Materials and Methods

Materials.

Isoflurane was obtained from Abbott Laboratories (North Chicago, IL) and propofol (2,6-diisopropylphenol) from Sigma-Aldrich (St. Louis, MO). Tetrodotoxin, riluzole, and 4-aminopyridine (4AP) were from Sigma-Aldrich,l-[³H]glutamate (42 Ci/mmol) fromAmersham Biosciences UK, Ltd. (Buckinghamshire, UK), and [¹⁴C]GABA (0.24 Ci/mmol) from PerkinElmer Life Sciences (Boston, MA).

Synaptosome Preparation.

Experiments were done in accordance with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals as approved by the Weill Medical College of Cornell University Institutional Animal Care and Use Committee. Male Sprague-Dawley rats (250–350 g) were anesthetized with 80% CO₂/20% O₂ and sacrificed by cervical dislocation. The brain was rapidly removed and placed on ice. The cerebral cortex was removed and homogenized in 10 volumes of ice-cold 0.32 M sucrose with a motor-driven (500 rpm) Teflon-glass homogenizer for 10 strokes, and the homogenate was centrifuged for 2 min at 4,000g. Crude rat cerebrocortical synaptosomes (supernatant) were demyelinated by centrifugation through 0.8 M sucrose (10 ml of supernatant layered onto 10 ml of 0.8 M sucrose) for 30 min at 36,000g (Dodd et al., 1981). The pellet containing isolated nerve terminals was resuspended in ice-cold 0.32 M sucrose for use within 2 h.

Glutamate and GABA Release.

Demyelinated rat cerebrocortical synaptosomes were loaded with 8 nMl-[³H]glutamate and 440 nM [¹⁴C]GABA for 15 min at 30°C in Krebs-HEPES buffer (composition: 140 mM NaCl, 5 mM KCl, 20 mM HEPES, 1 mM MgCl₂, 1.2 mM Na₂HPO₄, 5 mM NaHCO₃, 0.1 mM EGTA, and 10 mMd-glucose, pH 7.4 with NaOH), pelleted by centrifugation (10 min at 20,000g at 4°C), and resuspended in ice-cold 0.32 M sucrose. For concentration-effect experiments, the prelabeling procedure was followed by a 5-min incubation with 0.2 mMd-aspartate to reduce cytoplasmicl-[³H]glutamate and, thus, Ca²⁺-independentl-[³H]glutamate release via reverse transport (Lester et al., 1994). Synaptosomes were confined by Whatman GF/B glass fiber filter disks (Maidstone, UK) and superfused at 0.5 ml/min with Krebs-HEPES buffer (initially bubbled with 95% O₂/5% CO₂ for 10 min) at 36°C using a customized (Diagram FD1) Brandel SF12 superfusion apparatus (Gaithersburg, MD) set to collect 1-min fractions. Release was induced by pulses of either 30 mM K⁺ (with KCl replacing equivalent NaCl) for 3 min or 1 mM 4AP for 2 min. All pulses were induced in the presence (addition of 2 mM CaCl₂ yielding final free [Ca²⁺] of 1.9 mM; MaxChelator v2.10;http://www.stanford.edu/∼cpatton/maxc.html) or absence of Ca²⁺. At the end of each experiment, synaptosomes were lysed with 0.2 M perchloric acid, and radioactivity in the synaptosomes and each fraction was quantified by liquid scintillation spectrometry with dual isotope quench correction (Beckman Coulter LS 6000IC; Beckman Coulter, Inc., Fullerton, CA) using BioSafe II scintillation cocktail (RPI, Mt. Prospect, IL). The accuracy of the dual radiolabel assay has been verified previously in detecting changes in the release of each amino acid independently of the other (Westphalen and Hemmings, 2003).

• Download figure
• Open in new tab
• Download powerpoint

Figure FD1

Modified superfusion system integrating a “closed” design for use with volatile anesthetics. Isoflurane or isoflurane with secretagogue are contained in separate valved glass syringes such that the superfusate can be switched between three solutions.

Application of Anesthetics.

Isoflurane was added to glass syringes by dilution of saturated solutions (10–12 mM) in Krebs-HEPES buffer prepared 12 to 24 h before use. Propofol was added from concentrated solutions prepared in dimethyl sulfoxide and diluted in Krebs-HEPES buffer in glass tubes before addition to assay [maximum final [dimethyl sulfoxide] < 0.5% (v/v)]. The aqueous isoflurane concentration for initial experiments (1 mM) was approximately 3 times minimum alveolar concentration, the EC₅₀for suppression of movement in response to a painful stimulus (Taheri et al., 1991). The propofol concentration (>12 μM) was approximately 5 times the EC₅₀ for loss of righting reflex (Tonner et al., 1992). Isoflurane or propofol were introduced through Teflon tubing from glass syringes (closed system) or glass tubes (open system), respectively. Anesthetics were added 3 min before, during, and 1 min after stimulation to insure the presence of anesthetic throughout the concurrent release pulse. In parallel experiments, isoflurane and propofol concentrations exiting the synaptosome chamber in the superfusate were determined by gas chromatography (Ratnakumari and Hemmings, 1998) or high-performance liquid chromatography (Lingamaneni et al., 2001), respectively.

Data Analysis.

Release in each fraction was expressed as a fraction of synaptosomal content of labeled transmitter before each fraction collected (fractional release). The magnitude of release pulses was determined by subtracting baseline release (average of basal release before and after pulse) from cumulative fractional release values of the release pulse (sum FR; Garcia-Sanz et al., 2001). For analysis, sum FR data from each experiment were normalized to mean control release in the presence or absence of Ca²⁺. Concentration-effect data are expressed as a percentage of the within experiment control in the presence of Ca²⁺. Ca²⁺-dependent release is defined as release in the presence of Ca²⁺ minus release in the absence of Ca²⁺.

Mean sum FR values all followed Gaussian distributions, with some significantly differing in variance, as determined by a modification of the method of Kolmogorov and Smirnov and Bartlett's test, respectively. Concentration-effect data were fitted by least-squares analysis to estimate I_max, IC₅₀, and Hill slope with standard errors (Prism v. 3.02; GraphPad Software, San Diego CA). Significant differences between mean sum FR values and between inhibition curve parameters were determined by an unpaired t test with Welch correction for variances that were not assumed to be equal (Instat v. 3.0a; GraphPad Software).

Results

4-Aminopyridine-Evoked Release.

Stimulating preloaded rat cerebrocortical synaptosomes with 1 mM 4AP (Tibbs et al., 1989) induced Ca²⁺-independent release ofl-[³H]glutamate and [¹⁴C]GABA. In the presence of 1.9 mM free Ca²⁺, total release increased 2.4-fold (P < 0.001) and 3.7-fold (P < 0.001) over Ca²⁺-independent release, respectively (Fig.1A). In the absence of Ca²⁺, the release of glutamate and GABA occurs primarily by carrier-mediated reverse transport of cytoplasmic transmitter (Lester et al., 1994); the addition of Ca²⁺ triggers vesicular release, significantly increasing total release (Südhof, 1995). In the presence of Ca²⁺, 4AP evoked significantly (P< 0.001) more [¹⁴C]GABA thanl-[³H]glutamate.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Release of l-[³H]glutamate and [¹⁴C]GABA from rat cerebrocortical synaptosomes preloaded with l-[³H]glutamate and [¹⁴C]GABA in the presence (filled symbols) or absence (open symbols) of 1.9 mM free Ca²⁺. A, 2-min pulse of 1 mM 4AP (n = 12–18). B, 3-min pulse of 30 mM K⁺ (n = 20–26). C, 2-min pulse of 1 mM 4AP with (n = 7) or without (n= 7) isoflurane in the presence of 1.9 mM free Ca²⁺. Concentrations of isoflurane were measured from the synaptosome chamber. Application bars indicate secretagogue or anesthetic entry into synaptosome chamber; lag in measurements represents chamber dead space. Anesthetic concentrations reported under Resultsare the average of values at 28 and 29 min. Data are shown as the mean ± S.E.M.

Neither isoflurane (1.07 ± 0.04 mM) nor propofol (12.6 ± 3.0 μM) affected Ca²⁺-independent 4AP-evoked release of l-[³H]glutamate and [¹⁴C]GABA (Fig.2A). In the presence of Ca²⁺, isoflurane inhibited 4AP-evokedl-[³H]glutamate (P< 0.001) and [¹⁴C]GABA (P = 0.002) release (Fig. 2A). Propofol also inhibited 4AP-evokedl-[³H]glutamate (P = 0.010) and [¹⁴C]GABA (P = 0.024) release in the presence of Ca²⁺ (Fig. 2A).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Effects of isoflurane, propofol, and tetrodotoxin on 4AP-evoked (A), K⁺-evoked (B), and basal (C) release ofl-[³H]glutamate and [¹⁴C]GABA. Isoflurane (1.07 ± 0.04, 1.08 ± 0.04, 1.03 ± 0.13 mM, respectively), propofol (12.6 ± 3.0, 17.0 ± 0.5, 13.5 ± 1.2 μM, respectively), or tetrodotoxin (2 μM) were present during stimulation of release by 1 mM 4AP or 30 mM K⁺. Data are shown as the mean ± S.E.M. Statistical comparisons between experimental groups (★, P < 0.05; ★★,P < 0.01; ★★★, P < 0.001) or between amino acids (†, P <0.05; ††,P <0.01; †††, P <0.001) employed unpaired t tests (see Results for definedP values).

In the presence of Ca²⁺, isoflurane maximally inhibited (I_max)l-[³H]glutamate release to a significantly greater extent than [¹⁴C]GABA release (P = 0.023), with no significant difference in IC₅₀ values (Fig. 3A; Table1). Isoflurane completely inhibited Ca²⁺-dependentl-[³H]glutamate release but inhibited Ca²⁺-dependent [¹⁴C]GABA release by only 74%, as determined by the subtraction from maximum inhibition in the presence of Ca²⁺ of that in the absence of Ca²⁺. In the absence of Ca²⁺, there was no concentration-dependent effect of isoflurane onl-[³H]glutamate or [¹⁴C]GABA release (Fig. 3A).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Concentration-dependent effects of isoflurane (A), propofol (B), tetrodotoxin (C), and riluzole (D) on 1 mM 4AP-evoked release of l-[³H]glutamate and [¹⁴C]GABA. Data are shown as the percentage of control in the presence of Ca²⁺. Ca²⁺-independent controls are shown as the mean ± S.E.M. (n = 3–5). Parameters of fitted curves are presented in Table 1.

An apparent greater potency of propofol inhibition of Ca²⁺-dependentl-[³H]glutamate release than [¹⁴C]GABA release was not significant (P = 0.21), with no difference in efficacy (Fig. 3B). Inhibition of Ca²⁺-independent release of bothl-[³H]glutamate and [¹⁴C]GABA by propofol was also not significant (Fig. 3B).

Inhibition of 4AP-evoked release ofl-[³H]glutamate and [¹⁴C]GABA by tetrodotoxin or riluzole (Figs. 3, C and D) paralleled the effects of isoflurane. That is, tetrodotoxin (P < 0.001) or riluzole (P = 0.041) inhibited Ca²⁺-dependentl-[³H]glutamate release to a greater extent than [¹⁴C]GABA release, with no difference in IC₅₀ values (P = 0.26 and P = 0.71, respectively). Tetrodotoxin also inhibited Ca²⁺-independent release of bothl-[³H]glutamate and [¹⁴C]GABA (Figs. 2A and 3C), a finding reproduced by riluzole (Fig. 3D).

Potency for inhibition of Ca²⁺-independent versus Ca²⁺-dependent amino acid release differed for some drugs (Table 1). This disparity in potency probably reflects the different mechanisms associated with Ca²⁺-independent (reverse transport) and Ca²⁺-dependent (vesicular) release.

K⁺-Evoked Release.

In the absence of Ca²⁺, depolarization of preloaded synaptosomes by 30 mM K⁺ evoked the release ofl-[³H]glutamate and [¹⁴C]GABA. In the presence of 1.9 mM free Ca²⁺, release increased by 2.0-fold (P < 0.001) and 2.2-fold (P < 0.001) over Ca²⁺-independent release, respectively (Fig.1B). Neither isoflurane (1.08 ± 0.04 mM) nor propofol (17.0 ± 1.5 μM) affected K⁺-evoked release in the absence or presence of Ca²⁺ (Fig. 2B). Tetrodotoxin partially inhibited Ca²⁺-independent K⁺-evoked release ofl-[³H]glutamate (P = 0.010) and [¹⁴C]GABA (P = 0.033) and [¹⁴C]GABA release (P = 0.012) in the presence of Ca²⁺ (Fig. 2B).

Basal Release.

Addition of Ca²⁺ did not affect basal (unstimulated)l-[³H]glutamate release (P = 0.29) but increased basal [¹⁴C]GABA release (P = 0.017; Fig. 2C). Isoflurane (1.03 ± 0.13 mM) inhibited basall-[³H]glutamate release in the absence (P = 0.020) or presence (P = 0.14) of Ca²⁺ but significantly enhanced basal [¹⁴C]GABA release in the absence (P < 0.001) or presence (P = 0.017) of Ca²⁺ (Fig. 2C). Propofol enhanced basall-[³H]glutamate release (P = 0.031) but not basal [¹⁴C]GABA release (P = 0.71) in the presence of Ca²⁺ (Fig. 2C). Tetrodotoxin inhibited basall-[³H]glutamate release in the absence (P = 0.16) or presence (P = 0.18) of Ca²⁺ and significantly inhibited basal [¹⁴C]GABA release in the absence (P = 0.010) or presence (P = 0.006) of Ca²⁺ (Fig. 2C).

Equating Secretagogue Intensities.

The sum FR of [¹⁴C]GABA released by 1 mM 4AP in the presence of Ca²⁺ (0.11) was greater (P < 0.001) than that ofl-[³H]glutamate (0.055; Fig. 2A). To test the possibility that the lower sensitivity of [¹⁴C]GABA release to isoflurane was due to the greater stimulus intensity, the magnitude of [¹⁴C]GABA release was reduced to that approximatingl-[³H]glutamate release by using 0.1 mM 4AP (Fig. 4). The profile of inhibition of 0.1 mM 4AP-evoked [¹⁴C]GABA release by isoflurane was similar to that observed with 1 mM 4AP (Fig.5B; Table 1); maximum inhibition remained significantly lower (P < 0.001) than for 1 mM 4AP-evokedl-[³H]glutamate release (Fig. 3A), with no significant change in IC₅₀.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Concentration-dependent effects of 4AP and K⁺ on l-[³H]glutamate and [¹⁴C]GABA release in the presence of Ca²⁺. Data are shown as the mean ± S.E.M. (n = 4).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

A, effect of isoflurane (1.15 ± 0.04 mM) on the release of l-[³H]glutamate and [¹⁴C]GABA evoked by a reduced K⁺ stimulus (15 mM) in the absence or presence of Ca²⁺. Data are shown as the mean ± S.E.M. Statistical comparisons between experimental groups used unpaired t tests (★★,P < 0.01; see Results for definedP values). B, concentration-effect curves for inhibition by isoflurane of the release ofl-[³H]glutamate and [¹⁴C]GABA evoked by a reduced 4AP stimulus (0.1 mM) compared with data for 1 mM 4AP (from Fig. 3A).

The sum FR of controll-[³H]glutamate released by 1 mM 4AP (0.055; Fig. 2A) was also significantly less (P < 0.001) than that released by 30 mM K⁺ (0.12; Fig.2B) in the presence of Ca²⁺. This difference between K⁺-evoked and 4AP-evoked release was less marked for [¹⁴C]GABA release (0.13 versus 0.11;P = 0.056). To test the possibility that the lower sensitivity to isoflurane of K⁺-evokedl-[³H]glutamate release was due to the greater stimulus intensity, the magnitude of K⁺-evokedl-[³H]glutamate release was reduced to that released by 1 mM 4AP by using 15 mM K⁺ (Fig. 4). Isoflurane (1.15 ± 0.04 mM) also failed to inhibit 15 mM K⁺-evokedl-[³H]glutamate release (Fig. 5A). The reduced intensity of the K⁺stimulus also equalized [¹⁴C]GABA release with that evoked by 1 mM 4AP (Fig. 4), which was also unaffected by isoflurane in the presence of Ca²⁺ (Fig. 5A). In the absence of Ca²⁺, isoflurane slightly enhanced 15 mM K⁺-evoked [¹⁴C]GABA release (P < 0.001; Fig. 5A).

Isoflurane-Evoked Release.

Isoflurane (1.03 ± 0.13 mM) alone inhibited basal release ofl-[³H]glutamate but enhanced basal [¹⁴C]GABA release in the absence or presence of Ca²⁺ (Fig. 2C). At concentrations above 0.8 mM, isoflurane evoked [¹⁴C]GABA release (Fig.6). Above a threshold concentration of about 2 mM, isoflurane evokedl-[³H]glutamate release (Fig. 6). Statistical comparisons in the absence (2.09 ± 0.1 mM isoflurane;n = 6; Fig. 6A) or presence (2.07 ± 0.09 mM isoflurane; n = 6; Fig. 6B) of Ca²⁺ showed that isoflurane-evoked release of both l-[³H]glutamate (P = 0.079) and [¹⁴C]GABA (P = 0.68) release was Ca²⁺-independent. Isoflurane (2 mM) preferentially released [¹⁴C]GABA overl-[³H]glutamate in the absence (P = 0.0013) or presence (P = 0.015) of Ca²⁺. These effects of isoflurane were nonsaturable up to the highest concentration tested (2.4 mM).

• Download figure
• Open in new tab
• Download powerpoint

Figure 6

Concentration-dependent effects of isoflurane on basal release of l-[³H]glutamate and [¹⁴C]GABA in the absence (A) or presence (B) of 1.9 mM free Ca²⁺. Values for release evoked by 2 mM isoflurane are the mean ± S.E.M. (n = 6). Statistical comparisons between −Ca²⁺ and +Ca²⁺ (no significant differences) and betweenl-[³H]glutamate and [¹⁴C]GABA (†, P < 0.05; ††, P <0.01), determined by unpaired t tests, are shown. SeeResults for defined P values.

Discussion

The general anesthetics isoflurane and propofol significantly inhibited 4AP-evoked release of glutamate and GABA. Isoflurane preferentially inhibited 4AP-evoked glutamate release compared with GABA release over a range of clinical concentrations. This was due to greater efficacy rather than potency. The Na⁺channel blockers tetrodotoxin and riluzole also produced selective inhibition of 4AP-evoked release of glutamate versus GABA. This suggests that isoflurane, which inhibits neuronal voltage-gated Na⁺ channel currents (Rehberg et al., 1996;Rehberg and Duch, 1999; Lingameneni et al., 2000), selectively inhibits glutamate release via Na⁺ channel blockade (Tibbs et al., 1989). This conclusion is consistent with the recent finding that riluzole preferentially inhibits transmitter release from excitatory versus inhibitory neurons in hippocampal microcultures (Prakriya and Mennerick, 2000). The physiological difference(s) between glutamatergic and GABAergic nerve terminals that underlies this pharmacological selectivity remains to be determined.

In contrast to the secretagogue 4AP, neither isoflurane nor propofol affected glutamate or GABA release evoked by elevated K⁺. Previous studies support the conclusion that clinically relevant concentrations of volatile anesthetics do not affect K⁺-evoked excitatory amino acid release (Minchin, 1981; Kendall and Minchin, 1982; Lingamaneni et al., 2001) or GABA release (Minchin, 1981; Kendall and Minchin, 1982; Lecharney et al., 1995; Salford et al., 1997; Lingamaneni et al., 2001). Inhibition of K⁺-evoked release of glutamate (Miao et al., 1995; Larsen and Langmoen, 1998; Liachenko et al., 1999) or GABA (Larsen et al., 1998; Liachenko et al., 1999) by isoflurane has been reported. Interpretation of results obtained in brain slices (Liachenko et al., 1999), however, is confounded by indirect anesthetic effects mediated by postsynaptic GABA_A receptors (Buggy et al., 2000). Failure to compare anesthetic effects on release evoked by additional secretagogues (Lingamaneni et al., 2001) makes it difficult to interpret results obtained only with elevated KCl. Intravenous anesthetics, including propofol, do not affect K⁺-evoked glutamate (Lingamaneni et al., 2001) or GABA (Mantz et al., 1995; Lingamaneni et al., 2001) release from synaptosomes at clinical concentrations, a finding reproduced in this study. Greater sensitivity of 4AP-evoked versus K⁺-evoked glutamate release to volatile anesthetics and propofol has been reported in isolated nerve terminals prepared from several brain regions and species using a fluorometric assay of endogenous glutamate (Schlame and Hemmings, 1995; Lingamaneni et al., 2001). This differential sensitivity between secretagogues was not due to differences in stimulus intensity. When [K⁺]_o was reduced to 15 mM, which evoked glutamate and GABA release comparable to that evoked by 1 mM 4AP, isoflurane remained an ineffective inhibitor of K⁺-induced glutamate or GABA release.

Selective inhibition by isoflurane and propofol of 4AP-evoked over K⁺-evoked release implicates preferential blockade of voltage-gated Na⁺ channels over Ca²⁺ channels in their presynaptic actions. Repetitive depolarization due to blockade of A-type K⁺ channels by 4AP is amplified by activation of voltage-gated Na⁺ channels leading to the opening of voltage-gated Ca²⁺ channels coupled to transmitter release. Elevated K⁺ leads to synchronous activation of voltage-gated Ca²⁺channels followed by a plateau of residual Ca²⁺channel conductance (Tibbs et al., 1989). Blockade of Na⁺ channels is predicted to inhibit 4AP-evoked, but not K⁺-stimulated, transmitter release, as demonstrated in this study. In the presence of Ca²⁺, the Na⁺ channel blocker tetrodotoxin, like isoflurane, completely inhibited 4AP-evoked, but not K⁺-evoked, release. A role for Na⁺ channel blockade in the inhibition of glutamate release has been suggested previously for volatile anesthetics (Ratnakumari and Hemmings, 1998; Lingamaneni et al., 2001;Asai et al., 2002) and propofol (Ratnakumari and Hemmings, 1997).

Blockade of presynaptic voltage-gated Ca²⁺channels effectively inhibits evoked transmitter release by limiting Ca²⁺ influx (Wu and Saggau, 1997), which is essential for vesicular release (Südhof, 1995). Multiple Ca²⁺ channel subtypes can coexist in presynaptic terminals to regulate neurotransmitter release (Turner et al., 1993). Electrophysiological evidence indicates that certain voltage-gated Ca²⁺ channel subtypes are blocked by general anesthetics (Topf et al., 2003). In this study, anesthetics did not affect K⁺-evoked amino acid release, a process that involves Ca²⁺ influx primarily via voltage-gated Ca²⁺ channel opening. This may be explained by a high safety factor for inhibition of transmitter release by Ca²⁺ channel blockade or by reduced anesthetic sensitivity of the presynaptic Ca²⁺ channel subtypes coupled to transmitter release, possibly due to nerve terminal-specific modulation (Turner et al., 1993).

The Na⁺ channel blocker tetrodotoxin inhibited basal and Ca²⁺-independent K⁺-evoked glutamate and GABA release. Changes in the [Na⁺]_o/[Na⁺]_iratio alter Ca²⁺-independent reverse transport of transmitters that are coupled to Na⁺ cotransport (Lester et al., 1994). The disparity between inhibition of Ca²⁺-independent release by tetrodotoxin and by anesthetics may stem from the different potencies or efficacies by which they inhibit Na⁺ entry (Ratnakumari and Hemmings, 1997; Ratnakumari and Hemmings, 1998) and/or from anesthetic actions at other presynaptic targets (MacIver, 1997). Recent evidence suggests, however, that isoflurane or propofol do not directly affect presynaptic neuronal transporters of glutamate or GABA (Westphalen and Hemmings, 2003).

Isoflurane, like tetrodotoxin, inhibited basal glutamate release, apparently through Na⁺ channel blockade. In contrast, isoflurane enhanced basal GABA release, particularly at higher concentrations (>1.5 mM), whereas tetrodotoxin was inhibitory. This stimulatory effect of isoflurane occurred at higher than clinical concentrations (Taheri et al., 1991), was not saturable over the concentration range studied, and was Ca²⁺-independent. General anesthetics also evoke the release of norepinephrine (Pashkov and Hemmings, 2002), whereas halothane produces an increase in mini-excitatory postsynaptic potential frequency in a rat hippocampal slice preparation (Nishikawa and MacIver, 2000). These observations suggest that volatile anesthetics may stimulate a low level of spontaneous Ca²⁺-independent vesicular release from certain terminals by a mechanism resistant to inhibition by Na⁺ channel blockade. The mechanisms underlying Ca²⁺-independent vesicular transmitter release and biochemical differences between glutamatergic and GABAergic terminals remain unclear; whether these mechanisms are directly affected by anesthetics warrants further investigation.

Significantly more GABA was released compared with glutamate for a given concentration of 4AP (Tapia and Sitges, 1982). When the level of 4AP-evoked GABA release was equalized with that of glutamate by reducing the concentration of 4AP, selective inhibition of glutamate release by isoflurane was maintained, thus eliminating a difference in secretagogue intensity as a cause. Preferential stimulation of basal GABA release by isoflurane alone could also contribute to the lower maximal inhibition of evoked GABA release compared with that of glutamate. Selective inhibition of 4AP-evoked glutamate versus GABA release, however, was also produced by other Na⁺channel blockers that are not known to stimulate basal release.

Selective inhibition of amino acid release by general anesthetics and Na⁺ channel blockers and differential sensitivity to the secretagogue 4AP suggest fundamental physiological differences between glutamatergic and GABAergic nerve terminals. Mounting evidence supports the notion of distinct patterns of presynaptic ion channel distributions within and between neurons by demonstrating selective densities and/or function of K⁺ channels (Veh et al., 1995; Southan and Robertson, 1998) and Na⁺channels (Stuart et al., 1997; Martina et al., 2000). Differential involvement of a presynaptic phorbol ester receptor (Munc13-1) between glutamate and GABA terminals has also been reported (Augustin et al., 1999). Our findings, and those of Prakriya and Mennerick (2000), demonstrate the pharmacological implications of such transmitter-specific presynaptic specialization.

Inhibition of excitatory amino acid transmitter release, which may be enhanced by possible blockade of postsynaptic glutamate receptors (Perouansky and Antognini, 2003), appears to be an important mechanism of neuronal depression by clinical concentrations of volatile anesthetics. Concurrently, partial inhibition of inhibitory amino acid transmitter release may be balanced by potentiation of postsynaptic GABA_A receptors. The presynaptic effects of the intravenous anesthetic propofol are less potent in relation to its clinically relevant concentrations; its marked effects on GABA_A receptors (Hales and Lambert, 1991) may play a more important role in producing anesthesia. For isoflurane, selective depression of glutamate release, stimulation of spontaneous GABA release, and potentiation of postsynaptic GABA_A receptors provide complementary actions to depress excitatory and enhance inhibitory central nervous system transmission.