INTRODUCTION

TOP ABSTRACT INTRODUCTION Methods RESULTS DISCUSSION REFERENCES

Protons modulate the properties of many ion channels, including glutamate receptors. The activity of the N-methyl-D-aspartate (NMDA) subtype of glutamate receptor is negatively regulated by protons within the physiological range of extracellular pH (Tang et al. 1990; Traynelis and Cull-Candy 1990; Vyklicky et al. 1990). Polyamines such as spermine (Johnson 1996; McBain and Mayer 1994; Williams 1997) and neomycin (Lu et al. 1998) potentiate NMDA receptor function by decreasing the proton-mediated tonic inhibition of NMDA receptor activity (Gallagher et al. 1997; Kashiwagi et al. 1996; Lu et al. 1998). Conversely, one class of NMDA receptor antagonists, the phenylethanolamines (ifenprodil and CP-101, 606), act by increasing the sensitivity of NMDA channels to inhibition by protons (Mott et al. 1998). Therefore it appears that many drugs modulate NMDA receptor activity by altering the sensitivity of the receptor to inhibition by protons. In addition to modulating NMDA receptor function, protons have also been reported to inhibit -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor activity (Traynelis and Cull-Candy 1991). However, unlike NMDA receptors, a detailed mechanism of proton-mediated modulation of AMPA receptor function has not been elucidated.

AMPA receptors undergo rapid desensitization that could contribute to the decay of excitatory synaptic currents and perhaps play an important role in modulating synaptic plasticity (Jones and Westbrook 1996). Several reagents commonly described as positive modulators of AMPA receptor function, including cyclothiazide (CTZ) (Partin et al. 1994; Patneau et al. 1993; Yamada and Tang 1993), 4-[2-(phenylsulfonylamino)ethylthio]-2,6-difluoro-phenoxyacetamide (PEPA) (Sekiguchi et al. 1997), aniracetam (Ito et al. 1990; Vyklicky et al. 1991), organic nitrates (Toong et al. 1998), and wheat germ agglutinin (WGA) (Vyklicky et al. 1991) are known to reduce or block AMPA receptor desensitization. However, the mechanism by which these drugs modulate receptor desensitization is poorly understood.

In the present study, we investigated the effects of protons on AMPA receptor function in acutely isolated hippocampal pyramidal neurons. We found that protons inhibited AMPA receptor function by enhancing receptor desensitization; this is a novel mechanism that has not previously been identified. We also tested the hypothesis that positive modulators of AMPA receptors act by reducing this proton-mediated receptor desensitization. Our results indicate that the effects of CTZ and GT-21-005 (an organic nitrate), but not those of aniracetam and WGA, are in part contributed by a proton-mediated modulation of AMPA receptor desensitization.

	Methods

TOP ABSTRACT INTRODUCTION Methods RESULTS DISCUSSION REFERENCES

Preparation of acutely isolated hippocampal CA1 neurons

All the animals used in this work were handled in accordance with the regulations of the Medical Research Council of Canada. CA1 hippocampal pyramidal neurons were acutely isolated using the modified procedures of Wang and MacDonald (1995) except that protease (type XIV, Sigma) instead of papain was used. Briefly, Wistar rats 2-3 wk old were placed under halothane anesthesia and decapitated with a guillotine. Hippocampi were quickly removed and placed in a dish containing cold oxygenated external solution consisting of (in mM) 140 NaCl, 1.3 CaCl₂, 5.4 KCl, 25 HEPES, 33 glucose, 1 MgCl₂, and 0.0003 TTX (pH 7.4, osmolarity 320-335 mosmol 1¹). The hippocampi were cut into slices 300-500 µm thick by hand with a razor blade. The hippocampal slices were digested at room temperature (20-22°C) in the above solution, which also contained 1.5 mg/ml protease. The incubation medium was stirred with pure oxygen that was blown in at the bottom of the vessel. After 40 min of enzymatic digestion, slices were rinsed three times with enzyme-free solution. The slices were maintained in the external solution, bubbled with oxygen, and used for 8-10 h. The CA1 region was placed under a phase-contrast microscope and dissected out with a scalpel and then triturated with a fire-polished glass pipette. Data were obtained from pyramidal cells that were phase-bright with a clear outline.

whole-cell recording. Whole-cell recordings were performed with an Axopatch-1B amplifier (Axon Instruments) in voltage-clamp mode. Recording electrodes with resistances of 3-5 M were constructed from thin-walled borosilicate glass (1.5 mm diameter, WPI) using a two-stage puller (PP83, Narishige). Data were digitized, filtered (5kHZ), and acquired online using the pClamp6 program (Axon Instruments). The standard internal solution for the recording electrodes consisted of (in mM) 140 CsF, 35 CsOH, 10 HEPES, 2 MgCl₂, 2 tetraethylammonium, 11 EGTA, and 4 Na₂ATP (pH 7.3, osmolarity 300 mosmol l¹). The standard external solution was the same as that already described except that Mg²⁺ (3 mM) and AP5 (50 µM) were included to block NMDA receptors. After the whole-cell configuration was formed, the recorded cells were voltage-clamped at 60 mV and lifted into the stream of solution. Saturating concentrations of glutamate (3-10 mM) were applied to the neurons for 2 s to evoke responses. Under these recording conditions, the currents evoked by glutamate were completely blocked by GYKI 53655, which confirmed their identity as AMPA receptor responses.

outside-out patch recording. Outside-out patch recordings were carried out as described in Bai et al. (1999). Glutamate was applied by using theta tubing connected to a piezoelectric translator (PZS-100 driven by PZ-150; Burleigh, Fishers, NY). For solution exchange,  was <200 µs as determined by measuring the open-tip junction potentials. Currents were filtered at 5 kHz and digitized at 50 kHZ.

non-stationary variance analysis. Non-stationary variance analysis was used to estimate conductance and the open probability of the channels at the peak of the response (Sigworth 1980). Generally, 60 responses, evoked by applications of 10 mM glutamate for 100 ms at 5-s intervals, were recorded for each outside-out patch. During an epoch of 60 responses, an average of 10% run-down was observed. Data from patches showing >20% run-down were not included for analysis. Responses from each patch were divided into 10 to 12 groups (5 or 6 for each patch). After each of the five or six responses to the peak was aligned, the local means of each group were calculated to minimize distortion originating from run-down. Each individual response was subtracted from the local mean of the group so that the variance could be computed. Current responses of 90 ms from the peak were selected for analysis. The mean current was divided into 100 equally sized bins and the corresponding variances were pooled. The binned variance versus the mean current was plotted and fit with the equation ² = iI  I²/N + _base, where ² is the variance, I is the mean current, N is the number of channels activated at the peak, i is the single channel current, and _base is the background variance. Open probability at the peak, P_O,Peak was calculated by P_O,Peak = I_peak/(iN), where I_peak is the peak current; then, the single channel conductance was measured from  = i/(E  E_rev), where E is the holding potential and E_rev is the reversal potential (0 mV under our recording conditions) (Fig. 6).

solutions. Solutions with different pH values were prepared each day by adding 6N HCl or 10N NaOH, as appropriate, to the external solution. A series of solutions with pH values ranging from 8.0 to 5.0 was prepared. Solutions of individual pH values were divided into four tubes so that the following four sets of solutions could be prepared by adding glutamate or drugs. Set 1, control bathing solution; Set 2, glutamate-containing solution; Set 3, control bath solution + drug; Set 4, glutamate-containing solution + drug. The pH was rechecked and readjusted, if necessary, after the addition of drugs to ensure that the pH of the four sets of solutions was the same; generally, readjustment of pH was unnecessary. Some hydrophobic drugs were initially dissolved in DMSO and then diluted to the desired concentration in the external solution. The final concentration of DMSO that was applied to the cells was usually <0.1% except for solutions containing CTZ (0.5%) and aniracetam (1%). In these cases, the same amount of DMSO was added to the control and to the glutamate-containing solutions.

data analysis. Data were expressed as means ± SE. Values in parentheses refer to the number of cells used for statistical analysis. Concentration-response curves were fitted by the Hill equation I = I_max × (1/(1 + (EC₅₀/[ligand])ⁿ)), where I_max is the maximum response, EC₅₀ is the concentration of ligand that produces a half-maximal response, and n is the Hill coefficient. Statistical analyses were performed using either Student's t-test or two-way ANOVA. P values <0.05 were taken as an indication of significant difference.

	RESULTS

TOP ABSTRACT INTRODUCTION Methods RESULTS DISCUSSION REFERENCES

Protons inhibit whole-cell AMPA receptor currents in isolated CA1 neurons

We initially used whole-cell recordings to examine the effect of protons on AMPA receptor currents that were evoked by a saturating concentration of glutamate in isolated CA1 neurons. The peak (I_p) and steady-state (I_ss) currents both were significantly depressed with increasing concentrations of protons (Fig. 1A). At pH 5.0, both I_p and I_ss were almost completely blocked (Fig. 1, A-C). The values of pH that produced a 50% reduction in current (IC₅₀) were 6.1 ± 0.1 (I_p) and 6.2 ± 0.1 (I_ss) pH units (n = 19 neurons) (Fig. 1, B and C). Also, the ratio of steady-state to peak currents (I_ss/I_p) decreased with the increase in proton concentration (Fig. 1D), which suggests that desensitization of AMPA receptors was enhanced by protons.

View larger version (21K): [in this window] [in a new window]   Fig. 1. Protons inhibit -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor function. A: glutamate-evoked whole-cell currents of AMPA receptors at different values of pH. Note that decreasing pH reduced steady-state (Iss) as well as peak (Ip) glutamate-evoked currents. B: plot of Ip vs. pH from 19 neurons. The curve was fitted using the Hill equation. The 50% reduction in current (IC50) value of pH was 6.1 ± 0.1 and the Hill coefficient was 1.7 ± 0.1. C: fitting of Iss vs. pH with the Hill equation (n = 19 neurons). The IC50 value was 6.2 ± 0.1 and the Hill coefficient was 1.2 ± 0.1. D: Iss/Ip values at pH 6.0 and pH 8.0 normalized to the Iss/Ip value at pH 7.5 (n = 19 neurons).

Protons enhance AMPA receptor desensitization

We next examined the effect of acidification on the time course of the onset of desensitization in outside-out patch recordings of AMPA responses (Fig. 2A). The decay of AMPA currents during glutamate application was well fitted by a single exponential function. A comparison of the time constants of desensitization () between pH 8.0 and pH 6.0 showed that the onset of desensitization was enhanced by 30% (pH 8.0,  = 11.4 ± 1.2 ms; pH 6.0,  = 8.0 ± 0.8 ms; P < 0.01, n = 6) (Fig. 2B). This result also indicates that protons enhance the onset of AMPA receptor desensitization.

View larger version (30K): [in this window] [in a new window]   Fig. 2. Protons reduced AMPA receptor desensitization. A: AMPA receptor currents evoked by glutamate at pH 8.0 and pH 6.0 in an outside-out patch. Desensitization was well fitted by a single exponential. Bottom: normalized currents. The desensitization time constant was reduced by changing pH from 8.0 to 6.0. B: desensitization time constants pooled from 6 outside-out patches. **, P < 0.01 by paired t-test. C: current responses evoked by two different concentrations of glutamate at pH 8.0 and pH 6.0 from the same neuron. Note that the currents evoked by the test concentration (3 mM) decreased more rapidly at pH 6.0. D: pooled data from 8 neurons. Peak currents of the responses to test concentrations were normalized after subtracting steady-state currents evoked by conditioning concentrations. Note that protons enhanced pre-desensitization of the currents.

We also employed a pre-desensitization protocol to further corroborate the effects of protons on AMPA receptor desensitization. In this protocol, AMPA receptors were sequentially activated using two different concentrations of glutamate. An interval sufficient to permit recovery from desensitization was permitted between each pair of applications. The initial application (conditioning concentration) of glutamate, which lasted for 3 s, employed concentrations ranging from 1 to 1,000 µM and was immediately followed by a saturating concentration of glutamate (3 mM) that was applied for 1 s (test concentration) to assess degree of desensitization (Fig. 2C). The current evoked by the test concentration decreased as the conditioning concentration was increased (Fig. 2, C and D). Peak currents evoked by the test concentration were normalized, after the steady-state currents evoked by the conditioning concentration were subtracted, and were used to evaluate the extent of desensitization. The IC₅₀ value for the conditioning glutamate decreased from 25.8 ± 5.2 µM at pH 8.0 to 10.2 ± 3.0 µM at pH 6.0 (n = 8, P < 0.01) (Fig. 2D). These results further demonstrate that protons enhance AMPA receptor desensitization.

In addition to enhancing the time course of the onset of desensitization, protons might have altered the rate of recovery of receptors from desensitization. We examined this possibility in outside-out patches using a paired-pulse paradigm in which the first application of glutamate (10 mM, 100 ms) was followed by a second application of glutamate (10 mM, 20 ms) with increasing intervals ranging from 20 to 420 ms (Fig. 3A). The ratios of the amplitudes of the second and first responses (P2/P1) were plotted against the interpulse interval so that the rate of recovery from desensitization could be calculated (Fig. 3B). The time constants for recovery from desensitization, as determined by fitting each data set with a single exponential, were not altered when pH was changed from 8.0 ( = 99.8 ± 22.2 ms, n = 5) to 6.0 ( = 83.4 ± 13.1 ms, n = 5, P > 0.05), which suggests that protons did not change the rate of recovery from desensitization (Fig. 3B).

View larger version (20K): [in this window] [in a new window]   Fig. 3. Protons failed to change the rate of recovery from desensitization. A: current responses evoked by a paired-pulse paradigm with different intervals from the same outside-out patch at pH 8.0 (left) and pH 6.0 (right). Note that protons enhanced desensitization but did not change the rate of recovery from desensitization. B: recovery curves from experiments similar to those in A. Each data point represents the mean ± SE of the ratio P2/P1 where P1 and P2 are the peak current amplitudes of the first and second pulses, respectively (n = 5). Data points are fitted to single exponential functions.

Protons inhibit open probability of AMPA receptors

We next examined whether or not protons were able to reduce the amplitude of non-desensitized AMPA receptor-mediated currents. We tested this possibility by measuring the open probability of the non-desensitized state with nonstationary variance analysis (Fig. 4). Protons reduced the maximum open probability by 35.1 ± 4.6% (pH 8.0, 0.77 ± 0.06; pH 6.0, 0.51 ± 0.06; n = 8 patches, P < 0.01) without changing the conductance of AMPA receptors (pH 8.0,  = 14.1 ± 2.8 pS; pH 6.0,  = 13.1 ± 2.4 pS; n = 8 patches, P > 0.05).

View larger version (24K): [in this window] [in a new window]   Fig. 4. Protons reduce the maximum open probability of AMPA receptors as determined by non-stationary variance analysis. A and B: 5 macroscopic current responses recorded from an outside-out patch excised from a cultured hippocampal neuron at pH 8.0 (A) and pH 6.0 (B) are superimposed. C and D: composite current-variance plots of 50 responses at pH 8.0 (C) and pH 6.0 (D) from the same patch as in A and B. E: open probabilities at the peak (PO,Peak) from 8 patches at pH 8.0 and pH 6.0. Protons consistently reduce PO,Peak. F: conductances () from 8 patches at pH 8.0 and pH 6.0. Protons did not consistently change .

Protons did not change the deactivation of AMPA receptors

Several modulators of AMPA receptor desensitization, including cyclothiazide, aniracetam, and thiocyanate, also modulate AMPA receptor deactivation (Partin et al. 1996). We therefore used outside-out patches to examine whether or not protons modulate the deactivation kinetics of AMPA receptors. AMPA receptor deactivation was resolved using rapid (1 ms) applications of 10 mM glutamate (Fig. 4A). However, changing pH failed to alter deactivation kinetics (pH 8.0,  =3.3 ± 0.1 ms; pH 6.0,  =3.1 ± 0.2 ms; n = 8 patches, P > 0.05 by paired t-test) (Fig. 5).

View larger version (14K): [in this window] [in a new window]   Fig. 5. Protons do not affect deactivation of AMPA receptors. A: current responses evoked by the application of 10 mM glutamate for 1 ms from an outside-out patch at pH 8.0 and pH 6.0. Open tip junction potential is shown above the current traces. Time constants for deactivation were estimated from single-exponential fits starting from ~90 to 70% of the peak amplitude. Right: normalized currents. B: pooled time constants of deactivation at pH 8.0 and pH 6.0 from 8 patches (P > 0.05, paired t-test).

Voltage-independent modulation by protons

We also examined the voltage dependence of proton-induced inhibition as well as the pH dependence of the potency of glutamate. Inhibition by protons did not depend on holding potentials ranging from 80 to +40 mV (n = 5) (Fig. 6A), which argues against a simple open channel block. Furthermore, the potency for activation of I_p was not changed by protons (pH 8.0, EC₅₀ = 1051 ± 482 µM, n_H = 0.9 ± 0.1; pH 6.0, EC₅₀ = 1462 ± 407 µM, n_H = 0.8 ± 0.1; n = 6, P > 0.01) (Fig. 6Bb) whereas the potency for I_ss was reduced (pH 8.0, EC₅₀ = 66 ± 10 µM, n_H = 1.4 ± 0.1; pH 6.0, EC₅₀ = 119 ± 16 µM, n_H = 1.3 ± 0.2; n = 6, P < 0.05) (Fig. 4Bc).

View larger version (21K): [in this window] [in a new window]   Fig. 6. Voltage-independent inhibition of AMPA receptors by protons. Aa: glutamate-evoked AMPA receptor currents at different holding potentials from the same patch when solution pH was 8.0 (left) and 6.0 (right). Ab: voltage and peak current relationship (n = 5 patches). Ba: whole-cell currents evoked by different concentrations of glutamate from the same neuron at pH 8.0 (left) and pH 6.0 (right). Bb and Bc: concentration-response curves at pH 8.0 and pH 6.0. Note that the potency of glutamate was not changed for the peak currents (Bb) but was reduced for steady-state currents (Bc) when pH was changed from 8.0 to 6.0.

CTZ and organic nitrates reduce proton sensitivity of AMPA receptor desensitization

CTZ reduces AMPA receptor desensitization (Partin et al. 1994; Patneau et al. 1993; Yamada and Tang 1993). We therefore tested if the proton dependence of AMPA receptor-mediated current was altered by this drug. CTZ at a concentration of 100 µM dramatically reduced AMPA receptor desensitization (Fig. 7A). I_p and, especially, I_ss both were significantly enhanced by CTZ. CTZ shifted the I_ss versus the pH curve to the left by 1.0 pH unit (control, IC₅₀ = 6.2 ± 0.1, n_H = 1.6 ± 0.1; cyclothiazide, IC₅₀ = 5.2 ± 0.01, n_H = 1.8 ± 0.3; n = 12, P < 0.0001) (Fig. 7B). This result suggests that cyclothiazide reduces proton-mediated enhancement of AMPA receptor desensitization.

View larger version (25K): [in this window] [in a new window]   Fig. 7. Cyclothiazide (CTZ) and GT-21-005 block desensitization and reduce the pH sensitivity of glutamate-evoked currents. A: glutamate-evoked currents recorded from the same neuron at different values of pH before and after the application of CTZ. Note the CTZ-induced enhancement of both Ip and Iss; note also the change in pH sensitivity. B: fitting of the Iss vs. pH curve using the Hill equation. The curve was shifted to the left by CTZ (n = 12, P < 0.0001). C: desensitization of glutamate-evoked currents and sensitivity to pH was reduced by GT-21-005. D: fitting of the Iss vs. pH curves using the Hill equation. The curve was shifted to the left by GT-21-005 (n = 7, P < 0.01).

GT-21-005, an organic nitrate, substantially reduces AMPA receptor desensitization (Toong et al. 1998). At a concentration of 2 mM, GT-21-005 substantially reduced the desensitization of glutamate-evoked currents (Fig. 7C) and shifted the I_ss versus pH curve to the left by 0.6 pH units (control, IC₅₀ = 6.0 ± 0.1, n_H = 2.0 ± 0.1; GT-21-005, IC₅₀ = 5.4 ± 0.1, n_H = 2.0 ± 0.3; n = 7, P < 0.01) (Fig. 7D). This result suggests that this organic nitrate also reduces AMPA receptor desensitization by changing the pH sensitivity of desensitization.

Effects of WGA and aniracetam were independent of protons

WGA is a lectin that, with a relatively slow onset (minutes), irreversibly reduces AMPA receptor desensitization (Vyklicky et al. 1991). We therefore pretreated hippocampal slices with WGA (300 µg/ml) for 15 to 30 min and then dissociated CA1 neurons for recording. Pretreating the cells with WGA significantly reduced the desensitization of glutamate-evoked currents (Fig. 8A). However, the pH sensitivity of AMPA receptor desensitization was not changed by WGA (Fig. 8B). The IC₅₀ value was 5.8 ± 0.1 (n_H = 2.0 ± 0.3) for control (n = 9) and 5.7 ± 0.1 (n_H = 1.6 ± 0.1) for cells pretreated with WGA (n = 10; P > 0.05) (Fig. 8B).

View larger version (19K): [in this window] [in a new window]   Fig. 8. Wheat germ agglutinin (WGA)- and aniracetam-induced inhibition of desensitization were little effected by a change in pH. A: glutamate-evoked currents in solutions of different values of pH from a control neuron (top) and from a neuron pretreated with WGA (300 µg/ml) (bottom). Note the reduction of desensitization but lack of change in the pH sensitivity of these currents. B: fitting of Iss vs. pH curve by Hill equation. The curve was not altered by WGA (n = 9 for control neurons, n = 10 for neurons pretreated with WGA; P > 0.05). C and D: responses to glutamate in a single neuron before and after treatment with aniracetam (5 mM). No considerable change in the sensitivity of pH was observed.

The effect of aniracetam on the pH sensitivity of the currents was more complex (Fig. 8, C and D). Aniracetam (5 mM) appeared to reduce proton sensitivity between pH 7.0 and pH 6.0 but, at lower pH values, it had no effect (Fig. 8D). A comparison of the IC₅₀ values showed no significant difference (control, IC₅₀ = 5.9 ± 0.1, n_H = 1.1 ± 0.2; aniracetam, IC₅₀ = 5.7 ± 0.1, n_H = 1.8 ± 0.1; n = 5, P > 0.05). We were unable to examine the effects of higher concentrations of this drug because of its limited solubility in physiological solutions.

	DISCUSSION

TOP ABSTRACT INTRODUCTION Methods RESULTS DISCUSSION REFERENCES

In the present study, we provide evidence that protons enhance AMPA receptor desensitization without altering deactivation kinetics. Although it is likely that the mechanism underlying proton-mediated enhancement of desensitization is complex, our results do not support a simple slowing of recovery from desensitization because the change in extracellular pH from 8.0 to pH 6.0 enhanced receptor desensitization but failed to alter the rate of recovery from desensitization. The result that protons did not change the time constants of deactivation suggests that protons probably do not modulate the binding or unbinding of agonist. However, our results suggest that protons reduce the maximum open probability of non-desensitized receptors. This observation may provide one explanation for proton-induced enhancement of desensitization because, in some models, entry to desensitized states cannot proceed via the open, agonist-bound state of the receptor (Jones and Westbrook 1997). Therefore, a decrease in the probability of channel opening induced by protons might favor receptor desensitization.

The binding site for glutamate consists primarily of a segment of the N-terminus called the S1 segment as well as a second S2 segment located in the extracellular loop (Paas 1998). Both segments, and an additional region of the N-terminus, demonstrate substantial homology with prokaryotic periplasmic bacterial binding proteins (PBPs) and, by analogy, it has been postulated that these segments from two separate binding lobes are capable of physically trapping agonist (the "Venus flytrap" model) (Mano et al. 1996). Closure of the segments would perhaps stabilize the receptor in a high-affinity and non-conducting desensitized conformation. However, recent experiments (Abele et al. 1999) suggest that movements of these segments are much more subtle than are those reported for PBPs, which suggests instead that relatively small rotations of the lobes are associated with receptor desensitization (Abele et al. 1999; Swanson et al. 1997). In this respect, a single-site mutation within S1 can entirely block desensitization of the GluR3 subunit and various mutations within S2 can also alter desensitization (Stern-Bach et al. 1998). The flip and flop region, which is located toward the C-terminus from S2 but still within the extracellular loop, also controls desensitization, perhaps by altering rotation of the S2 segment. It seems likely that the proton sensor(s) of the AMPA channel is located in or near the S1 or S2 segment and that its occupancy may stabilize desensitized conformation of the AMPA channel.

Our results show that positive modulators of AMPA receptors, such as CTZ and GT-21-005, reduce the proton sensitivity of AMPA receptor desensitization whereas WGA and aniracetam do not. This evidence suggests that there are multiple ways in which drugs can alter desensitization. For example, aniracetam preferentially modulates desensitization of flop splice variants of AMPA receptor subunits (Johansen et al. 1995; Partin et al. 1996), which probably is a consequence of slowing channel closing. In contrast, CTZ preferentially reduces desensitization of flip splice variants (Partin et al. 1994) by stabilizing a non-desensitized agonist-bound closed state (Partin et al. 1996). GT-21-005 is a novel organic nitrate (Yang et al. 1996) and is a member of a large family of S-nitrates that have previously been described and that exhibit properties very different from the prototype nitrate ester nitroglycerin (Thatcher and Weldon 1998). The fact that GT-21-005 also reduces the proton-sensitivity of AMPA receptor desensitization suggests that it may act at the same site or sites as CTZ. In contrast, lectins such as concanavalin A and WGA probably act to reduce desensitization by binding to glycosylation sites on receptor subunits (Everts et al. 1997, 1999). CTZ and aniracetam are unlikely to interact with these glycosylation sites because their effects are in addition to those of lectins (Everts et al. 1997; Vyklicky et al. 1991).

Our study is important for understanding the roles attributed to AMPA receptors in some pathological conditions. Extracellular pH undergoes relatively small changes during synaptic transmission (Chesler and Kaila 1992). In contrast, large decreases in pH occur during intense seizure activity or ischemia and pH levels can decrease by 0.2 to more than 1.0 pH units (Chesler and Kaila 1992; Siesjo 1985; Silver and Erecinska 1992). The activation both of NMDA and of AMPA receptors potentially contributes to neuronal injury during these pathological conditions. Simultaneous acidification of the extracellular space would be expected to limit the degree of excitotoxicity by inhibiting NMDA receptor activity and enhancing AMPA receptor desensitization. Consistent with this suggestion, extracellular acidification, at a level observed during ischemia, reduces glutamate-mediated neuronal death (Giffard et al. 1990; Kaku et al. 1993).