Abstract
Ethanol actions on α-amino-3-hydyroxy-5-methyl-4-isoxazolepropionic acid (AMPA)-type glutamate receptors were studied using voltage-clamp recordings from mouse cortical and hippocampal neurons. During whole-cell recordings ethanol (EtOH) inhibited AMPA receptor-mediated currents in a dose-dependent manner at concentrations from 10 to 500 mM. The steady-state component of AMPA-activated current was more sensitive to EtOH than the peak component. To examine the effect of EtOH on a well resolved peak current component, patches were excised from cultured cortical neurons, to which AMPA and EtOH were applied using a piezoelectric solution application system. Under this condition, the peak current was not inhibited significantly by EtOH. To further study possible mechanisms of EtOH inhibition, kainate and AMPA were used to evoke currents in the absence and presence of cyclothiazide. Ethanol inhibition was stronger when receptors were activated by low than high kainate concentrations. Cyclothiazide reduced inhibition by EtOH regardless of the agonist used to activate the receptor. Finally, EtOH inhibition was reduced in a point mutated (L497Y) GluRAi receptor that lacks desensitization. These findings suggest that EtOH inhibits AMPA receptors by stabilizing the desensitized state. Our results can explain some of the variation observed in EtOH inhibition in previous studies, and support the idea that physiologically relevant concentrations of EtOH can have a strong effect on AMPA receptor function.
Glutamate is the major excitatory neurotransmitter in the central nervous system (CNS). It produces its physiological actions by activating several metabotropic and ionotropic receptors (for review, see Lovinger, 1997). The different ionotropic glutamate receptors are divided into three subtypes based on their primary amino acid sequence, and the agonists that best activate them. These three classes of receptor have been named the N-methyl-d-aspartate (NMDA), kainate, and α-amino-3-hydyroxy-5-methyl-4-isoxazolepropionic acid (AMPA) subtypes. Glutamate activation of AMPA receptors (AMPARs) is thought to mediate most fast synaptic excitatory neurotransmission in the brain, whereas transmission via kainate receptors contributes only a minor component.
Ethanol (EtOH) has been shown to act on several targets in CNS, and the glutamate receptors are among these sites of action. Ethanol decreases the function of all three classes of ionotropic receptors at concentrations in the physiologically relevant range (Lovinger et al., 1989; Weiner et al., 1999). Ethanol is also known to inhibit glutamatergic synaptic transmission (Lovinger et al., 1990). Although the majority of studies have focused on potent inhibition of NMDA receptors by EtOH, the effect on AMPARs is not as well studied. This is in part due to the fact that EtOH inhibition of AMPA receptors is rather weak in many neuronal preparations (Lovinger et al., 1989; Weiner et al., 1999) and AMPAR-mediated synaptic transmission is not greatly inhibited by EtOH at hippocampal synapses (Lovinger et al., 1990; Weiner et al., 1999). However, reasonably potent inhibition of AMPAR function by EtOH has been observed in heterologous expression systems and in some neuronal preparations when receptor function is studied using direct agonist application to isolated cells or oocytes (Dildy-Mayfield and Harris, 1992; Lovinger, 1993; Dildy-Mayfield and Harris, 1995; Wirkner et al., 2000). These studies demonstrated EtOH inhibition of AMPARs at concentrations ranging from those below the legal intoxication limit (10 mM or ∼50 mg/dl) to near lethal concentrations (100 mM or ∼500 mg/dl). Wirkner et al. (2000) found that EtOH inhibition of NMDA and AMPA receptors is likely caused by noncompetitive inhibition distinct from open channel block (Peoples et al., 1997). In oocytes, EtOH inhibition of AMPARs seems to vary with agonist concentration, with inhibition being greater at low agonist concentrations (Dildy-Mayfield and Harris, 1992). These results suggest that AMPA receptors might possess considerable EtOH sensitivity under certain conditions or in certain preparations, but little is known about the interactions with receptor-channel function.
AMPA receptors undergo profound desensitization during agonist exposures lasting milliseconds. With the exception of kainate, all of the widely used AMPAR agonists evoke strong desensitization of the receptor (Trussell et al., 1988; Tang et al., 1989). Kainate produces desensitization that is extremely rapid and nearly undetectable, because it is thought to be rapidly reversible, and hence is weak in comparison to the desensitization produced by other agonists (Patneau et al., 1993). The time constants of desensitization in response to AMPA or glutamate measured in outside-out membrane patches have been shown to range from 1 to 16 ms (Tang et al., 1989; Trussell and Fischbach, 1989; Hestrin, 1992; Barbour et al., 1994). It is thought that desensitization does not contribute to AMPAR-mediated transmission at intact synapses (Colquhoun et al., 1992; Hestrin, 1992; Diamond and Jahr, 1995), because of the rapid clearance of the transmitter from the synaptic cleft (Clements et al., 1992). However, this process may come into play in cases of unusually prolonged synaptic transmission, as observed in experiments in various CNS neurons, where the excitatory postsynaptic current decay reflects the rate of desensitization rather than deactivation (Trussell et al., 1993; Barbour et al., 1994; Otis et al., 1996; Maguire, 1999).
The fast time course of desensitization necessitates that rapid drug application should be used to properly resolve the peak current when AMPARs are activated by most agonists. Past studies examining EtOH inhibition of AMPAR function have not taken this factor into account, and there has been no attempt to determine whether EtOH alters desensitization or whether inhibition varies when receptors are activated under conditions that produce different degrees of desensitization. In the present study, we examined the effect of EtOH on AMPARs in acutely isolated as well as cultured CNS neurons. Our aim was to examine the effect of EtOH on different components of AMPAR-mediated current evoked by different agonists. Our findings suggest that EtOH inhibits AMPAR function with a pharmacologically relevant potency, and that inhibition is mainly due to stabilization of receptor desensitization. Some of the results described in this manuscript are in abstract form (Möykkynen et al., 2001).
Materials and Methods
Isolation of Hippocampal Cells. C57Bl/J6 mice (10–20 days old) were decapitated and the whole brains were moved to ice-cold high-sucrose solution (containing in 194 mM sucrose, 30 mM NaCl, 4.5 mM KCl, 1 mM MgCl2, 26 mM NaHCO3, 1.2 mM NaH2PO4, and 10 mM glucose, bubbled with 95% O2, 5% CO2 to achieve pH 7.4). Coronal brain slices (400 μm in thickness) were cut using a manual vibroslice (Camden Instruments Ltd., Leicester, UK). Slices were kept at room temperature in artificial cerebrospinal fluid (containing 124 mM NaCl, 4.5 mM KCl, 1 mM MgCl2, 26 mM NaHCO3, 1.2 mM NaH2PO4, 10 mM glucose, and 2 mM CaCl2, bubbled continuously with 95% O2, 5% CO2). Hippocampal neurons were isolated by incubating slices in pronase (0.4–0.6 mg/ml; Calbiochem, San Diego, CA) in artificial cerebrospinal fluid at 37°C for 20 to 30 min. Slices were transferred to trituration buffer (containing 20 mM NaCl, 130 mM N-methyl-d-glucamine, 2.5 mM KCl, 1 mM MgCl2, 10 mM HEPES, and 10 mM glucose, osmolarity adjusted with sucrose to 340 mOsM, pH adjusted to 7.4 with HCl), and the hippocampus was cut free of the rest of the tissue using a scalpel. Neurons were dissociated by gentle mechanical trituration with fire-polished Pasteur pipettes. Cells were then allowed to settle to the bottom of a 35-mm-diameter culture dish, which was then transferred to the stage of an inverted microscope.
Maintenance and Transfection of Human Embryonic Kidney (HEK) 293 Cells. HEK 293 cells were maintained in culture as described previously (McCool et al., 1996). Cells to be used for transfection were plated onto 35-mm-diameter culture dishes in medium containing Dulbecco's modified Eagle's medium with 4.5 g/l glucose, 4 mg/ml pyridoxine, 110 mg/l sodium pyruvate, 100 U/ml penicillin, 100 μg/ml streptomycin, and 2 mM l-glutamine. One to 2 days after plating, cells were transfected via calcium-phosphate precipitation using HEPES-buffered saline solution combined with 2.5 mM CaCl2 and the appropriate concentration of the cDNAs. Cells were transfected with either wild-type GluRAi (i, flip splice variant; 2–5 μg/dish) or the L497Y (numbering refers to full-length protein as in Stern-Bach et al., 1998) mutant GluRAi (2 μg/dish). cDNA encoding enhanced green fluorescent protein (EGFP; 1–2 μg/dish) was included in each transfection to mark successfully transfected cells. cDNAs were contained in mammalian expression vectors containing a cytomegalovirus promoter. Cells were exposed to cDNA for 1 day and then washed with standard feeding medium. Cells were then examined electrophysiologically beginning 16 to 24 h after washing. EGFP-positive cells were identified by epifluorescence.
Electrophysiological Recordings. During experiments the cells were continuously superfused with recording solution (containing 150 mM NaCl, 2.5 mM KCl, 2.5 mM CaCl2, 1 mM MgCl2, 10 mM HEPES, and 10 mM glucose, pH adjusted to 7.4 with NaOH, osmolarity to 340 mOsM with sucrose). Experiments were carried out at room temperature. Whole-cell and outside-out patch-clamp recording techniques were used (Hamill et al., 1981). Patch pipettes were pulled from 1.5-mm o.d. glass capillary tubing (WPI, Sarasota, FL) using a micropipette puller (model P-87; Sutter Instrument Co., Novato, CA). Pipette resistance was 2 to 4 MΩ, when filled with an internal solution containing 100 mM N-methyl-d-glucamine, 100 mM CH3SO3H, 40 mM CsF, 10 mM MgCl2, 10 mM HEPES, and 5 mM EGTA, pH adjusted with CsOH to 7.4, osmolarity 290 to 300 mOsM. Drugs and EtOH were diluted in recording solution and applied to whole cells with a multibarrel fast solution application system (Warner Instrument, Hamden, CT). In the majority of experiments, neurons were lifted clear of the bottom of the culture dish to facilitate solution exchange. Ethanol and agonist were applied simultaneously in most experiments. Preapplication of EtOH did not substantially alter experimental outcomes (see Results). To measure the solution exchange times in this system we applied 1 mM kainate to the cells to activate the AMPARs. The solution was then changed to a solution lacking Na+, which almost completely eliminated the kainate-evoked ion current. The exchange time for cells sitting in the bottom of the dish was 150 ± 60 ms and for lifted cells 22 ± 5 ms. Drug applications varied in duration from 100 ms to 8 s. Under the conditions used, the transient peak current component was observed only when the solution delivery pipettes were placed directly above the cells with the solution stream flowing directly onto the cell.
An Axopatch 1-D or Multiclamp amplifier and pClamp6.0 or pClamp8.0 software (Axon Instruments, Inc., Foster City, CA) were used to acquire and analyze current recordings. During experiments neurons were voltage clamped at –60 mV, unless otherwise mentioned. Currents were filtered with a 5-kHz low-pass filter and digitized at up to 100 kHz.
To better resolve EtOH effects on predesensitized peak currents we used a piezoelectric application system (model PZS-200; Burleigh Instruments, Inc., Victor, NY) to apply drugs to excised membrane patches with exchange times, measured for open tip potential change, being 4 ± 2 ms. Patches were pulled from cultured cortical neurons, which were made from mice at postnatal days 1 and 2 using methods described in Strack et al. (1999). The experiments were carried out after 7 to 14 days in culture. The drug application pipette was made from a theta tube glass capillary (Warner Instrument). The outer diameter was cut to about 150 μm. The recording electrode was placed at an angle of ∼90 degrees with respect to the drug application pipette. Current evoked by 100 μM AMPA in the absence and presence of 100 mM EtOH was examined using drug application durations of 1 and 700 ms.
In all experiments, each drug application was repeated at least twice and averaged. Control applications, agonist without EtOH, were measured at least twice between every pair of applications of agonist with EtOH. The effect of EtOH was calculated as percentage of inhibition relative to the averaged control values before and after EtOH application. EtOH concentration-response curves for were well fit using a single binding-site isotherm of the form Y = Bmax · X/(Kd + X).
AMPA concentration-response curves were fit with a standard four-parameter logistic equation of the form Y = Emin +(Emax – Emin)/[1 + (log EC50 – X)n], where Emin is the minimal response, Emax is the maximal response, EC50 is the concentration giving a half-maximal response, X is log concentration, and n is the Hill slope factor. No parameters were constrained during curve fitting.
Results are given as mean ± S.E.M. Statistical differences among groups were determined using ANOVA with post hoc comparisons using Bonferroni's test, and in some cases with a one sample t test. The criterion for statistical significance was p < 0.05.
All drugs were purchased from commercial sources. S-AMPA, kainate, (2S,4R)-4-methylglutamate (SYM 2081), and cyclothiazide were purchased from Tocris Cookson Inc (Ballwin, MO). Ethanol was purchased from Aaper Alcohol and Chemical (Shelbyville, KY). Concanavalin A was from Sigma-Aldrich (St. Louis, MO).
Results
Characterization of AMPA-Evoked Current in Isolated Hippocampal Pyramidal Neurons. Application of AMPA for 2.8 s to isolated hippocampal pyramidal neurons at a holding potential of –60 mV evoked inward currents of stable amplitude with characteristic fast peaks followed by a steady-state phase (Fig. 1A). Increasing concentrations of AMPA (1–1000 μM) were applied to determine the concentration-response relationship for this agonist (Fig. 1B). Concentration-response curves were fit with the logistic equation as shown in Fig. 1B and estimated EC40 (10 μM) and EC80 (100 μM) concentrations of AMPA from steady-state responses were used in later experiments.
Ethanol Inhibition of AMPA-Mediated Current. Ethanol inhibited AMPA-evoked current in a concentration-dependent manner in all isolated neurons measured. Figure 2, A and B, shows that EtOH produced prominent inhibition of the steady-state component of current activated by 100 μM AMPA as well as the peak and steady-state components of current activated by 10 μM AMPA. Inhibition by EtOH was reproducible with multiple drug applications, and reductions in current amplitude were fully reversed upon removal of EtOH from the cell (Fig. 2). Figure 2C also illustrates the stability of AMPA-evoked currents during prolonged recording from isolated hippocampal neurons.
Previous studies had suggested that EtOH inhibition of AMPARs in heterologous systems differed at different agonist concentrations (Dildy-Mayfield and Harris, 1992). We thus examined EtOH inhibition when the AMPAR was activated by either 10 (EC40), or 100 μM AMPA (EC80) in isolated hippocampal neurons (Fig. 3). EtOH inhibition of the agonist-activated peak current was significantly greater when receptors were activated by 10 μM compared with 100 μM AMPA (Fig. 3A). In contrast, no agonist concentration-dependent difference in inhibition of steady-state current was detected (Fig. 3B).
The dependence of EtOH inhibition on agonist concentration might reflect an interaction with the binding of agonist, but could also result from an interaction with channel gating properties that are dependent on agonist concentration. To begin to assess this latter possibility, we examined EtOH effects on several components of the whole-cell current, including the rapid-onset, transient peak of AMPA-evoked current, the subsequent exponential decay, and the steady-state component observed during continued agonist applications lasting 1.5 to 8 s, because these aspects of whole-cell current likely result from differential contributions of various channel state transitions such as desensitization/resensitization of the receptor channel. By comparing EtOH inhibition of these different current components in isolated hippocampal neurons we hoped to gain an initial idea about what channel properties are altered by EtOH. To see distinct peak and steady-state responses a reasonably high agonist concentration, 100 μM AMPA, was used. Both the peak and steady-state current components were inhibited by EtOH (see example responses in Fig. 2B). However, EtOH produced significantly greater inhibition of the steady-state than the peak response at all concentrations tested. For example, the inhibition by 100 mM ethanol of steady-state current was 35 ± 2% (mean ± S.E.M.; n = 13) and only 15 ± 1% for the peak response (Fig. 3C). Application of EtOH for at least 1 min before combined AMPA + EtOH application did not alter the relative inhibition of peak and steady-state current components. In a group of seven neurons 100 mM EtOH produced 19 ± 2% inhibition of peak current with EtOH preapplication or 14 ± 3% without preapplication, whereas inhibition of steady-state current averaged 30 ± 3% with and 33 ± 2% without EtOH preapplication. Steady-state/peak ratios for 100 μM AMPA with and without EtOH were also calculated for these recordings. The ratio decreased with increasing alcohol concentration (Fig. 3D).
The time course of the exponential decay of 100 μM AMPA-activated currents was determined using nonlinear curve fitting in neurons lifted from the bottom of the culture dish (to enhance solution exchange rate). Decay was well fit with a single exponential function both in the absence and presence of EtOH. The time constant (τ) for decay of current during agonist exposure was similar in the presence and absence of EtOH (for 50 mM EtOH experiments, baseline = 42 ± 7 ms, EtOH = 42 ± 7 ms; for 100 mM EtOH experiments, baseline = 35 ± 4 ms, EtOH = 33 ± 5 ms; p > 0.1, paired t tests).
Ethanol Does Not Alter AMPA-Induced Peak Current in Excised Membrane Patches. Membrane patches were excised from cultured cortical neurons to which drugs were applied using the piezoelectric applicator. The fast application system allowed drugs to be applied to patches with an exchange time of 4 ± 2 ms, faster than the time constant of desensitization. Thus, we could more accurately resolve the peak of AMPA-induced current using this system. Ethanol, even at the 100 mM concentration, did not affect the peak amplitude of current evoked by 100 μM AMPA (current in the presence of 100 mM EtOH averaged 99 ± 3% of the amplitude of control currents; Fig. 4A). The average time constant for current decay in the presence of agonist was 11.1 ± 1.7 ms, which is longer than the solution exchange time in our system (Fig. 4B), and is lower than that observed in the whole-cell mode. The decay time constant in the presence of agonist and 100 mM EtOH averaged 10.8 ± 1.8 ms, which was not significantly different from agonist alone (p > 0.05, paired t test, n = 5 patches). The steady-state component of current in excised membrane patches was too small to be resolved, as shown in Fig. 4A, and thus we could not estimate EtOH inhibition of this current component under this condition. EtOH inhibition of the steady-state current in whole-cell recordings from cultured cortical neurons was confirmed to be similar to that observed in isolated hippocampal neurons (∼40%) using the stepper motor driven solution applicator (data not shown).
We also examined recovery from desensitization of 100 μM AMPA-induced current in the absence and presence of 100 mM EtOH in isolated hippocampal neurons. This was accomplished using a two-pulse protocol in which AMPA was applied for 100 ms followed by a wash period of 200 to 1000 ms, and subsequent reapplication of AMPA (Fig. 5). When the effect of EtOH was examined, EtOH was applied throughout the duration of the recording protocol regardless of whether agonist was present. The peak amplitude of AMPA-induced current recovered with a time constant of 189 ms (95% confidence interval 175–204 ms, n = 9 cells) and this value increased to 233 ms (95% confidence interval 217–250 ms, n = 9 cells) in the presence of EtOH (Fig. 5B). Thus, EtOH produced a modest slowing of the rate of recovery from desensitization.
Ethanol Inhibition Varies with Kainate Concentration. Kainate activates AMPA receptors in a manner that produces extremely rapid desensitization and resensitization, resulting in steady-state desensitization that is less complete than that produced by AMPA. Application of kainate (30 μM–1 mM) to isolated hippocampal neurons produced current that showed no decay during agonist applications of up to 8 s. When receptors were activated by a maximally effective concentration of kainate (1 mM), relatively little inhibition was observed (only 23 ± 2% in the presence of 100 mM EtOH). This was considerably less than the inhibition produced by 100 mM EtOH on the steady-state component of current produced by 100 μM AMPA (Fig. 3). This result is not surprising given that overall receptor desensitization is less in the presence of the high kainate concentration in comparison to the high AMPA concentration. However, it is possible that EtOH might produce greater inhibition of responses to lower kainate concentrations where receptor occupancy is lower and more receptors desensitize from a single agonist-bound state (Patneau and Mayer, 1991). Indeed, it was previously reported that EtOH inhibition of kainate-evoked currents from Xenopus oocytes expressing rat hippocampal and cortical mRNA decreased with increasing agonist concentration (Dildy-Mayfield and Harris, 1992). To determine whether inhibition was sensitive to kainate concentration in isolated hippocampal neurons, we examined EtOH effects in the presence of 30 and 300 μM kainate, and compared the inhibition to that observed in the presence of 1 mM kainate. We observed a marked EtOH inhibition of low (30 μM) kainate-evoked currents (Fig. 6A). Figure 6B shows that inhibition produced by EtOH was dependent on both agonist and EtOH concentrations and was greater at low than high kainate concentrations. When we performed within-cell comparisons, we observed that 100 mM EtOH produced 44 ± 4% inhibition of 30 μM kainate-evoked currents, whereas the inhibition was only 23 ± 13% for currents activated by 300 μM kainate (p < 0.05, n = 6).
Kainate is also an agonist for kainate-type glutamate receptors, and therefore it was necessary to determine whether functional kainate receptors contributed to the current we observed. The currents activated by kainate in our isolated neurons did not exhibit the rapid decay during agonist application that is characteristic of kainate activation of kainate-preferring ionotropic glutamate receptors (Ozawa et al., 1998). Thus, it is unlikely that kainate receptors are involved in generation of the currents we measured. To be sure that kainate-activated currents could not be evoked in the isolated neurons, a selective kainate receptor agonist, SYM 2081 (Donevan et al., 1998), was applied to neurons in the presence of concanavalin A (Con A) to block rapid desensitization. SYM 2081 at a concentration of 2 μM along with 0.3 mg/ml Con A (SYM 2081 and Con A concentrations that produced near-maximal activation of kainate-preferring GluRs; Donevan et al., 1998) evoked currents with amplitudes smaller than 5 pA in the neurons examined (data not shown). Normal responses to 100 μM AMPA and EtOH inhibition of AMPA-induced current were observed in these same neurons. These findings suggest that kainate-activated current is mediated solely through AMPA receptors in the neurons examined.
Cyclothiazide Reduces EtOH Inhibition of AMPA-Activated Current. Cyclothiazide stabilizes the open state of AMPA receptors (Kessler et al., 1996), and one effect of this compound is to greatly reduce desensitization. If EtOH inhibits AMPAR function by stabilizing desensitization we would expect this inhibition to be reduced in the presence of cyclothiazide. We thus examined EtOH (50 and 100 mM) effects on current activated by 10 μM AMPA with and without 100 μM cyclothiazide in isolated hippocampal neurons. Cyclothiazide produced a profound enhancement of the peak AMPA-activated current, and eliminated the fast, transient current component (Fig. 7A). Ethanol produced greater inhibition of current activated by AMPA alone in comparison with current evoked by AMPA and cyclothiazide. Inhibition by 100 mM ethanol averaged 39 ± 6 and 21 ± 6% in the absence and presence of cyclothiazide, respectively (Fig. 7B). However, current in the presence of cyclothiazide, measured at –60 mV, had amplitudes in the nanoampere range, and thus were substantially larger in amplitude than those evoked by AMPA alone. It is, therefore, possible that reduced inhibition by EtOH might reflect problems with voltage or space clamp. To determine whether clamp-control problems contribute to the measured EtOH inhibition, we examined AMPA-activated current in the absence and presence of cyclothiazide at lower holding potentials so that current was reduced by reducing the driving force. Lowering the holding potential decreased the amplitudes of responses to a range of a few hundred picoamperes. No differences in ethanol inhibition were observed when comparing the results at –60 and –20 and –10 mV. EtOH (100 mM) inhibited currents 19 ± 5% at –20 to –10 mV and 19 ± 4% at –60 mV (n = 5). This finding indicates that EtOH inhibition is not dependent on current amplitude, and that clamp-control problems are unlikely to account for the differential inhibition of current in the absence and presence of cyclothiazide. These results support the idea that there is a difference in the EtOH sensitivity of AMPA receptors that undergo desensitization relative to those that do not.
Cyclothiazide also reduced EtOH inhibition of kainate-activated current (Fig. 7C). In the presence of 100 μM cyclothiazide, the inhibition of 30 μM kainate-activated current by 100 mM EtOH was only 21 ± 2%, whereas without cyclothiazide it was 44 ± 4%, suggesting that disruption of desensitization alters EtOH inhibition when kainate is used as agonist.
Ethanol Inhibition in WT and L497Y Mutant GluRAi. To further explore the relationship between desensitization of AMPA-type glutamate receptors and EtOH inhibition, we examined alcohol effects in wild-type GluRAi receptors in comparison with a mutant receptor (L497Y) that lacks desensitization (Stern-Bach et al., 1998). The receptors were expressed transiently in HEK 293 cells after transfection with the appropriate cDNA along with EGFP as a marker. Transfected cells were examined with whole-cell patch-clamp recording. Agonist and alcohol application was initiated after lifting cells from the bottom of the culture dish, as described above. Application of 100 μM AMPA produced currents that exhibited rapid desensitization. Application of EtOH produced dose-dependent inhibition of current activated by 100 μM AMPA in HEK 293 cells expressing wild-type GluRAi receptors (Fig. 8). Inhibition of the steady-state current component was greater than inhibition of peak current, as observed in neurons. Inhibition of the steady-state current component was comparable in magnitude to that observed in neurons (percentage of inhibition in wild-type GluRAi = 46 ± 2%) The current activated by 10 μM AMPA was insufficiently large in amplitude in these cells to allow measurement of inhibition by EtOH.
As previously reported, application of 10 or 100 μM AMPA to cells expressing L497Y receptors evoked relatively large amplitude currents that showed little or no evidence of desensitization. Ethanol inhibition was greatly reduced in the L497Y mutant receptor (Fig. 8E). Inhibition averaged only 22 ± 2% at 100 mM EtOH in the presence of 10 μM AMPA, and inhibition by 100 mM EtOH averaged 19 ± 1% when receptors were activated by 100 μM AMPA.
Discussion
The results reported at present provide additional evidence that EtOH inhibits AMPAR function. Indeed, inhibition of steady-state current activated by AMPA, and current activated by low concentrations of kainate, is comparable in magnitude to that observed at NMDA receptors in several preparations (Dildy-Mayfield and Harris, 1992; Nie et al., 1994; Martin et al., 1995; Wirkner et al., 2000). Our observations may also help to explain some aspects of EtOH inhibition of AMPARs heterologously expressed in Xenopus oocytes (Dildy-Mayfield and Harris, 1992). The steady-state component of current is the only component that is observed in Xenopus oocytes, given the slow superfusion of these large cells. Thus, the EtOH inhibition observed in the oocyte preparation almost certainly reflects changes in steady-state current. In previous studies, AMPARs were shown to be inhibited by EtOH if agonist was applied directly to cells, whether the cells were isolated or in brain slices (Dildy-Mayfield and Harris, 1992; Weiner et al., 1999; Wirkner et al., 2000). The drug application systems used for these studies had relatively slow exchange times, and applications lasted for seconds. Thus, the measured currents mainly reflect conditions under which desensitization has been initiated.
The observation that EtOH inhibits the steady-state component of AMPA-activated current to a greater extent than the peak current is consistent with the idea that EtOH promotes or stabilizes receptor desensitization. Previous studies have suggested that the steady-state current reflects openings from the desensitized state in preference to nondesensitized states. Indeed, it has been estimated that steady-state current in neurons reflects openings from 90% of desensitized and 10% of nondesensitized channels (Vyklicky et al., 1991). The desensitized receptor state is energetically the most favorable state of the agonist-bound receptor. Inhibition of peak current was minimal in outside-out membrane patches. These experiments are particularly important for interpretation of this finding because the solution exchange in this preparation is sufficiently fast to allow us to measure a large number of predesensitized receptors.
We observed a difference in EtOH inhibition between the peak and steady-state currents when 100 μM, but not when 10 μM AMPA, was used as agonist. When the low agonist concentration was used with a relatively slow agonist application system, a distinctive peak response could not be consistently produced. Peak response evoked by a low concentration of agonist is thus most likely composed of the mixture of openings from nondesensitized and desensitized states. Therefore, it is not surprising that we did not see any difference in EtOH inhibition between peak and steady-state response at the 10 μM AMPA concentration.
An interaction between EtOH and desensitization is also supported by our pharmacological experiments. Application of cyclothiazide greatly reduces desensitization, and this compound also significantly reduced EtOH inhibition of the receptor. Although it is possible that cyclothiazide counteracts EtOH effects by preventing other avenues of channel closing (e.g., deactivation/unbinding of agonist), our findings with cyclothiazide, together with the lack of effect of EtOH on initial current amplitude, argue for an interaction with desensitization rather than deactivation.
Evidence supporting an EtOH effect on AMPAR desensitization also comes from our experiments on the L497Y GluRAi mutant that lacks desensitization. Ethanol inhibition is clearly reduced, and nearly abolished at lower EtOH concentrations, in this mutant receptor. This contrasts with the wild-type GluRAi receptor that shows normal levels of desensitization, and inhibition by EtOH that is comparable to that observed in neurons. As observed in neurons, inhibition by EtOH was greater for the steady-state than for the peak current component in the wild-type GluRAi receptor. Inhibition of the nondesensitizing mutant receptor is similar in magnitude to that observed in the neurons in the presence of cyclothiazide (Figs. 7 and 8), supporting the idea that cyclothiazide reduces EtOH inhibition by reducing desensitization. It must be noted that a significant level of inhibition remains in this mutant receptor and in the presence of cyclothiazide. This could be due to EtOH effects that counteract the actions of cyclothiazide and the mutation, allowing some degree of desensitization. Alternatively, EtOH may inhibit receptors via an additional mechanism that does not involve desensitization. Thus, in addition to interactions with desensitization there may be other mechanisms involved in EtOH inhibition of AMPARs.
Interestingly, a recent preliminary report (Akinshola and Taylor, 2001) indicates EtOH inhibition of AMPARs is diminished by a point mutation in the large extracellular loop S2, within the flip/flop alternatively spliced region where the cyclothiazide binding site is thought to reside (Sommer et al., 1990; Partin et al., 1995). This finding provides additional support for an interaction of EtOH with AMPAR desensitization. In the recently characterized structure of the S1 and S2 regions of GluR2, this portion of the S2 loop lies in proximity to the residue equivalent to L497 in GluR1 (Sun et al., 2002). Thus, this region of the receptor may be part of a site for alcohol interaction with the protein. A direct interaction between cyclothiazide and EtOH within this region of the protein is also possible, such that the compounds can affect one another's actions at the receptor. This possible interaction will be an interesting subject for future studies.
The observation that EtOH inhibits kainate-induced responses in an agonist concentration-dependent manner is also consistent with an effect on desensitization. Kainate is a low-affinity partial agonist that produces rapid, but rapidly reversible, desensitization (Patneau et al., 1993). Thus, EtOH effects on desensitization could certainly underlie inhibition when this agonist activates the receptor. One possible explanation for differential EtOH inhibition at different agonist concentrations is that EtOH stabilizes desensitization occurring in the single agonist-bound state, thus increasing the proportional time the receptor spends in the low conductance state (at low agonist concentrations) relative to the situation in which desensitization occurs mainly from the high conductance (multiple-liganded) state (at high agonist concentrations). The observation that cyclothiazide reduces EtOH inhibition when kainate is the agonist is also consistent with stabilization of the desensitized state because cyclothiazide would reduce kainate-induced desensitization.
Our findings do not support an effect of EtOH on agonist affinity. We observed that EtOH inhibition of steady-state AMPA-induced current was similar regardless of the agonist concentration used. This finding suggests that EtOH inhibition is not dependent on receptor occupancy. Furthermore, because the steady-state current represents the highest affinity state of the receptor, we would expect to observe less, rather than more, inhibition of this component of current in relation to peak current if EtOH were only acting as a competitive antagonist. Our findings are consistent with those of Dildy-Mayfield and Harris (1992) who showed that EtOH did not alter the EC50 for kainate activation of GluRs in Xenopus oocytes.
Ethanol may produce its inhibitory action by altering the kinetics of the transition(s) leading to the desensitized state (i.e., desensitization) or slowing the transitions out of the desensitized state (i.e., resensitization). The measurements of current decay during agonist exposure indicate that the rate of desensitization seems unaltered in the presence of EtOH. Thus, our results are most consistent with stabilization of the desensitized state due to slowing of resensitization. The modest slowing of resensitization in the presence of EtOH (Fig. 5) supports this interpretation.
Many studies have shown that AMPA receptor-mediated synaptic responses are only weakly inhibited by EtOH at concentrations up to 100 mM (Lovinger et al., 1990; Weiner et al., 1999). In these same studies, it was shown that NMDA and kainate receptor-mediated transmission is inhibited to a much greater extent by EtOH. In the past, it was not clear why greater EtOH inhibition of AMPA receptors was observed when receptors were activated by agonist application to isolated cells than was observed at intact synapses. The findings in the present paper suggest a probable explanation for this apparent discrepancy, namely, that EtOH inhibition of the peak, presteady-state, component of AMPA-activated current is much less pronounced than inhibition of the steady-state current component. At the large majority of synapses, it seems that fast desensitization does not contribute to AMPAR-mediated synaptic responses (Colquhoun et al., 1992; Hestrin, 1992; Diamond and Jahr, 1995), because of the short duration of neurotransmitter presence in the synaptic cleft. Thus, the EtOH inhibition at an intact synapse would most likely reflect weak inhibition of predesensitized receptors. Of course, it is possible that the differential EtOH sensitivity of agonist-induced versus synaptic responses mediated by AMPARs might also involve a variety of factors, including differences in subunit, splice variant, or editing variant composition, post-translational modifications, age-related receptor modifications, and EtOH effects on presynaptic function.
These findings raise the question of whether EtOH inhibition of AMPARs plays any part in the neurophysiological effects of the drug in vivo. Although we cannot answer this question at present, our findings do suggest that EtOH might impact receptor function during conditions where AMPAR desensitization comes into play, such as during periods of prolonged extracellular glutamate release, as might happen during epileptiform activity (Jarvie et al., 1990) or during induction of long-term potentiation, or at those synapses where desensitization seems to play a role in shaping AMPAR-mediated synaptic responses (Barbour et al., 1994; Otis et al., 1996; Maguire, 1999). Since AMPAR desensitization has been shown to vary with different subunit compositions (Geiger et al., 1995), receptors from different brain areas and different developmental stages may show differences in EtOH sensitivity, which may explain some of the brain regional variability of EtOH effects on neurophysiology.
Acknowledgments
We thank Drs. Morris Benveniste and Kathryn Partin for helpful comments on the manuscript, and we also thank Dr. Partin for providing the L497Y mutant GluRAi construct. Cultured cortical neurons were prepared by Mark Maguire using facilities provided by the Vanderbilt John F. Kennedy Center for Research on Developmental Disabilities (Nashville, TN).
Footnotes
-
This study was supported by AA08986 (to D.M.L. and T.M.), the Sigrid Juselius Foundation (to D.M.L. and E.R.K.), the Finnish Foundation for Alcohol Studies (to T.M. and E.R.K.), and the National Institute on Alcohol Abuse and Alcoholism Division of Intramural Clinical and Basic Research.
-
DOI: 10.1124/jpet.103.050666.
-
ABBREVIATIONS: CNS, central nervous system; NMDA, N-methyl-d-aspartate; AMPA, α-amino-3-hydyroxy-5-methyl-4-isoxazolepropionic acid; AMPAR, AMPA receptor; EtOH, ethanol; HEK, human embryonic kidney; EGFP, enhanced green fluorescent protein; ANOVA, analysis of variance; Con A, concanavalin A; WT, wild-type.
- Received February 19, 2003.
- Accepted May 6, 2003.
- The American Society for Pharmacology and Experimental Therapeutics