Although the mechanisms underlying general anesthesia are not well elucidated, it is possible that the anesthetic state can be achieved by enhancing neuronal inhibition, by decreasing neuronal excitation, or by a combination of both. Since Nicoll  initially suggested that general anesthetics enhance inhibitory synaptic transmission, gamma-aminobutyric acid (GABA)-mediated synaptic inhibition has been investigated as a target site for these drugs. Enhancement of the GABA receptor channel response may be a primary action of volatile anesthetics , whereas various IV anesthetics modulate GABA-mediated inhibition [3,4].
Many studies have investigated the effects of general anesthetics on excitatory synaptic transmission. Richards  was the first to demonstrate that halothane depresses excitatory synaptic transmission in central nervous system (CNS) preparations in vitro. Halothane was later shown to depress glutamate receptor-mediated excitatory postsynaptic currents using patch-clamp techniques . MacIver et al.  demonstrated that volatile anesthetics depress glutaminergic synaptic transmission via reduction of presynaptic glutamate release in CA1 neurons of rat hippocampal slices.
These findings suggest that not all general anesthetics affect excitatory and inhibitory synaptic transmission in the same manner. There have been no reports comparing the effects of various anesthetics on both excitatory and inhibitory synaptic transmission using identical preparations. In the present study, we pharmacologically isolated excitatory and inhibitory synaptic pathways in the area CA1 of rat hippocampus in vitro and examined the effects of various IV and volatile anesthetics under these conditions.
The composition of the artificial cerebrospinal fluid (ACSF) was (mmol/L): NaCl 124, KCl 5, NaH2 PO4 1.25, CaCl2 2, MgSO4 2, NaHCO3 26, glucose 10, made with 18 M purity water. ACSF was precooled (8-10[degree sign]C) and equilibrated with 95% oxygen/5% carbon dioxide gas mixture before use (pH 7.35-7.45). The Mg2+-free ACSF was identical to the ACSF, except that MgSO4 was omitted. Pentobarbital, ketamine, and halothane were purchased from Dinabot (Osaka, Japan), Sigma (St. Louis, MO), and Takeda (Osaka, Japan), respectively. Isoflurane and propofol were kindly donated by Dinabot and Zeneca (Cheshire, UK), respectively. Most of chemicals used were obtained from Sigma. 6-Cyano-7-nitroquinoxaline-2,3-dione (CNQX) was obtained from Tocris Cookson (Bristol, UK).
The technique for the preparation of rat hippocampal slices was identical to the method previously described by Hirota and Roth . After approval had been obtained from the Animal Research Committee of Toyama Medical and Pharmaceutical University, male Wister rats (100-200 g) were deeply anesthetized with sevoflurane and decapitated. The brain was rapidly removed, and the dissected hippocampus was sliced in cold ACSF (8-10[degree sign]C) transversely to its long axis (400 [micro sign]m thick) with a microslicer (Dosaka EM, Osaka, Japan). Slices were placed onto nylon mesh at a liquid-gas interface in a recording chamber at 37[degree sign]C. A humidified gas mixture (95%O2/5% CO2) was applied to the chamber at a rate of 1 L/min, and ACSF was continuously perfused at a rate of 90 mL/h.
Square-wave paired-pulse stimuli (5-10 volt, 0.05 ms, 40-ms interval, 0.1 Hz) generated with a SEN-7203 stimulator (Nihon Kohden, Tokyo, Japan) were delivered to Schaffer collateral fiber via a nichrome bipolar electrode. Extracellular recordings were made with the 2 mol/L NaCl-filled glass microelectrodes (3-6 M Omega) placed in the area of CA1 cell bodies. Evoked responses were amplified with a MEZ-8301 amplifier (Nihon Kohden) and A/D conversions were made at a rate of 14,400 Hz. Data were stored on a hard disk of a Macintosh computer for later analysis.
IV anesthetics and GABAergic and glutaminergic drugs were dissolved in ACSF at required concentrations. Stock solutions of CNQX (10-3 mol/L) was prepared in dimethyl sulfoxide, and propofol was dissolved in 10% Intralipid (Pharmacia AB, Stockholm, Sweden) at a concentration of 10 mg/mL. These stock solutions were diluted in ACSF before they were perfused into the chamber. The concentrations of DMSO and Intralipid used in the experiments did not affect the field potentials. Halothane and isoflurane were applied as vapors to the chamber in a 95%O2/5% CO2 gas mixture using a vaporizer previously calibrated with an anesthetic gas analyzer (Capnomac; Datex, Helsinki, Finland). The concentrations of volatile anesthetics refer to the dial settings on the vaporizer. All drugs were applied for 20 min before recording to obtain stable effects. The Mg2+-free ACSF was used in experiments examining N-methyl-D-aspartate (NMDA) receptor-mediated responses because Mg2+ has been reported to block the NMDA channel at negative membrane potentials . The IV anesthetic concentrations applied to in vitro preparations were calculated based on the method previously described by Richards . The doses of pentobarbital, propofol, and ketamine required to anesthetize experimental animals ranged from 20 to 30 mg/kg , 10 to 24 mg/kg , and 44 to 250 mg/kg , respectively. Because the IV anesthetics can be diluted by the extracellular fluid (20%-30% of the total body weight), these amounts of pentobarbital, propofol, and ketamine should have maximal concentrations in the extracellular fluid in the ranges of 3-5 x 10 (-4) mol/L, 2-6 x 10-4 mol/L, and 2-10 x 10-4 mol/L, respectively. On the basis of these calculations, the concentration-response curves generated in preliminary experiments and the calculated 50% effective dose (ED50) values of anesthetics were then tested in the current study: pentobarbital 5 x 10-4 mol/L, propofol 5 x 10-4 mol/L, ketamine 10-3 mol/L, halothane 1.5 vol%, and isoflurane 2.0 vol%.
The extent to which inhibitory synaptic transmission contributes to depression of population spikes (PSs) was studied in the presence of the GABAA receptor antagonist bicuculline methiodide (BMI; 5 x 10-5 mol/L). Two types of ionotropic glutamate receptors were pharmacologically isolated using specific receptor antagonists . To assess the NMDA receptor-mediated PS (NMDA PS), BMI and the non-NMDA receptor antagonist CNQX were applied under Mg2+-free conditions. The GABAA receptor antagonist and the NMDA receptor antagonist DL-2-amino-5-phosphonovaleric acid (AP-5; 5 x 10-5 mol/L) were used to evaluate the non-NMDA receptor-mediated PS (non-NMDA PS).
PS amplitudes were measured for evaluation in a manner similar to that previously described . Measurements were made from the onset to the peak of the waveform. For data analysis, five evoked waves were collected and averaged. The effects of most of the anesthetics were determined on the first evoked PSs (PS1s) in the current experiments; however, the second evoked PSs (PS2s) were used for pentobarbital and propofol because our previous studies revealed that IV anesthetics produce greater effects on PS2 than on PS1 in identical preparations . Statistical significance of the data was determined using analysis of variance, followed by Student's t-test. A P value <0.05 was considered significant. Data are expressed as mean +/- SD.
In control conditions (no anesthetic), the amplitude of PS2 was larger than that of PS1. Pentobarbital and propofol decreased PS2 with a minimal change in PS1. Ketamine decreased both PS1 and PS2 to the same extent, whereas the volatile anesthetics isoflurane and halothane had greater effects on PS1 than on PS2 (Table 1, Figure 1). The effects were completely recovered on washout.
The administration of BMI induced multiple spikes and enhanced the PS1 and PS2 amplitudes to 107.6% +/- 6.4% and 106.8% +/- 6.1% of control (n = 6), respectively (Figure 1). The effects of pentobarbital and propofol were completely antagonized with BMI, whereas the effects of ketamine, halothane, and isoflurane were only partially antagonized (Figure 2). The results indicate that the inhibitory effects of pentobarbital and propofol are mainly due to GABAA-mediated mechanisms. Because other factors may be involved in the actions of ketamine and volatile anesthetics, we performed the following experiments on excitatory synaptic transmission.
The administration of CNQX reduced the amplitudes of PS1 and PS2 to 77.0% +/- 7.5% and 89.1% +/- 6.7% of control, respectively (n = 5). The combination of CNQX and BMI induced multiple spikes without significant changes in the amplitudes of PS1 and PS2. Under these conditions, PSs were elicited via NMDA receptor-mediated synaptic transmission. Pentobarbital and propofol had no significant effect on NMDA PSs, whereas ketamine, halothane, and isoflurane significantly depressed NMDA PSs (Figure 3A).
AP-5 alone had no consistent effect on the shape of PS, and the application of AP-5 and BMI in the absence of Mg2+ enhanced the PS1 and PS2 amplitudes to 121.7% +/- 8.7% and 120.1% +/- 6.6% of control, respectively (n = 5). The non-NMDA PSs were not affected in the presence of pentobarbital, propofol, and ketamine, but they were significantly decreased in the presence of the volatile anesthetics (Figure 3B).
The hippocampus is a highly laminated limbic structure with well defined afferents, efferents, and neurotransmitters, and it may be one of the major target sites for general anesthetics . Thus, the hippocampal slice preparation is an ideal model for the study of anesthetic action on synaptic transmission in the CNS. In the present study, we evoked PSs in area CA1 of the rat hippocampus via stimulation of the Schaffer collateral fibers. This pathway is monosynaptic and is inhibited via recurrent inhibitory interneurons. Because the PS reflects the number and synchrony of pyramidal cells that generate action potentials, the gradual decline of neural excitability (run-down) is expected to be minimal.
The concentrations of IV anesthetics tested in the current study were higher than the plasma concentrations in clinical settings. Because the doses of IV anesthetics required to anesthetize experimental animals are 10-100 times higher than those for humans [4,10,11], the different sensitivities to IV anesthetics among species (see Methods) could be involved. It may be attributed to limitations of the in vitro preparations: absence of the blood-borne factors from ACSF and/or lack of certain inputs and outputs that normally exist in the intact brain. The same degree of IV anesthetic concentrations [10,11] have been used for in vitro electrophysiological studies in brain slice preparations.
GABA is the major inhibitory neurotransmitter in hippocampus. The GABA (A) receptor (the BMI-sensitive receptor) is a ligand-gated ion channel consisting of a chloride channel complex. When the postsynaptic GABAA receptor is activated, chloride ions move into the postsynaptic cell, causing an increased membrane conductance that inhibits postsynaptic action potential discharge and decreases excitatory synaptic responses. Glutamate is the major excitatory neurotransmitter in the hippocampus. Glutamate receptors are divided into two functional subtypes (NMDA and non-NMDA) by their specific agonists. Glutamate released into the synaptic cleft activates both postsynaptic NMDA and non-NMDA receptors and induces an excitatory postsynaptic current. The fast component of excitatory postsynaptic current is due to the non-NMDA receptor, and the slower component is attributed to the NMDA receptor.
In the current study, we demonstrated for the first time in the same preparation that general anesthetics act differently on inhibitory and excitatory synaptic events in the CNS. Pentobarbital and propofol alter GABAA receptor-mediated inhibitory synaptic transmission but not NMDA and non-NMDA receptor-mediated excitatory synaptic transmission. Our results agree with previous reports [3,4]. Pentobarbital can block NMDA receptor-mediated currents in isolated single hippocampal neurons . Although the excitatory synaptic transmission might be altered, the current experiments provide evidence that pentobarbital has a primary action on inhibitory, rather than excitatory, synaptic transmission.
We found that ketamine depresses NMDA PS but not non-NMDA PS, which indicates that this anesthetic inhibits excitatory synaptic transmission via NMDA receptors, as previously reported . The effects of ketamine on GABA receptors are controversial. Tang and Schroeder  and Brockmeyer and Kendig  reported that ketamine does not attenuate GABAergic synaptic transmission in the spinal cord. We observed, however, that BMI partially antagonized the actions of ketamine, which suggests that the action of the anesthetic could, in part, be a result of enhancement of GABAergic inhibitory synaptic transmission. Our results are consistent with other studies in the cervical ganglion  and hippocampus . Because ketamine has different regional actions on the NMDA receptor , it could produce different actions on GABAergic transmission in the hippocampus compared with the spinal cord.
The results demonstrate that volatile anesthetics alter GABAA, NMDA, and non-NMDA receptor-mediated pathways, which suggests that these drugs exert inhibitory actions as a result of both enhancement of the inhibitory synaptic transmission and reduction of excitatory synaptic transmission. Postsynaptic GABAA receptors are considered a main target of general anesthetics , and Mimic et al.  identified the specific sites on GABAA receptors that are critical for modulation by volatile anesthetics. Recent studies, however, propose that volatile anesthetics attenuate the glutamate receptor-mediated synaptic transmission in the CNS [5-7,12].
Our experiments were based on the fact that the Schaffer collateral input to CA1 pyramidal neurons of the hippocampus is generated via glutamate-mediated monosynaptic excitatory synaptic transmission in combination with GABAA ergic recurrent in-hibition. It has also been reported that volatile anesthetics modulate the GABAB receptor-mediated inhibition in the hippocampus  and that volatile anesthetics can depress postsynaptic sodium channels  and calcium channels . Thus, a number of receptor-mediated pathways and/or postsynaptic events may be involved in the actions of volatile anesthetics.
In conclusion, we have shown that general anesthetics can produce different actions on GABAergic inhibitory and glutaminergic excitatory synaptic transmission in the CNS. Volatile anesthetics modulate both excitatory and inhibitory synaptic activities, whereas IV anesthetics produce more specific actions on inhibitory synaptic events. These result support the hypothesis of drug- and site-specific mechanisms of general anesthesia [14,25].
We thank Zeneca (Cheshire, UK) for the gift of propofol and Dinabot (Osaka, Japan) for the gift of isoflurane and an isoflurane vaporizer.
1. Nicoll RA. The effects of anaesthetics on synaptic excitation and inhibition in the olfactory bulb. J Physiol (Lond) 1972;223:803-14.
2. Nakahiro M, Yeh JZ, Brunner E, Narahashi T. General anesthetics modulate GABA receptor channel complex in rat dorsal root ganglion neurons. FASEB J 1989;3:1850-4.
3. Albertson TE, Walby WF, Joy RM. Modification of GABA-mediated inhibition by various injectable anesthetics. Anesthesiology 1992;77:488-99.
4. Albertson TE, Tseng CC, Joy RM. Propofol modification of evoked hippocampal dentate inhibition in urethane-anesthetized rats. Anesthesiology 1991;75:82-90.
5. Richards CD. On the mechanism of halothane anesthesia. J Physiol (Lond) 1973;233:439-56.
6. Perouansky M, Baranov D, Salman M, et al. Effects of halothane on glutamate receptor-mediated excitatory postsynaptic currents. Anesthesiology 1995;83:109-19.
7. MacIver MB, Mikulec AA, Amagasu SM, et al. Volatile anesthetics depress glutamate transmission via presynaptic actions. Anesthesiology 1996;85:826-34.
8. Hirota K, Roth SH. Sevoflurane modulates both GABAA
and GABA (B
) receptors in area CA1 of rat hippocampus. Br J Anaesth 1997;78:60-5.
9. Ascher P, Nowak L. Electrophysiological studies of NMDA receptors. Trends Neurosci 1987;10:284-8.
10. Richards CD. On the mechanism of barbiturate anaesthesia. J Physiol (Lond) 1972;227:749-67.
11. Oshima E, Richards CD. An in vitro investigation of the action of ketamine on excitatory synaptic transmission in the hippocampus of the guinea-pig. Eur J Pharmacol 1988;148:23-33.
12. Narimatsu E, Tsai YC, Gerhold TD, et al. A comparison of the effect of halothane on N-methyl-D-aspartate receptor-mediated excitatory synaptic transmission in the hippocampus. Anesth Analg 1996;82:843-7.
13. Hirota K, Roth SH, Fujimura J, et al. GABAergic mechanisms in the action of general anesthetics. Toxicol Lett 1998;100:203-7.
14. Kendig JJ, MacIver MB, Roth SH. Anesthetic actions in the hippocampal formation. Ann N Y Acad Sci 1991;625:37-53.
15. Miljkovic Z, MacDonald JF. Voltage dependent block of excitatory amino acid currents by pentobarbital. Brain Res 1986;376:396-9.
16. Yamamura T, Harada K, Okamura A, et al. Is the site of action of ketamine anesthesia the N-methyl-D-aspartate receptor? Anesthesiology 1990;72:704-10.
17. Tang AH, Schroeder AL. Spinal cord depressant effects of ketamine and etoxadrol in the cat and the rat. Anesthesiology 1973;39:37-43.
18. Brockmeyer DM, Kendig JJ. Selective effects of ketamine on amino acid -mediated pathways in neonatal rat spinal cord. Br J Anaesth 1995;74:79-84.
19. Little HJ, Atkinson HD. Ketamine potentiates the responses of the rat superior cervical ganglion to GABA. Eur J Pharmacol 1984;98:53-9.
20. Gage PW, Robertson B. Prolongation of inhibitory postsynaptic currents by pentobarbitone, halothane and ketamine in CA1 pyramidal cells in rat hippocampus. Br J Pharmacol 1985;85:675-81.
21. Aida S, Fujiwara N, Shimoji K. Different regional effects of ketamine spontaneous and glutamate-induced activities of single CNS neurons in rats. Br J Anaesth 1994;73:388-94.
22. Mihic SJ, Ye Q, Wick MJ, et al. Site of alcohol and volatile anaesthetic action on GABAA
and glycine receptors. Nature 1997;389:385-9.
23. Rehberg B, Xiao Y-H, Duch DS. Central nervous system sodium channels are significantly suppressed at clinical concentrations of volatile anesthetics. Anesthesiology 1996;84:1223-33.
24. Study RE. Isoflurane inhibits multiple voltage-gated calcium currents in hippocampal pyramidal neurons. Anesthesiology 1994;81:104-16.
25. MacIver MB, Roth SH. Barbiturate effects on hippocampal excitatory synaptic responses are selective and pathway specific. Can J Physiol Pharmacol 1987;65:385-94.