The causes of volatile anesthetic-induced cerebral vasodilation include direct effects on smooth muscle and indirect effects via changes in metabolic rate and release of mediators from vascular endothelium and brain parenchyma. The role of nitric oxide and the relative importance of neuronal and endothelial nitric oxide synthase (nNOS and eNOS, respectively) are unclear.
Rat brain slices were superfused with oxygenated artificial cerebrospinal fluid. Hippocampal arteriolar diameters were measured using computerized videomicrometry. Vessels were preconstricted with prostaglandin F2alpha (PGF2alpha; halothane group) or pretreated with 7-nitroindazole sodium (7-NINA, specific nNOS inhibitor, 7-NINA + halothane group) or N-nitro-L-arginine methylester (L-NAME; nonselective NOS inhibitor, L-NAME + halothane group) and subsequently given PGF2alpha to achieve the same total preconstriction as in the halothane group. Increasing concentrations of halothane were administered and vasodilation was calculated as a percentage of preconstriction.
Halothane caused significant, dose-dependent dilation of hippocampal microvessels (halothane group). Inhibition of nNOS by 7-NINA or nNOS + eNOS by L-NAME similarly attenuated halothane-induced dilation at 0.6, 1.6, and 2.6% halothane. The dilation (mean +/- SEM) at 1.6% halothane was 104 +/- 10%, 65 +/- 6%, and 51 +/- 9% in the halothane, 7-NINA + halothane and L-NAME + halothane groups, respectively. The specificity of 7-NINA was confirmed by showing that acetylcholine-induced dilation was not inhibited by 7-NINA but was converted to constriction by L-NAME.
At clinically relevant concentrations, halothane potently dilates intracerebral arterioles. This dilation is mediated, in part, by neuronally derived nitric oxide. Endothelial NOS does not play a major role in halothane-induced dilation of hippocampal microvessels.
VOLATILE anesthetics cause cerebral vasodilation, which may lead to increased cerebral blood flow (CBF), cerebral blood volume, and intracranial pressure. 1,2When there is decreased intracranial compliance or already established intracranial hypertension, further increases in intracranial pressure may cause cerebral ischemia and herniation of brain tissue. The mechanisms of volatile anesthetic-induced cerebral vasodilation are understood incompletely and may include direct effects on vascular smooth muscle and indirect effects via release of mediators from endothelial and parenchymal cells and changes in neuronal activity or cerebral metabolic rate. 1–3The relative importance of these factors is unknown. In particular, the role of nitric oxide (NO) and the relative importance of NO derived from neuronal and endothelial NO synthases (NOS; nNOS, type I NOS; eNOS, type III NOS, respectively) are unclear. In vitro studies using isolated large cerebral arterial rings have shown that vascular relaxation caused by volatile anesthetics does not depend on the presence of an intact endothelium. 4,5In contrast, in vivo studies of the pial circulation have shown that inhibition of NOS with the nonselective NOS inhibitor, N -nitro-L-arginine methylester (L-NAME), decreases volatile anesthetic-induced dilation. 3,6–8A study that evaluated laser Doppler flow in the parietal cortex of nNOS gene-deficient mice and wild-type controls suggested that eNOS may be involved in isoflurane-induced hyperemia at low anesthetic concentrations (1.2 and 1.8%), with nNOS involved at higher concentrations (2.4%). 9These studies described changes in large cerebral arteries and superficial cerebral vessels, but did not provide information about the mechanism of anesthetic-induced effects on intraparenchymal arterioles specifically.
We previously used an in vitro rat brain slice preparation to show that halothane causes potent, dose-dependent dilation of hippocampal arterioles 10and that halothane and isoflurane cause dilation that is region-specific and agent-specific. 11At equipotent doses, the two agents cause similar dilation in hippocampal, but not in neocortical, vessels, in which halothane produces a greater degree of dilation. 11
The objectives of this investigation were to assess the contribution of NO to the resting tone of hippocampal intraparenchymal arterioles, to define the role of NO in the mechanism of halothane-induced microvascular dilation, and to determine the relative importance of NO derived from eNOS and nNOS. An in vitro rat brain slice preparation was used to assess the effects of halothane on hippocampal arterioles in the presence and absence of NOS inhibition. The NOS inhibitors used were the nonselective NOS inhibitor L-NAME and the selective nNOS inhibitor 7-nitroindazole sodium (7-NINA).
Materials and Methods
All experimental procedures used in this investigation were reviewed and approved by the Animal Use and Care Committee of the Medical College of Wisconsin, with protocols completed in accordance with the Guiding Principles in the Care and Use of Laboratory Animals of the American Physiologic Society and in accordance with National Institutes of Health (NIH) guidelines. All animals used in this investigation were housed within the animal facilities of the Medical College of Wisconsin, accredited by the American Association for the Accreditation of Laboratory Care.
Adult male Sprague-Dawley rats, weighing 200–300 g, were placed in an animal holding chamber and underwent inhalational induction of anesthesia using 2% halothane (Anaquest Inc., Madison, WI) in oxygen. After achieving sufficient anesthesia, a midline thoracotomy was performed and 20 ml NaCl, 0.9%, was infused into the left ventricle while simultaneously making a right atrial incision for blood drainage. The animals were then decapitated and the brains rapidly removed and rinsed with nutrient medium (artificial cerebrospinal fluid, aCSF) of the following composition (mM): NaCl: 124; KCl: 5; CaCl2: 2.4; MgCl2: 1.3; glucose: 10; KH2PO4: 1.24; NaHCO3: 26. Nutrient medium was prepared daily and equilibrated with 95% oxygen and 5% carbon dioxide (95% O2–5% CO2) to achieve a pH of 7.4. All measurements of cerebral microvessel diameters were performed within 5 h of the tissue slice preparation.
Brains were cut freehand into blocks containing the hippocampus. A vibratome mechanical tissue slicer (OTS-3,000-03; FHC, Brunswick, ME) was used to immediately section the block into coronal slices approximately 280-μm thick. Throughout the slicing procedure, tissues were continuously bathed in the oxygenated aCSF at room temperature. Subsequently, the slices were transferred to a Plexiglas holding chamber (M & G Plastic Specialists, West Allis, WI) and maintained at interface with oxygenated aCSF at the same temperature. Individual slices were then transferred for evaluation to a recording chamber mounted on an inverted halogen transillumination microscope (Nikon Diaphot 200; Yokohama, Japan).
The recording chamber consisted of a central recording–superfusion compartment and a laterally placed elevated chamber to allow gentle vacuum suction. Nylon mesh beneath the brain slice allowed for circulation of superfusate under and around the slice. Flow through the recording chamber was at a rate of 2.0 ml/min, completely exchanging the volume in the chamber in less than 2 min. The chamber temperature was continuously monitored and maintained at 34°C using a thermoelectric Peltier device coupled to a sensing thermistor. The slices were maintained in this chamber, continuously superfused with the oxygenated aCSF for approximately 1 h before initiation of the experimental protocol. During this equilibration period an intracerebral microvessel was located. The aCSF that superfused the brain slices was equilibrated with a mixture of oxygen, carbon dioxide, and air sufficient to maintain the bath at a pH of 7.35–7.45, a PCO2of 35–45 mmHg, and a PO2in the range 210 ± 30 mmHg. Gas analysis (Radiometer ABL3, Copenhagen, Denmark) of the superfused fluid was obtained during the equilibration period and at least every 30 min during the experimental protocol. The PO2was also continuously measured with an in-line flow-through PO2electrode (DO-166FT; Lazar Research Laboratories, Los Angeles, CA).
A vessel (range, 10–25 μm in diameter) was located within the parenchyma of the hippocampus. The integrity and diameter were continuously monitored using videomicroscopy. Arterioles were differentiated from venules by the presence and the characteristics of the vascular smooth muscle. Equipment consisted of an inverted halogen transillumination microscope, a 40× objective (Olympus WPlanFL 160/0, Tokyo, Japan), and a 2.25× video projection lens (Nikon CCTV/Microscope Adapter, Yokohama, Japan). The image was transmitted to a video camera (CCD 72; Dage MTI, Michigan City, IN) and displayed on a video monitor (Sony HR Trinitron, Tokyo, Japan). Vessel diameter changes were recorded on videotape using a VHS video recorder (Magnavox, Rebersburg, PA) and analyzed using a computerized imaging analysis system (Metamorph Imaging System, Universal Imaging Corp., West Chester, PA) with an IBM-compatible computer.
There were three experimental groups: the halothane group, the 7-NINA + halothane group, and the L-NAME + halothane group. The protocol is represented schematically in figure 1. In all groups, vessel diameter was measured at the end of the equilibration period. In the halothane group, vessels were preconstricted with prostaglandin F2α(PGF2α; Sigma Chemical Co., St. Louis, MO), given at a bath concentration of 1 μM, to produce a constriction of approximately 11–14% from baseline. 10,11In the other two groups, vessels were pretreated with either 7-NINA (Tocris Cookson Inc., Ballwin, MO), given at a bath concentration of 2 μM (7-NINA + halothane group), or L-NAME (Sigma Chemical Co., St. Louis, MO), given at a bath concentration of 50 μM (L-NAME + halothane group). In initial dose-finding experiments, we found that 7-NINA caused a maximal constriction of 6–8% at a dose of 2 μM, whereas L-NAME caused a much greater degree of constriction at doses more than 50 μM. Although other authors using the same preparation have given nonselective NOS inhibitors to produce a preconstriction of approximately 30%, 12,13we were reluctant to use higher doses of L-NAME because of the possibility of mechanisms unrelated to NOS inhibition. 14After 7-NINA or L-NAME administration, vessels were further constricted with PGF2α, given at a bath concentration of 0.5–0.8 μM, to achieve the same total preconstriction as that in the halothane group. We chose doses of 7-NINA and L-NAME that caused a similar level of constriction (6–8%) and used PGF2αto ensure the same total level of preconstriction so that the response to halothane could be compared in vessels that have the same initial vascular tone. All drugs were diluted in aCSF and infused directly into the recording chamber.
After preconstriction, increasing doses of halothane were given, dissolved in aCSF for 30 min at each dose (fig. 1). Halothane was volatilized into the aCSF by passing the oxygen, carbon dioxide, and air mixture through a vaporizer (Model F100; OH Medical Products, Airco Inc., Madison, WI). The time necessary for equilibration of halothane concentrations in the recording chamber was determined during previous experiments. 10The halothane vaporizer was set at dial settings of 0.4, 0.6, 1.8, and 2.6%. At the end of each equilibration period, a 1-ml sample of aCSF was taken from the recording chamber for measurement of halothane concentrations by gas chromatography (Sigma 3B; Perkin-Elmer, Norwalk, CT). A glass coverslip was placed over the chamber to ensure a consistent relation between the vaporizer dial settings and the measured halothane concentrations. Measured aqueous concentrations (mM) were converted to partial pressures (percent of one atmosphere, %), yielding mean values of 0.4, 0.6, 1.6, and 2.6%. 15In the rat, these partial pressures are equivalent to minimum alveolar concentrations (MAC) of 0.3, 0.5, 1.5, and 2.3 (MAC value for the rat = 1.1%). 16
To evaluate the specificity of 7-NINA and L-NAME at the doses used in this investigation, the dilator effect of acetylcholine was assessed during control conditions (acetylcholine group, n = 6) and in the presence of 7-NINA (7-NINA + acetylcholine group, n = 6) and L-NAME (L-NAME + acetylcholine group, n = 7). In the acetylcholine group, vessels were preconstricted with 1 μM PGF2α. In the 7-NINA + acetylcholine and the L-NAME + acetylcholine groups, vessels were pretreated with 2 μM 7-NINA or 50 μM L-NAME and further constricted with 0.5–0.8 μM PGF2α. After preconstriction, acetylcholine was given at bath concentrations of 10 μM and 100 μM in sequence for 20 min at each dose.
Because vasomotor changes may be nonuniform, microvessel diameters were derived as an average of 10–13 measurements taken every 6–10 μm along approximately 80 μm of vessel length. The percentage constriction of the cerebral arteriole from the baseline diameter was calculated using the following equation:
where DBLis the baseline diameter at the start of the experiment and DNis the new diameter. The percentage dilation of the cerebral arteriole was calculated as a percentage of the total amount of preconstriction using the following equation:10,11
where DN is the new diameter and DP is the diameter after preconstriction (fig. 1).
Statistical analysis was performed using one-way analysis of variance (between groups) and repeated-measures one-way analysis of variance (within groups). In both cases, follow-up multiple comparisons were made using the Duncan multiple range test. Differences were considered statistically significant when the P value was less than 0.05. All data are expressed as the mean ± SEM.
A total of 50 hippocampal slices were obtained from 24 animals. Brain slices were excluded from the analyses if microvessels could not be adequately visualized or if the luminal diameters of the microvessels were not clearly discernible during the experiment. The number of slices that completed the experimental protocol and baseline data for the experimental groups are shown in table 1. The arteriolar diameters before preconstriction did not differ significantly between groups. 7-NINA and L-NAME caused similar microvascular constriction in the 7-NINA + halothane and L-NAME + halothane groups. The total preconstriction after administration of PGF2αwas also similar in all groups.
Halothane alone caused significant and dose-dependent microvascular dilation (halothane group). However, this was significantly less in the presence of 7-NINA and in the presence of L-NAME (fig. 2). Halothane, 0.6% and 2.6%, caused dilation of 63 ± 7% and 143 ± 18%, respectively. Halothane, 0.6%, caused dilation of 37 ± 7% and 36 ± 6% in the presence of 7-NINA and L-NAME, respectively. Halothane, 2.6%, caused dilation of 85 ± 11% and 65 ± 8% in the presence of 7-NINA and L-NAME, respectively. 7-NINA and L-NAME caused similar attenuation of halothane-induced dilation at 0.6, 1.6, and 2.6% halothane. At 0.4% halothane, halothane-induced dilation was significantly less in the presence of L-NAME than in the presence of halothane alone or in the presence of 7-NINA (35 ± 4%, 30 ± 7%, and 14 ± 3% dilation at 0.4% halothane in the halothane, 7-NINA + halothane, and L-NAME + halothane groups, respectively).
The mean bath concentration of halothane 1 h after the anesthetic was discontinued was 0.07–0.08 mM in all groups. These aqueous concentrations are equivalent to mean partial pressures of 0.24–0.27%. Vessel diameters had decreased toward their preconstricted control values (mean dilation 15 ± 5%, 3 ± 4%, and 10 ± 11% in the halothane, halothane + 7-NINA, and halothane + L-NAME groups, respectively), and significant microvascular dilation (compared with baseline) was not observed in any group.
In experiments to evaluate the specificity of 7-NINA and L-NAME, baseline diameters did not differ significantly between groups (mean diameters 15.8–16.8 μm). 7-NINA and L-NAME caused similar constriction in the 7-NINA + acetylcholine and L-NAME + acetylcholine groups (7.8 ± 1% and 6.6 ± 0.9%, respectively). The total preconstriction after administration of PGF2αwas also similar in all groups (total preconstriction of 11.1 ± 1.9%, acetylcholine group; 13.1 ± 1.3%, 7-NINA + acetylcholine group; 11.8 ± 1.4%, L-NAME + acetylcholine group). Vessel diameters were significantly greater than the control diameters at both concentrations of acetylcholine in the acetylcholine and acetylcholine + 7-NINA groups (percentage change from the preconstricted diameter 9.4 ± 2.1% and 9.8 ± 1.5%, respectively; 100 μM acetylcholine;fig. 3). There was no significant difference between the effect of acetylcholine in the presence and absence of 7-NINA. In contrast, the dilator response to acetylcholine was converted to constriction in the L-NAME + acetylcholine group. In this group, vessel diameters were significantly different from the control diameters and from those in the acetylcholine and 7-NINA groups at both concentrations of acetylcholine (percentage change from the preconstricted diameter 15.3 ± 1.8%; 100 μM acetylcholine;fig. 3).
Nitric oxide is recognized as an ubiquitous molecule that has an important role in many physiologic and pathologic processes. In the brain, NO is continuously produced by eNOS and nNOS. 17Once formed, NO diffuses into vascular smooth muscle in which it activates soluble guanylate cyclase (GC). This increases the synthesis of cyclic guanosine 3′,5′-monophosphate (cGMP) and leads to vasodilation. 17In this investigation, we assessed the effects of halothane on hippocampal arterioles in the presence and absence of L-NAME and 7-NINA. L-NAME is an L-arginine analog that causes a prolonged, competitive inhibition of eNOS and nNOS. 18,19In contrast, 7-nitroindazole 20(7-NI) and its more water-soluble sodium salt, 7-NINA, 21cause a selective inhibition of nNOS.
The unchanged acetylcholine-induced dilation in the presence of 7-NINA suggests that when 7-NINA is given in our preparation and at the dosage used in our study, it does not inhibit eNOS. In contrast, the dilator response to acetylcholine was converted to constriction in the presence of L-NAME, suggesting a direct smooth muscle effect of acetylcholine. Fergus and Lee 12also found that the dilator response to acetylcholine was inhibited by N ω-nitro-L-arginine (L-NA) (nonselective NOS inhibitor) but not by 7-NI (10 μM). In contrast, the dilator response to NMDA was inhibited by L-NA and 7-NI, showing that NMDA-induced microvascular dilation is mediated, in part, by NO derived from nNOS. 12Taken together, these findings suggest that 7-NINA selectively inhibited nNOS. Traditionally, eNOS is considered to be present in vascular endothelial cells and nNOS is considered to be present in neurons, perivascular nerves, and astrocytes. 22–24This delineation may not be entirely correct because nNOS may also be present in the endothelium. 25However, endothelium-derived nNOS does not appear to mediate endothelium-dependent dilation. 12,17
The role of NO in volatile anesthetic-induced cerebral vasodilation is controversial. In isolated cerebral arteries, relaxation caused by halothane and isoflurane is not affected by removal of the endothelium 4,5or, in the case of isoflurane, by the nonselective NOS inhibitor, N G-monomethyl-L-arginine. 5The results of studies to determine the effects of halothane on GC and cGMP content depend on whether the direct actions of halothane or its effects on agonist-induced increases in these variables were evaluated. Although halothane increases basal cGMP levels in isolated canine middle cerebral arteries, it is the particulate (non–NO-dependent), rather than the soluble (NO-dependent), GC activity that increases. 26Halothane decreases endothelium-dependent relaxation of isolated carotid arteries and the aorta. 4,27Halothane also decreases NO-induced relaxation and NO-induced increases in cGMP content in the isolated aorta. 27In in vitro preparations of brain tissue and vascular endothelial and smooth muscle cells, volatile anesthetics have minimal effect on NOS activity, 28GC activity, 29and cGMP content. 30,31
In contrast to in vitro studies, administration of nonselective NOS inhibitors in vivo, either topically or systemically, decreases volatile anesthetic-induced cerebral vasodilation. 6,7Similarly, studies of regional CBF found that L-NAME attenuates anesthetic-induced hyperemia. 3,8Okamoto et al. found that isoflurane-induced increases in laser Doppler flow were decreased by L-NA at all concentrations in wild-type mice, but only at lower concentrations in nNOS gene-deficient (knockout) mice. 9The authors concluded that eNOS (with or without nNOS) may be involved in alterations of CBF at lower concentrations of isoflurane, with nNOS only involved in high-dose isoflurane-induced hyperemia. 9
A role for NO in volatile anesthetic-induced cerebral vasodilation cannot be excluded by in vitro studies that use isolated arteries. These vessels are removed from their neuronal and glial framework and therefore cannot be under the influence of NO derived from nNOS. 22–24In addition, the role of NO may be different in vessels of different size. 32Although volatile anesthetic-induced pial vessel dilation may depend, in part, on NO derived from eNOS, 7,9it may also, similar to the response to hypercapnia, 33be influenced by NO derived from nNOS.
We previously used the brain slice microvessel preparation to demonstrate that vasomotor responses to alterations in oxygen 34and carbon dioxide 35levels are well-preserved. The model used in this investigation is the first to evaluate the effects of anesthetic agents on intraparenchymal arterioles and on the relation between these vessels and their surrounding neurons and glial cells. The findings of our study suggest that NO is involved intimately in halothane-induced dilation of these vessels. The lack of a difference between the effects of L-NAME and 7-NINA at halothane concentrations of 0.6, 1.6, and 2.6% suggests that, at these concentrations, NO derived from nNOS but not eNOS is involved. In addition, the greater inhibitory effect of L-NAME at 0.4% halothane suggests a role for eNOS at this concentration. These findings agree with those of Okamoto et al. , 9who suggested that cerebral hyperemia at low concentrations of isoflurane is modulated by eNOS. However, our results contrast with those of Koenig et al., 7who suggested a more important role for eNOS in isoflurane-induced pial arteriolar dilation. The cause of this discrepancy is unclear, but may be related to regional differences in vascular reactivity (intraparenchymal vs. pial arterioles), agent-specific differences (halothane vs. isoflurane), differences in the experimental model used (in vitro vs. in vivo microvessel diameter), and differences in the technique used to inhibit endothelium-dependent dilation (NOS inhibitor vs. endothelial injury).
It is important to interpret these results with caution. Although microvascular constriction by NOS inhibitors strongly suggests basal production of NO in our preparation, the relative baseline activities of eNOS and nNOS and the relative degree of inhibition by nonselective NOS inhibitors are unknown. An important stimulus for the synthesis of NO by eNOS in vivo is intraluminal blood flow and pressure, 36which are absent in our preparation. Nevertheless, agonist-induced release of NO occurs, as shown by acetylcholine-induced dilation inhibited by L-NA, 12L-NAME, or a period of hypoxia–reoxygenation. 37
Administration of NOS inhibitors in vivo causes cerebral vasoconstriction 6and decreased CBF. 3,8,9,38This may inhibit the response to anesthetics and other vasodilators because of a nonspecific vasoconstrictor effect, rather than because of a deficiency of NO. We addressed this difficulty in our study by ensuring the same total preconstriction in all three experimental groups. Smith et al. 8evaluated the role of NO in the cerebrocortical laser Doppler flow response to halothane in rats. They found that L-NAME significantly inhibited this response; however, the use of vasoactive agents to restore baseline cerebrocortical flow and mean arterial pressure in a subset of the L-NAME–treated rats resulted in hyperemia to halothane, which was not different from that observed during control conditions. 8The discrepancy between our findings and those of Smith et al. 8may reflect regional differences in vascular reactivity or methodological differences (vessel diameter vs. erythrocyte flow, absent vs. present intraluminal blood flow, no baseline anesthesia vs. baseline barbiturate anesthesia, use of vasoconstrictors vs. vasodilators to equalize vascular tone).
It has been suggested that NO may act, not as a classic mediator, but as a “permissive modulator 8” or provide a tonic, background effect on the cerebral vasculature in the presence of anesthetic agents. 38This means that, although volatile anesthetics do not directly stimulate NOS or cause an increased production or release of NO, 28,29the presence of a certain critical level of NO or cGMP may be necessary for maximal dilation. The finding that halothane continued to cause some vasodilation despite NOS inhibition supports the hypothesis that NO represents one of several mediators or modulators that are involved in volatile anesthetic-induced cerebral vasodilation. 1–3The residual anesthetic-induced dilation that was not inhibited by NOS inhibitors may be caused by incomplete acute NOS inhibition, but may also reflect direct smooth muscle effects, effects on local neuronal activity or neuronal-vascular coupling, stimulation of particulate (non–NO-dependent) GC, 26or involvement of other endothelial or parenchymal mediators, such as prostaglandins. 3Just as anesthetic agents may differ with respect to regional effects on CBF and metabolism, 3,11so too the mechanism of cerebrovascular effects may vary depending on the agent and region evaluated and the dosage of anesthetic used. 9Endothelium-dependent responses may exhibit wide regional heterogeneity, 32,39reflecting variations in the distribution of NOS, 22neural control of NO release, 40and sensitivity of vascular endothelial and smooth muscle cells. 39
In conclusion, halothane-induced dilation of intraparenchymal cerebral arterioles appears to depend, in part, on the synthesis of NO. These results suggest that the NO that contributes to this dilation is derived mainly from nNOS. The source of nNOS for regulating cerebrovascular tone in the presence of halothane may be perivascular or astrocytic, rather than brain parenchymal, because the latter may involve a greater diffusion distance. In this in vitro brain slice preparation, NO derived from eNOS does not appear to play an important role. Future studies using this model will investigate the effects of volatile anesthetic agents on intracerebral microvessels in nNOS gene-deficient mice.
The authors thank David A. Schwabe, Dale C. Ekbom, and Nicole Beauvais for expert technical assistance.