Role of Endothelium in the Action of Isoflurane on Vascular Smooth Muscle of Isolated Mesenteric Resistance Arteries

With approval from the Kyushu University Animal Care and Use Committee (Fukuoka, Japan), as previously detailed, 17,18either endothelium-intact (+E) or endothelium-denuded (−E) strips (250–400 μm long and 150–200 μm wide) were prepared from the third- or fourth-order branches (approximately 150–200 μm in diameter) of Sprague-Dawley rat (male, 250–320 g) mesenteric arteries, which are known to substantially contribute to systemic vascular resistance. 19,20 

Isometric force was measured by attaching the strip to a strain gauge transducer (UL-2 type; Shinko, Tokyo, Japan) as previously detailed. 10,18In brief, the strip was horizontally mounted in a chamber attached to the stage of an inverted microscope, and the resting tension was adjusted to obtain a maximal response to KCl. The solution was changed by infusing it rapidly into one end while aspirating simultaneously from the other end. All experiments were carried out in 4-(2-hydroxyethyl) piperazine-1-ethanesulfonic acid (HEPES)-buffered physiologic salt solution containing guanethidine (3 μm) to prevent norepinephrine outflow from the sympathetic nerve terminals. Endothelial intactness was verified by the ability of acetylcholine (ACh, 1 μm) to cause complete (≥ 90%) relaxation during contractions induced by norepinephrine (10 μm). Conversely, functional removal of endothelium was verified by the inability of ACh (10 μm) to cause significant (≥ 10%) relaxation during contractions induced by norepinephrine (10 μm). All experiments were carried out at 35°C to prevent early deterioration of the strips.

The ionic concentrations of the HEPES-buffered physiologic salt solution were as follows: NaCl, 138 mm; KCl, 5.0 mm; MgCl², 1.2 mm; CaCl², 1.5 mm; HEPES, 10 mm; and glucose, 10 mm. The pH was adjusted with NaOH to 7.35 at 35°C. The high K⁺(40-mm) solutions were prepared by replacing NaCl with KCl iso-osmotically.

The following materials were purchased: norepinephrine, ACh, N  ^G-nitro l-arginine (LNNA), indomethacin, nordihydroguaiaretic acid (NDGA), superoxide dismutase (SOD), and U46619 (9,11-dideoxy-11α, 9 α-epoxymethanoprostagladin F2α) (Sigma Chemical, St. Louis, MO). HEPES and tetraethylammonium (TEA) (Nacalai Tesque, Kyoto, Japan); ketanserin tartrate (Research Biochemicals International, Natick, MA); and isoflurane (Dainihon Pharmaceutical, Osaka, Japan). Losartan, BQ-123 (Cyclo[D-α-aspartyl-L-propyl-D-valyl-L-leucyl-D-tryptophyl]), and BQ-788 (N-[N-[N-[2,6-dimethyl-1-piperidinyl]carbonyl]-4- methyl-L-leucyl]-1-(methoxycarbonyl)-D-tryptophyl]-D-nor-leucine monosodium) were obtained from Banyu Pharmaceutical, Tokyo, Japan. All other reagents were of the highest grade commercially available.

We first examined the effects of isoflurane (1–5%) on contractile responses to either norepinephrine or KCl in both the +E and −E strips. In our preliminary experiments with the +E strips, it was difficult to obtain constant response to lower concentrations (< 1 μm) of norepinephrine for a long period, but it was easy to obtain constant response to 10 μm (maximum) norepinephrine for more than 4 h. Therefore, in the +E strips, we studied the anesthetic effects on the maximal response to 10 μm norepinephrine. In our previous experiments with this artery, both inhibition of the endothelial vasodilator pathways (i.e.,  nitric oxide [NO] and endothelium-derived hyperpolarizing factor [EDHF] pathways) and endothelial denudation similarly shifted the concentration-response curve for norepinephrine contraction to the left, decreasing the 50% effective concentration value from approximately 1.5 μm to approximately 0.5 μm. 10The mean amplitude of the 10-μm norepinephrine contraction in the +E strips was almost equal to that of the 0.5-μm norepinephrine contraction in the unrubbed strips treated with inhibitors of the NO and EDHF pathways. 10Therefore, we examined the anesthetic effects on the submaximal response to 0.5 μm norepinephrine in the −E strips, which was also constant for the long period (greater than 4 h). We used 40 mm KCl to activate voltage-gated Ca²⁺channels and to eliminate the influence of EDHF. The responses to 40 mm KCl were also constant for the long period (greater than 4 h) in both the +E and −E strips.

Three minutes was sufficient for the responses to either 10 μm norepinephrine or KCl to reach a plateau, and reproducible responses to them were obtained with 7-min intervals. Thus, in experiments with 10 μm norepinephrine or KCl, either norepinephrine or KCl was applied for 3 min at 7-min intervals. However, in experiments with a low concentration (0.5 μm, 45% effective concentration) of norepinephrine in the −E strips, it took approximately 5 min for the norepinephrine response to reach a plateau after exposure to isoflurane, because of strongly inhibited development of force. In addition, with this longer application time, reproducible responses were more easily obtained with a longer interval (17 min) than with the short interval (7 min). Therefore, 0.5 μm norepinephrine was applied for 5 min at 17-min intervals in the −E strips, as done previously. 10After the response to either norepinephrine or KCl became constant with these protocols, isoflurane was applied for 5 or 15 min before and during subsequent applications of either stimulant, until a steady-state effect was obtained. We also investigated the time course for the recovery of vascular smooth muscle cells from the effects of isoflurane after washout of isoflurane from the chamber.

In the above experiments, we found that isoflurane enhances the norepinephrine response in an endothelium-dependent manner. Since both the NO and EDHF signaling pathways appear to be activated in the contractile response to norepinephrine in this artery, 10the enhanced contractile response to norepinephrine might be due to inhibition of the NO or EDHF pathway or both. In addition, although isoflurane did not increase the basal force level, it is theoretically possible that isoflurane enhances the contractile response to norepinephrine by stimulating the release of an endothelium-derived vasoconstricting factor. 10It has been suggested that cyclooxygenase products (e.g. , thromboxane A², prostaglandin F2α), lipoxygenase products, endothelin-1 (ET), angiotensin-II (AT-II), serotonin (5-HT), histamine, adenosine triphosphate, adenosine diphosphate, and superoxide anions all act as endothelium-derived vasoconstricting factors. 21We thus decided to examine the involvement of these endothelial substances in the vasoconstricting action of isoflurane. In the −E strips, U46619 (thromboxane–prostaglandin endoperoxide analogue), ET-1, AT-II, and 5-HT—but not histamine, adenosine 5′-triphosphate, and adenosine 5′-diphosphate—caused significant contractions in the presence or absence of norepinephrine (0.5 μm, approximately 45% effective concentration). 10We therefore evaluated the effects of inhibitors of the NO, EDHF, cyclooxygenase, and lipoxygenase pathways; blockers of ET-1, AT-II, and 5-HT receptors; and SOD on the vasoconstricting action, although we did not examine the involvement of histamine, adenosine 5′-triphosphate, or adenosine 5′-diphosphate in the vasoconstricting action.

To examine the involvement of NO, EDHF, cyclooxygenase, and lipoxygenase products in the vasoconstricting action, we evaluated the effects of a cocktail application of 100 μm LNNA (NO synthesis inhibitor), 10 mm TEA (nonselective K⁺channel blocker known to inhibit the endothelium-mediated hyperpolarization 22), 10 μm indomethacin (cyclooxygenase inhibitor 21), and 1 μm NDGA (lipoxygenase inhibitor 21) on the enhanced response to norepinephrine by 3% isoflurane. The concentrations of these inhibitors and the preincubation times were decided on the basis of the results of our recent experiments in this artery 10and others reported in the literature. 21,23In other words, the strips were pretreated with these inhibitors for a time considered sufficient for all of them to exert their maximal effects (60 min), and the inhibitors were then applied throughout the remainder of the experiment. Because these inhibitors significantly enhanced the norepinephrine (10 μm) response, the norepinephrine concentration was decreased to produce approximately the same level of tension as the control (10 μm) norepinephrine–induced contraction, as previously done. 10,17Such low concentrations of norepinephrine were detected in each strip (see Results).

To evaluate the involvement of ET-1, AT-II, and 5-HT in the vasoconstricting action, we examined the effects of a cocktail application of 1 μm BQ-123 (selective ET^Areceptor antagonist), 1 μm BQ-788 (selective ET^Breceptor antagonist), 1 μm losartan (selective AT-II type 1 receptor antagonist), and 0.03 μm ketanserin (5-HT²/5-HT^1Creceptor antagonist) on the enhanced response to norepinephrine by 3% isoflurane. As we recently reported with regard to this artery, 10the concentrations of BQ-123, BQ-788, losartan, and ketanserin and the preincubation time (25 min) were determined by their ability to eliminate the maximal contractions induced by ET-1 (10 nm), AT-II (3 μm), or 5-HT (3 μm) in the −E strips.

To evaluate the involvement of superoxide anions in the vasoconstricting action, we examined the effects of SOD (15–150 U/ml), a scavenger of superoxide anions, on the enhanced response to norepinephrine by 3% isoflurane. The concentrations of SOD and its preincubation time (25 min) were decided on the basis of findings reported in the literature, 24,25in which the ability of SOD to inhibit endothelium-dependent vasoconstriction was demonstrated in isolated arteries.

We also evaluated the ability of isoflurane to inhibit the NO and EDHF pathways by testing its effects on both the NO- and EDHF-mediated vasorelaxations caused by ACh. In this artery, the relaxation caused by low concentrations (≤ 0.1 μm) of ACh is exclusively mediated by NO, whereas the relaxation caused by higher concentrations (≥ 0.3 μm) of ACh is mediated by both NO and EDHF. 10Thus, in this experiment, we evaluated (1) the effects of isoflurane on the NO-mediated (i.e. , LNNA-sensitive) relaxation caused by 0.03 μm ACh in the +E strips precontracted with norepinephrine and (2) its effects on the EDHF-mediated (i.e. , LNNA-resistant, TEA-sensitive) relaxation caused by 1 μm ACh in the LNNA-treated, unrubbed strips precontracted with norepinephrine.

Isoflurane was delivered via  a calibrated isoflurane vaporizer (Forawick; Muraco Medical, Tokyo, Japan) in line with the air gas aerating the HEPES-buffered solutions. Each solution was equilibrated with isoflurane for at least 15 min before introduction to the chamber, which was covered with a thin glass plate to prevent the equilibration gas from escaping into the atmosphere. Using gas chromatography, we previously determined the concentrations of isoflurane in the physiologic salt solution produced by 0.5, 1.0, 1.5, and 3.0% isoflurane under exactly the same experimental conditions, 18,26and the obtained values were within 92% (91.8–98.5%) of theoretical values predicted by the partition coefficient of isoflurane in Krebs solution (0.55) at 37°C. An excellent linear relation was obtained between the aqueous concentrations of isoflurane (y) and its concentrations (vol%) in the gas mixture (x): y =−0.0068 + 0.21x; r [regression coefficient]= 0.998). 26Therefore, the concentrations produced by 1, 2, 3, and 5% isoflurane in the physiologic salt solution can be predicted as 0.21, 0.42, 0.63, and 1.05 mm, respectively. The aqueous concentration produced by 3% isoflurane (i.e. , 0.63 mm) is almost equal to a recently reported concentration of isoflurane (0.65 mm) 9in blood sampled from this rat under steady-state anesthesia with 1.5% (1 MAC in this rat 27) isoflurane.

The effects of isoflurane on norepinephrine contraction, KCl contraction, or ACh relaxation were assessed at the points at which its effects reached a maximum. Changes in force were expressed as percentage of the reference. Because the relation between actual concentrations of isoflurane in the solutions and anesthetic concentrations (vol%) in the gas mixture is theoretically linear, the anesthetic concentrations on the x-axis are displayed as vol% for the isoflurane concentration–response relations, as done previously. 10,18 

All the results are expressed as mean ± SD. In each series of experiments, the effect of each concentration of isoflurane was examined once in a single strip prepared from one animal. However, it was not always possible to test the effects of all four concentrations of isoflurane in each strip because the high concentrations (≥ 3%) of isoflurane produced prolonged inhibition of vascular reactivity, as shown in Results. Accordingly, although the number of preparations is equal to the number of animals at each data point, the number of total animals required for each series of experiments is not necessarily equal to the number of preparations for each data point. The number of preparations for each data point (n) and that of total animals required for each series of experiments (N) are both described in Results.

The data were statistically assessed by analysis of variance, Scheffé F test, and Student t  test, as appropriate. Comparisons of the time-dependent effects of isoflurane among concentrations and those of the effects of isoflurane between treatments (short application vs.  long application) were made by two-factor (time, concentration; concentration, treatment) analysis of variance for repeated measures or factorial analysis. When overall differences were detected, individual comparisons among or between groups at each concentration were performed by means of the Scheffé F test. Comparisons of the effects of isoflurane within each concentration group or each treatment group were made by one-factor (time or concentration) analysis of variance for repeated measures, and post hoc comparisons were made with use of the Scheffé F test for multiple comparisons. Similarly, the effects of SOD on the vasoconstricting action of isoflurane were assessed by one-factor (concentration) analysis of variance for repeated measures, followed by the Scheffé F test. The effects of pharmacologic inhibitors on norepinephrine response and the effects of isoflurane (3%) on norepinephrine response after treatment with the pharmacologic inhibitors were assessed by means of a two-tailed, paired Student t  test. In the Discussion, comparisons between the current data on isoflurane and our previous data on sevoflurane 10at each data point were made with a two-tailed, unpaired Student t  test. A level of P < 0.05 was considered significant. All the statistical analyses were performed on a computer with use of Statview J 4.02^®software (Abacus Concepts, Berkeley, CA).

In the +E strips, isoflurane (≥ 2%), applied for 5 min before and during subsequent applications of norepinephrine (10 μm), initially (i.e. , 5 min after application of isoflurane) enhanced the contractile response to norepinephrine (fig. 1). However, the enhancement was subsequently attenuated during application of isoflurane, reaching a steady state where the magnitude of norepinephrine response was not significantly different from that before application of isoflurane (fig. 1). However, at this steady state, the rate of rise in force in response to norepinephrine was consistently inhibited by isoflurane (fig. 1Ab). In addition, rhythmic oscillations observed in the contractile response to norepinephrine were consistently inhibited during application of isoflurane (fig. 1Ab). The contractile response to norepinephrine was strongly inhibited after washout of isoflurane (≥ 2%) from the chamber (fig. 1). This inhibition was prolonged for 5–75 min, depending on the preparations (fig. 1); the mean recovery time for the 5% isoflurane-induced postwashout inhibition of norepinephrine response was 40.0 ± 21.4 min (n = 8; N = 13). Comparison of the time-dependent effects of isoflurane between two different protocols regarding time for isoflurane application (i.e. , short vs.  long application protocols) indicated that the postwashout inhibition was indeed triggered by the washout of isoflurane but was not due to emergence of late-onset inhibition (fig. 1Bb).

In the −E strips, isoflurane (≥ 3%), applied for 15 min before and during subsequent applications of 0.5 μm (approximately 45% effective concentration) norepinephrine, inhibited the contractile response to norepinephrine (fig. 2). It took more than 30 min for the inhibitory effect of isoflurane to reach maximum and steady state (fig. 2). The inhibition was prolonged after washout of isoflurane from the chamber, and it took more than 15 min for vascular smooth muscle cells to completely recover from the inhibition (fig. 2). Identical results were obtained in other experiments in which isoflurane (1–5%) was applied for 15 min before and during subsequent applications of 10 μm (maximum) norepinephrine (fig. 2).

Figure 2Cshows comparison of the effect of isoflurane on contractile response to norepinephrine in the presence or absence of endothelium in two different protocols with regard to time for pretreatment with isoflurane (i.e. , 5 vs.  15 min). As shown in figure 1, when isoflurane was applied to the +E strips for 5 min before and during a subsequent application of norepinephrine, the norepinephrine response was significantly enhanced during application of isoflurane (fig. 2C). However, as shown in figure 2C, when isoflurane was applied to the +E strips for the longer time (15 min) before and during a subsequent application of norepinephrine, the norepinephrine response was not significantly enhanced during application of isoflurane. In contrast, when isoflurane was applied to the −E strips for 5 min before and during a subsequent application of norepinephrine (0.5 and 10 μm), the norepinephrine response was not significantly inhibited during application of isoflurane (fig. 2C). However, when isoflurane was applied to the −E strips for the longer time (15 min) before and during a subsequent application of norepinephrine (0.5 and 10 μm), the norepinephrine response was significantly inhibited during application of isoflurane (fig. 2C).

In the +E strips, lower concentrations (≤ 3%) of isoflurane, applied for 5 min before and during subsequent applications of KCl (40 mm), did not significantly influence the contractile response to KCl (fig. 3). However, as shown in figure 3A, a low concentration (1%) of isoflurane enhanced the contractile response to KCl in some of the +E strips, although its mean effect did not reach statistical significance. Only higher concentrations (≥ 3%) of isoflurane significantly inhibited the KCl response (fig. 3).

In the −E strips, isoflurane (≥ 2%), applied for 5 min before and during subsequent applications of KCl (40 mm), significantly inhibited the contractile response to KCl in a concentration-dependent manner (fig. 3).

In contrast to the difference between the +E and −E strips in time dependence of the effect of isoflurane on norepinephrine response, the time dependence of the inhibitory effect of isoflurane on KCl response observed in the +E strips was similar to that observed in the −E strips (fig. 3). Namely, the onset of isoflurane-induced inhibition of KCl response was immediate in both the +E and −E strips; 5 min was sufficient for its inhibitory effect on KCl response to reach maximum and steady state, regardless of the presence of endothelium (fig. 3). In addition, no significant inhibition of the KCl response was observed after washout of isoflurane in either +E or −E strips (fig. 3).

The treatment with LNNA + TEA + NDGA + indomethacin enhanced the response to 10 μm norepinephrine (186 ± 35.2% of the control;P < 0.05; n = 5; N = 5). In the unrubbed strips treated with these inhibitors, isoflurane (3%) still enhanced the contractile response to a lower concentration of norepinephrine (3.4 ± 3.0 μm), the amplitude of which was not different from that of the control response to 10 μm norepinephrine (P > 0.05; n = 5; N = 5) (fig. 4A).

The treatment with losartan + ketanserin + BQ123 + BQ788 did not influence the response to 10 μm norepinephrine (83.9 ± 15.7% of the control;P > 0.05; n = 5; N = 5). Isoflurane (3%) still enhanced the response to 10 μm norepinephrine in the unrubbed strips after treatment with these receptor blockers (fig. 4A).

The treatment with SOD did not influence the response to 10 μm norepinephrine (103.4 ± 5.7, 107.8 ± 13.8, and 100.0 ± 29.8% of the control after treatment with 15, 50, and 150 U/ml SOD, respectively;P > 0.05; n = 5; N = 5). Isoflurane (3%) failed to enhance the response to 10 μm norepinephrine in the unrubbed strips after treatment with the higher concentrations of SOD (≥ 50 U/ml) (fig. 4B).

Isoflurane (≥ 1%) inhibited relaxation caused by 0.03 μm ACh in the +E strips precontracted with 10 μm norepinephrine (fig. 5). Isoflurane (≥ 3%) also inhibited relaxation caused by 1 μm ACh in the unrubbed strips precontracted with 10 μm norepinephrine after exposure to LNNA (100 μm, 60 min) (fig. 5).

The effects of isoflurane on vascular response to norepinephrine, KCl, or ACh observed in this study were substantially identical to those of sevoflurane that we previously observed in this resistance artery. 10Thus, as we previously proposed with regard to sevoflurane, 10the direct (nonneural) action of isoflurane on vascular smooth muscle tissues (including endothelium) does not appear to contribute to systemic hypotension during isoflurane anesthesia. 9,11,28However, the direct action of isoflurane on vascular smooth muscle tissues may contribute to the prolonged systemic hypotension after isoflurane anesthesia. 9 

The current results indicate that isoflurane has opposing actions on contractile response to norepinephrine, i.e. , an endothelium-dependent enhancing and an endothelium-independent inhibitory action. The former emerges immediately after application of isoflurane, remains during its application, and disappears soon after its removal. By contrast, the latter gradually emerges after application of isoflurane and persists after its removal. The norepinephrine response during application of isoflurane is determined by the net balance between these two components. As a result, in the +E strips, the norepinephrine response is enhanced immediately after application of isoflurane, subsequently normalized to the control level at steady state during its application, and notably inhibited after its washout. This proposal is substantially identical to our previous proposal concerning sevoflurane, 10although there are some differences in these anesthetics in terms of their action on norepinephrine response.

In our experiments, the aqueous concentrations produced by 3% isoflurane (0.63 mm) and 5% sevoflurane (0.67 mm) are almost equal to the respective anesthetic concentrations in blood sampled from this rat under steady-state anesthesia at 1 MAC (1.5% isoflurane 27and 2.8% sevoflurane 29): 0.65 mm and 0.66 mm, respectively. 9Thus, in the following discussion, we compare the current data on 3% isoflurane with our previous data on 5% sevoflurane. 10In the +E strips, the norepinephrine response was initially enhanced after application of either isoflurane or sevoflurane; however, this enhancement was subsequently attenuated during application of each anesthetic, presumably because of the emergence of the endothelium-independent (smooth muscle) component. As a result, at steady state (15 min after application of either anesthetic), the magnitude of norepinephrine response returned to the control level during exposure to isoflurane (fig. 1), whereas it was still enhanced (P < 0.05; approximately 40%) during exposure to sevoflurane. 10Namely, during exposure to isoflurane, the smooth muscle component eventually counteracted the endothelial component, preventing the magnitude of norepinephrine response from being enhanced. By contrast, during exposure to sevoflurane, the endothelial component predominated over the smooth muscle component throughout, enhancing the norepinephrine response. 10However, in the −E strips, the inhibition of the magnitude of the norepinephrine response observed 15 min after application of isoflurane was not significantly different from that observed 15 min after application of sevoflurane 10(P > 0.05). Furthermore, in the +E strips, the rate of rise in force in response to norepinephrine was consistently inhibited during exposure to isoflurane (fig. 1), but it was influenced little during exposure to sevoflurane. 10However, in the −E strips, the rate of rise in force in response to norepinephrine was similarly inhibited by both anesthetics (fig. 2). 10Thus, it seems reasonable to postulate that the endothelium-dependent vasoconstricting action of sevoflurane is more potent than that of isoflurane in comparison at the equi-MAC concentration.

Isoflurane inhibited the KCl response at only higher concentrations (≥ 3%) in the +E strips but at lower concentrations (≥ 2%) in the −E strips. In addition, a low concentration (1%) of isoflurane appeared to enhance the KCl response in the +E strips, although its degree did not reach statistical significance. These indicate that the endothelium-dependent vasoconstricting component exists also in its action on KCl response. The endothelial function is considered globally inhibited during 40-mm KCl depolarization because of the lack of endothelial cell hyperpolarization due to K⁺channel opening and the reduced driving force for transmembrane Ca²⁺influx into endothelial cells. Thus, if the endothelium-dependent action is in part mediated by the Ca²⁺influx, the action will be attenuated during the depolarization, a circumstance possibly explaining the observed inability of isoflurane to significantly enhance the KCl response in the +E strips. Unlike in the norepinephrine response, neither the postwashout inhibition nor the prolonged inhibition was observed in the action of isoflurane on KCl response, indicating that the mechanisms of inhibition of the norepinephrine response are at least in part different from the mechanisms of inhibition of the KCl response. The observed action of isoflurane on KCl response was also identical to the previously observed action of sevoflurane on KCl response 10; no significant difference was observed in the effect on KCl response in either +E or −E strips between 3% isoflurane and 5% sevoflurane (P > 0.05).

Isoflurane enhanced the norepinephrine response after inhibition of the NO and EDHF pathways, indicating that the enhanced norepinephrine response by isoflurane is at least in part independent of either the NO or EDHF pathway. Since isoflurane inhibited both the NO-mediated and EDHF-mediated relaxation caused by ACh, isoflurane presumably inhibits some cellular mechanisms specifically involved in the NO- or EDHF-mediated responses induced by ACh but not by norepinephrine or β-adrenergic agonists (e.g. , endothelial phospholipase C activity). Although it has been suggested that isoflurane stimulates the release of vasoactive prostanoids from endothelium, 30,31isoflurane still enhanced the norepinephrine response after inhibition of both cyclooxygenase and lipoxygenase in the current study, suggesting that the action is at least in part independent of either enzyme. In addition, isoflurane still enhanced the norepinephrine response after blockade of ET-1, AT-II, or 5-HT receptors, suggesting that neither receptor agonist is involved in the action.

In this study, isoflurane failed to enhance the norepinephrine response after treatment with SOD. Since it is believed that the exogenously applied SOD (molecular weight, ≥ 33,000) does not enter the endothelial cells, 21,32–34its mechanism of action is probably the extracellular inactivation of superoxide anions (·O^2̄). 21,35Therefore, the extracellular ·O^2̄, presumably released from the endothelial cells, appears to be involved in the vasoconstricting action. In aerobic cells, the ·O^2̄can be generated from oxygen by action of xanthine oxidase or of various cellular enzymes such as cyclooxygenase, lipoxygenase, cytochrome P⁴⁵⁰, or flavin enzymes. 33,36Vascular endothelial cells, which contain these enzymes, also can generate the ·O^2̄. 37,38In addition, the ·O^2̄was previously shown to have a direct vasoconstrictor action and act as an endothelium-derived vasoconstricting factor. 24,35In this study, both cyclooxygenase and lipoxygenase inhibitors did not affect the enhanced norepinephrine response by isoflurane. Thus, in this artery, isoflurane may enhance the norepinephrine response by acting on some endothelial enzyme system other than the cyclooxygenase or the lipoxygenase system and thereby stimulating the endothelial release of ·O^2̄.

Because the ·O^2̄can inactivate the NO and activation of the NO pathway is involved in the norepinephrine response in this artery, 10,39it is also conceivable that isoflurane enhances the norepinephrine response by generating the ·O^2̄via  some endothelial mechanisms and thereby inactivating the NO. However, this seems unlikely because isoflurane still enhanced the norepinephrine response after inhibition of the NO synthesis. The amount of the generated ·O^2̄might not be sufficient for the ·O^2̄to inactivate the NO but might be sufficient for the ·O^2̄to serve as an endothelium-derived vasoconstricting factor.

In conclusion, in isolated rat mesenteric resistance arteries, isoflurane at clinical concentrations significantly influences contractile response to the sympathetic neurotransmitter norepinephrine, and its action consists of an endothelial and a smooth muscle component. The former enhances the norepinephrine response, whereas the latter inhibits it. During application of isoflurane, the endothelial component predominates over or counteracts the smooth muscle component, enhancing the norepinephrine response or preventing the norepinephrine response from being inhibited. Only the smooth muscle component persists after washout of isoflurane, leading to the prolonged inhibition of the norepinephrine response.

The authors thank Masae Yamakawa, B.S. (Department of Anesthesiology and Critical Care Medicine, Kyushu University Hospital, Fukuoka, Japan), for her kind assistance in this work.