Abstract
This study examined the mechanism for hyperexcitability after ethanol withdrawal from isolated neonatal rat spinal cord. Ethanol (65–130 mM, 30 min) significantly depressed the glutamate receptor-mediated population excitatory postsynaptic potential (pEPSP) underlying the monosynaptic reflex. On washing with drug-free solution the response recovered to levels significantly above control. Minimum ethanol exposure time required for induction of withdrawal hyperexcitability was approximately 15 min. A second application of ethanol after washout depressed the pEPSP to an extent similar to the first, and a second wash did not elevate response significantly more than the initial wash. Ethanol-induced hyperexcitability thus develops with a time course of minutes and plays a role in determining apparent initial ethanol potency. Both α-amino-3-hydroxy-5-methyl-4-isoxazolepropionate and N-methyl-d-aspartate receptor-mediated components of the pEPSP were necessary for the expression of hyperexcitability on withdrawal but not for its induction. Butanol withdrawal also was associated with hyperexcitability, methanol was not. The case with octanol is uncertain because of slow recovery from this more lipophilic agent. Hyperexcitability on ethanol withdrawal was specific to the glutamate receptor-mediated pEPSP and not generalized to other evoked potentials. These results may be relevant to rapid and/or very rapid acute functional tolerance and to ethanol withdrawal.
In the course of studies on the properties of ethanol as a general anesthetic agent in spinal cord (Wong et al., 1997) we observed that the MSR and its underlying pEPSP recovered to levels above control on washout. In behavioral studies ethanol displays the properties of tolerance, manifested as a decrease in potency over time, and dependence, manifested as hyperexcitability on withdrawal. Tolerance is a complex phenomenon with more than one component; at least some components develop with a rapid time course measured in minutes after a single exposure (Pearson et al., 1997). Dependence also displays a short-term component after a single administration of ethanol (Buck et al., 1997). The present studies were carried out to determine the pharmacology and temporal properties of hyperexcitability after ethanol withdrawal in the spinal cord, to test whether it displayed the characteristics of tolerance as well as withdrawal and to examine whether it could be induced by exposure to other short-chain alcohols or was a phenomenon unique to ethanol.
Methods
Spinal cords from 2- to 7-day-old rats were removed after decapitation under enflurane anesthesia in a protocol approved by Stanford’s animal care and use committee. Isolated spinal cords were perfused at 4 ml/min with ACSF at 27–28°C, equilibrated with 95% O2/5% CO2, pH 7.3 to 7.4. The temperature is physiological for rats of this age when not under the mother. The pre-equilibrated ACSF was delivered from glass syringes mounted on an infusion pump set to a constant rate of approximately 4 ml/min. ACSF was of the following composition (mM): NaCl, 123; KCl, 5; NaH2PO4, 1.2; MgSO4, 1.3; NaHCO3, 26; CaCl2, 2; glucose, 30.
To evoke and record population-evoked potentials from the spinal cord, suction electrodes were arranged to stimulate a lumbar dorsal root and record from the corresponding ipsilateral ventral root (monosynaptic reflex and slow ventral root potential), or from an ipsilateral ventral root offset one to two segments from the dorsal root being stimulated (pEPSP). Single stimuli 0.2 ms in duration, 9 V nominal intensity, were delivered to the dorsal root at a constant frequency of 1/50 s throughout the experiment. Responses were digitized, averaged in groups of 5 and stored for later analysis.
Ethanol was obtained from commercial sources (Gold Shield Chemical Company, Hayward, CA) as the 95% pure compound, diluted to the desired concentration in ACSF and delivered to the preparation from syringes as outlined above. Concentrations of ethanol in the bath were verified by gas chromatography of the vapor phase in equilibrium with the solution in the chamber. Vapor phase partial pressures as per cent of one atmosphere v/v were converted into ACSF concentrations in millimolar by use of a saline-gas partition coefficient of 2650 at 37°C to give the equivalent saline partial pressure in milliliters per liter, correcting for the difference from absolute temperature and dividing by Avogadro’s number.
Four to twelve preparations were exposed to a single concentration of ethanol in each type of experiment. The following measurements were made: monosynaptic reflex amplitude; pEPSP area; the area under the curve of the first 8 s of the slow ventral root potential. In some experiments on the pEPSP, the selective glutamate receptor antagonists CPP and CNQX were used to block NMDA and AMPA receptor-mediated components, respectively (Tauck and Kendig, 1996). In untreated preparations the pEPSP was divided for further analysis into two components, an early point 20 ms after the stimulus maximally sensitive to CNQX and a later point 80 ms after the stimulus maximally sensitive to CPP, and the amplitudes at these points were measured separately. Data on amplitude and area were normalized to individual control values and analyzed by t test.
Results
As we have reported previously (Wong et al., 1997), ethanol depressed the monosynaptic reflex amplitude; the depression was significant at concentrations 200 mM and above. On washing for 30 to 60 min the response recovered to levels above control in some preparations (fig. 1A and 2A). The recovery was significantly above control at ethanol concentrations of 17.5, 70 and 557 mM (P < .05, fig 2B).
Monosynaptic reflex amplitude reflects both changes in the underlying glutamate-mediated excitatory synaptic transmission and in the excitability of the motor neurons. To discriminate between these bases for the increase in MSR amplitude, the pEPSP was monitored by recording one to two segments away from the stimulating electrode, where it appears as a graded depolarization uncontaminated by the large compound action potential of the MSR. At 130 mM ethanol, the pEPSP was depressed and recovered to levels significantly above control on washing (P < .05 at 30–60 min wash) (figs. 1B and3A). Time-matched control preparations without exposure to ethanol, although displaying a slight upward trend in pEPSP area over time, were not significantly different from their starting levels at the equivalent time period (fig. 3A).
Figure 3B shows the concentration-response curves for pEPSP depression and increase on washout. The pEPSP was depressed significantly at concentrations of 65 (P < .05), 130 and 300 mM (P < .01) but not at 32.5 mM. The pEPSP area after 60 min washout was significantly above control at 65 (P = .023) and 130 mM (P = .0024) but not at 32.5 or 300 mM (fig. 3B).
To relate the phenomenon to tolerance, four preparations were exposed to two applications of ethanol separated by a period of wash (fig.4). Although the second application tended to produce less depression than the first, the difference was not significant. After the second wash period, the pEPSP increased to a slightly higher level than after the first wash period, but again the difference was not significant.
A limited set of experiments was carried out to determine the minimum duration of exposure to ethanol required to generate an increase in the pEPSP on withdrawal of the alcohol. At 7 and 12 min exposures to 130 mM ethanol, depression of the EPSP area was not at steady state and was less than at 30 min; there was no increase above control on washout. At 15 min and 20 min exposure the extent of depression with 130 mM ethanol was similar to that at 30 min, and there was an increase above control on washout similar to experiments with 30 min exposure. Ten minutes exposure to 300 mM ethanol did not result in an increase above control on washout.
We previously have described the pEPSP as being made up of two components, one mediated via AMPA and a smaller one mediatedvia NMDA receptors (Tauck and Kendig, 1996). Both are sensitive to ethanol (Wong et al., 1997). To ascertain which components are required for the induction or expression of the potentiation observed on ethanol washout, preparations were treated with the selective NMDA receptor antagonist CPP or the AMPA/kainate antagonist CNQX before and during ethanol exposure. In the presence of CPP the AMPA receptor-mediated component did not recover beyond the level achieved previous to ethanol exposure (fig.5A). However, when CPP was washed off, the entire response recovered above the level that obtained before exposure to any pharmacologic intervention (fig. 5A). In the presence of CNQX the results were similar; recovery after ethanol exposure was only to the level observed with CNQX before ethanol exposure (fig. 5B). Recovery after CNQX was incomplete but still trending upward when the experiments were terminated. The isolated spinal cord has a reliable survival time of approximately 4 hr; counting the recovery period before measurements were begun these preparations were beginning to approach that time. Longer wash might have revealed an increase to levels above control or a leveling off as the limits of viability were reached. In separate experiments application and washout of CNQX and CPP, separately or together resulted in EPSP recovery only to control levels in the absence of ethanol exposure. These results suggest that both AMPA and NMDA receptor-mediated responses need to be present for the expression of the potentiation of the response after ethanol. NMDA receptors are not essential for induction, and AMPA receptors may not be.
To further document the roles of the respective components in expressing potentiation after ethanol, response amplitude was measured separately at latencies corresponding to maximal contribution by AMPA and NMDA receptors as outlined under “Methods.” Both components responded similarly to ethanol exposure, rising to levels above control on washing with ethanol-free solution (fig.6A).
Comparable studies were carried out with three other short-chain alcohols at a single concentration close to their respective anesthetic potencies in adult rats (Fang et al., 1997b). The results are shown in figure 6. All agents decreased the amplitudes of both components of the EPSP. Methanol (327 mM) was not associated with an increase in the EPSP on washing (fig. 6B), whereas 14 mM butanol was (P < .05) (fig. 6C). There was no increase after octanol exposure (60 μM), but recovery after this lipophilic agent was slow and the data points were still trending upward at the termination of the experiments.
Two other responses in the spinal cord, the slow ventral root potential and the dorsal root potential, were depressed by ethanol and recovered to control levels but not above. Data are shown for the slow ventral root potential (fig. 7).
Discussion
Depression of the pEPSP reflects a decrease in glutamate-mediated synaptic transmission. In the spinal cord the monosynaptic pEPSP has two components, a large fast AMPA-mediated response and a smaller depolarization mediated by NMDA receptors (Tauck and Kendig, 1996). Both are sensitive to ethanol (Wong et al., 1997). The present studies cannot distinguish between presynaptic depression of transmitter release and postsynaptic actions on the motor neurons. Other studies have shown that ethanol is capable of altering transmitter release as a function of intracellular calcium concentration (Rabe and Weight, 1988; Martin and Swartzwelder, 1992) as well as directly modifying responses to agonist stimulation of both AMPA (Dildy-Mayfield et al., 1996a, b; Morrisett and Swartzwelder, 1993) and NMDA receptors (Lovinger et al., 1990; Lovinger et al., 1989; Peoples and Weight, 1995;Dildy-Mayfield et al., 1996b; Morrisett and Swartzwelder, 1993). Indirect depression is also possible via enhancement of GABAA (Mihic et al., 1994; Weineret al., 1997) and/or glycine-mediated inhibition (Masciaet al., 1996; Aguayo et al., 1996), which might directly inhibit motor neurons and also enhance voltage-dependent magnesium block of NMDA receptors (Schummers et al., 1997). The intact isolated spinal cord cannot be used to distinguish between these possibilities by use of GABAA or glycine antagonists. Bicuculline, picrotoxin and strychinine, when applied in concentrations sufficient to block their respective receptors, induce seizure-like behavior. The base line becomes unstable and evoked response magnitude cannot be measured.
The hyperresponsiveness seen on ethanol withdrawal does not represent a generalized increase in spinal cord excitability, because it was not shared by two other evoked potentials in spinal cord, namely the slow ventral root potential and the dorsal root potential. The slow ventral root potential also is generated by motor neurons, because it is observed in single motor neurons recorded in ruptured patch mode (Lozier AP and Kendig JJ, unpublished data). The hyperresponsiveness we observed with ethanol is thus specific for glutamate postsynaptic potentials and appears to require both AMPA and NMDA receptors for its expression but not for its induction. It is thus clearly different from hyperresponsiveness after naloxone-precipitated withdrawal frommu opioids (Feng and Kendig, 1995a, 1996). The latter is a phenomenon of the slow ventral root potential and is blocked by an NMDA receptor antagonist (Feng and Kendig, 1995b, 1996). The restriction of the phenomenon to alkanols with a chain length greater than one carbon also indicates specificity of the effect.
Spinal cord hyperexcitability after ethanol withdrawal may be related to either or both of two behavioral phenomena, namely tolerance and withdrawal. Increase in excitability certainly can be related to withdrawal. That the phenomenon represents induction of a tolerant state is less certain. Reapplication of ethanol after a first exposure and washout did not reveal significant loss of potency. However, this may be because of the slow pharmacokinetics of the isolated spinal cord, in which drug access to and removal from sites of action is limited by diffusion across considerable distances. In the intact isolated spinal cord an ethanol exposure of 15 min is necessary and sufficient both to reach steady-state depression and to induce the hyperexcitable state revealed on withdrawal. This result suggests that induction of the latter is already well advanced by the time steady-state depression is reached and thus participates in the determination of the initial potency of ethanol in this preparation. Our results cannot discriminate between very rapid acute functional tolerance, with a time course of 5 to 7 min, which can be observed in isolated neurons (Pearson et al., 1997), and a more slowly developing acute functional tolerance demonstrated in behavioural experiments in mice, with a time course of approximately 30 min (Erwin and Deitrich, 1996).
There are several possible explanations for the increase in EPSP amplitude. Selective loss of GABAA or glycine-mediated inhibition, possibly because of a neurotoxic effect of high ethanol concentrations, is not likely. Because block of either form of inhibition by bicuculline or strychnine results in an increase in the slow ventral root potential (Gibbs and Kendig, 1992), such a loss would have been observed as a generalized, not a selective, increase in excitability. In behavioral studies sensitivity to the convulsant effects of strychnine is not altered during ethanol withdrawal (Gonzalez, 1993). In studies on hippocampal slices acutely exposed to ethanol, NMDA receptor-mediated EPSPs in the CA1 region display a rapid tolerance to ethanol developing during 15 min (Groveret al., 1994). NMDA receptor-mediated afterdischarges in dentate gyrus are enhanced after ethanol withdrawal (Morrisett, 1994). In neither of these studies was the mechanism for the phenomenon identified. Several studies also have been carried out in hippocampal slices from mice made tolerant to ethanol, which display hyperexcitability (Reynolds et al., 1990). The protocol for these studies includes days-long exposure to ethanol in vivo, and thus may represent a different phenomenon from the effects of short-term ethanol exposure in vitro. In recordings from hippocampal slices from tolerant mice, glutamatergic synaptic transmission is enhanced in the withdrawal period as are calcium currents (Whittington et al., 1995). This hyperresponsiveness was not correlated with changes in GABAA (Whittington et al., 1992)- or GABAB-mediated inhibition (Molleman and Little, 1995) but was reduced by a dihydropyridine calcium channel antagonist (Ripley et al., 1996). Although an NMDA receptor antagonist protected against withdrawal hyperexcitability while present (Ripley and Little, 1995a), the antagonist provided no protection if sufficient time was allowed for washout of antagonist from the brain (Ripley and Little, 1995b). The latter result is in accord with the results of the present study and suggests that NMDA receptors contribute to the expression of hyperexcitability on withdrawal but are not essential for its induction during the period of ethanol exposure.
These studies were motivated by an interest in ethanol as a general anesthetic agent (Wong et al., 1997). If ethanol resembles clinical inhalation agents, its potency in preventing movement in response to a noxious stimulus, which is a commonly accepted anesthetic endpoint, is determined by actions in the spinal cord. Because we wished to calibrate the appropriate ethanol general anesthetic potency in rats of the age used for the isolated spinal cord preparation, we carried out determinations of ethanol potency in 7-day-old compared with mature rats. There was a very large age-related potency difference, in which the general anesthetic blood concentration of ethanol required to prevent movement in adult rats was approximately 100 mM, whereas the corresponding value in neonates was more than twice that level, approximately 240 mM (Fang et al., 1997a). Age-related potency differences also occur with clinical general anesthetic agents, but only about a 20% increase in concentration is required for neonatal versus adult rats. Ethanol age-related potency differences in sleep time, which corresponds to regaining righting reflex, have also been reported; one study attributes them to a more pronounced tolerance in the young rat (Silveri and Spear, 1996). We could not directly compare spinal cord evoked responses in neonatal and adult rats in the present study, since intact cords from animals older than 7 days cannot be maintained in vitro.
Of the many effects of ethanol that could be related to age-dependent tolerance, actions on calcium-dependent kinases might be prominent candidates. A considerable literature documents ethanol effects on protein kinase C in particular (MacDonald, 1995). Mutant mice lacking the γ-isoform of protein kinase C are less sensitive than wild-type mice to ethanol, and GABAA receptor sensitivity to ethanol is also reduced in these mice (Harris et al., 1995). Mice lacking Fyn-kinase are more sensitive to ethanol than wild-type mice, and in hippocampal slices prepared from these mice NMDA receptor-mediated responses display no acute tolerance (Miyakawaet al., 1997). In this case ethanol is proposed to enhance tyrosine phosphorylation of an NMDA receptor in wild-type mice but not in mice lacking the kinase (Miyakawa et al., 1997). In cultured Purkinje neurons low ethanol concentrations increase calcium signals evoked by quisqualate application in cells 10 days in culture, but reduce the signal in cells 21 days in culture (Gruol and Curry, 1995). Protein kinases and other calcium-sensitive proteins known to play a developmental role are more prominent in the nervous systems, including spinal cord, of immature animals (Igwe and Filla, 1995;Mullany et al., 1996).
The present study describes and characterizes hyperexcitability of a glutamate AMPA and NMDA-mediated evoked potential after ethanol exposure in the spinal cord. The phenomenon requires both NMDA and non-NMDA receptors for its expression but is not blocked by NMDA and AMPA-kainate receptor antagonists during ethanol exposure. It is specific to this response and is not observed with the single-carbon alkanol methanol. The phenomenon may play a role in determining apparent initial ethanol potency in the isolated spinal cord, and may be related to tolerance and/or withdrawal in vivo.
Acknowledgments
We are indebted to the laboratory of E. I. Eger II at the University of California at San Francisco for carrying out the studies on ethanol potency in very young rats. Pompiliu Ionescu, MD of that laboratory provided the analysis of ethanol bath concentrations.
Footnotes
-
Send reprint requests to: Joan J. Kendig, Department of Anesthesia, Stanford University School of Medicine, Stanford, CA 94305.
-
↵1 Supported by National Institutes of Health grants NS13108 and GM47818 (to J.J.K.).
- Abbreviations:
- MSR
- monosynaptic reflex
- pEPSP
- population excitatory postsynaptic potential
- AMPA
- α-amino-3-hydroxy-5-methyl-4-isoxazolepropionate
- NMDA
- N-methyl-d-aspartate
- CPP
- 3-((R)-2-carboxypiperazin-4-yl)-propyl-1-phosphonic acid
- CNQX
- 6-cyano-7-nitroquinoxaline-2,3-dione
- ACSF
- artifical cerebrospinal fluid
- GABA
- γ-aminobutyric acid
- Received August 11, 1997.
- Accepted December 29, 1997.
- The American Society for Pharmacology and Experimental Therapeutics