The Action of Sevoflurane on Vascular Smooth Muscle of Isolated Mesenteric Resistance Arteries (Part 1): Role of Endothelium

With approval from the Kyushu University Animal Care and Use Committee, male Sprague-Dawley rats (250–320 g) were anesthetized with diethyl ether. The mesenteric tissue was then exposed, rapidly excised, and immediately placed in a dissecting chamber filled with oxygenated 4-(2-hydroxyethyl) piperazine-1-ethanesulfonic acid (HEPES)–buffered physiologic salt solution. A transverse strip (250–400 μm in length, 150–200 μm in width) was prepared from the third- or fourth-order small mesenteric artery (≈ 150–200 μm in diameter) with the endothelium intact (+E) or removed (−E). The details of this preparation were reported previously. 20,22Briefly, the endothelium was removed by gently rubbing the intimal surface with the round surface of a small pin. 20These branches are known to substantially contribute to systemic vascular resistance. 8,23 

Isometric tension was measured by attaching the strip to a strain gauge transducer (UL-2 type; Shinko Co., Tokyo, Japan), as previously detailed. 20,22Briefly, the strip was horizontally mounted in a chamber (0.9 ml) 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 performed in the presence of guanethidine (3 μm) 24to prevent norepinephrine outflow from the sympathetic nerve terminals. Endothelial intactness was verified by the ability of acetylcholine (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 acetylcholine (10 μm) to cause significant (≥ 10%) relaxation during contractions induced by norepinephrine (10 μm). Because the thin strips from the small arteries begin to deteriorate a few hours after setup, 20all experiments were performed 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; 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 isoosmotically.

Norepinephrine, acetylcholine, N  ^G-nitro-l-arginine (LNNA), methylene blue, hemoglobin (bovine), indomethacin, nordihydroguaiaretic acid (NDGA), phenidone (1-phenyl-3-pyrazolidinone), superoxide dismutase (SOD), and U46619 (9,11-dideoxy-11α, 9α-epoxymethanoprostagladin F2α) were purchased from Sigma Chemical (St. Louis, MO). HEPES, ouabain (g-strophanthin), and tetraethylammonium were purchased from Nacalai Tesque (Kyoto, Japan). Sevoflurane was purchased from Kodama Pharmaceutical (Osaka, Japan). BQ-123 (Cyclo(D-α-aspartyl-L-propyl-D-valyl-L-leucyl-D-tryptophyl), BQ-788 (N-[N-[N-[2,6-dimethyl-1-piperidinyl]carbonyl]-4-methyl-L-leucyl]-1-(methoxycarbonyl)-D-tryptophyl]-D-norleucine mono- sodium), and losartan were obtained from Banyu Pharmaceutical (Tokyo, Japan). Ketanserin tartrate was purchased from Research Biochemicals International (Natick, MA). All other reagents were of the highest grade commercially available.

Oxyhemoglobin (HbO²) was prepared by reducing commercial bovine hemoglobin containing 75% methemoglobin, as previously described. 25Pure hemoglobin (oxyhemoglobin) was prepared by adding to a 1-mm solution of commercial hemoglobin in distilled water, a 10-fold molar excess of the reducing agent, sodium dithionite (Na²S²O⁴). Sodium dithionite was then removed by dialysis using a cellulose tubing against 200 volumes of distilled water for 2 h at 4°C. The identification and final concentrations of oxyhemoglobin were determined spectrophotometrically.

We first evaluated the effects of sevoflurane (1–5%) on contractile responses to either norepinephrine (0.5–10 μm) or KCl (40 mm) in both the +E and −E strips. In most of these experiments, 3 min was enough for the responses to various concentrations of norepinephrine or KCl to reach a plateau, and reproducible responses to them were easily obtained with 7-min intervals. Thus, in most of these experiments, either norepinephrine or 40 mm K⁺was applied for 3 min at 7-min intervals. However, only in experiments with a low concentration (0.5 μm) of norepinephrine in the −E strips, approximately 5 min were necessary for the norepinephrine response to reach a plateau after exposure to sevoflurane because of notably inhibited development of force. However, 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. After the response to either norepinephrine or KCl became constant with these protocols, sevoflurane was applied for 5 or 15 min before and during subsequent applications of either stimulant until steady state effect was obtained.

In these experiments, we found that sevoflurane enhances the response to norepinephrine in an endothelium-dependent manner. We thus decided to address the aforementioned possibility that the enhanced norepinephrine response is caused by inhibition of the NO or EDHF pathways. In this attempt, we first characterized the endothelium-mediated relaxation caused by either acetylcholine or histamine, using inhibitors of the NO or EDHF pathways. Specifically, we evaluated the effects of LNNA (NO synthesis inhibitor), tetraethylammonium (K⁺channel blocker known to inhibit the endothelium-mediated hyperpolarization 12,26), ouabain plus Ba²⁺(a combination of a Na⁺–K⁺adenosine triphosphatase (ATPase) inhibitor and a K⁺channel blocker recently reported to inhibit the EDHF-mediated response in this artery 27), and KCl (40 mm) depolarization (established intervention to eliminate the influence of EDHF) on those relaxations. Furthermore, because acetylcholine has been suggested to stimulate the endothelial production of prostacyclin that may cause the release of NO or hyperpolarize the VSM membrane, 26,28we also evaluated the effects of indomethacin (cyclooxygenase inhibitor) on the acetylcholine relaxation. In these experiments, we found that both the acetylcholine-induced and the histamine-induced vasorelaxations consist of the NO-mediated and EDHF-mediated components.

We next characterized the contractile response to norepinephrine, using the aforementioned inhibitors of the NO or EDHF pathways. Specifically, we evaluated the effects of endothelial denudation, LNNA, tetraethylammonium, and ouabain plus Ba²⁺on the norepinephrine response. The concentrations of these inhibitors and the preincubation time necessary for them to exert their maximal effects were determined in the aforementioned experiments. Concentration–response curves for norepinephrine contraction in the presence and absence of either LNNA or LNNA plus tetraethylammonium were constructed for the strips with and without endothelium. After the response to 10 μm norepinephrine became constant, lower concentrations of norepinephrine were applied for a period ranging from 3 to 5 min (until such response reached a plateau) at 7-min intervals. The effects of ouabain plus Ba²⁺were also investigated on the norepinephrine (10 μm) response in the LNNA-treated, unrubbed strips. We additionally evaluated the effects of indomethacin (3–10 μm 28), NDGA (0.3–3 μm, 28lipoxygenase inhibitor), and phenidone (10–30 μm 28lipoxygenase/cyclooxygenase inhibitor) on the norepinephrine response. In these experiments, we found that both the NO and the EDHF signaling pathways are involved in the contractile response to norepinephrine.

To evaluate the ability of sevoflurane to inhibit the NO or EDHF pathways, we tested its effects on both the NO-mediated and the endothelium-mediated vasorelaxations caused by either acetylcholine or histamine and found that sevoflurane inhibits both. Several attempts were made to gain access to the mechanisms causing its inhibitory action on the endothelium-mediated relaxation. First, we attempted to evaluate the effects of sevoflurane on endothelium-independent relaxation caused by low concentrations of K⁺, which has recently been proposed as the EDHF. 27We next attempted to evaluate the effects of sevoflurane on the endothelium-dependent relaxation caused by either A23187 or ionomycin, both of which were previously reported to cause the endothelium-dependent hyperpolarization without acting on the endothelial receptors. 26We finally attempted to confirm the previous observation 11,13that SOD (100–300 U/ml) attenuates the inhibitory action of sevoflurane on the NO-mediated relaxation.

These results indeed suggested the possibility that sevoflurane enhances the norepinephrine contractile response by inhibiting the NO or EDHF pathways in this artery. Therefore, we then evaluated the effects of various inhibitors of these pathways, including LNNA, methylene blue, oxyhemoglobin, and tetraethylammonium, on the enhanced response to norepinephrine by sevoflurane. However, sevoflurane still enhanced the norepinephrine response after treatment with these inhibitors. Therefore, we next evaluated a possibility that sevoflurane stimulates the release of an endothelium-derived vasoconstricting factor (EDCF) and thereby enhances the norepinephrine response. Cyclooxygenase products (e.g. , thromboxane A², prostaglandin F^2α), lipoxygenase products, endothelin-1 (ET-1), angiotensin-II (AT-II), serotonin (5-HT), histamine, adenosine triphosphate (ATP), adenosine diphosphate, and superoxide anions all have been suggested to act as EDCFs. 28In our preliminary experiments, in the −E strips, U46619 (thromboxane–prostaglandin endoperoxide analogue), ET-1, AT-II, and 5-HT all caused significant contractions, both in the presence and the absence of norepinephrine (0.5 μm ≈ 45% effective concentration [EC⁴⁵]); however, histamine (≤ 300 μm; n = 4), ATP (≤ 10 μm; n = 4) or adenosine diphosphate (≤10 μm, n = 4) did not cause any significant contraction in either presence or absence of the submaximal concentration (0.5 μm) of norepinephrine. We therefore did not evaluate the involvement of histamine, ATP, or adenosine diphosphate in the enhanced norepinephrine response by sevoflurane.

Although sevoflurane did not significantly increase the basal force level in these experiments, there still exists the possibility that sevoflurane enhances the norepinephrine response by stimulating the EDCF release. It should be theoretically possible that the agent enhances the force level only in the presence of agonist stimulation (e.g. , in the increased Ca²⁺level). In support of this idea, pretreatment with subthreshold concentrations of U46619, ET-1, 5-HT, and AT-II, which themselves did not increase the basal force level, were indeed capable of significantly (P < 0.05; paired Student t  test) enhancing the response to norepinephrine (10 μm, maximum) in the +E strips, just as observed with sevoflurane (0.1 μm U46619, 166.4 ± 36.7% of control; 0.1 nm ET-1, 165.4 ± 15.3% of control; 0.3 μm 5-HT, 129.7 ± 17.1% of control; 1 nm AT-II, 137.0 ± 18.0% of control [all values are mean ± SD; four rats]). We therefore decided to evaluate the involvement of cyclooxygenase and lipoxygenase products, ET-1, AT-II, 5-HT, and superoxide anion in the enhanced response to norepinephrine by sevoflurane. Specifically, we evaluated the effects of indomethacin, NDGA, phenidone, BQ-123 (selective ET^Areceptor antagonist), BQ-788 (selective ET^Breceptor antagonist), losartan (selective AT-II type 1 receptor antagonist), ketanserin (5-HT²–5-HT^1Creceptor antagonist), and SOD (scavenger of superoxide anions) on the sevoflurane-induced enhancement of norepinephrine response. Because blockade of both ET^Aand ET^B receptor subtypes is necessary to eliminate ET-1 response, 29the effect of BQ-123 plus BQ-788 was evaluated on the action of sevoflurane. The concentrations of BQ-123 plus BQ-788, losartan, and ketanserin, and the preincubation time, were determined by their ability to eliminate the maximal contractions induced by ET-1 (10 nm), AT-II (3 μm), and 5-HT (3 μm) in the −E strips (n = 4). The concentrations of indomethacin, NDGA, phenidone, and SOD, and the preincubation time for these inhibitors were decided based on previous literature. 11,13,28,30In all experiments with these inhibitors, the strips were pretreated with the inhibitors for time considered to be sufficient for exertion of their maximal effects, and the inhibitors were then applied throughout the remainder of the experiment. The concentrations of these inhibitors and the pretreatment time for the inhibitors used in this study are summarized in the Results section (table 1). Some of these inhibitors significantly enhanced the contractile response to norepinephrine (10 μm). In such cases, the norepinephrine concentration was decreased to produce approximately the same level of tension as the control (before application of the inhibitors) 10-μm norepinephrine–induced contraction. Such a low concentration of norepinephrine was determined in each strip. The mean concentration of norepinephrine used in each experimental condition is described in the Results section (table 1). In the final series of experiments, we investigated the time course for the recovery of VSM cells from the effects of sevoflurane after washout of sevoflurane from the experimental chamber (i.e. , after removal of sevoflurane from the extracellular environment).

Sevoflurane was delivered via  a calibrated sevoflurane vaporizer (Sevotec 3; Ohmeda, Steeton, West Yorkshire, UK) in line with the air gas aerating the HEPES-buffered solutions. Each solution was equilibrated with sevoflurane 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 reported concentrations of sevoflurane in the physiologic salt solution produced by 0.9, 1.9, 2.8 (1 minimum alveolar concentration [MAC] in the Sprague-Dawley rat 31) and 3.7% sevoflurane during exactly the same experimental condition 7; and the obtained values were within 95% (95.4–98.5%) of theoretical values predicted by the partition coefficient of sevoflurane in water (0.36 at 37°C). Excellent linear relation was obtained between the aqueous concentrations of sevoflurane (y) and its concentrations (vol%) in the gas mixture (x) (y = 0.0028 + 0.13 x; r = 0.9993). 7Therefore, the concentrations produced by 3 and 5% sevoflurane in the physiologic salt solution in our experiments can be predicted as 0.40 and 0.67 mm, respectively. The latter concentration is almost equal to a recently reported concentration of sevoflurane in blood sampled from this rat during steady state anesthesia with 2.8% (1 MAC in this rat 31) sevoflurane;i.e. , 0.66 mm. 5 

The acetylcholine relaxation, histamine relaxation, and norepinephrine contraction were assessed at the points at which relaxing or constricting effects reached a maximum. Similarly, the effects of sevoflurane on acetylcholine relaxation, histamine relaxation, norepinephrine contraction, or KCl contraction were assessed at the points at which effects reached a maximum. Changes in force were expressed as the percent value of the reference. The concentration–response data for acetylcholine relaxation, histamine relaxation, or norepinephrine contraction, and some effects of sevoflurane, were fitted according to a logistic model described by De Lean et al.  32The 50% effective concentration (EC⁵⁰) or the 50% inhibitory concentration (IC⁵⁰) values were derived from the least-squares fit using the aforementioned model. Because the relation between actual concentrations of sevoflurane 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 percent for the sevoflurane concentration–response relations.

All results are expressed as the mean ± SD, except the results shown in the figures that are expressed as the mean ± SEM for their clarity. The n denotes the number of preparations. The statistical assessment of the data was made by an analysis of variance (one- or two-factor), a Scheffé F test, and a Student t  test (paired or unpaired), as appropriate. Comparisons of the effects of acetylcholine, histamine, norepinephrine, and sevoflurane among (or between) treatments were made by two-factor (concentration, treatment) analysis of variance for repeated measures or factorial analysis of variance. When overall differences were detected, individual comparisons among (or between) groups at each concentration were performed by the Scheffé F test. Comparisons of the effects of these agents within each treatment group were made by one-factor (concentration) analysis of variance for repeated measures, and post hoc  comparisons were made using the Scheffé F test for multiple comparisons. Similarly, statistical assessment of the data on the time-dependent effects of sevoflurane was made by one- or two-factor analysis of variance, with repeated measures or factorial followed by the Scheffé F test. All other necessary comparisons between two data were made by a two-tailed, paired or unpaired, Student t  test after confirming the equal population variances between the groups. A level of P < 0.05 was considered significant.

Sevoflurane (≥ 2%) enhanced submaximal (0.5–1 μm) and maximal (10 μm) contractile responses to norepinephrine in the +E strips; the responses to 0.5, 0.6, and 1 μm norepinephrine were enhanced to 128.4 ± 8.4%, 124.6 ± 15.3%, and 122.9 ± 11.4% of control, respectively, by 3% sevoflurane (P < 0.05; n = 7; data for 10 μm norepinephrine; see fig. 1). However, in the −E strips, sevoflurane (≥ 3%) inhibited submaximal (0.5 μm; ≈ EC⁴⁵) and maximal (10 μm) contractile responses to norepinephrine (fig. 1). Sevoflurane did not influence contractile response to KCl in the +E strips, whereas it inhibited the response to KCl in the −E strips (fig. 1).

Acetylcholine and histamine both produced endothelium-dependent relaxation in the presence of either high K⁺(40 mm) or norepinephrine (10 μm;fig. 2A). LNNA (100 μm; 60 min) eliminated the acetylcholine relaxation in the presence of KCl (fig. 2A); however, LNNA (100–300 μm; 60 min) only partly inhibited the relaxation caused by either acetylcholine or histamine in the presence of norepinephrine (fig. 2A). Such LNNA-resistant relaxation caused by either acetylcholine or histamine was strongly inhibited or eliminated by tetraethylammonium (10–30 mm), ouabain (1 mm) plus Ba²⁺(30 μm), or KCl (40 mm) (fig. 2A). Indomethacin (10 μm; 60 min) did not influence the acetylcholine relaxation (fig. 2A). Sevoflurane inhibited both the LNNA-sensitive and -resistant components of the endothelium-dependent vasorelaxation induced by either acetylcholine or histamine in the presence of norepinephrine (fig. 2B).

The relaxant responses to the low concentrations (10–20 mm) of K⁺were not constantly observed in our preparations. In addition, both A23187 (≤ 10 μm) and ionomycin (≤ 10 μm) did not cause any significant relaxation in the LNNA (100 μm)-treated, +E strips precontracted with 10 μm norepinephrine (n = 4). We therefore did not further evaluate the effects of sevoflurane on the relaxation induced by either the low concentrations of K⁺or those Ca²⁺ionophores.

Superoxide dismutase (100–300 U/ml) did not influence the sevoflurane (3%)–induced inhibition of the acetylcholine (0.03 μm)–induced, LNNA-sensitive relaxations in the +E strips precontracted with 10 μm norepinephrine (P < 0.05; n = 4): control inhibition (64.9 ± 24.5%); inhibition in the presence of 100 U/ml SOD (74.4 ± 17.5%); inhibition in the presence of 300 U/ml (71.9 ± 32.6%).

Norepinephrine produced concentration-dependent (+E, ≥ 1 μm; −E, ≥ 0.3 μm) contraction in both the +E and the −E strips (figs. 3A and B). The LNNA treatment (100 μm; 60 min) enhanced the contractile response to norepinephrine in the +E strips (figs. 3A and B). The treatment with either tetraethylammonium (10 mm) or ouabain (1 mm) plus Ba²⁺(30 μm) further enhanced the contractile response to norepinephrine in the LNNA-treated, +E strips (figs. 3A and B). However, neither LNNA (100 μm) nor LNNA plus tetraethylammonium (10 mm) influenced the concentration–response curve for norepinephrine contraction in the −E strips (P ≥ 0.05;fig. 3C). The treatment with indomethacin (3–10 μm; 60 min) modestly attenuated the response to norepinephrine in the +E strips; the norepinephrine (10 μm) contraction was inhibited to 88.9 ± 5.8% and 72.0 ± 6.6% of control by 3 and 10 μm indomethacin, respectively (P < 0.05; n = 5 or 6). Both NDGA (0.3–3 μm; 15 min) and phenidone (10–30 μm; 25 min) also attenuated the contractile response to 10 μm norepinephrine in the +E strips (P < 0.05; n = 5–7; 0.3 μm NDGA, 71.8 ± 28.6% of control; 1 μm NDGA, 61.6 ± 34.9% of control; 10 μm phenidone, 85.0 ± 15.2% of control; 30 μm phenidone, 64.9 ± 11.5% of control).

In the LNNA-treated strips, sevoflurane did not influence the maximal response to 10 μm norepinephrine that had notably been enhanced after exposure to LNNA (fig. 4). However, in the LNNA-treated strips, sevoflurane still enhanced the submaximal response to lower concentrations of norepinephrine (0.86 ± 0.53 μm), the amplitude of which was not different from that of the control response to 10 μm norepinephrine (94.9 ± 29.4% of the control;P > 0.05; n = 7;fig. 4). The sevoflurane (3%)–induced enhancement of the response to the low concentrations of norepinephrine after exposure to LNNA was not significantly different from that of the submaximal response to either 0.5, 0.6, or 1 μm norepinephrine in the LNNA-untreated strips (P > 0.05; data shown previously).

Sevoflurane still enhanced the contractile response to norepinephrine after treatment with either LNNA plus tetraethylammonium, LNNA plus tetraethylammonium plus methylene blue plus oxyhemoglobin, indomethacin, NDGA, or phenidone (table 1). Because all of these treatments enhanced (P < 0.05) the response to 10 μm norepinephrine, the effects of these treatments on the vasoconstricting action of sevoflurane were studied for contractions induced by lower concentrations of norepinephrine, the amplitude of which was not significantly different from that of the control 10 μ m norepinephrine-induced contraction (P > 0.05; n = 5 or 6;table 1).

Antagonists of either ET-1 (ET^Aand ET^B), AT-II type 1, or 5-HT receptors also failed to prevent the enhanced response to norepinephrine by sevoflurane (table 1). SOD (100–300 U/ml) also did not influence the sevoflurane-induced enhancement of norepinephrine response (table 1).

Constant responses to either norepinephrine or KCl were observed in either +E or −E strips for more than 4 h in our experiments. As described previously, in the +E strips, the response to norepinephrine was enhanced during application of sevoflurane; however, it was significantly inhibited after washout from the chamber (fig. 5A). This inhibition was prolonged for 15–75 min, depending on the preparations (fig. 5A). Comparison of the time-dependent effects of sevoflurane between two different protocols regarding time for sevoflurane application (i.e. , short vs.  long application protocols) indicated that the “postwashout inhibition” was triggered by the washout of sevoflurane, but not due to emergence of some “late-onset” inhibition (fig. 5A, d). Identical postwashout inhibitions were caused by 5% sevoflurane in the +E strips treated with LNNA, LNNA plus tetraethylammonium, or LNNA plus tetraethylammonium plus indomethacin (P > 0.05; n = 4–6).

In the −E strips, sevoflurane, applied after contractile responses to norepinephrine (0.5 μm) became constant, inhibited the contractile response to norepinephrine (fig. 5B). Fifteen minutes was necessary for the inhibitory effect of sevoflurane to reach maximum and steady state (fig 5B). The inhibition was again prolonged after washout of sevoflurane from the chamber, and it took more than 30 min for VSM cells to completely recover from the inhibition (fig. 5B).

In contrast, the onset of sevoflurane-induced inhibition of KCl contraction in the −E strips was immediate; five minutes was enough for its inhibitory effect on contractile response to KCl to reach maximum and steady state (fig. 5C). No significant inhibition of the KCl contraction was observed after washout of sevoflurane in either +E or −E strip (fig. 5C).

The actions of sevoflurane on isolated “endothelium-intact” resistance arteries observed in this study, as summarized below, do not appear to contribute to systemic hypotension previously observed during sevoflurane anesthesia. 1–5First, in the presence of endothelium, sevoflurane did not inhibit, but enhanced, contractile response to the sympathetic neurotransmitter norepinephrine. Second, in the presence of endothelium, sevoflurane did not influence contractile response to KCl, which is mediated by activation of voltage-gated Ca²⁺channels. Finally, sevoflurane inhibited both the NO-mediated and the EDHF-mediated vasorelaxations. Rather, our data suggest a possibility that, in the endothelium-intact mesenteric arterial bed, sevoflurane causes an increase in vascular tone and thereby a decrease intestinal blood flow through direct actions on VSM cells. However, no significant changes in intestinal blood flow or vascular resistance were observed in this rat anesthetized with sevoflurane. 2,3Therefore, sevoflurane probably has some “indirect” vasodilating actions on this arterial bed, counteracting the direct vasoconstricting action. Centrally and peripherally attenuated neural excitatory activity may account for such vasodilating actions. Sevoflurane was recently reported to attenuate this mesenteric arterial tone by inhibiting norepinephrine outflow from the nerve terminals and thereby causing the “in situ  hyperpolarization.”5 

It was recently reported that the systemic hypotension persists after anesthesia with sevoflurane in this rat. 5Recovery of mean arterial pressure was incomplete (≈ 70–85%) during the postanesthesia period after inhalation of 0.5–1.0 MAC (0.22–0.66 mm in blood) sevoflurane;i.e. , after a 15–30 min washout period during which its blood concentration became negligible (0.01–0.02 mm). 5In addition, the in situ  hyperpolarization of VSM cell membrane induced by sevoflurane disappeared during the postanesthetic period. 5In our study, the concentrations produced by 3–5% sevoflurane in the physiologic salt solution would be approximately 0.40–0.67 mm. Thus, the persistent hyporesponsiveness to norepinephrine after washout of sevoflurane observed in our study may contribute to the prolonged systemic hypotension observed after sevoflurane anesthetic in this rat. 5 

The direct action of sevoflurane on contractile response to norepinephrine in this artery appears to consist of two distinct components;i.e. , an endothelial and a smooth muscle component. The former enhances the norepinephrine response, whereas the latter inhibits it. In the presence of intact endothelium, the endothelial component predominates over the smooth muscle component, enhancing the norepinephrine response. Sevoflurane might inhibit the norepinephrine response in a situation in which the endothelial function is impaired. The endothelial vasoconstricting component immediately emerged after application of sevoflurane and quickly disappeared after its removal, whereas the smooth muscle vasodilatory component emerged relatively gradually after application of sevoflurane and persisted after its removal. However, such prolonged inhibition was not observed in the action of sevoflurane on KCl response. This indicates that the mechanisms causing the inhibitory action of sevoflurane on norepinephrine contraction are at least in part different from those on KCl contraction.

Although sevoflurane enhanced the norepinephrine contraction in this resistance artery, previous studies using isolated arteries yielded conflicting results regarding this issue. 6,7,11,13,33In endothelium-intact canine mesenteric arteries, sevoflurane, depending on timing of its application, did not influence or enhance the norepinephrine contraction, 11,33whereas it attenuated the response to norepinephrine or phenylephrine in either the presence or the absence of endothelium in rabbit mesenteric arteries or rat aorta. 6,7,13The observed differences could be caused by either species or regional differences. Unlike in the rat mesenteric arteries, the smooth muscle vasodilating component might predominate over the endothelial vasoconstricting component in those vascular beds.

Sevoflurane inhibited both the LNNA-sensitive and the LNNA-resistant components of either acetylcholine-induced or histamine-induced relaxation. Either tetraethylammonium or ouabain plus Ba²⁺, both of which have been proposed to inhibit the action of EDHF, 27,34inhibited the latter component. Therefore, the LNNA-resistant component is presumably mediated by EDHF, suggesting the ability of sevoflurane to inhibit the EDHF pathway in addition to the NO pathway. The high concentrations (10–30 mm) of tetraethylammonium, known to act as a nonspecific K⁺channel blocker, 35,36probably inhibited K⁺channels of both endothelial and VSM cells involved in the EDHF response. 27Conversely, ouabain plus Ba²⁺inhibited Na⁺–K⁺ATPase and Ba²⁺-sensitive K⁺channels of VSM cells involved in the EDHF responses. 27In addition, the high concentration (1 mm) of ouabain, known to inhibit K⁺channels in VSM, 34might also inhibit those K⁺channels.

It was previously proposed that volatile anesthetics, including sevoflurane, interfere with the EDHF-mediated relaxation in addition to the NO-mediated relaxation, 12,17although, to our knowledge, there has been no electrophysiologic evidence showing that the anesthetics inhibit the EDHF-mediated hyperpolarization. In addition, previous studies 12,17evaluated the anesthetic effects only on the acetylcholine relaxation. However, because the volatile anesthetics have been suggested to inhibit acetylcholine signaling by interfering with the muscarinic receptor, 37the anesthetic might globally inhibit the acetylcholine relaxation by acting on the muscarinic receptor in those studies 12,17; it was therefore unclear whether the anesthetic inhibited some mechanisms specifically involved in the EDHF response. However, in this study, sevoflurane inhibited both the acetylcholine- and the histamine-induced, LNNA-resistant (i.e. , presumed EDHF-mediated) vasorelaxations. This may imply that sevoflurane interferes with the EDHF-mediated response by acting on sites distal to the receptor stimulation. To our knowledge, there is no evidence that volatile anesthetics interfere with the histaminergic receptor. To further investigate the site for the inhibitory action of sevoflurane on the EDHF-mediated response, as detailed in Materials and Methods, we attempted to evaluate the effects of sevoflurane on vasorelaxation caused by either Ca²⁺ionophores or the low concentrations of K⁺. However, we could not perform these experiments because of the reasons described in Results.

The observed lack of effect of SOD on the inhibitory action of sevoflurane on NO-mediated relaxation is contradictory to previous studies 11,13that propose that sevoflurane impairs the NO-mediated relaxation by generating superoxide free radicals and thereby inactivating NO. Although the previous studies were performed in a hyperoxic environment (95% O²), our experiments were performed in a nonhyperoxic environment (i.e. , air). Therefore, sevoflurane might fail to generate a significant amount of superoxide free radicals in our study. Our results are rather consistent with a previous study, 38which showed the inability of sevoflurane to chemically interact with NO and its ability to directly inhibit NO production in endothelial cells.

The observed effects of LNNA, tetraethylammonium, and ouabain plus Ba²⁺on contractile response to norepinephrine indicate that both the NO–cyclic 3′,5′-guanosine monophosphate and EDHF pathways are involved in the norepinephrine response, consistent with previous proposals. 12,15,16In addition, as discussed previously, the observed effects of sevoflurane on the endothelium-mediated relaxation suggest its ability to inhibit both the NO and the EDHF pathways in this artery. It is therefore conceivable that sevoflurane enhances the contractile response to norepinephrine by inhibiting the NO or the EDHF pathway. However, sevoflurane still enhanced the norepinephrine response after inhibition of these pathways. Sevoflurane may inhibit some cellular mechanisms specifically involved in the NO-mediated or EDHF-mediated responses induced by acetylcholine or histamine but not by norepinephrine. Although LNNA eliminated the sevoflurane-induced enhancement of the maximal response to norepinephrine (10 μm), we do not believe that inhibition of the NO pathway accounts for the enhanced response to norepinephrine by sevoflurane. Because the response to 10 μm norepinephrine was already notably enhanced, possibly saturated, after exposure to LNNA (at application of sevoflurane), sevoflurane probably failed to further enhance such nearly saturated response to norepinephrine. Our results also indicate that the vasoconstricting action of sevoflurane is not specific to higher (e.g. , 10 μm) concentrations of norepinephrine. Although we do not deny the involvement of inhibition of the NO pathway in the enhanced norepinephrine response by sevoflurane, we believe that such involvement is minimal, if there is any. We rather believe that some mechanism other than inhibition of the NO pathway is mainly attributable to the enhanced response to norepinephrine.

Comparison of the effects of sevoflurane on KCl (40 mm) response between +E and −E strips indicate the presence of an endothelium-dependent vasoconstricting component in its action on KCl response. Because endothelial function is globally inhibited in the presence of KCl (40 mm) depolarization because of a lack of hyperpolarization caused by K⁺channel opening and a resultant reduced driving force for transmembrane Ca²⁺influx into endothelial cells, the endothelium-mediated action of sevoflurane could be inhibited to some extent in the strips depolarized with KCl. Thus, sevoflurane would fail to enhance the KCl response in the presence of endothelium. However, because sevoflurane still exerts the endothelium-dependent vasoconstricting action in the KCl-depolarized strips, the aforementioned idea is supported: that the enhanced contractile response by sevoflurane is not caused by inhibition of the EDHF pathway.

Volatile anesthetics were previously proposed to stimulate the release of EDCFs such as cyclooxygenase products or oxygen-derived free radicals and thereby enhance contractile responses. 39,40However, the observed lack of effect of indomethacin, phenidone, NDGA, or SOD on the vasoconstricting action of sevoflurane indicates that cyclooxygenase products, lipoxygenase products, or oxygen-derived free radicals are not involved in the action. In addition, sevoflurane still enhanced the norepinephrine response after blockade of ET-1, AT-II, or 5-HT receptors, suggesting that neither of these receptor agonists is involved in the vasoconstricting action. Sevoflurane may enhance the norepinephrine response by stimulating the release of some other unidentified EDCFs or by inhibiting some unidentified endothelium-mediated vasodilatory mechanisms activated by norepinephrine. Further investigations are necessary to elucidate the mechanisms.

Sevoflurane significantly influences contractile response to norepinephrine in isolated mesenteric resistance arteries, and its action consists of an endothelial and a smooth muscle component. The former enhances the norepinephrine response, whereas the latter inhibits it. In the presence of intact endothelium, the endothelial component predominates over the smooth muscle component, causing enhancement of the norepinephrine response. However, only the smooth muscle component persists after washout of sevoflurane, leading to the prolonged inhibition of the norepinephrine response and possibly contributing to the prolonged systemic hypotension after sevoflurane anesthesia. 5Such persistent hyporesponsiveness to norepinephrine is caused by an effect on some cellular mechanism specifically involved in the norepinephrine response, but not in the KCl response. Although our results indicate significant involvement of both the NO and the EDHF pathways in the norepinephrine response and the ability of sevoflurane to inhibit both the NO-mediated and the EDHF-mediated responses to either acetylcholine or histamine, sevoflurane still enhanced the norepinephrine response after inhibition of these pathways. This suggests that the enhanced response to norepinephrine by sevoflurane is not caused by inhibition of either the NO or the EDHF pathway, and also that sevoflurane inhibits some cellular mechanisms specifically involved in the NO-mediated or EDHF-mediated responses induced by acetylcholine or histamine (but not by norepinephrine). Finally, cyclooxygenase products, lipoxygenase products, oxygen-derived free radicals, ET-1, AT-II, or 5-HT are not involved in the vasoconstricting action of sevoflurane.

The authors thank Professor Takeo Itoh (Nagoya City University, Nagaya, Japan) and Dr. Walter A. Boyle III (Washington University, St. Louis, Missouri) for their helpful comments regarding this work, and Ms. Masae Yamakawa (Kyushu University) for her kind assistance in this work.