs, that stimulate NO release, such as acetylcholine and serotonin, cause direct vasoconstriction at high doses or in the presence of L-arginine antagonism, and so they were not used to test our hypothesis. Bradykinin primarily stimulates endothelial cell Ca2+influx to trigger NO production, but it has been reported to cause vasodilation by several mechanisms. [17,31,41,42]Bradykinin may induce prostacyclin production [43]and stimulate endothelium-dependent hyperpolarizing factor [44]to produce vasodilation. This factor may endogenously activate K+channel opening. [45]In the related study, [26]we found that indomethacin, an inhibitor of prostacyclin production, and tetraethylammonium, a nonspecific K+channel blocker, did not alter a bradykinin-induced increase in CF or release of NO but that adding 100 [micro sign]m L-NAME with indomethacin and tetraethylammonium inhibited increases in effluent CF and NO induced by 1 nM bradykinin. Although L-NAME completely blocked release of NO to 10 nM bradykinin, it did not completely block the increase of CF. In the presence of all three inhibitors of vasodilation, a bradykinin-induced increase in CF could not be completely abolished. Figure 8indicates that bradykinin concentrations up to 100 nM incrementally increase [L-cit]. The related study [26]suggested the major mechanism of vasodilation by up to 10 nM bradykinin depends on NO release but that higher bradykinin concentrations exert an additional mechanism of vasodilation, perhaps a direct effect on vascular smooth muscle.

In conclusion, this study shows that coronary effluent release of NO and L-cit by bradykinin is associated with an increase in CF, and the given concentrations of HAL, ISO, and SEV do not alter NO production, NO release, or NO-mediated vasodilator effects in intact guinea pig hearts. We question whether the lower levels of volatile anesthetics exert a physiologically significant effect to reduce coronary vascular resistance via the NO pathway.

The authors thank Kimberly A. Stommel, B.S., for conducting the high-performance liquid chromatography experiments and Enis Novalija, M.D., James S. Heisner, B.S., and Richard L. Paulsen, B.S., for their assistance with this project.

Furchgott RF, Zawadzki JV: The obligatory role of endothelial cells in the relaxation of arterial smooth muscle by acetylcholine. Nature 1980; 288:373-6
Palmer RM, Ferrige AG, Moncada S: Nitric oxide release accounts for the biological activity of endothelium-derived relaxing factor. Nature 1987; 327:524-6
Ignarro LJ, Byrns RE, Buga GM, Wood KS: Endothelium-derived relaxing factor from pulmonary artery and vein possesses pharmacologic and chemical properties identical to those of nitric oxide radical. Circ Res 1987; 61:866-79
Kelm M, Schrader J: Control of coronary vascular tone by nitric oxide. Circ Res 1990; 66:1561-75
Knowles RG, Palacios M, Palmer RM, Moncada S: Formation of nitric oxide from L-arginine in the central nervous system: A transduction mechanism for stimulation of the soluble guanylate cyclase. Proc Nat Acad Sci U S A 1989; 86:5159-62
Rapoport RM, Murad F: Agonist-induced endothelium-dependent relaxation in rat thoracic aorta may be mediated through cGMP. Circ Res 1983; 52:352-7
Bassenge E, Busse R: Endothelial modulation of coronary tone. Prog Cardiovas Dis 1988; 30:349-80
Stone DJ, Johns RA: Endothelium-dependent effects of halothane, enflurane, and isoflurane on isolated rat aortic vascular rings. Anesthesiology 1989; 71:126-32
Flynn NM, Buljubasic N, Bosnjak ZJ, Kampine JP: Isoflurane produces endothelium-independent relaxation in canine middle cerebral arteries. Anesthesiology 1992; 76:461-7
Brendel JK, Johns RA: Isoflurane does not vasodilate rat thoracic aortic rings by endothelium-derived relaxing factor or other cyclic GMP-mediated mechanisms. Anesthesiology 1992; 77:126-31
Greenblatt EP, Loeb AL, Longnecker DE: Marked regional heterogeneity in the magnitude of EDRF/NO-mediated vascular tone in awake rats. J Cardiovas Pharmacol 1993; 21:235-40
Uggeri MJ, Proctor GJ, Johns RA: Halothane, enflurane, and isoflurane attenuate both receptor and non-receptor-mediated EDRF production in rat thoracic aorta. Anesthesiology 1992; 76:1012-7
Johns RA, Tichotsky A, Muro M, Spaeth JP, Le Cras TD, Rengasamy A: Halothane and isoflurane inhibit endothelium-derived relaxing factor-dependent cyclic guanosine monophosphate accumulation in endothelial cell-vascular smooth muscle co-cultures independent of an effect on guanylyl cyclase activation. Anesthesiology 1995; 83:823-34
Yoshida KI, Okabe E: Selective impairment of endothelium-dependent relaxation by sevoflurane: Oxygen free radicals participation. Anesthesiology 1992; 76:440-7
Nakamura K, Terasako K, Toda H, Miyawaki I, Kakuyama M, Nishiwada M, Hatano Y, Mori K: Mechanisms of inhibition of endothelium-dependent relaxation by halothane, isoflurane, and sevoflurane. Can J Anaesth 1994; 41:340-6
Blaise G, To Q, Parent M, Lagarde B, Asenjo F, Sauve R: Does halothane interfere with the release, action or stability of endothelium-derived relaxing factor/nitric oxide? Anesthesiology 1994; 80:417-26
Loeb AL, O'Brien DK, Longnecker DE: Halothane inhibits bradykinin-stimulated prostacyclin production in endothelial cells. Anesthesiology 1994; 81:931-8
Sill JC, Nelson OR, Uhl C: Isoflurane-, halothane- and agonist-evoked responses in pig coronary arteries and vascular smooth muscle cells, Mechanisms of Anesthetic Action in Skeletal, Cardiac, and Smooth Muscle. Edited by TJJ Blanck, DF Wheeler. New York, Plenum, 1991, 257-69
Hart JL, Jing M, Bina S, Freas W, Van Dyle RA, Muldoon SM: Effects of halothane on EDRF/cGMP-mediated vascular smooth muscle relaxations. Anesthesiology 1993; 79:323-31
Greenblatt EP, Loeb AL, Longnecker DE: Endothelium-dependent circulatory control - A mechanism for the differing peripheral vascular effects of isoflurane versus halothane. Anesthesiology 1992; 77:1178-85
Crystal GJ, Kim SJ, Salem MR, Khoury E, Gurevicius J: Nitric oxide does not mediate coronary vasodilation by isoflurane. Anesthesiology 1994; 81:209-20
Johns RA: Endothelium, anesthetics, and vascular control. Anesthesiology 1993; 79:1381-91
Stowe DF, Ebert TE: Neural and endothelial control of the peripheral circulation: Implications for anesthesia: Part II. Endothelium mediated effects in the normal and diseased circulation. J Cardiovasc Anesthesiol 1996; 10:159-71
Wang YX, Zhou T, Chua TC, Pang CC: Effects of inhalation and intravenous anesthetic agents on pressor response to NG-nitro-L-arginine. Eur J Pharmacol 1991; 198:183-8
Sigmon DH, Florentino-Pineda I, Van Dyke RA, Beierwaltes WH: Halothane impairs the hemodynamic influence of endothelium-derived nitric oxide. Anesthesiology 1995; 82:135-43
Fujita S, Roerig DL, Bosnjak ZJ, Stowe DF: Effects of vasodilators and perfusion pressure on coronary flow and simultaneous release of nitric oxide from guinea pig isolated hearts. Cardiovasc Res 1998; 38:655-67
Stowe DF, Monroe SM, Marijic J, Bosnjak ZJ, Kampine JP: Comparison of halothane, enflurane, and isoflurane with nitrous oxide on contractility and oxygen supply and demand in isolated hearts. Anesthesiology 1991; 75:1062-74
Shibuki K, Okada D: Endogenous nitric oxide release required for long-term synaptic depression in the cerebellum. Nature 1991; 349:326-8
Ueeda M, Silvia SK, Olsson RA: Nitric oxide modulates coronary autoregulation in the guinea pig. Cir Res 1992; 70:1296-303
Brown IP, Thompson CI, Belloni FL: Role of nitric oxide in hypoxic coronary vasodilatation in isolated perfused guinea pig heart. Am J Physiol 1993; 264:H821-9
Vanhoutte P, Boulanger C, Illiano S, Nagao T, Vidal M, Mombouli J: Endothelium-dependent effects of converting-enzyme inhibitors. J Cardiovasc Pharm 1993; 22(suppl 5):S10-6
Rockett KA AM, Cowden WB, Clark IA: Killing of plasmodium falciparum in vitro by nitric oxide derivatives. Infect Immun 1991; 59:3280-3
Green LC WD, Glogowski J, Skipper PL, Wishnok JS, Tannernbaum SR: Analysis of nitrate, nitrite, and [(15) N] nitrate in biological fluids. Analytical Biochem 1982; 126:131-8
Kikuchi K, Hayakawa H, Nagano T, Hirata Y, Sugimoto T, Hirobe M: New method of detecting nitric oxide production. Chem Pharm Bull 1992; 40:2233-5
Westenberger U, Thanner S, Ruf HH, Gersonde K, Sutter G, Trentz O: Formation of free radicals and nitric oxide derivative of hemoglobin in rats during shock syndrome. Free Rad Res Comm 1990; 11:167-78
Malinski T, Taha Z: Nitric oxide release from a single cell measured in situ by a porphyrinic-based microsensor. Nature 1992; 358:676-8
Tsukahara H, Gordienko DV, Goligorsky MS: Continuous monitoring of nitric oxide release from human umbilical vein endothelial cells. Biochem Biophysic Res Comm 1993; 193:722-9
Chelly EJ, Doursout MF, Lechevalier T, Liang YY, Chelly F, Hartley C, Kilbourn RG: Cardiac and regional hemodynamic interactions between halothane and nitric oxide synthase activity in dogs. Anesthesiology 1996; 85:142-9
Eskinder H, Hillard CJ, Flynn N, Bosnjak ZB, Kampine JP: Role of guanylate cyclase-cGMP systems in halothane-induced vasodilation in canine cerebral arteries. Anesthesiology 1992; 77:482-7
Klip A, Britt BA, Elliott ME, Walker D, Ramial T, Pegg W: Changes in cytoplasmic free calcium caused by halothane: Role of the plasma membrane and intracellular Ca2+stores. Biochem Cell Biol 1986; 64:1181-9
Rubanyi GM, Vanhoutte PM: Superoxide anions and hyperoxia inactivate endothelium-derived relaxing factor. Am J Physiol 1986; 250:H822-7
Cowan CL, Cohen RA: Different mechanisms of relaxation of pig coronary artery to bradykinin and cromakalim are distinguished by potassium channel blockers. J Pharmacol Exp Therap 1992; 260:248-53
Lamontagne D, Koenig A, Bassenge E, Busse R: Prostacyclin and nitric oxide contribute to the vasodilator action of acetylcholine and bradykinin in the intact rabbit coronary bed. J Cardiovasc Pharmacol 1992; 20:652-7
Nakashima M, Mombouli JV, Taylor AA, Vanhoutte PM: Endothelium-dependent hyperpolarization caused by bradykinin in human coronary arteries. J Clin Invest 1993; 92:2867-71
Suzuki H, Chen G: Endothelium-derived hyperpolarizing factor (EDHF). An endogenous potassium-channel activator. Trends Pharmacol Sci 1990; 5:212-5