Donald S. Faber and Manfred R. Klee
Research Institute on Alcoholism, 1021 Main Street, Buffalo, New York, 14203 and Max Planck Institute for Brain Research, 6 Frankfure/M-Niederrad, Deutschardenstr. 46, West Germany.

* Portions of this work were supported by NINCDS Grant # NS12132. The authors acknowledge the services of the Analytical Laboratory of the Research Institute on Alcoholism and thank Dr. Cedric M. Smith for his critical comments on the manuscript.

INTRODUCTION

Numerous investigations have revealed effects of ethanol on both neuronal excitability and synaptic transmission (cf., 1 for review). Since these studies have utilized a variety of invertebrate and vertebrate model systems and the concentrations employed have often been beyond those associated with moderate intoxication or ataxia, it is not surprising that a myriad of actions have been described for ethanol and that various mechanisms have been postulated to underlie its effects. Ideally, one would like to correlate actions of ethanol on single neurons and neuronal networks with stereotyped changes in behavior. However, while psychophysical tests,indicate the vertebrate central nervous system is most sensitive to ethanol (1), the majority of the neurophysiological studies have rather utilized peripheral vertebrate or isolated invertebrate preparations. Nevertheless, these investigations have yielded a great deal of neurophysiological information about the cellular mechanisms of action of ethanol. Such mechanisms generally fall in one of the following categories: 1) a specific reduction in neuronal excitability through alterations in the voltage dependent ionic conductances underlying action potential generation (3-6); 2) indirect effects on excitability consequent to changes in passive membrane permeabilities and resting membrane potential (3,4,7,8,9); and 3) pre- or post-synaptic changes in the efficacy of synaptic transmission, e.g. altered transmitter release or a change in the transmitter sensitivity of the post-synaptic membrane-bound receptors (10-18).

In the past few years we have analyzed the actions of ethanol in two model neural systems (4,12,15,19) both of which provide the opportunity to obtain stable intracellular recordings from identified neurons. The first is the abdominal ganglia of the marine mollusk Aplysia californica. This preparation is ideal for determining drug effects on membrane potential and resistance and spike electrogenesis, the latter being the membrane property altered most consistently by ethanol in different nervous systems. The results obtained from this preparation are interesting not only in elucidating mechanisms of action but also in emphasizing the fact that neurons are not all alike and have different reactions to ethanol. Presumably, such variations underlie the difficulties in establishing both the specific actions of ethanol and its selectivity within the central nervous system. More recently, we have studied its effects on the goldfish medullary network involving the Mauthner cell (M-cell). The results obtained from this preparation can be correlated with changes in a relatively simple behavior mediated by the M-cell, the startle reflex. Furthermore, there is a striking correspondence between the minimal brain ethanol concentrations necessary to alter the functional organization of this network on the one hand and the behavior of the fish on the other hand (15,19). That is, a relatively specific neuronal effect can be demonstrated at a "physiological" dose level, and additional effects occurring at higher brain ethanol levels can be evaluated in terms of progressive changes in behavior. In this paper, we review and compare the basic actions of ethanol as revealed by these two preparations, in order to develop a more comprehensive understanding of its effects on central nervous system networks.

THE INVERTEBRATE SINGLE-CELL APPROACH

The experiments presented here involved intracellular recordings from large neurons in the isolated visceral and pleural ganglia of the marine mollusk Aplysia californica. These nerve cells have diameters between 100 and 800 pm and the locations and characteristics of about thirty identifiable cells in the abdominal ganglion have been described extensively (20). Cells can be further distinguished on the basis of their responses to putative neurotransmitters (20-25), their spontaneous activity and endogenous rhythms (20) and their electrogenic properties (26-30). Therefore, precise investigations of a specific cell or group are possible, as are comparative studies within the same ganglion.

Because of the size and unipolar structure of the cells, not only voltage recording but also current and voltage clamp techniques can easily be used in this preparation. Voltage clamping allows the direct study of the conductance changes and current flows underlying action potential generation at the level of the cell soma, and the control of this process. Furthermore, characteristics of synaptic transmission can be studied both by stimulating presynaptic inputs and by analyzing the responses of single cells to iontophoretic applications of transmitter candidates. By combining these different approaches the actions of drugs such as ethanol on a broad spectrum of membrane properties can be specified.

Effects on Membrane Potential and Resistance

Effects of ethanol on the resting membrane potential (RMP) of Aplysia neurons were variable, different effects occurring in different cell groups (4,12). Ethanol depolarized one class of cells, hyperpolarized a second, and had no significant effect on the RMP of a third! Within each cell group, moreover, the results were consistent. As will be discussed below these different cell groups have other distinguishing membrane characteristics as well. All the effects of ethanol described in this and the following sections were concentration-related and were completely reversible.

The magnitude of the depolarizations produced by 4% ethanol ranged from 5 to 20 mV. The strongest depolarizations were in cell L-11 (Figure 1), which characteristically had a RMP of around -30 to -40 mV. Within the initial ten minutes of perfusion with 4% ethanol the RMP of this cell was reduced by about 50% and no spikes were generated. In other cells which were consistently depolarized the size of the potential shift was less and it did not result in an inactivation of the spike generating mechanism. In fact, despite these depolarizations most of the cells were less excitable and an increase in the threshold for spike initiation was found with both orthodromic and antitlromic stimulations and with trans-membrane current injections.

Fig. 1. Differential effects of ethanol on resting membrane potential of Aplysia neurons. A,B: simultaneous intracellular recordings were obtained from cells L-11 (upper records) and R-15 (lower records). A: ethanol (4%) was added to the ASW at the arrow and shortly thereafter L-11 was depolarized until finally the spike generating mechanism was inactivated. In contrast, R-15 hyperpolarized and the membrane potential oscillations underlying its bursting behavior ceased. B: after 15 minutes total exposure to ethanol, the perfusion fluid was switched back to ASW at the arrow. Within minutes the cells began to recover from the effects of ethanol.

Ethanol-induced depolarizations have also been described for frog muscle fibers, and an increase of resting sodium permeability was postulated by Knutsson and Katz (9) as the mechanism underlying this action. We have directly confirmed this hypothesis by comparing the effects of ethanol in the standard artificial sea water (ASW) perfusing the ganglion and in sodium-free sea water. In four cells which were depolarized an average of-12 mV (range 8-17 mV) by 4% ethanol, perfusion with Na+ free ASW resulted in a membrane hyperpolarization which averaged 15 mV (range 7-24 mV). This somewhat large increase in RMP agrees with the relatively high PNa/PK ratio reported by Carpenter (31) for these neurons. In three of the four cells, the depolarization produced by ethanol was completely blocked in the sodium-free solution. In the fourth cell ethanol produced the same magnitude depolarization in the sodium deficient ASW as in normal ASW, but the absolute membrane potential was 24 mV greater in the former than in the latter. These results are, therefore, consistent with the hypothesis that ethanol increased resting NaP

In contrast with the above results another class of cells that was consistently hyperpolarized by ethanol in the concentration range of 2 to 4% could be identified (Figure 1). These hyperpolarizations were dose dependent and reached a maximum of 12 mV. Figure 1 illustrates this effect on the bursting pacemaker cell R-15 (20,32). The membrane potential of this cell normally oscillates with a depolarizing phase triggering a train of spikes and in turn being followed by a hyperpolarizing wave. In 4% ethanol the oscillations were dampened, and the membrane was hyperpolarized by about 5 to 10 mV. These cells could also be distinguished by the fact that they are all depolarized by acetylcholine (ACh) in contrast to the previous group of cells which are inhibited by ACh.

The alterations in membrane potential produced by ethanol were not associated with any significant changes in membrane resistance, which is in contrast to the small increases in resting membrane conductance reported for squid giant axon (3) and frog muscle fibers (8). Generally, effects of ethanol on membrane resistance are not striking and probably do not contribute significantly to its actions.

Effects on Action Potential Generation

In all Aplysia neurons investigated, ethanol (2 to 5%) produced reductions in spike amplitude of 20 to 50% (Figure 2 and references 4 and 12). This effect was mainly due to a decreased rate of rise of the action potential. Since the inward current underlying spike initiation in molluscan neurons is carried by both Nal' and Ca++, (29,30) we tested the possibility that both conductance mechanisms are altered by ethanol. This was the case; ethanol reduces the size of action.- potentials remaining in either sodium-free or calcium deficient ASW. Furthermore, when the sodium channels were specifically blocked by the addition of 30 MM tetrodotoxin (TTX) to the ASW, 4% ethanol reduced the TTX-insensitive component of the spike by the same amount as in sodium-free ASW. That is, ethanol appears to block both the sodium and calcium components of the action potential.

Fig. 2. Effects of ethanol on the action potential of the giant cell in the left pleural ganglion. In all records the simultaneously recorded action potential (upper traces) and its electrically differentiated representation (lower traces) are shown. Al: control recordings. A2: after 10 min exposure to 2% ethanol, spike amplitude decreased and its duration increased. A3 and Bl: progressive recovery of the spike after washing in ethanol-free ASW. B2: the effects of ethanol on spike generation were more pronounced when its concentration was raised to 5%. B3: again, washing the preparation with ASW reversed the effects of ethanol.

Fig. 3. The effect of two doses of ethanol on the I/V relationship obtained during voltage clamp experiments. Dose-dependent reduction of the inward current system by 2 and 4% ethanol and unchanged amplitude of the outward current in the giant cell
+60R-2. Abscissa: absolute membrane potential; holding mV potential was 40 mV. Ordinate: current, outward current is positive (from reference 4).

Voltage clamp experiments are required to discriminate quantitatively the effects of ethanol on the different ionic currents underlying spike electrogenesis. Previous experiments using squid giant axon had produced conflicting results; Armstrong and Binstock (3) reported that ethanol reduced on the transient voltage-dependent increase in sodium conductance underlying spike initiation, while Moore et al., (6) found a similar reduction in the delayed potassium current involved in membrane repolarization as well. Our Aplysia experiments were undertaken to resolve this difference and to determine the effects of ethanol on the early inward calcium current not seen in the squid axon. As shown in Figure 3, these studies revealed a relatively specific effect of ethanol on the early inward current. In addition, by altering the Na+ and Ca."' concentrations in the ASW, it was possible to demonstrate that both components of this current were equally blocked by 2 to 4% ethanol.

These results confirm the conclusion drawn from the effects of ethanol on spike amplitude and are consistent with the decreased neuronal excitability seen in many systems after ethanol application. Furthermore, since the sodium and calcium currents can be dissected pharmacologically and presumably involve different membrane channels (30), the evidence that ethanol blocks both channels may speak against a specific interaction with the membrane gates controlling channel activation. Rather, it seems reasonable to suggest that it produces a change in membrane structure which consequently alters the characteristics of the channels. In this context the membrane expansion theory of Seeman (33,34) and others merits consideration. The essence of this theory is that anesthetics such as ethanol adsorb to hydorphobic regions of excitable membranes, thereby expanding these regions of membrane proteins and distorting the ionic conductance channels. This theory, however, does not explain why some channels are altered and others such as the voltage dependent K+ ones, are not.

Fig. 4. Ethanol slows the recovery of the inward current system from the inactivation produced by a brief conditioning depolarization. The interval between conditioning and test depolarizing pulses was varied from 2 to 100 msec. Ordinate: inward current evoked by the test
pulse expressed as a fraction of the maxim mum current. Abscissa: interval between conditioning and test pulses (4).

Experiments on the kinetic properties of the early inward currents revealed an additional significant action of ethanol. It prolonged the recovery of this system after a conditioning stimulus which evoked a maximal increase in Na+ and Ca++ conductance (Figure 4). In terms of membrane excitability, this observation implies an increase in the relative refractory period and reduction in the maximal firing rate of the neuron. This finding is consistent with recent'experiments on cat spinal motoneurons in which one pronounced effect of ethanol was to increase the critical interval at which the cell could follow paired antidromic stimuli (Figure 5; reference 35). Other changes produced by doses of 200-500 mg/kg body weight include membrane hyper-polarizations of 2-10 mV and a slight reduction in spike overshoot. The latter is probably due to the hyperpolarization rather than to a specific action on the voltage-dependent Nat' channels since there is no effect on the rising phase of the action potential. At appreciably higher dose levels reductions in both mono- and polysynaptic excitatory postsynaptic potentials (EPSPs) are also seen, which is in agreement with previous reports (36,37).

Fig. 5. Increase in the critical interval of cat motoneurons after ethanol injections and impaired spike transmission from the initial segment (IS) to the somadendritic (SD) membrane. A-C: intracellular recordings from a motoneuron. Paired antidromic stimuli were used. A: control responses; stimuli were at the critical interval such that the IS-SD delay was increased for the second response. B: 90 sec. after an i.v. ethanol injection (500 mg/kg), the SD component of the second spike failed. Same stimulus interval as in A. C: 8'45" after the ethanol injection the critical interval is still increased by 20%. In all records voltage responses and their electrically differentiated representations are superimposed.

Effects on Synaptic Transmission

Barker (10) has proposed a comprehensive theory for the action of anesthetics based on his observations that in various invertebrate preparations all anesthetics tested (including ethanol) specifically reduced EPSPs involving increased postsynaptic cation conductances. All other PSPs remained unchanged. A similar conclusion was reached by Chase (13) who found ethanol depressed EPSPs in another mollusk, but he apparently did not study its effects on inhibitory PSPs (IPSPs). In our experiments (Bergmann et al., in preparation) blocking effects on synaptic transmission have been observed with ethanol concentrations (0.5-4.0%) appreciably lower than those necessary to alter membrane potential and spike electrogenesis. However, we have not found the specificity described by Barker. To the contrary, ethanol blocks both excitatory and inhibitory postsynaptic responses (Figures 6 and 7, column B). Furthermore, we have demonstrated that these effects are mediated postsynaptically by investigating the responses of these neurons to iontophoretically applied ACh. Aplysia neurons have been classified as ACh-D- and H-cells depending upon whether they are depolarized or hyperpolarized by ACh, and there are both different receptors mediating these responses and different ionic conductances involved. Specifically, D- responses are due to an increase in membrane conductance to either Na+ and Ca++ ions or to Cl- ions, while H responses can be either Cl- or I dependent. The fact that there are Cl- dependent D- and H- responses reflects variations in the Cl- equilibrium potential relative to resting membrane potential. Despite this evidence for different ACh receptors, we have found that ethanol blocks most ACh responses tested in a dose-dependent manner, as is illustrated in Figures 6 and 7, column A, for one excitatory and two inhibitory responses.

Fig. 6. Reduction of ACh-D-responses and EPSPs by ethanol on a RB neuron. In each line depolarizing responses to iontophoretic applications of ACh (A) with decreasing currents (17.4 to 1.74 x 10-7 A) are shown on the left (column A) and EPSPs are shown on the right (column B). From top to bottom control responses (C) and responses recorded during perfusion with 0.5, 2.0 and 4.0% ethanol are shown. As shown in the lower record the blocking action L.J- can be overcome with a double application of the maximal ACh stimulus.

Fig. 7. Ethanol reduces the Cl- and K+ components of the ACh-H-responses of a medial cell in the pleural ganglion. Al: with increasing intensities of ACh iontophoresis a two-component H -response is recorded at a resting potential of -40 mV. B: when resting potential is shifted to -70 mV, the early C1- component is reversed to a depolarizing response while the late K+ component is still seen as a hyper-polarization. A2, B2: 2% ethanol reduces both components, its effect on the early C1- component being stronger. A3, B3: recovery after 30' washing in ethanol-free ASW.

Not all PSPs are blocked by ethanol, and we have found some EPSPs which are extremely resistant to its action. In fact, Woodson et al., (18) have recently focused on one such EPSP and have demonstrated that the only detectable action of ethanol is to accelerate the decay of post-tetanic potentiation. Therefore, it appears that Barker's hypothesis is untenable and that ethanol has selective effects on both EPSPs and IPSPs. Finally, it should be pointed out that these effects of ethanol on cholinergic responses are the opposite of those generally described for sympathetic ganglia (38) and the vertebrate motoneuron junctions with Renshaw cells (39) and motor endplates (16,17), at which ethanol may rather enhance synaptic transmission.

EXPERIMENTS UTILIZING THE M-CELL NETWORK AS A CNS MODEL

The Mauthner cells are a pair of neurons found in the medulla of many fish and amphibia. Since the pioneering studies of the late 50's and early 60's (40-43) our knowledge of the physiology and function of this cell and its associated network has expanded greatly, and the preparation now constitutes an ideal model for the study of many important neurobiological problems. Some relevant background information on the goldfish M-cell, which served as our experimental model (Figure 8) is reviewed here. The M-cell is easily recognized morphologically on the basis of its position and relatively large soma (ca 40 pm diameter) and two main dendrites, one running laterally and the other ventrally. It has the largest and fastest conducting axon in the goldfish spinal cord by a factor of at least 2. Furthermore, as the M-cell antidromic action potential generates a characteristic extracellular negative field which can be as large as 30-50 mV in the vicinity of the cell's axon hillock (Figure 9-A1) and the intracellularly recorded spike is correspondingly smaller (Figure 9-A2), (40), the cell soma and lateral dendrite can be easily identified electrophysiologically.

Fig. 8. The schematic model of the M-cell network. The M-cell,receives an input from a pool of inhibitory interneurons (I) and excitatory inputs from the VIIIth nerves and both ipsi- and contralateral posterior lateral line nerves (P. lat. line Nvs.). These afferent inputs also project to the inhibitory interneuronal pool, as do axon collaterals from both M-cells.

Afferent excitatory synapses from the ipsilateral vestibular nerve and the posterior lateral line nerves to the lateral dendrites have been extensively studied (44-48), as has the recurrent collateral network mediating somatic inhibition of the M-cell (Figure 8; reference 41, 47-50). Korn and Faber (49,50) have recently identified the interneurons which are activated by M-cell collaterals and in turn feed inhibition back onto the M-cell. They have further shown that the excitatory afferent inputs to the M-cell also converge onto these interneurons (47). Consequently, these cells, which appear functionally similar to spinal Renshaw cells, mediate both collateral and afferent inhibition of the Mauthner neuron. A remarkable feature of this network is the presence of electrical inhibitions mediated by the interneurons and the M-cell. A discussion of these interesting phenomena is not pertinent to this chapter; however, it is important to point out that spike activity in the inhibitory interneurons produces both electrical and chemical inhibition of the M-cell (51). The former is characterized by an extracellular postivity which is named the EHP and is recorded in the vicinity of the M-cell axon hillock (Figure 9-A1). The EHP then, can be used as a sign of the activation of these neurons, e.g. by M-cell axon collaterals, while the chemical component of the inhibition can be recorded intracellularly from the M-cell either as an IPSP or as an increased membrane conductance (Figure 9-B1). The latter is most easily demonstrated as the reduction in the amplitude of a second antidromic spike when paired spinal stimuli are used (41). It is generally accepted that activation of one Mauthner cell results in contraction of the trunk and tail musculature on the contralateral side of the body through a monosynaptic excitation of spinal motoneurons. The M-cell mediates the goldfish startle or tail flip reflex (42,52,53), which is a basic escape reflex initiated by auditory, lateral line or visual inputs. Since the excitatory transmission from the eighth nerve to the M-cell is so powerful, and there is generally a one-to-one relationship between the firing of the M-cell and that of the spinal cord motoneurons (43) the controls on this reflex are centered in the medulla. Specifically, the control is exerted by the inhibitory network described above (Figure 8). It has therefore been possible to directly compare the effects of ethanol on the neurons in this network with any alterations in its functional output (15,19).

Behavioral Effects of Ethanol

Behavioral effects of 1% ethanol (w/v) on freely swimming goldfish were initially determined by gross observations. Brain ethanol concentrations in these fish reached an equilibrium level of approximately 10 µg/mg brain weight in approximately two hours. During the first thirty to sixty minutes the fish underwent a phase of hyperexcitability characterized by hyperreflexia, including an enhanced startle reflex, poorly coordinated swimming and gulping of air at the water surface. These behavioral effects correlated with brain ethanol levels of 3-5 Vg/mg brain weight. As the ethanol concentration increased, the fish sank to the bottom of the aquarium, became grossly ataxic and overturned, with the loss of their righting reflexes.

Fig. 9. Ethanol blocks both the electrical and chemical components of collateral inhibition of the M-cell. In all records the M-cell was activated antidromically by spinal cord stimulation. Al, A2: extracellular recordings from the axon cap. Al: the negative M-cell spike is followed by the positive EHP (arrow) which mediates the electrical inhibition. A2: 35 min after starting perfusion with 1% ethanol, the EHP was strongly reduced. Final brain ethanol level 20 min later was 4.7 pg/mg brain wt. Bl, B2: intracellular recordings from the M-cell soma illustrating that ethanol blocks the collateral IPSP as well. Bl: control. With paired antidromic stimuli the increased conductance associated with the IPSP following the first stimulus either blocks the second spike or reduces its amplitude by about 50%. B2: after 60 nziry perfusion with 1% ethanol this conductance change was almost completely abolished. Brain ethanol concentration was 3.45 pg/mg. Cl, C2: intracellular recordings from the M-cell in another experiment illustrating the reversibility of the ethanol effect. Cl: the IPSP was partially blocked during perfusion with 1% ethanol for 45 min. C2: the solution perfusing the gills was then switched to an ethanol-free one. 25' later the collateral inhibition was markedly increased. Voltage calibrations in A2 and B2 apply to Al and Bl, respectively, the time scale in B2 for Al, A2 and Bl, and voltage calibration in C2 for Cl.

Effects of M-cell Excitability and Collateral Inhibition

In these experiments ethanol (1 to 2%) was applied through the fluid perfusing the gills and the development of its action was observed during two hours of continuous intracellular recordings from the M-cell. Final brain ethanol concentrations were determined with the techniques of gas chromatography.

Ethanol at a concentration of 3-5 pg/mg of brain tissue reduces both the electrical (Figure 9-Al, -A2) and chemical (Figure 9-B1, -B2) components of the collateral inhibition. Also, as illustrated in Figure 9-C1, -C2 this blockage is reversed when ethanol is removed from the perfusion fluid. In contrast to this action on collateral inhibition, the M-cell resting membrane potential (generally between 70-80 mV) remains unchanged even at levels as high as 20 pg/mg and only minor changes in its antidromic action potential occurred at the lower ethanol levels. No effects were seen on the action potential recorded from the M-cell's axon. However, when the brain ethanol concentration was in the range of 6.5-15 ug/mg brain weight minor effects on excitability were occasionally seen with intracellular recordings at the level of the M-cell soma. Specifically, the safety factor for transmission of the antidromic impulse to the axon hillock was reduced and invasion sometimes failed (Figure 10-A). Under these conditions a full action potential could be restored by pairing a depolarizing current with the antidromic axon spike (Figure 10-B-D), results which suggest an increase in threshold at the axon hillock. This differential effect of ethanol on spike electrogenesis at the axon hillock as opposed to the axon itself is similar to its differential action on the initial segment and some-dendritic membranes of motoneurons. It might be related to a change in the diameter of the involved regions or to basic regional differences in membrane properties.

Fig. 10. Ethanol may reduce the safety factor of spike transmission from the M-cell axon to the axon hillock. A-D: intracellular M-cell records obtained after a  freely-swimming fish reached the behavioral level of "overturn" in 1% ethanol. Respiration was continued with a 1% ethanol solution during surgery and recordings A with antidromic stimuli only a brief axon spike was recorded paired stimuli were used). B: when a depolarizing current pulse (upper trace) was paired with the antidromic stimulus the axon spike was then capable of evoking a full-sized M-cell spike (superimposed records with and without the current pulse). C,D: expanded M-cell antidromic action potentials and their electrically differentiated representations, respectively. Superimposed records obtained with and without a depolarizing pulse to facilitate transmission to the axon hillock. Voltage calibration in B pertains for A and time scale in C is for D as well.

Fig. 11. Evidence that while ethanol blocks collateral inhibition of the M-cell, it does not block afferent inhibitions mediated by the same interneurons. A-D: intracellular M-cell recordings obtained from one experiment during ethanol exposure. A: double antidromic stimuli demonstrate the absence of a collateral IPSP. B,C: stimulation of the ipsilateral VIIIth nerve produced a mono-synaptic EPSP supra threshold for the M-cell (B) and a conditioning antidromic stimulus did not inhibit its effectiveness. D: there is a marked conductance increase when a threshold VIIIth nerve stimulus is used, as indicated by the reduction in the test antidromic action potential. This is due to the inhibitory input to the M-cell. Final ethanol level was 3.1 pg/mg. E-H: M-cell intracellular records from another experiment demonstrate that the afferent inhibition is not blocked even at ethanol levels as high as 13.8 pg/mg. E,F: controls. E: paired antidromic stimuli demonstrating the magnitude of the collateral inhibition. F: interposing a subthreshold VIIIth nerve stimulus between the antidromic stimuli added an additional inhibitory component, as is seen by the further reduction in the test antidromic spike. G-H: records obtained during perfusion with 2% ethanol. The collateral IPSP has disappeared. G: while the VIIIth nerve input still produces a marked inhibition of the M-cell. H: in all traces 2 or more superimposed records are shown; calibrations in D and H also pertain for A -C and E-G, respectively.

Site of Action of Ethanol

The blockage of collateral inhibition by ethanol was the only effect observed at the low concentrations. In fact, no other synaptic inputs to the M-cell were reduced by ethanol, even at higher concentrations. Figure 11 illustrates the fact that neither the excitatory nor inhibitory inputs from the ipsilateral eighth nerve were depressed at concentrations equal to or greater than those which blocked the collateral inhibition. Two sets of findings, therefore, indicate that the depression of collateral inhibition actually occurs at the level of the synapse between the M-cell axon collaterals and the inhibitory interneurons (Figure 9): 1) the blockage of the EHP unambiguously demonstrates a failure of excitation of the interneurons, and 2) the lack of an effect on the afferent inhibitions mediated by the same interneurons is proof that excitation of the interneurons can still produce an inhibition of the M-cell.

Mechanism Underlying Ethanol's Action on Collateral Inhibition

Three possible mechanisms by which ethanol could depress synaptic transmission from M-cell axon collaterals to the inhibitory interneurons have been considered: (1) a presynaptic effect on spike propagation in the axon collaterals, (2) a depression of transmitter release, and (3) a reduction in the sensitivity of the subsynaptic receptors of the inhibitory interneurons to the M-cell transmitter. The first possibility seems unlikely since depression of M-cell excitability only occurred at higher ethanol levels. Nevertheless, a selective effect on the thinner axon collaterals cannot be ruled out. The results of the experiments described below strongly suggest that ethanol acts presynaptically to impair transmitter release. We routinely used a repetition rate of 1/3-1/7 seconds in these experiments. Under these conditions, a stable maximal IPSP can be observed (Figure 12-A; note the IPSP is reversed, i.e. depolarizing, in this figure since a KC1 electrode was used for the intracellular recordings). At the lower repetition rate, no effect of 1% ethanol was seen after twenty-five minutes of perfusion (Figure 12-8) but there was typically a 50-70% reduction ten to fifteen minutes later (Figure 12-C). After one hour exposure, the IPSP appeared to be abolished (Figure 12-D). However, we observed at that time a small IPSP could be restored by switching to a lower stimulus frequency and that a full-sized one could be evoked for one stimulus after a one minute period of rest without spinal cord stimulation (Figure 12-E); a second stimulus seven seconds later was completely ineffective (Figure 12-E). As the action of ethanol progressed further, the effectiveness of such one minute rest periods diminished appreciably (Figure 12-F) and longer rest periods in the range of five minutes were necessary (Figure 12-G). Again, however, only the first in a train of stimuli at a 1/7 second repetition rate evoked an IPSP. Finally, after eighty to ninety minutes total perfusion time with ethanol, such rest periods were completely ineffective and the IPSP was completely blocked (Figure 12-H and -I). At that time, the ethanol level in the experiment illustrated in Figure 12 was 4.51 pg/mg brain weight, and we estimate that it was no more than 2-3 pg/mg brain weight during the first hour of exposure.

Fig. 12. Time course of the development of the ethanol effect on the collateral IPSP. The records were obtained with KCLcontaining microelectrode. Three or more superimposed traces of the M-cell responses to spinal cord stimulation at a rate of 1/7 seconds are shown in A-I. A: control. The antidromic action potential elicited by a single shock to the M-axon is followed by a depolarizing IPSP which causes the cell to initiate a second spike. B -I: records obtained at the indicated times after starting perfusion with 1% ethanol. The traces illustrated in E, F and H were preceded by one minute rest periods during which no stimulation was employed, and those illustrated in G and I were preceded by five minute rest periods. B-D: gradual reduction and, finally, blockage of the IPSP over a fifty-eight minute period. E: at fifty-nine minute ethanol, a one minute rest restored the IPSP for the first stimulus in a train with a repetition rate of 1/7 seconds. F: diminished effectiveness of the one minute rest ten minutes later. G: at seventy-four minutes, a five minute rest restored the IPSP for one stimulus to a magnitude greater than that in control. H: eighty-one minutes after starting perfusion with ethanol, the one minute rest was completely ineffective and no IPSP was evoked. I: after an additional five minute rest, an IPSP less than 10% the amplitude of that illustrated in control could be evoked by the first stimulus in the 1/7 second train (2).

The evidence that such prolonged rest periods are capable of transiently restoring synaptic transmission suggests that ethanol acts on the process of transmitter release from the M-cell axon collaterals, and does not appear readily compatible with the effects on collateral excitability or the sensitivity of the postsynaptic neurons to the M-cell transmitter. This conclusion is consistent with the biochemical studies of Kalant et al., (54), which indicated that incubation with ethanol inhibits acetylcholine release from rat cerebral cortex slices. The correlation is further strengthened by evidence that the transmitter released by the goldfish M-cell (Faber and Klee, unpublished observations) is ACh, as is also the case with the hatchet fish M-cell (55). It is also interesting that Weakly (56) similarly concluded that barbiturates act presynaptically to depress transmitter release in the mammalian spinal cord.

Our preliminary behavioral observations indicate that the initial phase of hyperexcitability following exposure to ethanol occurs at the same brain ethanol levels which rather selectively depress excitatory synaptic transmission from the M-cell collaterals onto the interneurons exerting a major inhibitory control of the startle reflex. In addition, higher anesthetic levels may be associated with direct effects on excitability as well. Clearly, more quantitative behavioral experiments on the effects of ethanol and additional experiments on the other components of the neural network involved in the startle reflex are needed. Nevertheless, it is clear that this preparation indeed offers a promising approach to the general problem of correlating physiological and behavioral aspects of drug actions.

DISCUSSION

Comparison of the results obtained with the Aplysia and M-cell models provides a framework for some general conclusions concerning the actions of ethanol. One is that the mechanisms of synaptic transmission appear more sensitive to ethanol than either spike electrogenesis or the membrane properties controlling resting membrane potential. This conclusion is based largely on the evidence that cholinergic transmission is depressed in both systems by ethanol concentrations significantly lower than those which alter the other processes. Furthermore, the actions on synaptic transmission are selective and not all PSPs are blocked or reduced by ethanol. We, therefore, suggest that effects on electrical excitability are secondary to those on synaptic transmission. The latter would be primarily responsible for the more specific ethanol effects which are manifested behaviorally as the hyperexcitability, loss of motor coordination, etc., and the former would contribute to its general depressant properties.

Ethanol enhances cholinergic transmission both in the cat spinal cord (motoneurons to Renshaw cells, reference 39) and at the frog neuromuscular junction (16,17). It is not clear why a depression occurs at some cholinergic junctions and a facilitation at others. Possible explanations include regional differences in the characteristics of transmitter synthesis and release and differences in postsynaptic receptors. One interesting experiment would be to determine if the synaptic transmissions mediated in the spinal cord by M-cell axon collaterals are also blocked by ethanol or if it has different actions on different presynaptic processes of the same cell.

One of the most intriguing aspects of our results is that in both the cat spinal motoneuron and goldfish Mauthner cell ethanol did not produce a generalized depression of spike electrogenesis. Rather the M-cell axon hillock and the motoneuron soma-dendritic membranes were preferentially depressed. This selective action would contribute to delaying or blocking synaptic activation of the neurons and would reduce their maximum firing rates. The mechanism underlying this selectivity is not clear; it may relate to the lack of myelination of these membrane regions or to a lower density of voltage-dependent Na+ channels.

The above observations point directly to a situation which neurophysiology has recently come to face. Namely, the membrane properties of neurons differ and there is no standard cell which can be used for neuropharmacological investigations. The opposing effects of ethanol on RMP of different Aplysia neurons is another clear demonstration of this problem. It is therefore almost impossible to evaluate the significance of these RMP changes, and the general consequences of such a situation are clear; in order to determine the mechanisms of action of a substance such as ethanol it is first necessary to identify the systems responsible for its behavioral effects. Then, drug action should be studied in this system and concentration levels maintained in that range with which the behavioral change is obtained. In this context, as described above, the M-cell and its associated network offers an ideal model system in which all of the basic tools available can be applied to such problems. It is therefore gratifying that the results agree: behavioral hyperexcitability is correlated with a disinhibition of the M-cell due to a specific effect on one synpatic system, and a more general anesthetic action is correlated with decreased neuronal excitability. In fact, we consider ourselves lucky that ethanol has such a clear action on this system. More often the systems best suited for such a combined approach turn out to be those which do not react to the drug being studied!

REFERENCES

1. Waligren, H. and Barry, H.: Cellular basis of ethanol action on the nervous system, Actions of Ethanol, Volume 1, Biochemical, Physiological and Psychological Aspects. Edited by Waligren, H. and Barry, H., Elsevier, New York, p. 209-273 (1970).
2. Smith, C.: The pharmacology of sedative/hypnotics: Sites and mechanisms of action, Handbook of Experimental Pharmacology: Drug Addiction. Edited by Martin, W.N., Springer-Verlag, New York (in press).
3. Armstrong, C.M. and Binstock, L.: The effects of several alcohols on the properties of the squid giant axon. J. gen. Physiol. 48:265-277 (1964).
4. Bergmann, M.C., Klee, M.R. and Faber, D.S.: Different sensitivities to ethanol of three early transient voltage clamp currents of Aplysia neurons. Pflugers Arch. ges. Physiol. 348:139-153 (1974).
5. Moore, J.W.: Effects of ethanol on ionic conductances in the squid axon membrane. Psychosom. Med. 28:450-457 (1966).
6. Moore, J.W., Ulbricht, W. and Takata, M.: Effect of ethanol on the sodium and potassium conductances of the squid axon membrane. J. gen. Physiol. 48:279-295 (1964).
7. Gallego, A.: On the effect of ethyl alcohol upon frog nerve. J. Cell. comp. Physiol. 31:97-106 (1948).
8. Knutsson, E.: Effects of ethanol on the membrane potential and membrane resistance of frog muscle fibres. Acta physiol. scand. 52:242-253 (1961).
9. Knutsson, E. and Katz, S.: The effect of ethanol on the membrane permeability to sodium and potassium ions in frog muscle fibres. Acta pharmacol. (kbh.) 25:54-64 (1967).
10. Barker, J.L.: Selective depression of postsynaptic excitation by general anesthetics, Molecular Mechanisms of Anesthesia, Progress in Anesthesiology, Vol. 1. Edited by Fink, B.R., Raven Press, New York, p. 134-153 (1975).
11. Barker, J.L. and Gainer, H.: Pentobarbital: Selective depression of excitatory postsynaptic potentials. Science 182:720-722 (1973).
12. Bergmann, M.C., Faber, D.S. and Klee, M.R.: Reduction of the early inward sodium and calcium currents of Aplysia neurons by ethanol. Pflugers Arch. 332:R66 (1972).
13. Chase, R.: The suppression of excitatory synaptic responses by ethyl alcohol in the nudibranch mollusc. Tritonia Diomedia. Comp. Biochem. Physiol. 506:37-40 (1975).
14. Davidoff, R.A.: Alcohol and presynaptic inhibition in an isolated spinal cord preparation. Arch. Neurol. 28: 60-63 (1973).
15. Faber, D.S. and Klee, M.R.: Ethanol suppresses collateral inhibition of the goldfish Mauthner cell. Brain Research 104:347-353 (1976).
16. Gage, P.W.: The effect of methyl, ethyl and N-propyl alcohol on neuromuscular transmission in the rat. J. Pharmacol. exp. Ther. 150:236-243 (1965).
17. Okada, K.: Effects of alcohols and acetone on the neuromuscular junction of frog. Jap. J. Physiol. 17:245-261 (1967).
18. Woodson, P.B.J., Traynor, M.E., Schlapfer, W.T. and Barondes, S.H.: Increased membranes fluidity implicated in acceleration of post-tetanic potentiation by alcohols. Nature 260:797-799 (1976).
19. Faber, D.S. and Klee, M.R.: Effects of ethanol on collateral inhibition of the goldfish Mauthner cell. Soc. for Neurosci., 5th Ann. Mtg. (1975).
20. Frazier, W.T., Kandel, E.R., Kupfermann, I., Waziri, R. and Coggeshall, R.E.: Morphological and functional properties of identified neurons in the abdominal ganglion of Aplysia californica. J. Neurophysiol. 30:12881351 (1967).
21. Ascher, P.: Inhibitory and excitatory effects of dopamine on Aplysia neurons. J. Physiol. (Lond.) 225:173-209 (1972).
22. Blankenship, J.A., Wachtel, H. and Kandel, E.C.: Ionic mechanisms of excitatory, inhibitory, and dual synaptic actions mediated by an identified interneuron in abdominal ganglion of Aplysia. J. Neurophysiol. 34:76-92 (1971).
23. Faber, D.S. and Klee, M.R.: Strychnine interactions with acetylcholine, dopamine and serotonin receptors in Aplysia neurons. Brain Research 65:109-126 (1974).
24. Gerschenfeld, H.M. and Paupardin-Tritsch, D.: Ionic mechanisms and receptor properties underlying the responses of molluscan neurones to 5-hydroxytryptamine. J. Physiol. (Lond.) 243:427-456 (1974).
25. Kehoe, J.S.: Three acetylcholine receptors in Aplysia neurones. J. Physiol. (Lond.) 225:115-146 (1972).
26. Alving, B.O.: Differences between pacemaker and non-pacemaker neurons of Aplysia on voltage clamping. J. gen. Physiol. 54:512-531 (1969).
27. Smith, T. G., Barker, J.L. and Gainer, H.: Requirements for bursting pacemaker potential activity in molluscan neurones. Nature 253:450-452 (1975).
28. Faber, D.S. and Klee, M.R.: Membrane characteristics of bursting pacemaker neurons in Aplysia. Nature New Biol. 240:29-31 (1972).
29. Geduldig, D. and Gruener, R.: Voltage clamp of the Aplysia giant Neurone: Early sodium and calcium currents. J. Physiol. (Lond.) 211:217-244 (1970).
30. Geduldig, D. and Junge, D.: Sodium and calcium components of action potentials in the Aplysia giant neurone. J. Physiol. (Lond.) 199:347-365 (1968).
31. Carpenter, D.O.: Membrane potential produced directly by the Nal" pump in Aplysia neurons. Comp. Biochem. Physiol. 35:371-385 (1970).
32. Strumwasser, F.: The demonstration and manipulation of a circadian rhythm in a single neuron, Circadian Clocks. Edited by Aschoff, J., North-Holland Publishing Company, Amsterdam (1965).
33. Seeman, P.: The membrane actions of anesthetics and tranquilizers. Pharmacol. Rev. 24:583-655 (1972).
34. Seeman, P.: The membrane expansion theory of anesthesia, Molecular Mechanisms of Anesthesia, Progress in Anesthesiology, Vol. 1. Edited by Fink, B.R., Raven Press, New York, p. 243-251 (1975).
35. Klee, M.R., Lee, K.C. and Park, M.R.: Changes in membrane properties of cat motorneurons due to ethanol. Exp. Brain Res. 23 supp1:108 (1975).
36. Eidelberg, E. and Wooley, D.F.: Effects of ethyl alcohol upon spinal cord neurones. Arch. int. Pharmacodyn. 185:388-396 (1970).
37. Lathers, C.M. and Smith, C.M.: Ethanol effects on muscle spindle afferent activity and spinal reflexes. J. Pharmacol. exp. Ther. 197:126-134 (1976).
38. Larrabee, M.G. and Posternak, J.M.: Selective action of anesthetics on synapses and axons in mammalian sympathetic ganglia. J. Neurophysiol. 15:91-114 (1952).
39. Meyer-Lohmann, J., Hagenah, R., Hellweg, C. and Benecke, R.: The action of ethyl alcohol on the activity of individual Renshaw cells. Naunyn: Schmiedeberg's Arch. Pharmacol. 272:131-142 (1972).
40. Furshpan, E.J. and Furukawa, T.: Intracellular and extracellular responses of the several regions of the Mauthner cell of the goldfish. J. Neurophysiol 25:732771 (1962).
41. Furukawa, T. and Furshpan, E.J.: Two inhibitory mechanisms in the Mauthner neurons of goldfish. J. Neurophysiol. 26:140-176 (1963).
42. Wilson, D.M.: Function of giant Mauthner's neurons in the lungfish. Science 129:841-842 (1959).
43. Diamond, J.: The Mauthner cell, Fish Physiology, Vol. V. Edited by Hoar, W.S. and Randall, D.J., Academic Press, New York, p. 265-346 (1971).
44. Robertson, J.D., Bodenheimer, T.S. and Stage, D.E.: The ultrastructure of Mauthner cell synapses and nodes in goldfish brains. J. Cell Biol. 19:159-199 (1963).
45. Furshpan, E.J.: "Electrical transmission" at an excitatory synapse in a vertebrate brain. Science 144:878-880 (1964).
46. Korn, H. and Faber, D.S.: Inputs from the posterior lateral line nerves upon the goldfish Mauthner cell. I. Properties and synaptic localization of the excitatory component. Brain Research 96:342-348 (1975).
47. Faber, D.S. and Korn, H.: Inputs from the posterior lateral line nerves upon the goldfish Mauthner cell. II. Evidence that the inhibitory components are mediated by interneurons of the recurrent collateral network. Brain Research 96:349-356 (1975).
48. Nakajima, Y.: Fine structure of the synaptic endings on the Mauthner cell of the goldfish. J. Comp. Neur. 156: 375-402 (1974).
49. Korn, H. and Faber, D.S.: An electrically mediated inhibition in the goldfish medulla. J. Neurophysiol. 38: 452-471 (1975).
50. Faber, D.S. and Korn, H.: A neuronal inhibition mediated electrically. Science 179:577-578 (1973).
51. Faber, D.S. and Korn, H.: Evidence that single vertebrate CNS neurons can mediate both electrical and chemical inhibitions. Soc. for Neurosci., 6th Ann. Mtg. (1976).
52. Eaton, R.C. and Farley, R.D.: Mauthner neuron field potential in newly hatched larvae of the zebrafish. J. Neurophysiol. 38:502-512 (1975).
53. Zottoli, S.J.: Correlation of the startle reflex and Mauthner cell auditory responses in unrestrained goldfish. J. exp. Biol. (in press).
54. Kalant, H., Israel, Y. and Mahon, M.A.: The effect of ethanol on acetylcholine synthesis, release and degradation in brain. Canad. J. Physiol. Pharmacol. 45:172-176 (1967).
55. Weakly, J.N.: Effect of barbiturates on 'Quantal' synaptic transmission in spinal motoneurones. J. Physiol. (Lond.) 204:63-77 (1969).
56. Spira, M., Model, P.G. and Bennett, M.V.L.: Cholinergic transmission at a vertebrate central synapse. J. Cell Biol. 47:199a (1970).