Inhibition by Volatile Anesthetics of Endogenous Glutamate Release from Synaptosomes by a Presynaptic Mechanism

After institutional approval, synaptosomes were prepared by a modification [8]of the method of Dunkley et al. [4]Adult male (150 g) Sprague-Dawley rats were obtunded with CO²and killed by decapitation. Brains were removed and immediately chilled in ice-cold sucrose solution (0.32 M sucrose, 1 mM ethylenediamine tetraacetic acid, pH 7.5). Cerebral cortices from two brains were removed and homogenized in 20 ml sucrose solution using a motor-driven polytetrafluorethylene-glass homogenizer (Potter-Elvehjem) at 900 rpm for ten up-and-down strokes. The homogenate was centrifuged at 1,000g for 10 min, and the pellet was re-homogenized in 20 ml sucrose solution and centrifuged at 1,000g for 10 min. The supernatant fractions from the two centrifugations were combined and centrifuged at 15,000g for 25 min. The resulting pellet was washed and resuspended in 8 ml sucrose solution (P²fraction). Aliquots (2 ml) of the P²fraction were loaded onto discontinuous gradients consisting of three 2.5-ml layers of filtered Percoll density gradient medium (23%, 10%, and 3%) in sucrose solution plus 0.25 mM dithiothreitol. The gradients were centrifuged at 25,000g for 6 min. The synaptosomal fraction was collected from the 23/10% Percoll interface and diluted about fivefold in buffered saline-glucose (140 mM NaCl, 5 mM KCl, 5 mM NaHCO³, 1 mM MgCl², 10 mM glucose, and 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, pH 7.4), which was equilibrated with 95% Oxygen²/5% CO². The synaptosomes were centrifuged at 23,000g for 10 min and resuspended in 10 ml buffered saline-glucose. The protein concentration of the synaptosomal preparation was determined by the method of Bradford [9]using bovine serum albumin as a standard. The preparation was divided into aliquots consisting of 1 mg protein and centrifuged at 10,000g for 7 min. The synaptosomes were stored as pellets on ice for as long as 4 h before use.

Endogenous glutamate release was measured by a modification of the continuous fluorometric assay described by Nicholls et al, [5]In this system, released glutamate is oxidized by L-glutamate dehydrogenase, which is coupled to the reduction of the oxidized form of nicotinamide adenine dinucleotide phosphate (NADP sup +) to NADPH. NADPH gives rise to a strong fluorescence signal at 510 nm when excited at 340 nm. This method allows the detection of glutamate at sub-micromolar concentrations. An inherent delay is introduced by the time required for the reaction to occur, so that the traces do not reflect the actual glutamate concentration. [10,11]Corrections for this delay were not incorporated into the assay and consequently the initial rates of glutamate release associated with depolarization may not be accurately reflected.

The synaptosome pellets (corresponding to 1 mg protein) were resuspended in 3 ml incubation medium (122 mM NaCl, 3.1 mM KCl, 0.4 mM KH²PO⁴, 5 mM NaHCO³, 1.2 mM MgSO⁴, 20 mM N-tris[hydroxymethyl]-methyl-2-aminoethanesulfonic acid, 10 mM D-glucose, 16 micro Meter bovine serum albumin (essentially free of fatty acid), 1 mM NADP sup +, 100 U L-glutamate dehydrogenase, 1.3 mM CaCl², pH 7.4 with NaOH). The sample was placed in a cuvette in a fluorescence spectrofluorometer (SLM Aminco 8000, SLM Instruments, Inc., Urbana, IL) equipped with a magnetic stirrer and a temperature-regulated (37 degrees Celsius) cuvette holder. The stirred samples were equilibrated at 37 degrees Celsius for 4 min, and then data acquisition was started. The excitation wavelength was 340 nm, and the monitored emission wavelength was 510 nm.

After recording basal glutamate release (which was Calcium sup 2+ -independent and is the result of nonvesicular leakage [3]), a secretogogue was added. The kinetics of glutamate release evoked by 4-AP or KCl are biphasic, with a rapid phase complete within 2 s and a more extensive slow phase. [10]The rate of glutamate release from 0-60 s, which was essentially linear with respect to time, was measured and corrected for the stable basal release rate measured before secretogogue addition. The fluorescence signal was calibrated by the addition of 10 nmol L-glutamate to a reaction mixture without synaptosomes or to the reaction mixture at the end of an experiment once release had reached a stable rate. The latter method required considerably more time to perform and therefore was not used routinely, to minimize the time from preparation of synaptosomes to completion of the release assay. The addition to the assay of 0.1% (volume in volume) dimethyl sulfoxide, which was used as a drug vehicle, had no effect on basal glutamate release or on glutamate release evoked by 4-AP, veratridine, KCl, or ionomycin (data not shown).

Synaptosomes (corresponding to 10 mg protein) were resuspended in 5 ml buffered saline-glucose containing 5 micro Meter fura-2 acetoxymethyl ester and incubated at 37 degrees Celsius for 40 min. The mixture was divided into 1-ml aliquots, which were centrifuged at 10,000g for 7 min. The pellets were washed with dye-free saline-glucose and were stored on ice until use. For free intrasynaptosomal [Calcium²+] measurements, synaptosomal pellets were resuspended in 3 ml incubation medium and incubated in a stirred quartz cuvette at 37 degrees Celsius in the absence or presence of halothane for 5 min followed by the addition of 1 mM 4-AP. Synaptosomal [Calcium²+] was calculated from the ratio of the fluorescence intensity of fura-2 at an emission wavelength of 510 nm with excitation wavelengths of 340 and 380 nm (switched every 2 s). The signal ratio was converted into free intrasynaptosomal [Calcium²+] using software (SLM Instruments, Inc.) based on the method of Grynkiewicz et al. [12]The maximal ratio (saturating [Calcium²+]) was obtained by lysing the synaptosomes with 6.2 mM Triton X-100, and the minimal ratio (0 [Calcium²+]) was obtained by complexing free Calcium²+ with 7.7 mM ethyleneglycol-bis-(beta-aminoethyl ether) N,N,N',N'-tetraacetic acid.

Volatile anesthetics were added to cuvettes by adding aliquots of saturated water solutions. Immediately after the addition of anesthetic, cuvettes were sealed with a polytetrafluorethylene stopper and wrapped with Parafilm. Subsequent additions were made through a hole in the stopper that just accommodated the needle of a 10-micro liter Hamilton syringe. Actual anesthetic concentrations in the cuvette were determined by gas chromatography. [13]After equilibration, a 0.5-ml aliquot of the synaptosome solution was withdrawn from the cuvette with a gas-tight syringe and extracted with 1 ml n-heptane. The n-heptane extract (7 micro liter) was injected into a gas chromatograph (GC-8A, Shimadzu Corp., Kyoto, Japan) equipped with a thermal conductivity detector. Separation was achieved on a 1.8 m/6 mm ID glass column packed with Porapak Q (Supelco, Bellefonte, PA). The column temperature was 210 degrees Celsius, the injector temperature was 230 degrees Celsius and carrier gas (Helium) flow was 40 ml *symbol* min sup -1.

NADP sup +, NADPH, L-glutamate, L-glutamate dehydrogenase (Proteus sp), bovine serum albumin (essentially fatty acid free), Sodium (plus/minus)-pentobarbital, 4-AP, ionomycin, tetrodotoxin, veratridine, fura-2 acetoxymethyl ester, and dimethyl sulfoxide were obtained from Sigma (St. Louis, MO). Percoll density gradient medium (15-30-nm diameter silica particles coated with nondialyzable polyvinylpyrrolidone) was obtained from Pharmacia (Uppsala, Sweden) and was filtered through a 0.45-micro meter filter before use. beta-Phorbol 12, 13-dibutyrate (PDBu) was from LC Laboratories (Woburn, MA). Enflurane and isoflurane were obtained from Anaquest (Madison, WI) and halothane (thymol-free) from Halocarbon Products (North Augusta, SC). Ro 31-8220 was kindly provided by Roche Products (Hertfordshire, UK).

Statistical differences between mean control and experimental values were determined by Student's unpaired t test (for single test agents) or by analysis of variance with the Newman-Keuls multiple-range test (for multiple test agents) (PHARM/PCS Pharmacologic Calculation System, version 4.2, Springer-Verlag, New York, NY). [14]Concentration-effect data were analyzed using a graded dose-response program that carries out linear regression analysis on data between 20% and 80% of the maximal response (PHARM/PCS Pharmacologic Calculation System, version 4.2). [14].

Glutamate release was monitored by continuous fluorometry in the presence of added NADP sup + and L-glutamate dehydrogenase. [5]The following control experiments demonstrated the suitability of this system for studying the effects of anesthetics on neurotransmitter release (data not shown). (1) The fluorescence yield of the assay was linearly proportional to the amount of glutamate or the amount of NADPH added to the incubation mixture. (2) The release of glutamate from synaptosomes stimulated by KCl-induced depolarization was proportional to the amount of synaptosomal protein present in the assay to 2 mg. (3) Halothane (1 mM) did not affect the activity of the coupling enzyme L-glutamate dehydrogenase and did not interfere with the fluorescence yield of NADPH in our assay conditions; a previous investigation demonstrating inhibition of L-glutamate dehydrogenase by halothane [15]used suprapharmacologic concentrations (20 vol%, or [nearly equal] 6 mM).

Synaptosomal glutamate release can be evoked by 4-AP, a Potassium sup + channel blocker that destabilizes membrane potential and causes repetitive spontaneous Sodium sup + channel-dependent depolarizations that closely mimic electric stimulation. [3,6]The addition of 4-AP to synaptosomes resulted in a marked stimulation of glutamate release that was Calcium²+ -dependent (Figure 1). The ability of 4-AP to evoke glutamate release was potentiated by the phorbol ester PDBu (1.37 plus/minus 0.11 nmol *symbol* min sup -1 *symbol* mg sup -1 without versus 2.43 plus/minus 0.22 nmol *symbol* min sup -1 *symbol* mg sup -1 with 1 micro Meter PDBu; P < 0.01; n = 3). The rate of formation of NADPH, which is proportional to the amount of glutamate released, is essentially linear over the first 60 s period in these assay conditions. The rate of Calcium²+ -dependent glutamate release during this interval (early phase) can be quantified and used to screen for pharmacologic effects on this phase of release. After about 2 min, the rate of glutamate release stabilized at a rate somewhat greater than the basal rate; this release is Calcium²+ independent and is probably attributable to reversal of the Sodium sup + /glutamate transporter. [3].

The early phase of 4-AP-evoked glutamate release was inhibited by halothane in a dose-dependent manner (Figure 2). However, halothane did not affect the slow rate of Calcium²+ -independent basal glutamate release (0.28 plus/minus 0.07 nmol *symbol* min sup -1 *symbol* mg sup -1 without versus 0.24 plus/minus 0.03 nmol *symbol* min sup -1 *symbol* mg sup -1 with 0.9 mM halothane). Halothane inhibition of 4-AP-evoked release was noncompetitive with respect to 4-AP concentration (data not shown). A plot of initial glutamate release rate versus halothane concentration (Figure 3) showed that halothane inhibited glutamate release in the absence or presence of PDBu. The concentration of halothane resulting in 50% inhibition (IC⁵⁰) of 4-AP-evoked glutamate release was 0.5 plus/minus 0.2 mM. Halothane remained an effective inhibitor of 4-AP-evoked glutamate release despite selective inhibition of protein kinase C by 3.3 micro Meter Ro 31-8220 (compound 3 in Davis et al. [16]) (Figure 4).

Glutamate release evoked by 4-AP requires Sodium sup + channel activation, and is sensitive to inhibition by the specific Sodium sup + channel blocker tetrodotoxin. [6]Halothane caused no further inhibition of the residual 4-AP-evoked glutamate release in the presence of 0.3 micro Meter tetrodotoxin, a maximally effective concentration (tetrodotoxin-resistant glutamate release: 0.16 plus/minus 0.02 nmol *symbol* min sup -1 *symbol* mg sup -1 without halothane; 0.15 plus/minus 0.02 nmol *symbol* min sup -1 *symbol* mg sup -1 with 0.9 mM halothane; n = 3).

Enflurane and isoflurane also inhibited 4-AP-evoked glutamate release. As with halothane, these anesthetics were effective at aqueous concentrations that correspond to 1 or 2 MAC (Table 1). In contrast to the volatile anesthetics, a clinically effective concentration of the intravenous anesthetic pentobarbital did not significantly inhibit 4-AP-evoked glutamate release (Table 1).

Veratridine, a neurotoxin that directly activates Sodium sup + channels, also depolarizes synaptosomes and evokes neurotransmitter release. [3]Veratridine-evoked glutamate release, like 4-AP-evoked glutamate release, was significantly inhibited by 0.9 mM halothane (Figure 5), and by lower concentrations of halothane (data not shown).

Depolarization of synaptosomes with 10 mM KCl results in submaximal glutamate release and increase in intracellular free [Calcium sup 2+], whereas 30 mM KCl produces essentially maximal glutamate release and increase in [Calcium²+]. [10]Depolarization of synaptosomes by 10 or 30 mM external KCl evoked glutamate release that was not inhibited significantly by halothane in the absence (Table 2) or presence (data not shown) of 1 micro Meter PDBu. The early phase of KCl-evoked glutamate release in synaptosomes was not affected by 0.65 or 0.9 mM halothane (Table 2), whereas the second, slower phase of glutamate release was slightly greater in the presence of 0.9 mM halothane (data not shown). This second phase is thought to result from reversal of electrogenic Sodium sup + -coupled glutamate transport as a result of prolonged depolarization and does not involve vesicular release. [3].

Ionomycin is an ionophore that evokes vesicular glutamate release by permitting Calcium²+ influx into synaptosomes. Halothane significantly stimulated Calcium²+ -dependent ionomycin-evoked glutamate release (Figure 6and Table 2).

To determine whether the effects of halothane on glutamate release were mediated by changes in [Calcium²+], intrasynaptosomal free [Calcium²+] was measured in conditions similar to those used to determine glutamate release. Halothane did not affect resting intrasynaptosomal [Calcium²+] before or after addition of CaCl²to the external medium (Table 3). Halothane significantly depressed the increase in the plateau phase of intrasynaptosomal [Calcium²+] induced by 4-AP in the absence (79% inhibition) or presence (73% inhibition) of PDBu (Figure 7and Table 3). A less marked effect of halothane was observed on the KCl-induced increase in intrasynaptosomal [Calcium²+] (40% inhibition; Table 3). In contrast, halothane potentiated the increase in [Calcium²+] induced by ionomycin by 3.5-fold (Table 3).

This study demonstrates a direct inhibitory effect of volatile halogenated anesthetics at clinically relevant concentrations on the release of the excitatory neurotransmitter glutamate evoked by 4-AP. Inhibition of the release of glutamate, the major excitatory neurotransmitter in the mammalian central nervous system, is consistent with the generalized neuronal depression characteristic of general anesthesia. [1]Our results suggest that depression of glutamate release may not be a universal anesthetic mechanism, however, because the barbiturate pentobarbital, a structurally unrelated anesthetic compound, did not inhibit 4-AP-evoked glutamate release at anesthetic concentrations. Inhibition of glutamate release, in various combinations with postsynaptic glutamate receptor antagonism, [17]gamma-aminobutyric acid A receptor potentiation, [18]and Calcium²+ channel antagonism, [19]may contribute to the molecular mechanisms of volatile anesthetic action, depending on the specific anesthetic agent and synaptic physiologic characteristics.

The basal rate of glutamate release before secretogogue addition was not affected by halothane in this study. A previous study using mouse cortical synaptosomes demonstrated a 14% increase in basal glutamate release by 0.75 mM halothane at 32 degrees Celsius during a 15-min assay. [20]However, the interpretation of these results is complicated by the use of such a long assay interval and by the low assay temperature.

The mechanism of the inhibitory effect of halothane on glutamate release was investigated in more detail. In general, exocytotic neurotransmitter release from synaptic vesicles involves multiple steps after invasion of the presynaptic nerve terminal by an action potential. [21]The plasma membrane depolarizes as a result of voltage-dependent Sodium sup + channel activation. This elicits Calcium sup 2+ entry through voltage-dependent Calcium²+ channels. Presynaptic Calcium²+ entry is then coupled to the exocytosis of synaptic vesicles. Each of these steps is subject to modulation by cellular regulatory mechanisms, such as protein phosphorylation, [22]and by specific drugs and neurotoxins. [3]Heterogeneity in presynaptic mechanisms, including Calcium²+ -secretion coupling, has been demonstrated in the differential release of glutamate, cholecystokinin-8 and norepinephrine from synaptosomes in response to varying [Calcium²+]. [23]Any of these physiologic steps involved in neurotransmitter release are potential targets for presynaptic anesthetic effects.

Because volatile anesthetics have been found to affect the activity of purified protein kinase C in vitro, [24,25]the role of protein kinase C in the inhibition of glutamate release by halothane was examined. The protein kinase C activator PDBu potentiates 4-AP-evoked glutamate release from synaptosomes by a mechanism that is thought to involve facilitation of membrane depolarization by phosphorylation and inhibition of Potassium sup + channels. [7]However, 4-AP-evoked glutamate release was sensitive to inhibition by halothane even in the absence of PDBu. Furthermore, halothane remained an effective inhibitor of glutamate release in the presence of Ro 31-8220, a potent and selective inhibitor of protein kinase C. [14]These data indicate that the inhibition of glutamate release by halothane is not attributable to inhibition of protein kinase C.

Halothane did not inhibit glutamate release significantly when evoked by direct synaptosome depolarization with KCl, which circumvents 4-AP-induced Potassium sup + channel blockade and Sodium sup + channel-mediated depolarization, or by ionomycin, a Calcium²+ ionophore that circumvents voltage-dependent Calcium²+ channels. In fact, ionomycin-evoked glutamate release was enhanced by halothane, an effect that may involve a direct interaction between halothane and ionomycin to allow more Calcium²+ influx. These findings suggest that the volatile anesthetic-sensitive step(s) in neurotransmitter release involve(s) membrane depolarization steps proximal to Calcium²+ influx. The observation that ionomycin-evoked glutamate release is not inhibited by halothane suggests that Calcium²+ -secretion coupling or other aspects of synaptic vesicle exocytosis are not important targets for the inhibition of 4-AP-evoked glutamate release by halothane.

Although neuronal voltage-dependent Calcium²+ channels are inhibited by volatile anesthetics (at somewhat higher concentrations than those used here), [18]inhibition of Calcium²+ influx by this mechanism is apparently not involved in the effect of halothane on 4-AP-evoked glutamate release, because KCl-evoked glutamate release (which also depends on Calcium²+ influx through voltage-dependent Calcium²+ channels) was insensitive to halothane in our assay. These data also implicate a target proximal to Calcium²+ influx and imply that synaptic vesicle exocytotic mechanisms are relatively insensitive to volatile anesthetic effects. Experiments with electro-permeabilized adrenal chromaffin cells have also demonstrated insensitivity of Calcium²+ -secretion coupling to volatile anesthetic effects. [26].

Although halothane did not significantly inhibit KCl-evoked glutamate release as measured in this study, halothane inhibited both 4-AP- and KCl-evoked increase in [Calcium²+] in synaptosomes. The resting intracellular synaptosomal [Calcium²+] measured in this study was 370 nM. This is close to the value of 350 nM reported by Komulainen and Bondy, [27]intermediate between the values of 485 nM and 263 nM reported for light and heavy Percoll synaptosome fractions, respectively, by Verhage et al., [28]and higher than the values near 200 nM reported by others. [10,29]The higher values obtained in this study are probably attributable to variations in the preparation and fura-2 loading of the synaptosomes or the use of a greater extracellular [Calcium²+] compared with those of the other studies and may reflect the presence of light synaptosomes not containing mitochondria in our preparation. [28]Despite the greater basal [Calcium²+], marked secretogogue-induced increases in [Calcium²+] were observed that are comparable to those reported by others. [10,27]The observation that halothane inhibited the increase in [Calcium²+] produced by depolarization with 30 mM KCl suggests that halothane is capable of inhibiting voltage-dependent Calcium²+ channels in synaptosomes. This effect was not likely mediated by an effect on the Sodium sup + /Calcium²+ exchanger, the Sodium sup +, Potassium sup + adenosine triphosphatase, or the Calcium²+ adenosine triphosphatase, because basal intrasynaptosomal [Calcium²+] was not affected.

The observation that halothane is more effective in the inhibition of the 4-AP-evoked rather than the KCl-evoked intrasynaptosomal [Calcium²+] increase despite comparable secretogogue-induced [Calcium²+] increases indicates different mechanisms for the halothane effects, and argues against the sole involvement of Calcium²+ channel blockade in the effect of halothane on 4-AP-evoked release. The absence of a significant effect of halothane on KCl-evoked glutamate release, despite its inhibition of KCl-evoked [Calcium²+] increase, suggests that the synaptosomal KCl-activated Calcium²+ channels inhibited by halothane may not be closely coupled to glutamate release. Two pertinent observations support this interpretation. First, the N-type Calcium²+ channel antagonist omega-conotoxin GVIA inhibits KCl-evoked increase in [Calcium²+] in synaptosomes by 20% [30]without affecting glutamate release. [29]Second, P-type Calcium²+ channels appear to be relatively insensitive to inhibition by volatile anesthetics, [31]and may represent the noninactivating voltage-dependent Calcium²+ channels that are coupled to synaptosomal glutamate release. [29].

Glutamate release evoked by 4-AP- or veratridine-induced depolarization of synaptosomes requires the action of Sodium sup + channels. [3]Because 4-AP- or veratridine-evoked glutamate release is sensitive to inhibition by halothane, whereas KCl- or ionomycin-evoked glutamate release is not, presynaptic Sodium sup + channels represent a potential site of halothane action. Previous studies have shown that volatile anesthetics inhibit axonal conduction, [32]axonal firing threshold, [33]and axonal Sodium sup + channel gating and conductance [34-36]; however, these effects generally occurred at higher anesthetic concentrations than anesthetic effects on synaptic transmission. [1]Our results suggest that the Sodium sup + channels present in the presynaptic nerve terminal may be more sensitive to volatile anesthetic effects than Sodium sup + channels present in axons, possibly because of a different isozymic composition. [37]These data are consistent with the recent finding that central nervous system Sodium sup + channels are suppressed by clinically relevant concentrations of general anesthetics. [38]Further studies are required to determine the precise mechanism(s) of volatile anesthetic inhibition of synaptosomal glutamate release, because only indirect conclusions regarding the mechanism(s) involved can be drawn from the present data.

Several methodologic considerations are relevant to the interpretation of the data from this study. First, synaptosomes are a heterogeneous preparation of nerve endings derived from a variety of neuron types. Thus, the biochemical changes measured in this system are an average of those occurring in a population and may reflect different regulatory and modulatory mechanisms, and different pharmacologic responses. Second, the neurotransmitter content of synaptosomes depends on the brain region(s) used in their preparation. Glutamate is the dominant neurotransmitter released by cerebrocortical synaptosomes, but the release of other amino acid neurotransmitters, peptides and catecholamines can also be measured. [3]It is possible that the release of other neurotransmitters, that demonstrate different Calcium sup 2+ release sensitivities, may differ in their responses to various anesthetics. Third, the use of chemical secretogogues to induce neurotransmitter release does not precisely mimic release evoked by electric stimulation. This is particularly true of KCl-evoked and veratridine-evoked release, which may result in nonphysiologic inactivation of Sodium sup + and Calcium²+ channels, as well as of ionomycin-evoked release, which leads to very high and nonlocalized increases in intrasynaptosomal [Calcium²+]. By producing only transient depolarizations, 4-AP appears to mimic the physiologic changes produced by electric stimulation most closely, although its action at Potassium sup + channels could obscure anesthetic effects at this potential target.

In summary, we have demonstrated the ability of clinically relevant concentrations of volatile anesthetics to inhibit 4-AP-evoked release of the excitatory neurotransmitter glutamate from presynaptic nerve terminals. Analysis of the effects of anesthetics on 4-AP-evoked glutamate release from synaptosomes in vitro should provide a useful model for the further pharmacologic characterization of presynaptic anesthetic mechanisms.

The authors are indebted to Thomas J. J. Blanck, M.D., Ph.D., and Talvindar S. Sihra, Ph.D., for helpful discussions.