INHALATION ANAESTHETICS EXHIBIT PATHWAY- SPECIFIC AND DIFFERENTIAL ACTIONS ON HIPPOCAMPAL SYNAPTIC RESPONSES IN VITRO

Transcription

1 Br. J. Anaesth. (1988), 60, INHALATION ANAESTHETICS EXHIBIT PATHWAY- SPECIFIC AND DIFFERENTIAL ACTIONS ON HIPPOCAMPAL SYNAPTIC RESPONSES IN VITRO M. B. MACIVER AND S. H. ROTH General anaesthetics can depress synaptic transmission and neuronal excitability in the central nervous system (CNS) [1,2]. Although it is possible that CNS depression is the major action underlying anaesthesia, evidence exists for anaesthetic-induced facilitation of excitatory transmission [3-5] and various patterns of " activated " EEG recordings have been observed during anaesthesia [6,7]. Furthermore, anaesthetics produce agent-specific effects on CNS electrical activity in vivo [8], and a single anaesthetic state, or common neurophysiological mechanism, has not been observed for all anaesthetics. Concentration-dependent and anaesthetic-specific differential effects have also been reported on a number of invertebrate and isolated mammalian peripheral nervous system preparations [9]. These differential actions do not support a traditional "unitary" theory of anaesthesia [2,8-11]; rather, actions at multiple and selective membrane sites appear likely [12-1]. Recent studies of anaesthetic actions on vertebrate CNS neurones in vitro have not revealed differences in action between agents. For example, it was reported that a number of general anaesthetics (including inhalation agents and barbiturates) hyperpolarize neurones of the spinal cord and hippocampus [1], and a good correlation M.BRUCE MACIVER*, B.SC, M.SC, PH.D.; SHELDON H. ROTH+, B.sc, M.sc, PH.D.; Departments of Pharmacology & Therapeutics and Anaesthesia, Faculty of Medicine, The University of Calgary. Accepted for Publication: November 17, * Present address: Department of Anesthesia, Stanford University School of Medicine, Stanford, California, U.S.A. + Address for correspondence: Department of Pharmacology & Therapeutics, Faculty of Medicine, The University of Calgary, 3330 Hospital Drive N.W., Calgary, Alberta, Canada T2NN1. SUMMARY The effects of halothane, isoflurane andenflurane were compared on three CNS excitatory synaptic pathways in vitro, to determine whether selective actions described in vivo result from differential effects on anatomically distinct cortical pathways and neurone populations. Halothane ( vo/%) depressed postsynaptic excitability of CA1 pyramidal neurones in response to activation of stratum radiatum synaptic inputs, and concentration-dependent excitatory ( vo/%) and depressant ( vo/%) actions were observed on dentate granule neurone excitability and perforant path evoked synaptic responses. In contrast, isoflurane increased CA 1 neurone excitability ( vol%) and produced postsynaptic depression of dentate neurones ( vol%). Enflurane also increased CA1 excitability ( vo/%), but depressed synaptic responses at equivalent concentrations, and produced mixed excitatory ( vo/%) and depressant ( vo/%) effects on dentate synaptic responses. Differentia/ actions were also observed for the three anaesthetics on stratum oriens excitatory inputs to CA 1 neurones, and on antidromic responses. A good correlation (r = 0.992) exists between the membrane/buffer partition coefficients of these anaesthetics and their half-maximal concentrations for depression of synaptic responses; however, this correlation does not reflect the different, anaesthetic-specific actions observed. The results indicate that inhalation anaesthetics act at multiple and selective hydrophobic recognition sites which are heterogenously distributed on different synaptic pathways.

2 VOLATILE ANAESTHETICS AND CNS SYNAPTIC TRANSMISSION existed between concentrations required to produce hyperpolarization and those needed for anaesthesia. These authors suggested that membrane hyperpolarization produces a depression of CNS excitability which results in anaesthesia, and that all general anaesthetics share this common mechanism of action [15-17]. Halothane, isoflurane and enflurane, however, have been shown to depolarize hippocampal neurones [18], and the major common action of these anaesthetics appears to be depression of spontaneous discharge, via actions on spike initiation. The controversy concerning common v. differential actions of general anaesthetics prompted us to re-investigate the effects of inhalation agents on synaptic responses of mammalian CNS neurones in vitro. 681 CA1 DG sep METHODS Preparation Experiments were conducted on 73 hippocampal slices prepared from male Sprague Dawley rats, as previously described [19]. A single slice from each rat was used and only one anaesthetic was studied on each preparation. Rats were anaesthetized with ether, the heart stopped by a blow to the back of the thorax and the brain rapidly removed and placed in pre-cooled (810 C) oxygenated artificial cerebrospinal fluid (ACSF; see Materials). Transverse slices (00 um) of hippocampus were cut using a vibratome (Campden Instruments, U.K.), following dissection of the dentate-hippocampal formation. The angles used to prepare slices for studies of CA1 and dentate granule (DG) pathways were approximately 5-10 on either side of the striation pattern of alveus fibres, to follow the lamellar organization of synaptic pathways in the rat hippocampal formation (fig. 1). Slices were placed on a nylon mesh screen at the gas/liquid interface in a Plexiglass tissue chamber. Oxygenated (95 % oxygen-5% carbon dioxide and prewarmed (35 C) ACSF was continuously perfused through the chamber at a rate of ml min"1. Slices were incubated for 1 h without electrical stimulation, followed by a 1-2 h period of control stimulation/ recording to establish the stability of each preparation. All preparations used in this study exhibited field potential amplitudes of at least 5 mv (peak to peak) and control amplitude variability less than 5 % during the initial data acquisition period, and following washout of anaesthetic. FIG. 1. Diagram of the rat hippocampal formation showing the angle used to prepare slices for CA1 pyramidal and dentate granule (DG) synaptic pathways. Stimulating electrodes were placed in stratum oriens to activate commissural/ associational (c/a), or stratum radiatum to activate Schaffercollateral (sc) inputs to CA1 pyramidal neurones. Dentate granule evoked field potentials were produced by stimulating perforant path (PP) fibres. Antidromic responses were produced by stimulating the alveus (alv) or mossy fibre (mf) pathways, for CA1 or DG responses, respectively, sep = Septal; tern = temporal; fim = firnbria; fiss = hippocampal fissure; sub = subiculum; sh = septo-hippocampal fibres. Electrode placement Bipolar nichrome stimulating electrodes were placed on perforant path fibres, stratum radiatum, or stratum oriens to activate excitatory synaptic inputs to DG or CA1 pyramidal neurones, respectively (fig. 1). Antidromic responses were produced by stimulation of the alveus or mossyfibre pathways. Extracellular glass recording electrodes (sodium chloride 2 mol litre"1, 2-10 Mfl) were placed in the cell body regions of DG and CA1 areas to record stimulus-evoked field potentials. Paired stimulus pulses of ms duration (10-80 ua) were delivered at 0.1 Hz.

3 682 Stimulus intensity was varied to determine input/ output relationships (E-S analysis) at a fixed interstimulus interval of ms [20-22]. Recorded signals were amplified (x 1000), filtered (1 Hz to 10 KHz, bandpass), and digitally stored for later analysis (0 us resolution on a PDP 11/23 UNIX system). Population spike (PS) amplitudes were measured from threshold to peak negativity and excitatory post-synaptic potential (EPSP) rise times (dv/dt) were determined from the initial EPSP slope by linear regression analysis of digital sample points between 20 and 80 % of the EPSP peak positivity [19]. All records shown are single sweeps of 208 sample points. Administration and analysis of anaesthetics The inhalation agents were applied as vapours to the tissue chamber via the pre-warmed and humidified oxygen-carbon dioxide gas stream above the slices, using calibrated commercial vaporizers. Concentrations, expressed as volume percent (vol %), refer to settings on the vaporizer dials. To reach equilibrium, vapour concentrations were applied for a minimum of 30 min. The onset times for drug effects were often less than 1 min, and maximum effects reached a plateau within 15 min. Accurate measurements of concentrations in the perfusate were made using a Hewlett- Packard 5880A gas chromatograph fitted with a glass column (packed with Poropak Q) and flame ionization detector. The carrier gas was nitrogen delivered at a flow rate of approximately 0 ml min" 1. Temperatures of the injector, oven and detector were 155, 10 and 170 C, respectively. Aqueous test samples (20 ul) were taken from the recording chamber with a gas-tight Hamilton syringe and injected directly onto the column. Peak area values from the test samples were plotted against those of known concentrations of standards, to determine bath concentrations. Values were determined as peak area ratios relative to dextrose, which was used as an internal standard in the ACSF. Column retention times were approximately 3-5 min for the inhalation anaesthetics and 7.5 min for dextrose. Materials Adult male Sprague-Dawley rats ( g) were obtained from the University of Calgary Biosciences Vivarium. Anaesthetics were obtained from the following suppliers: halothane from Halocarbon (Ontario) Ltd (Mississauga, Ont.), enflurane and isoflurane from Ohio Medical BRITISH JOURNAL OF ANAESTHESIA Anaesthetics (Pointe Claire, Que.). The vaporizers were manufactured by Ohio Medical Products (Madison, Wis.). The ACSF physiological solution [23] had the following composition (mmol litre" 1 ): NaCl 13, KC1.5, CaCl 2 1.6, KH 2 PO , MgSO 2, NaHCO 3 16, dextrose 10; modified by adjusting calcium and potassium concentrations to values measured in vivo [2]. RESULTS Field potential responses Typical field potential recordings from the three synaptic pathways are shown in figure 2. Stratum radiatum (RAD) to CA1 evoked responses exhibited the largest PS amplitudes, with interstimulus intervals of ms producing maximal paired pulse potentiation. Stratum oriens (OR) to CA1 evoked responses were smaller in amplitude and an interpulse interval of ms was optimal for paired pulse potentiation. Stimulation of perforant path (PP) inputs to dentate granule neurones produced the largest EPSP amplitudes, although PS amplitudes were usually smaller than CA1 responses. Dentate EPSP responses had faster rise times and a shorter half-time for decay, compared with CA1 responses, and interstimulus intervals of ms produced maximal paired pulse potentiation. Halothane. Low concentrations of halothane ( vol %) produced depression of PS responses for RAD inputs to CA1 neurones (upper recordings in figure 2A). PS depression occurred with little or no effect on field EPSP amplitudes (compare responses to the first v. second stimuli). Depression of PS responses did not appear to result from altered paired pulse potentiation, since facilitation of EPSP responses was not altered. The latency to onset and initial slope of the EPSP was not changed in the presence of halothane concentrations less than 1.5 vol %; however, higher concentrations depressed EPSP responses and complete block of transmission occurred at 3.0 vol %, at all stimulus intensities (see below (E-S relationships)). EPSP responses evoked by OR inputs to CA1 were depressed by the lowest effective concentrations of halothane ( vol %; middle recordings of figure 2A). Depression of EPSP responses was accompanied by reduced PS amplitudes. Onset latency and paired pulse potentiation of the EPSP were not altered by

4 VOLATILE ANAESTHETICS AND CNS SYNAPTIC TRANSMISSION 683 A: HALOTHANE B = '«>FUIRANE C: ENFLURANE FIG. 2. A: Effects produced by halothane on stratum radiatum (RAD), stratum oriens (OR) inputs to CA1 neurones, and perforant path (PP) inputs to dentate granule neurone evoked field potentials. Each superimposed series of records shows control (C) responses before (solid line) and after washout (light dotted line) together with two or three concentrations of anaesthetic (numbers refer to vol %). B: Effects produced by isoflurane on evoked field potentials, c: Effects of enflurane on evoked responses. Calibration: 20 ms and 2.0 mv for A, B and C. Data in each series of records are from separate preparations.

5 68 BRITISH JOURNAL OF ANAESTHESIA concentrations which completely blocked PS responses ( vol%). The lower recordings in figure 2A illustrate the different profile of effects produced by halothane on PP inputs to dentate granule neurones, compared with the CA1 responses described above. Low concentrations ( vol %) increased PS amplitudes, whereas EPSP responses were either unchanged or slightly depressed. Latencies for onset of both the EPSP and PS were increased at concentrations which enhanced PS amplitudes. Concentrations of halothane greater than 1.5 vol% depressed both EPSP and PS responses. Paired pulse facilitation of the EPSP was not observed at concentrations greater than 2.0 vol% (compare responses to the first v. second stimuli at 2.5 vol% infigure 2A). Isoflurane. The effects produced by isoflurane differed from the actions of halothane on both the RAD and PP synaptic pathways (compare figure 2A with 2B). LOW concentrations of isoflurane ( vol %) increased RAD to CA1 PS amplitudesj whereas EPSP responses were unchanged or slightly depressed (upper recordings in figure 2B). Increased PS amplitudes were accompanied by decreased latencies of the spike; but onset latencies, rise time of the EPSP, and paired pulse potentiation were not consistently altered at these concentrations. Concentrations greater than 1.0 vol % produced depression of PS responses, accompanied by increased latencies and broadening of the negative spike (e.g. 1.5 vol % in figure 2B ). Concentrations greater than 2.0 vol % depressed synaptic (EPSP) responses and above 3.0 vol % completely blocked postsynaptic discharge, although paired pulse facilitation of EPSP was still apparent. In contrast to the biphasic actions on RAD inputs, isoflurane produced only depression of OR inputs to the same postsynaptic population (CA1). Concentrations from 0.5 to.0 vol % depressed both EPSP and PS responses, and this was accompanied by increased latencies (middle recordings in figure 2B). Paired pulse potentiation was observed at concentrations up to.0 vol % for this pathway. The actions of isoflurane on PP inputs to DG neurones are shown in the lower recordings of figure 2B. All effective concentrations depressed PS responses, with only minimal effects on the EPSP. Even in the presence of concentrations which markedly depressed PS amplitudes (e.g. 2.5 vol%), EPSP amplitudes, rise times and onset latencies were only slightly altered (compare EPSP and PS responses to the first and second stimuli, respectively). Concentrations greater than 3.0 vol % were required to depress EPSP responses in this pathway; this is well above the values needed to alter EPSP amplitudes for either of the inputs to CA1 neurones. Enflurane. Alterations in field potential produced by enflurane are illustrated infigure 2c, and marked differences in effect were observed compared with halothane and isoflurane. RAD inputs to CA1 neurones were depressed by enflurane over the concentration range vol % (upper recordings in figure 2c). EPSP responses were depressed at the lowest concentrations, and reduced PS amplitudes appeared to result from depression of synaptic input (see below (E-S relationships)). Paired pulse potentiation and onset latencies for EPSP were not markedly altered by concentrations less than.0 vol %, but an increased latency for PS responses accompanied depression of spike amplitudes. Low concentrations of enflurane (below 2.5 vol%) did not alter EPSP responses for OR inputs to CA1 (middle recordings of figure 2c). PS amplitudes were depressed by concentrations as low as 0.5 vol % and complete depression occurred above.0 vol %. Concentrations greater than 2.0 vol % produced intermittent burst discharges from CA1 neurones, as previously described [19]. This seizure-like activity occurred both spontaneously and in response to stimulation of RAD and OR inputs to CA1 neurones (e.g. at.0 vol % in middle recordings of figure 2c). Burst firing, however, did not occur in response to each stimulus, and non-burst responses were chosen for the other records shown in figure 2c, for comparison with the non-burst responses in the presence of halothane and isoflurane. Enflurane produced concentration-dependent biphasic actions on PP inputs to DG neurones (lower recordings in figure 2c), similar to the effects produced by halothane. Concentrations from 0.25 to 1.0 vol % produced increased PS amplitudes accompanied by prolonged spike latencies (e.g. at 0.5 vol % in figure 2c) and increased EPSP onset latencies. Note that the EPSP amplitude and rise time were depressed by concentrations which increased PS responses. Higher concentrations of enflurane ( vol%) produced further reduction in EPSP

6 VOLATILE ANAESTHETICS AND CNS SYNAPTIC TRANSMISSION r 100. Enflu responses, resulting in depression of PS amplitudes. Paired pulse potentiation was not altered by concentrations less than.0 vol%. Concentration-response relationships Concentration-response curves were different for each of the three volatile anaesthetics on the synaptic pathways studied (fig. 3). Note the dissimilar biphasic effects (enhancement/depression) v. monophasic responses on different pathways. Facilitation of PS responses were observed at the lowest effective concentrations for all three anaesthetics on one of three pathways for each agent. The major common effect on all pathways was depression of PS amplitudes in the presence of higher concentrations; halothane was most potent, producing half-maximal (ED 60 ) depression of PS amplitudes at 0.5, 0.8 and 1.5 vol% for oriens, radiatum and perforant path inputs, respectively; isoflurane produced half-maximal depression at 1.25,1.6 and 1.8 vol % and enflurane was the least potent, with ED 50 values of 1.65, 2.0 and 2.15 vol %. The ED 50 values for depression of stratum radiatum inputs to CA1 neurones corresponded to perfusate concentrations of 0., 0.6, and 0.9 mmol litre" 1 for halothane, isoflurane and enflurane, respectively, measured using gas chromatographic analysis (see Methods). Depression of CA1 antidromic responses required 5 to 10 times greater concentrations for minimal effects. Half-maximal depression of antidromic PS amplitudes could not be achieved with halothane and enflurane, limited by the maximum range of the vaporizers (fig. 3). Similar effects were produced by these agents on antidromic responses of dentate granule neurones (data not shown), except in the presence of enflurane ( vol %), which increased dentate antidromic PS amplitudes [19]. 60? Concn (volz) FIG. 3. Concentration-response curves for halothane, isoflurane and enflurane actions on stratum oriens (O)> stratum radiatum ( ), perforant path ( ) and CA1 antidromic (A) responses. Population spike (PS) amplitudes are expressed as percent of control for various concentrations (vol%) of anaesthetic; each point is the mean + SEM from 3 to 7 experiments; effects were measured at equilibrium. 10 E S relationships The relationship between the rise times (dv/dt) of the positive synaptic potential (EPSP) and the amplitude of the negative population spike (PS) describes an input/output function for a synaptic pathway [20,21,25]. The intercept on the EPSP axis provides a measure of postsynaptic discharge threshold [22,23,26]. To determine whether the inhalation anaesthetics altered postsynaptic excitability to a greater extent than synaptic function, E-S relationships for stratum radiatum and perforant path responses were studied.

7 686 BRITISH JOURNAL OF ANAESTHESIA 1 CA1 DG - Halothane r - Isofluran 1 </> a Enflurane EPSP (vs -1 ) FIG.. E S curves for stratum radiatum inputs to CAl pyramidal neurones and perforant path inputs to dentate granule (DG) neurones in the presence of various concentrations (vol %) of halothane, isoflurane and enflurane. Control (C) responses are also shown: pre-anaesthetic (O) and post-recovery (#). Data for each graph are from separate preparations. As shown in figure, a steep relationship between the EPSP and PS was apparent at low stimulus intensities, and PS responses approached a maximum at the higher intensities [20,22]. Data in each graph are from separate preparations. Each point represents the EPSP rise time and amplitude of the associated PS, measured at the second single stimulus intensity, from a trial of 2 test stimuli at eight to 10 different intensities. If a stimulus failed to elicit a PS response or if burst discharges were produced (i.e. in the presence of enflurane), the point was not plotted. Halothane. Population spike responses for RAD inputs to CAl appeared to be most sensitive to the depressant effects of halothane (upper left.

8 VOLATILE ANAESTHETICS AND CNS SYNAPTIC TRANSMISSION 687 figure ). Stimuli which produced equivalent EPSP responses failed to elicit PS amplitudes comparable to control values, indicating that depression of postsynaptic excitability was the major action on this pathway. Depression of excitability was also apparent from the shift to the right of the EPSP axis intercept (threshold), such that larger EPSP responses were required to produce minimal discharge of the CAl population. In the upper right of figure, the effects of halothane on PP inputs to dentate granule neurones are shown. Low concentrations (e.g. 0.5 vol %) produced a shift to the left in the EPSP axis intercept and increased PS amplitudes were observed for comparable control EPSP responses, indicating that a decreased threshold for discharge of granule neurones could account for the enhancement of field potentials at concentrations of vol %. In the presence of higher concentrations (e.g. 1.5vol%), depression of synaptic input resulted in decreased PS amplitudes, even though discharge threshold was further reduced. Isoflurane. Isoflurane produced opposite effects, compared with halothane, on E-S profiles for RAD and PP pathways (fig. ). Low concentrations (e.g. 0.5 vol%) enhanced PS amplitudes for equivalent control EPSP responses and also appeared to decrease the discharge threshold of the CAl population. Higher concentrations (e.g. 2.0 vol %) produced a further reduction of discharge threshold; however, at higher stimulus intensities, PS amplitudes were reduced. Isoflurane produced two opposing effects on CAl postsynaptic excitability: enhanced excitability, such that smaller EPSP dv/dt responses could discharge the CAl population, and depression of the level of discharge at stimuli above threshold. The major action of isoflurane on PP inputs to dentate neurones was a depression of PS amplitudes (middle right of figure ). Even in the presence of high concentrations (e.g. 2.5 vol %) EPSP responses were only slightly depressed (10%), while PS amplitudes were reduced by approximately 80% of control. This is opposite to the effects of halothane, which depressed EPSP responses but enhanced the excitability of dentate neurones. Isoflurane did not alter the EPSP axis intercept, suggesting that the depression of granule neurone discharge did not result from an increase in PS threshold. Enflurane. All effective concentrations of enflurane produced a dramatic reduction of discharge threshold for CAl neurones (lower left of figure ), which may account for the spontaneous and stimulus-evoked seizure-like burst firing previously described for this anesthetic [5,19]. The dominant effect on RAD inputs to CAl, however, was depression of synaptic responses (EPSP), resulting in an overall attentuation of field potential amplitudes. Enflurane appeared to act on at least two independent sites, depressing synpatic transmission and enhancing CAl neurone excitability. PP to dentate neurone EPSP responses were also depressed by enflurane (lower right of figure ), but the major effect occurred on postsynaptic excitability. In the presence of low concentrations (e.g vol %) depressed synaptic responses were accompanied by increased PS amplitudes, which appeared to involve a reduction in discharge threshold (left shift of the EPSP axis intercept). Higher concentrations (e.g. 3.0 vol%) produced a further reduction of EPSP responses and marked depression of PS amplitudes. Enflurane did not produce dramatic reductions in discharge threshold of granule neurones, and burst firing of dentate cells was not observed [19]. DISCUSSION Halothane, isoflurane and enflurane were capable of depressing field potential responses in the three synaptic pathways studied; however, the concentration-dependent effects were different for each anaesthetic. This is apparent when comparing the biphasic (facilitation/depression) v. monophasic actions on population spike amplitudes (fig. 3). Furthermore, depression of field potential amplitudes could result from opposing actions on synaptic responses (EPSP) and postsynaptic excitability (PS) of CAl pyramidal and dentate granule neurones (fig. ). These varied and opposing effects suggest that inhalation anaesthetics act via different mechanisms at selective cellular (membrane) sites. We propose that inhalation anaesthetics interact with discrete membrane "recognition" sites which can discriminate between agents on the basis of molecular structure. Landau and colleagues [27] have demonstrated that ether and methoxyflurane can produce biphasic actions on miniature end-plate currents at the frog neuromuscular junction. They concluded in their report that "the effect of

9 688 such structurally non-specific drugs comprises a combination of discrete and distinguishable events which are specific and dose dependent for each of the agents". The results of the present study complement their observations and extend the hypothesis to mammalian nervous tissue. Pathway-specific actions of inhalation anaesthetics In addition to the obvious morphological differences between CAl pyramidal neurones and dentate granule cells, there are also several important differences in membrane physiology [28]. Granule cells exhibit a relatively high resistance to convulsant-induced burst firing, which has been attributed to a lack of inward calcium currents [29]. CAl pyramidal neurones, in contrast, exhibit a low threshold for convulsantinduced seizure activity, and large inward calcium currents have been recorded [30]. Therefore, distribution of membrane calcium channels may differ between the two populations. Given that glutamate is a likely transmitter in both PP and RAD circuitry, the differential actions of anaesthetics could also be accounted for by the differences in distribution of glutamate receptor sub-types in CAl and DG regions [31]. In the present study, marked differences in anaesthetic effects on synaptic responses and postsynaptic excitability were observed between the different pathways and neurone populations. Halothane blocked transmission in the stratum radiatum by depressing CAl neurone excitability, but perforant path inputs to dentate granule neurones were blocked via a reduction in synaptic strength sufficient to overcome a halothaneinduced increase in granule neurone excitability (fig. ). Enflurane enhanced the discharge of CAl neurones, but depressed the stratum radiatumevoked EPSP. In contrast, the major effect of enflurane on dentate granule neurones was a depression of discharge. EPSP responses for radiatum and oriens inputs to CAl neurones were more sensitive to isoflurane than EPSP responses for perforant path inputs to dentate neurones (fig. 2B), yet depression of discharge was observed on both populations of neurones (fig. ). It is possible that anaesthetic recognition sites are differentially distributed on the same population of neurones. Isoflurane, for example, enhanced the discharge of CAl neurones by reducing threshold (fig. ), and this resulted in larger population spike amplitudes following stimulation of RAD inputs. This reduction in BRITISH JOURNAL OF ANAESTHESIA threshold for CAl discharge in response to RAD stimulation was not generalized to OR inputs, since only a depression of stratum oriens-evoked population spikes was observed (fig. 2B). Enflurane also reduced the discharge threshold of CAl neurones when RAD inputs were tested, but an increase in OR-evoked population spikes was not observed, even though EPSP amplitudes were not depressed and spontaneous burst discharges were generated by the CAl neurone population. These anaesthetics produced different effects on the postsynaptic discharge of CAl neurones, dependent on which synaptic input brought the population to threshold. A differential distribution of anaesthetic sites in basilar and apical dendrites of CAl neurones could account for the different effects on postsynaptic excitability. Concentration-dependent biphasic responses Anaesthetics produce unique patterns of excitatory and depressant effects in vivo, and specific brain regions are differentially affected by the same anaesthetic [5,8]. Similar patterns of concentration-dependent biphasic actions have also been reported for anaesthetics on relatively simple invertebrate neuronal preparations [9,32-3]. Previous studies of general anaesthetic effects on isolated mammalian CNS preparations have not described biphasic excitatory and depressant actions. Depression of excitatory synaptic responses [2], enhancement of synaptic inhibition [35,36], or decreased postsynaptic excitability [3,37] were the dominant actions reported for isolated mammalian neurones. Recent studies using the rat hippocampal slice preparation [1,15] suggest an enhanced potassium conductance as the major common action leading to depression of CNS excitability [17], and direct depression of CAl discharge (that is, via sodium channel blockade) also appears to be an important action for the inhalation anaesthetics [18]. However, a single common effect (e.g. enhanced potassium conductance) or "unitary" mechanism, per se, cannot account for the selective and pathwayspecific effects observed in the present study. A summary of the differential effects on radiatum, oriens and perforant path field potential responses is provided in table I, together with a comparison of the actions of barbiturates on the three synaptic pathways [38]. The only agent which produced a common action on all three pathways and two neurone populations was phenobarbitone as indicated by a "universal"

10 VOLATILE ANAESTHETICS AND CNS SYNAPTIC TRANSMISSION 689 TABLE I. Differential effects of anaesthetics on hippocampal synoptic transmission. RAD = Stratum radiatum inputs to CA1; OR = stratum oriens inputs to CAI pyramidal neurones; PP = perforant path inputs to dentate granule (DG) neurones. Anti. = Antidromic field potential amplitude; Excit. = excitability from the EPSP axis intercept of Input/Output curves increased excitability means a smaller EPSP dv/dt was able to produce discharge of the postsynaptic neurones. Black arrows = major effect; open arrows = minor effect; direction of arrow = an increase (up) or decrease (down) in response relative to pre-anaesthetic control responses. Concentration-dependent biphasic effects are indicated ( )) with the arrow on the left for the lowest effective concentration. = Lack of effect Effect on field potential responses RAD OR PP CAI DG Anaesthetic Pentobarbitone [381 Phenobarbitone [381 Halothane Isoflurane Enflurane EPSP u i PS tl II u EPSP depression of EPSP, PS and antidromic field potentials. Pentobarbitone, halothane, isoflurane and enflurane produced pathway-specific actions (compare biphasic and monophasic actions on EPSP and PS of RAD, OR and PP synaptic responses). The anaesthetics also altered the postsynaptic excitability of CAI and DG neurone populations in a differential manner; for example, halothane increased the excitability of DG but depressed CAI neurones (table I). The other agents studied did not alter DG neurone excitability, but produced variable effects on CAI discharge threshold. As indicated above, the marked decrease in CAI discharge threshold produced by enflurane may account for its seizureinducing properties. Several of the recorded responses (OR, EPSP and PS; CAI antidromic) were "universally" depressed by all five anaesthetics (table I), although the degree of depression varied between agents; that is the major action of pentobarbitone on OR evoked responses was depression of PS amplitudes, in contrast to phenobarbitone which produced depression of synaptic input (EPSP) to a greater extent than CAI discharge (PS). All other responses were differentially affected by the anaesthetics; for example pentobarbitone and isoflurane produced biphasic effects on RADevoked PS responses, phenobarbitone and halothane produced marked depression, and enflurane only marginally depressed this response (table I). The barbiturates and halothane depressed CAI discharge, but enflurane and isoflurane increased the excitability of CAI neurones (table I). Thus the anaesthetics studied exhibited some common I 1 PS I I 1 I EPSP tl I PS i u i u Anti. Excit. I t T Anti. n Excit. actions, but also a number of distinctly different effects. The particular profile of effects observed was dependent on any combination of the anaesthetic, concentration, synaptic pathway, neurone population, and the response (EPSP or PS) measured. Correlation between synaptic actions and anaesthetic potency Major support for a "unitary" theory of anaesthetic action originates from the observed correlation between anaesthetic potencies and partition coefficients for membrane/buffer or solvent/water systems [39]. A correlation (r = 0.992) exists between Log membrane/buffer partition coefficients and Log ED 50 concentrations required to depress stratum radiatum synaptic transmission in the present study. The correlation was not markedly changed by including data for pentobarbitone and phenobarbitone (r = 0.983) (fig. 5A), and depression of stratum radiatum transmission was better correlated with predicted in vivo potencies than were the minimum effective concentrations for producing hyperpolarization reported by Nicoll and Madison [1]. The correlation was improved (r = 0.991) by omitting data for phenobarbitone (fig. 5B), which is not an effective anaesthetic in vivo [35]. Membrane recognition sites and the mechanisms of action of anaesthetics The results of the present study are best explained by anaesthetic interactions with hydrophobic microdomains in neuronal membranes, thus accounting for the correlation between t

11 690 BRITISH JOURNAL OF ANAESTHESIA o 0. - E r = 0.6 o g A B '. ENF t is \HAL PB* HAL #-'PS HAL PB Pm 'ENF /' ENF / BPH»ISO OPH In vivo potency (mmol litre' 1 ) FIG. 5. Stratum radiatum to CA1 pyramidal neurone PS responses (ED 50 ) (mmol litre" 1 ) compared with: A membrane/ buffer partition coefficients (P m ); B: in vivo anaesthetic potencies (circles) and with minimum effective concentrations for hyperpolarization of spinal cord neurones ( ) (data from [1]). ENF = Enflurane; ISO = isoflurane; HAL = halothane; PB = pentobarbitone; PH = phenobarbitone. potency and membrane/buffer partition coefficients. However, a high degree of selectivity has been demonstrated for anaesthetic recognition sites, as indicated by the different potencies for stereoisomers of pentobarbitone [32,38] and the differential (biphasic, monophasic, burstinducing) effects of structural isomers such as enflurane and isoflurane. ACKNOWLEDGEMENTS This study was supported by the Medical Research Council of Canada (SHR) and the Alberta Heritage Foundation for Medical Research. We thank Dr C. D. Richards (University of London, U.K.) for providing the brain slice chamber, and Dr D. P. Harris for assistance with computer analysis and programming. We also thank Drs T. P. Hicks and B. H. Bland for criticism of and comments on the manuscript. REFERENCES 1. Nicoll RA, Madison DV. General anesthetics hyperpolarize neurons in the vertebrate central nervous system. Science 1982; 217: Richards CD. Actions of general anaesthetics on synaptic transmission in the CNS. British Journal of Anaesthesia 1983; 55: Morris ME. Facilitation of synaptic transmission by general anaesthetics. Journal of Physiology 1978; 28: Morris ME. General anesthetics and intracellular free calcium ions. In: Roth SH, Miller KW, eds. Molecular and Cellular Mechanisms of Anesthetics. New York: Plenum, 1986; Stevens JE, Fuginaga M, Oshima E, Mori K. The biphasic pattern of the convulsive property of enflurane in cats. British Journal of Anaesthesia 198; 56: Rosner RS, Clark DL. Neurophysiological effects of general anesthetics. Anesthesiology 1973; 39: Stockard J, Bickford R. The neurophysiology of anesthesia. In: Gordon E, ed. A Basis and Practice of Neuroanesthesia. Amsterdam: Excerpta Medica, 1975; 3-^ Winters WD. A review of the continuum of drug-induced states of excitation and depression. Progress in Drug Research 1982; 26: Judge SE. Effect of general anaesthetics on synaptic ion channels. British Journal of Anaesthesia 1983; 55: Richards CD, Martin K, Gregory S, Keightley CA, Hesketh TR, Smith GA, Warren GB, Metcalfe JC. Degenerate perturbations of protein structure as the mechanism of anaesthetic action. Nature (London) 1978; 276: Roth SH. Membrane and cellular actions of anesthetic agents. Federation Proceedings 1980; 39: Halsey MJ. A reassessment of the molecular structurefunctional relationships of the inhaled general anaesthetics. British Journal of Anaesthesia 198; 56: 9S-25S. 13. Wardley-Smith B, Halsey MJ. Mixtures of inhalation and i.v. anaesthetics at high pressure. British Journal of Anaesthesia 1985; 57: Hay don DA, Urban BW. The actions of some general anaesthetics on the potassium current of the squid giant axon. Journal of Physiology (London) 1986; 373 : Carlen PL, Gurevich N, O'Beirne M. Electrophysiological evidence for increased calcium-mediated potassium conductance by low-dose sedative hypnotic drugs. In: Rubin EP, Weiss D, Putney JW, eds. Calcium in Biological Systems. New York: Plenum, 198; Krnjevic K. Excitable membranes and anesthetics. In: Fink BR, ed. Cellular Biology and Toxicity of Anesthetics. Baltimore: Williams and Wilkins, 1972; Krnjevic K. Cellular and synaptic effects of general anesthetics. In: Roth SH, Miller KW, eds. Molecular and Cellular Mechanisms of Anesthetics. New York: Plenum, 1986; Yoshimura M, Higashi H, Fujita S, Shimoji K. Selective

12 VOLATILE ANAESTHETICS AND CNS SYNAPTIC TRANSMISSION 691 depression of hippocampal inhibitory potentials and spontaneous firing by volatile anesthetics. Brain Research 1985; 30: Maclver MB, Roth SH. Enflurane-induced burstfiringof hippocampal CA1 neurones. British Journal of Anaesthesia 1987; 59: Andersen P, Sundberg SH, Sveen O, Swann JW, Wigstrom H. Possible mechanisms for long-lasting potentiation of synaptic transmission in hippocampal slices from guinea pigs. Journal of Physiology (London) 1980; 302: Abraham WC, Bliss TVP, Goddard GV. Heterosynaptic changes accompany long-term but not short-term potentiation of the perforant path in the anaesthetized rat. Journal of Physiology (London) 1985; 363: Balestrino M, Aitken PG, Somjen GG. The effects of moderate changes of extracellular K and Ca on synaptic and neural function in the CA1 region of the hippocampal slice. Brain Research 1986; 377: Richards CD, Russell WJ, Smaje JC. The action of ether and methoxyflurane on synaptic transmission in isolated preparations of mammalian cortex. Journal of Physiology (London) 1975; 28: Krnjevic K, Morris ME, Reiffenstein RJ. Stimulationevoked changes in extracellular K and Ca in pyramidal layers of the rat hippocampus. Canadian Journal of Physiology and Pharmacology 1982; 60: Richards CD, White AE. The actions of volatile anaesthetics on synaptic transmission in the dentate gyrus. Journal of Physiology (London) 1975; 252: Lomo T. Patterns of activation in a monosynaptic cortical pathway: the perforant path input to the dentate area of the hippocampal formation. Experimental Brain Research 1971; 12: Landau EM, Richter J, Cohen S. The mean conductance and open-time of the acetylcholine receptor channels can be independently modified by some anesthetic and convulsant ethers. Molecular Pharmacology 1979; 16: Turner DA, Schwartzkroin PA. Passive electronic structure and dendritic properties of hippocampal neurons. In: Dingledine R, ed. Brain Slices. New York: Plenum, 198; Fricke RA, Prince DA. Electrophysiology of dentate gyrus granule cells. Journal of Neurophysiology 198; 51: Benardo LS, Masukawa LM, Prince DA. Electrophysiology of isolated hippocampal pyramidal dendrites. Journal of Neurosdence 1982; 2: Collingridge GL. Long term potentiation in the hippocampus: Mechanisms of initiation and modulation by neurotransmitters. Trends in Pharmacological Sciences 1985; 6: 07^ Roth SH, Tan K-S, Maclver MB. Selective and differential effects of barbiturates on neuronal activity. In: Roth SH, Miller KW, eds. Molecular and Cellular Mechanisms of Anesthetics. New York: Plenum, 1986; Weston GA, Roth SH. Differential actions of volatile anaesthetic agents on a single isolated neurone. British Journal of Anaesthesia 1986; 58: Maclver MB, Roth SH. Anesthetics produce differential actions on the discharge activity of a single neuron. European Journal of Pharmacology 1987; 139: Olsen RW, Fischer JB, Dunwiddie TV. Barbiturate enhancement of gamma-aminobutyric acid receptor binding and function as a mechanism of anesthesia. In: Roth SH, Miller KW, eds. Molecular and Cellular Mechanisms of Anesthetics. New York: Plenum, 1986; Gage PW, McKinnon D, Robertson B. The influence of anesthetics on postsynaptic ion channels. In: Roth SH, Miller KW, eds. Molecular and Cellular Mechanisms of Anesthetics. New York: Plenum, 1986; Macdonald RL, Skerritt JH, Werz MA. Barbiturate and benzodiazepine actions on mouse neurons in cell culture. In: Roth SH, Miller KW, eds. Molecular and Cellular Mechanisms of Anesthetics. New York: Plenum, 1986; Maclver MB, Roth SH. Barbiturate effects on hippocampal excitatory synaptic responses are selective and pathway specific. Canadian Journal of Physiology and Pharmacology 1987; 65: Roth SH. Physical mechanisms of anaesthesia. Annual Review of Pharmacology and Toxicology 1979; 19: