Presynaptic depression of synaptic transmission mediated by activation of metabotropic glutamate receptors in rat neocortex

Conventional intracellular recordings were obtained from layer II-III neurons in adult rat neocortical brain slices. Excitatory and inhibitory (I) postsynaptic potentials (PSPs) were evoked prior to and during bath application of agonists and antagonists of metabotropic glutamate receptors (mGluRs). In the presence of the selective mGluR agonist 1S,3R-1-aminocyclopentane-1,3- dicarboxylic acid (1S,3R-ACPD; 5– 200 microM), both excitatory and inhibitory components of the evoked PSPs were reversibly reduced. PSPs were significantly, but less effectively, decreased by L-2-amino-4-phosphonobutyric acid. Exposure to putative mGluR antagonists, alpha-methyl-4-carboxyphenylglycine or L- 2-amino-3-phosphonopropionic acid, did not inhibit the 1S,3R-ACPD- mediated effect. In the presence of 6,7-dinitroquinoxaline-2,3-dione and D-2-amino-5-phosphonovaleric acid, 1S,3R-ACPD reversibly depressed directly evoked neocortical IPSPs; however, quisqualic acid (1–10 microM) did not mimic this effect. Analysis of spontaneous PSPs and paired-pulse facilitation indicated a presynaptic locus of action for 1S,3R-ACPD at mGluRs. These findings indicate that a specific mGluR subtype(s) may modulate both excitatory and inhibitory synaptic transmission in the adult rat neocortex via a presynaptic reduction of transmitter release.

Despite their marked expression in the neocortex, the role of mGluRs in normal synaptic transmission in this brain region is unclear. Our goal was to examine the effects of mGluR activation on synaptic transmission in the rat frontal neocortex. Conventional intracellular recording techniques were used to record PSPs evoked by intracortical stimulation.
Our results show that synaptic transmission is reversibly depressed during mGluR activation, and the data suggest a presynaptic locus of action.
Preliminary reports of some ofthese data have been published elsewhere (Burke and Hablitz, 1992).

Materials and Methods
Slices of frontal neocortex were obtained from adult Sprague-Dawley rats of both sexes (125-250 gm). Rats were decapitated under ether or ketamine (100 mg/kg) anesthesia, and the brains were quickly dissected out and placed in ice-cold saline for 30-60 sec. The tissue was blocked with a razor blade to separate the hemispheres and to remove the caudal and ventral portions of the brain. The tissue block was attached to a Teflon chuck with cyanoacrylate (Superglue). Four or five coronal slices (500 pm thick) of frontal neocortex were prepared on a Vibroslice (Campden Instruments), and incubated in a holding chamber at room temperature for at least 2 hr. Single brain slices were transferred to an interface-type recording chamber, and maintained with constant saline perfusion (1 ml/min). The chamber was warmed slowly to the recording temperature of 34 rt 1°C. The extracellular solution consisted of the following (in mM): NaCI, 125; KC], 3.5; NaH? PO,,1.25;CaCI,,2.5;MgSO,,1.3;NaHCO,,26;glucose,10. This solution was continuously perfused with a mixture of 95% OL and 5% CO, to attain a steady state level of oxygenation and to maintain a pH of 7.4.
Conventional intracellular recording electrodes were pulled from tilament-containing borosilicate glass tubing (1.5 mm o.d.; A-M Systems, Inc.), and filled with 4 M potassium acetate (adjusted to pH 7.2 with acetic acid). Electrode resistances ranged from 50 to I IO Mfi. Recordings were acquired with the use of an Axoclamp 2-A amplifier (Axon Instruments) in bridge mode. Voltage signals were digitized on line with PCLAMP 5.5 software (Axon Instruments) or were recorded onto videotape via a Neuro-Corder DR-384 (NeuroData Instruments) to be digitized at a later time. Records were stored on disk and analyzed with PCLAMP software. Spontaneous PSPs were digitized and subsequently analyzed using Strathclyde Electrophysiological Software (courtesy of J. Dempster, University of Strathclyde, Glasgow). Neuronal membrane potential was monitored continuously on a Gould RS3200 oscillograph recorder (Gould Inc.). A mono-or bipolar stimulating electrode was placed intracortically in layers IV-V slightly lateral to the recording electrode.
After impalement of layer II-III neurons (N = 89 from 70 rats), the resting membrane potential was allowed to stabilize for a minimum of 20 min. Neuronal input resistance was monitored in bridge mode with a -0.4'nA constant amplitude current pulse of 100 msec duration. All measurements were performed at the original resting membrane potential for each neuron. Constant amplitude current pulses of variable duration were delivered through the stimulating electrode to activate fibers projecting to layer II-III neurons. Postsynaptic potentials (PSPs) in response to increasing stimulus durations (40-I 80 set) were evoked at IO-15 set intervals before, during, and after bath application of putative mGluR agonists and antagonists. In the presence of 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) and D(-)2-amino-5-phosphonovaleric acid (D-APS), the effects of mGluR activation on excitatory amino acid antagonist-resistant PSPs were investigated. The voltage dependence of these directly evoked inhibitory PSPs was determined by injecting constant depolarizing current through the recording electrode. Bridge balance was monitored continuously.
To examine paired pulse facilitation (PPF) of neocortical PSPs, stimulus parameters (i.e., intensity and duration) that evoked primarily excitatory PSPs were used. Interpulse intervals of IO-200 msec were used. Facilitation was calculated as the ratio between the amplitude of the second (test) response and the first (conditioning) response in a pair. During drug application, PPF was tested with the control intensity and with an increased stimulus intensity (to maintain the amplitude of the first PSP at control levels). Control records were compared to those obtained during mGluR activation.
Each trace shown is the average of two responses to a given stimulus. All values are expressed as mean t SD, except where noted. Statistical analysis consisted of paired and unpaired t tests. A significance level of 0.05 was used.

Results
PSPs evoked during the control period were compared to those evoked during bath application of pharmacological agents. Three points of measure were analyzed: peak PSP amplitude in response to minimum stimulation, peak amplitude to maximum stimulation, and the peak amplitude ofthe delayed depolarizing response to maximum stimulation. Stimulus-response curves were determined for each measurement. Minimum stimulation (40 set) produced a purely excitatory PSP, whereas the response to maximal stimulation (180 set) was an EPSP-IPSP complex. The early component of the compound PSP was predominantly excitatory, while the late depolarizing response was primarily a GABA, receptor-mediated Cl conductance (Connors et al., 1988;Sutor and Hablitz. 1989a,b;Hablitz and Sutor, 1990). trans-ACPD reduces evoked synaptic transmission in the adult rat neoc0rte.v The evoked PSP to minimum stimulation before and during bath application of 50 FM trans-ACPD is shown in Figure 1A.
In this example, only a slight reduction in the peak amplitude was observed (to 96.6% of control). As shown in Figure lB, however, the response to maximal stimulation was more notably depressed. In this cell, the peak amplitude to maximum stimulation was reduced to 84.8% of control. The late depolarization was similarly reduced to 77.9% of control. No change in resting membrane potential (RMP; -89 mV) was noted in this cell during trans-ACPD application. Neuronal input resistance (R,,) prior to and during drug application was 24.9 MO and 27.8 MQ, respectively. In Figure 1, C and D present input-output curves for the peak amplitudes of the early response and the late depolarization, respectively. These plots show that synaptic potentials were reduced at all but the weakest stimulus strengths. Mean values for these data with 20-50 WM trans-ACPD (N = 8) are presented in Table 1. Figure 1. trans-ACPD decreases synaptic transmission in the adult neocortex. Intracellular recordings of evoked synaptic activity in a layer II-III neuron are shown in A and B. A, Minimum electrical stimulation produces a small PSP that was marginally inhibited by 50 FM trans-ACPD (96.6% of control). B, In the same neuron, more intense stimulation produced an EPSP-IPSP complex. IPSP is depolarizing due to the high RMP (-89 mV). Both excitatory and inhibitory components are reduced by truns-ACPD (to 84.8% and 77.9% of control, respectively). C and D, Plot ofresponse amplitude as a function of stimulus intensity for the peak and late depolarizing responses, respectively, during control (0)  of lS,3R-and IR,3S-ACPD. Some studies have shown that IS,3R-ACPD is the active isomer at mGluRs (Irving et al., 1990;Schoepp et al., 199 1 b), whereas others report that both enantiomers are full agonists for mGluRs coupled to PI hydrolysis and intracellular Ca'+ mobilization . We therefore tested the effects ofbath-applied lS,3R-and lR,3S-ACPD on synaptic transmission in the neocortex.
Bath application of lS,3R-ACPD (N = 33) resulted in a reduction of evoked synaptic responses in the adult neocortex. As shown in Figure 2, all components of the evoked PSPs were markedly depressed during 100 PM lS,3R-ACPD application. In Figure 2A, the peak amplitude of an evoked PSP is plotted against time during lS,3R-ACPD application (27 min) and its subsequent washout. In this cell, the response to minimum stimulation was reduced to 34.7% of control; the peak response and the late depolarization to maximum stimulation were reduced to 47.2% and 4.5.2%, respectively (Fig. 2B). During drug application, R,, was slightly increased (16.8 vs 20.1 MQ), and the RMP was depolarized from -88 mV to -84 mV. Upon washout of lS,3R-ACPD, the evoked PSPs returned to their control amplitudes; R,, also returned to baseline. RMP repolarized by 1 mV with washout. Catania et al. (1992) and Lonart et al. (1992) each describe a rapid and long-lasting desensitization of mGluR-mediated PI hydrolysis with extended exposure to mGluR agonists. We observed no desensitization of the response to lS,3R-ACPD, as this agonist could be reapplied to the same neocortical brain slice (N = 3, data not shown), producing similar results each time. Additionally, no desensitization ofthe lS,3R-ACPD-mediated effects was observed during application periods up to 115 min (N = 4).
Unlike the active isomer, 1 R,3S-ACPD did not depress synaptic transmission, nor were effects on R,, or RMP observed. Stimulus duration (psec) In Figure 2C, the peak PSP amplitude to minimal stimulation was 99.4% of control; the peak response to maximum stimulation was 100.4% of control during 1 R,3S-ACPD application, and the late depolarization was 89.7% of control. As shown in Table 1, synaptic and membrane properties varied less than 10% from control values in the presence of 200 FM I R,3S-ACPD (N = 3).

Dose dependence qf IS,SR-ACPD &ects
The dose-response curves for the three measured components of the evoked PSPs are shown in Figure 3. Responses have been normalized to the maximum percentage reduction produced by 200 WM lS,3R-ACPD (N = 12) for each of the three PSP measurements (Table 1). The EC,,, for the mGluR-mediated reduction of the peak response to minimum stimulation by lS,3R-ACPD was 30.23 KM. The peak response and the late depolarizing response to maximum stimulation had similar EC,, values of 33.57 KM and 32.66 PM, respectively. Seventeen of 40 cells tested with 5-200 FM lS,3R-ACPD depolarized slightly during drug application (depolarizations ranged from 1 to 4 mV). Statistically significant differences in RMP and R,, were noted only for 100 and 200 WM lS,3R-ACPD (Table 1). Pooled data from all experiments indicate a statistically significant trend toward depolarization (-88.6 * 3.1 and -87.7 & 3.0 mV, control vs drug, respectively) and an increased R,,, measured at the original RMP (20.7 ? 4.3 and 22.4 + 4.6 MO, control vs drug, respectively) during the application of 1 S,3R-ACPD.
L-API reduces postsvnaptic potentials L-2-Amino-4-phosphonobutyric acid (L-AP4) has been reported to antagonize some responses mediated by mGluR activation (Schoepp and Johnson, 1988;Zheng and Gallagher, 1992).  control and wash The maximum reductions observed were: for minimum stimulation, 68.3% (0)  RMP was -89 mV. Evoked PSPs to minimum (left truces) and maximum (right trace$ stimulation were reversibly inhibited by lS,3R-ACPD. C, As in B, but recordings from another neuron; RMP was -88 mV. lR,3S-ACPD did not reduce synaptic transmission in the adult rat neocortex. However, L-AP4 also activates a presynaptic glutamate receptor that depresses synaptic transmission (Harris and Cotman, 1983;Cotman et al., 1986;Forsythe and Clements, 1990;Baskys and Malenka, 199 1;Rainnie and Shinnick-Gallagher, 1992;Calabresi et al., 1993) via a G-protein<oupled mechanism (Trombley and Westbrook, 1992). This electrophysiologically defined receptor has been termed the AP4 receptor (Collingridge and Lester, 1989;Monaghan et al., 1989). Recently, L-AP4 has been shown to activate the mGluR4 subtype, which is negatively coupled to the CAMP cascade (Thomsen et al., 1992;Tanabe et al., 1993) and it has been suggested that the AP4 receptor and the mGluR4 are the same glutamate receptor subtype (Nakanishi, 1992; Thomsen et al., 1992;Schoepp and Conn, 1993;Tanabe et al., 1993; but see Trombley and Westbrook, 1992). We wished to determine the effects of bath-applied L-AP4 on synaptic transmission and passive membrane properties in the adult neocortex prior to examining its putative antagonism of mGluRs. Bath application of L-AP4 (200-1000 FM, N = 4) consistently reduced evoked PSPs, but had no effect on R,, or RMP (Table 1). An example of the L-AP4-mediated effect is shown in Figure 4. The minimum response was reduced to 65.7% of control by 200 PM L-AP4 (Fig. 4A). In Figure 4B, the peak response to maximum stimulation was reduced to 7 1.9%, and the late depolarization to maximum stimulation was similarly reduced to 58.5% of control. The input-output relationships for the early and late PSP measurements are shown in Figure 4, C and D, respectively. Responses were reduced at all but the smallest stimulus strengths. The effects of mGluR agonists on synaptic and membrane properties of neocortical neurons are summarized in Table 1. Putative mGluR antagonists Recent studies have indicated that (R,S)-a-methyl-4-carboxyphenylglycine (MCPG), or its active form, (+)-MCPG, is a competitive mGluR antagonist (Bashir et al., 1993;Eaton et al., 1993;Frenguelli et al., 1993;Jane et al., 1993). Bath application of (+)-MCPG (250-1000 PM, N = 6) for 30 min did not affect synaptic transmission in neocortical slices. When 200 PM lS,3R-ACPD was bath applied to MCPG-treated slices, little antagonism of the mGluR-mediated synaptic depression was observed. PSPs to minimum stimulation were 92.0 ? 22.8% of control amplitude in MCPG; these PSPs were reduced to 56.4 +-26.9% with ACPD application (Fig. 5A). In the presence of MCPG, the peak PSP amplitude to maximum stimulation was 103.6 f 11.9% of control. lS,3R-ACPD reduced these responses to 58.7 f 13.1% (Fig. 5B). During MCPG application, the late depolarization to maximum stimulation was 102.3 t 15.6% of control amplitude, and was reduced to 59.4 -t 24.1% (Fig. 5C). These changes were statistically significant. Thus, MCPG does not antagonize mGluR-mediated synaptic depression in the adult neocortex. Effective MCPG concentrations Intracellular recordings of responses to intracortical stimulation were obtained before and during bath application of L-AP4; RMP was -90 mV. A and B, L-AP4 (200 ELM) reduced PSPs evoked by minimum (A, to 65.7% of control) and maximum stimulation (B, to 7 1.9% and 58.5% of control, peak and late response, respectively). C and D,'Inputoutput relationships for peak amplitude and late depolarizing response, respectively, before (0)  were blocked. This suggests that MCPG is an effective antagonist at postsynaptic mGluRs in the adult neocortex, but may be less effective (or inactive) at presynaptic mGluRs (see Frenguelli et al., 1993).
We found that bath application of L-AP3 (200-1000 PM, N = 3) was without effect on synaptic transmission in the neocortex. When slices were bathed in L-AP3 for at least 30 min prior to applying 1 S,3R-ACPD (100-200 PM), no antagonism of the mGluR-mediated synaptic depression was observed. These concentrations of ACPD were selected because of their reliability in reducing PSP amplitude and because of the general ineffectiveness of lower ACPD concentrations (see Table 1). Data are given as mean k SD. Each neuron served as its own control. All measurements were performed at each neuron's original resting membrane potential (RMP). y Neuronal input resistance (R,,) was tested with ~0.4 nA constant current pulses of 100 msec duration. )' The percentage change in peak PSP amplitude in response to minimum intracortical stimulation (40 Bsec). ' The percentage change in peak PSP amplitude in response to maximum intracortical stimulation (180 wet). *I The percentage change in peak amplitude of the delayed depolarizing response to maximum intracottical stimulation (180 rsec). This response was primarily a GABA, receptor-mediated 0 conductance (Connors et al., 1988;Sutor and Hablitz, 1989a,b;Hablitz and Sutor, 1990). * Significant differences between control and drug conditions (p i 0.05, paired Student's t test).
PSPs to minimum stimulation were 103.1 * 8.4% of control amplitude in L-AP3; these PSPs were reduced to 5 1.8 ? 17.8% with ACPD application (Fig. 54). In the presence of L-AP3, the peak PSP amplitude to maximum stimulation was 99.0 + 8.4% of control. lS,3R-ACPD reduced these responses to 47.1 & 5.9% (Fig. 5B). During L-AP3 application, the late depolarization to maximum stimulation was 93.1 k 11.8% of control amplitude; this response was reduced to 26.1 + 7.0% (Fig. 5C). Differences between PSP amplitudes in the presence of L-AP3 and in the combined presence of L-AP3 and ACPD were significant for responses evoked by both minimum and maximum stimulation. L-AP3 also did not antagonize postsynaptic effects of mGluR activation (Burke and Hablitz, unpublished observations). Thus, L-AP3 is not an effective agonist at neocortical mGluRs mediating electrophysiologic responses.
Adenosine receptors do not mediate the reduction in synaptic transmission by ACPD Activation of hippocampal adenosine receptors has been shown to decrease synaptic transmission via a presynaptic mechanism (Yoon and Rothman, 199 1;Prince and Stevens, 1992;Thompson et al., 1992). Recent reports have demonstrated interactions between mGluRs and adrenergic receptors that increase CAMP accumulation in hippocampus (Winder and Conn, 1993) and decrease accumulation in the cerebral cortex (Cartmell et al., 1993). To investigate the possibility that endogenous adenosine might play a role in the effects of mGluR activation in the neocortex, 3,7-dimethyl-1 -propargylxanthine (DMPX; 50-100 FM, N = 2) was bath applied at least 30 min prior to lS,3R-ACPD application. DMPX is a selective AZ receptor antagonist with an EC,,, of 11 * 4 FM; the EC,,, for the A, receptor is 45 f 5 PM (Seale et al., 1988). DMPX did not affect evoked synaptic transmission (data not shown). When lS,3R-ACPD (200 KM) was bath applied with DMPX, typical reductions in the peak amplitudes of the synaptic responses were observed. PSPs to minimum stimulation were reduced to 48.1 t 24.6% of control amplitudes (Fig. 5A). The peak amplitude to maximum stimulation was reduced to 40.1 f 13.2% of control (Fig. 5B), and the late depolarizing response to maximum stimulation was reduced to 35.6 & 6.4% ofcontrol with lS,3R-ACPD (Fig. 5C). These data indicate that endogenous adenosine does not play a role in the synaptic depression mediated by lS,3R-ACPD in the neocortex. mGluR activation reduces direct!l, evoked IPSPs After blocking the excitatory components of evoked PSPs with the excitatory amino acid receptor antagonists 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) and D( -)-2-amino-5-phosphonovaleric acid (D-APS), we tested the effects of mGluR activation on IPSPs evoked by direct stimulation of interneurons. CNQX (10 FM) blocks the non-NMDA receptor-mediated component, and D-AP5 (20 PM) blocks the NMDA receptor-mediated component of neocortical EPSPs (Sutor and Hablitz, 1989a;Hablitz and Sutor, 1990). In Figure 6, A and B show control responses and those recorded during bath application of the EAA antagonists, respectively. The resulting PSPs were evoked at the same stimulus intensities used during the control period. PSPs evoked during EAA receptor antagonism reversed at membrane potentials near ~75 mV (Fig. 6C), indicating that they were Cl--mediated GABAergic IPSPs (Connors et al., 1988;Sutor and Hablitz, 1989a;Hablitz and Sutor, 1990). Bath application of 50 WM lS,3R-ACPD (Fig. 60) reduced the peak amplitude of the response to maximum stimulation to 56.4% of control. Results from all cells tested are shown in Figure 6F. The effect of lS,3R-ACPD on directly evoked IPSPs was reversible upon washing out.

Presynaptic locus qf action
Changes in synaptic efficacy producing a change in PSP amplitude might arise from two sources: (1) changes in the sensitivity of the postsynaptic site and/or (2) changes in the output of the presynaptic terminals. We investigated whether lS,3R-ACPD blocked spontaneous PSPs at concentrations that blocked evoked PSPs by about 50%. In contrast to the effects on evoked synaptic transmission, lS,3R-ACPD (50-200 FM, N = 12) did not decrease the amplitude of spontaneous PSPs. Under control conditions, mean spontaneous PSP amplitude was 1.49 ? 0.13 mV (mean -t SEM); in the presence of lS,3R-ACPD, the mean amplitude essentially was unchanged, 1.43 + 0.08 mV (Fig.  6A). Consistent with a presynaptic site of action, drug application increased the mean spontaneous PSP interval from 17.54 +-7.14 set (mean 1-SEM) to 42.4 1 * 11.73 set (Fig. 6B). These observations suggest that lS,3R-ACPD induces a presynaptic decrease in neurotransmitter release. The effects of lS,3R-ACPD were reversible upon washing. Paired pulse facilitation (PPF) is a phenomena seen at many chemical synapses and is generally thought to reflect a presynaptic enhancement in neurotransmitter release (Zucker, 1989). Manipulations that reduce transmitter release have been shown to increase PPF (Mallart and Martin, 1967;Katz and Miledi, 1968;Harris and Cotman, 1983). In the following set of experiments, a stimulus strength that evoked an EPSP was used. Interpulse intervals of lo-200 msec were tested, and the ratio of the second response amplitude to that of the first response was calculated. Control records were compared to those obtained during ACPD application. Exposure to 100 FM lS,3R-ACPD reduced the synaptic response by 39.5 + 20.9% (N = 4), and enhanced PPF at all interstimulus intervals (data not shown). Because there is normally a greater percentage facilitation at smaller PSP amplitudes, PPF also was tested with an increased stimulus intensity that matched the conditioning PSP amplitude to that recorded during the control period. Under these conditions, facilitation remained increased at all interpulse intervals (data not shown). Statistically significant results were observed for intervals between 10 and 120 msec.

Discussion
Previous studies have shown that ACPD reduces synaptic transmission in a variety of brain regions (Baskys and Malenka, 199 1;Crepe1 et al., 1991;Desai and Conn, 1991;Lovinger, 199 1;Calabresi et al., 1992Calabresi et al., , 1993Desai et al., 1992;Rainnie and Shinnick-Gallagher, 1992;Lovinger et al., 1993;Swartz et al., 1993). Our data demonstrate that metabotropic glutamate receptor activation in the adult rat neocortex produces a dose-dependent, reversible depression of synaptic transmission. Furthermore, these results indicate a presynaptic site of action for agonists working at mGluRs, presumably acting to decrease neurotransmitter release. Pharmacological profiles additionally suggest mGluR subtype-specific actions in the neocortex.
Locus of mGluR action Activation of neocortical mGluRs by 1 S,3R-ACPD reduced the three measured components of evoked PSPs. Decreased synaptic transmission mediated by lS,3R-ACPD could occur as the result of antagonism of a postsynaptic glutamate receptor, activation of a postsynaptic receptor that reduces R,, of the cell, or the activation of a presynaptic receptor that reduces neurotransmitter release. The pharmacological profile of lS,3R-ACPD makes antagonism of postsynaptic glutamate receptors unlikely. This compound is specific for mGluRs with no appreciable ionotropic glutamate receptor activity even at mM concentrations (Schoepp et al., 199 la;Sacaan and Schoepp, 1992). Additionally, under conditions of EAA receptor antagonism, directly evoked IPSPs were reversibly inhibited by lS,3R-ACPD. This observation is consistent with the finding that mGluR responses are not antagonized by ionotropic EAA-receptor antagonists (Palmer et al., 1988;Recasens et al., 1988;Schoepp and Johnson, 1988) and indicates that lS,3R-ACPD does not reduce synaptic transmission in the neocortex by antagonizing ionotropic glutamate receptors.
In most neurons examined,'application of lS,3R-ACPD did not decrease neuronal input resistance. To the contrary, higher concentration of lS,3R-ACPD tended to increase R,, in layer II-III neurons. Pooled data indicate a statistically significant trend toward a slight increase in R,, and a small depolarization with lS,3R-ACPD application. However, lS,3R-ACPD-mediated depression of evoked synaptic transmission was observed in neurons that showed slight changes in R,, and RMP, as well as those showing no change. Thus, it is unlikely that the changes in evoked PSPs are the result ofalterations in passive membrane properties secondary to the activation ofa postsynaptic receptor.
Our findings support the hypothesis that lS,3R-ACPD acts via the activation ofpresynaptic mGluRs. The reduction in both ionotropic glutamate receptor-mediated transmission and GA-BAergic transmission by lS,3R-ACPD strongly suggests a presynaptic locus of action, with mGluR receptors present on both excitatory and inhibitory terminals. Direct evidence that lS,3R-ACPD decreases synaptic transmission via a presynaptic action was obtained from investigations of spontaneous PSPs and paired-pulse facilitation of neocortical PSPs. At concentrations that reduced evoked PSPs by about 50%, lS,3R-ACPD did not reduce.spontaneous PSP amplitude. However, the frequency of spontaneous PSPs was markedly decreased. These results are consistent with a presynaptic locus of action: lS,3R-ACPD reduces transmitter release without changing postsynaptic responsiveness to transmitter. A postsynaptic receptor blocker would reduce responses produced by both stimulus-evoked and spontaneous release of transmitter. In the neocortex, lS,3R-ACPD enhances facilitation associated with paired-pulse stimulation. The observed increase in facilitation is an expected consequence ofdecreasing transmitter release (Mallart and Martin, 1968). It is presumed that PPF is influenced by presynaptic Ca'+ ion availability (Zucker, 1989). Manipulations of extracellular Ca'+ concentrations that affect synaptic transmission has been shown to inversely affect PPF (Katz and Miledi, 1968;Mallart and Martin, 1968;Creager et al., 1980). Harris and Cotman (1983) suggest that this mechanism may be responsible for the actions of L-AP4 in the hippocampus. Recent reports indicate that mGluR activation by a variety of agonists can depress Ca'+ currents (Lester and Jahr, 1990;Sayer et al., 1992;Trombley and Westbrook, 1992;Swartz et al., 1993). The effects of ACPD and L-AP4 on synaptic transmission and PPF are consistent with a presynaptic decrease in a Ca'+ current involved in neocortical transmitter release (see Sayer et al., 1992;Swartz et. al., 1993).
Pharmacological prqfile.for neocortical mGluRs The activity of trans-ACPD is reportedly restricted to the lS,3Rstereoisomer (Irving et al., 1990;Schoepp et al., 1991;but see Manzoni et al., 1992). In the adult rat neocortex, lS,3R-ACPD reversibly reduced synaptic transmission in a dose-dependent manner. In the presence of 200 FM 1 R,3S-ACPD (a concentration of lS,3R-ACPD that reduced PSPs by >SO%), all evoked responses were within 90% ofcontrol values. Thus, the selective, presynaptic mGluR properties of trans-ACPD reside in the lS,3R-stereo conformation. L-AP4 has been touted as a putative mGluR antagonist (Schoepp and Johnson, 1988;Zheng and Gallagher, 1992) as well as an agonist at presynaptic glutamate autoreceptors (Harris and Cotman, 1983;Cotman et al., 1986;Forsythe and Clements, 1990;Baskys and Malenka, 1991;Rainnie and Shinnick-Gallagher, 1992;Trombley and Westbrook, 1992;Calabresi et al., 1993). Our findings that L-AP4 reduces synaptic transmission in the neocortex are consistent with the latter reports and, further, the suggestion that a mGluR subtype may be the electrophysiologically defined AP4 receptor (Nakanishi, 1992;Thomsen et al., 1992;Schoepp and Conn, 1993;Tanabe et al., 1993). However, two types of presynaptic glutamate receptors may exist, one that is sensitive to L-AP4 and ACPD, and another one that is sensitive to ACPD, but not to L-AP4 . Additionally, Sahara and Westbrook (1993) observed that ACPD and L-AP4 modulate Ca'+ currents via distinct G-protein-coupled receptors in hippocampal cultures. Such complexity could arise from regional variations in expression of mGluR subtypes and differential coupling to effector mechanisms. It is possible that more than one subtype is responsible for the presynaptic effects of 1 S,3R-ACPD, which would explain the lower efficacy of L-AP4 versus lS,3R-ACPD in the rat neocortex. Further investigations with subtype-specific agonists and antagonists will be required to elucidate the identity and the specific functions of the mGluR subtypes.
While L-glutamic acid is the most likely endogenous agonist for the mGluRs, quisqualic acid is a potent agonist for a number of mGluR subtypes (Nakanishi, 1992;Schoepp, 1993;Schoepp and Conn, 1993). We evaluated the possible activation of mG1uRs by quisqualic acid in slices in which,ionotropic EAA receptors were blocked. Under these conditions, we failed to observe statistically significant reductions in IPSP amplitude. However, quisqualic acid did enhance R,, in all cells tested. It is possible that quisqualic acid does not activate the mGluR(s) responsible for presynaptically reducing PSP amplitude, but that this compound is a potent agonist for mGluRs located postsynaptically and therefore capable of increasing R,, in neocortical neurons in layers II-III.
Our studies with the putative mGluR antagonists MCPG and L-AP3 indicate that both compounds were poor antagonists of lS,3R-ACPD-mediated synaptic depression in the neocortical slice. MCPG, however, was effective at blocking the direct postsynaptic effects of lS,3R-ACPD (Burke and Hablitz, unpublished observations). Thus, MCPG may be an effective antagonist at postsynaptic mGluRs in the adult neocortex, but is less effective at presynaptically located mGluRs (see Frenguelli et al., 1993). The lack of pre-or postsynaptic antagonism by L-AP3 is consistent with most electrophysiological findings (Stratton et al., 1990;Charpak and Gahwiler, 199 1;Desai et al., 1992;Hu and Storm, 1992;Lovinger et al., 1993). This suggests that the mGluRs that are responsible for the reduction of synaptic transmission in the neocortex are not the same mGluRs that are L-AP3 sensitive and linked to PI hydrolysis (Schoepp and Johnson, 1989;Houamed et al., 199 1;Desai et al., 1992) intracellular Ca' + mobilization (Irving et al., 1990) or increased CAMP accumulation (Winder and Conn, 1993). However, our studies do not identify the mGluR-mediated second messenger system(s) involved in this form of synaptic modulation.
The functional role of the mGluRs in the adult neocortex is still unclear. It is obvious from this study that activation of these receptors has the functional consequences of presynaptic inhibition and postsynaptic excitation. Under "normal" conditions, these receptors may function at a low level to increase the signal to noise levels of synaptic responses. In pathologic situations, excessive mGluR activation may result, causing either excitatory or inhibitory consequences, depending on the receptor subtypes that are activated.