Metabotropic Glutamate Receptor Enhancement of Spontaneous ...

Metabotropic Glutamate Receptor Enhancement of Spontaneous IPSCs in Neocortical Interneurons FU-MING ZHOU AND JOHN J. HABLITZ Department of Neurobiology, University of Alabama at Birmingham, Alabama 35294 Zhou, Fu-Ming and John J. Hablitz. Metabotropic glutamate receptor enhancement of spontaneous IPSCs in neocortical interneurons. J. Neurophysiol. 78: 2287–2295, 1997. Using neocortical layer I neurons as a model for GABAergic interneurons, we have studied g-aminobutyric acid-A (GABAA ) receptor-mediated spontaneous inhibitory postsynaptic currents (IPSCs) and modulation by metabotropic glutamate receptors (mGluRs). In the presence of 0.5 mM tetrodotoxin (TTX) and ionotropic glutamate receptor antagonists and under symmetrical Cl 0 conditions, the mean amplitude of miniature IPSCs (mIPSCs) was Ç50 pA at a holding potential of 070 mV with individual events ranging from 10 to 400 pA. Averaged mIPSCs had a 10–90% rise time of Ç0.6 ms. The decay was double exponential. The fast component had a time constant of Ç4 ms and comprised Ç40% of the total amplitude. The slow component had a time constant of Ç22 ms. The frequency of spontaneous IPSCs (sIPSCs), recorded in the absence of TTX, was increased by bath application of the mGluR agonist 1S,3R-1aminocyclopentane-1,3-dicarboxylic acid (ACPD; 10–100 mM) or the group I mGluR selective agonist quisqualic acid (Quis; 0.5–1 mM). Under identical conditions, mIPSCs were not affected. The kinetics of sIPSCs and mIPSCs were not altered by ACPD or Quis. Quis (1 mM) induced an inward current of Ç70 pA at a holding potential of 070 mV, whereas ACPD (40–200 mM) induced a smaller inward current. This current was linear over the voltage range 070 to /30 mV and reversed polarity near 0 mV. In currentclamp recordings, both Quis and ACPD induced a depolarization and action potential firing in layer I and deeper layer interneurons. We conclude that neocortical layer I neurons receive GABAA receptor-mediated inhibitory synaptic inputs. Activation of mGluRs, possibly mGluR1 and/or mGluR5, causes an enhancement of inhibitory synaptic transmission by directly depolarizing cortical GABAergic interneurons through the opening of nonselective cation channels.

INTRODUCTION

In the mammalian brain, g-aminobutyric acid (GABA) can bind to GABAA and GABAB receptors (Howe et al. 1987; see Mody et al. 1994 for review). The former receptor is associated with a Cl 0 channel complex, whereas the latter is linked to K / channels via G-protein pathways. The influx of Cl 0 through GABAA receptors and the associated increase in membrane conductance decrease neuronal excitability. Molecular biological studies have identified numerous distinct GABAA receptor subunits. Different cell types may express GABAA receptor subunits with different properties (see McKernan and Whiting 1996 for review). A balance of excitation and inhibition in cortical neuronal circuits is critical for normal brain function. A slight reduction of cortical inhibition can produce epileptiform activity (Connors and Amitai 1993). In the mammalian cerebral

cortex, GABAergic interneurons are the principal source of inhibitory drive onto pyramidal cells (Kawaguchi and Kubota 1996; Thomson et al. 1996). Inhibition of inhibitory interneurons could produce disinhibition and alter the equilibrium between inhibition and excitation in the network. Characterization of inhibitory inputs to inhibitory interneurons is essential for understanding the operation of cortical circuitry. Many anatomic and electrophysiological studies have described the synaptic connections between inhibitory neurons and their targets (Buhl et al. 1995; Miles 1990; Thomson et al. 1996). Inhibitory postsynaptic potentials have been recorded in both hippocampal and neocortical interneurons (McCormick et al. 1985; Morin et al. 1996). Due to the difficulty in recording from GABAergic neurons in the hippocampus and cerebral cortex, detailed descriptions of the biophysical properties of GABAA receptors on these cortical interneurons are lacking. Metabotropic glutamate receptors (mGluRs) are a large family of G-protein coupled receptors that have been implicated in diverse functions of the mammalian CNS. Based on their structural homology, signal transduction mechanism and pharmacological profiles, mGluRs currently are classified into three groups (Pin and Bockaert 1995). Group I includes mGluR1 and mGluR5 receptors the most potent agonist of which is quisqualic acid (Quis). Group II includes mGluR2 and mGluR3 where (2S,1 *R,2 *R,3 *R)-2-(2,3-dicarboxycyclopropyl)-glycine (DCG-IV) is the most potent agonist. Group III comprises mGluRs 4 and 6–8, with L-2amino-4-phosphonobutyric acid being the most potent agonist. In the hippocampus and cerebellum, it has been reported that activation of mGluRs enhances spontaneous GABAergic synaptic transmission (Gereau and Conn 1995; Llano and Marty 1995; McBain et al. 1994; Poncer et al. 1995; Sciancalepore et al. 1995). Histochemical studies have shown that mGluRs are distributed widely in the neocortex, and group I mGluRs are enriched in presumed GABAergic neurons (Baude et al. 1993; Reid et al. 1995; Romano et al. 1995; Shigemoto et al. 1992), indicating that mGluRs may be able to modulate interneuron activities in the cerebral cortex. However, effects of mGluRs on spontaneous inhibitory postsynaptic currents (IPSCs) in the neocortex have not been examined. Therefore, we hypothesized that activation of mGluRs may act directly on cortical interneurons and increase the inhibitory outputs from these inhibitory interneurons. Immunohistochemical studies have demonstrated that most neocortical layer I neurons are GABA-positive, indicating they are inhibitory neurons (Gabbott and Somogyi 1986; Li and Schwark 1994). Our recent studies on the postnatal

0022-3077/97 $5.00 Copyright q 1997 The American Physiological Society

9k20

/ 9k1d$$oc27 J292-7

10-29-97 14:18:38

neupa

LP-Neurophys

2287

2288

F.-M. ZHOU AND J. J. HABLITZ

development, morphology, and electrophysiology of layer I neurons (Zhou and Hablitz 1996a–d, 1997) indicate that these cells are inhibitory interneurons. Therefore we have used layer I neurons as a model system for analysis of mGluR effects on IPSCs in interneurons. The results indicate that GABAergic interneurons receive a prominent inhibitory input that can be enhanced by mGluRs. Preliminary results have appeared in an abstract (Zhou and Hablitz 1996e).

line via a tape recorder (Neuro-corder, model DR-384, Neuro Data), and stored on videotape in PCM form. Off-line, signals were captured with a detection threshold typically at 6 pA and digitized at 0.1–0.4 ms per point using a Digidata 1200 interface (Axon Instruments) controlled by SCAN software (provided by Dr. J. Dempster, University of Strathclyde, Glasgow, UK). Analysis of the records also was achieved using the SCAN program. Automatic measurements of the rise time, decay time, and decay time constant of individual synaptic currents were verified and corrections made when necessary.

METHODS

Database and analysis

Slice preparation Brain slices were prepared as previously described (Zhou and Hablitz 1996a). All animals were housed and handled according to approved guidelines. Briefly, 15- to 19-day-old rats were decapitated, and brains were dissected free in õ1 min. The isolated brain was immersed immediately in ice-cold saline. Slices (200mm thick) were cut from frontal cortex (Paxinos and Watson 1986) on a Vibratome and stored at room temperature (22 { 17C) for ¢1 h before recording. Slices were used °8 h after preparation. For slicing, solutions contained (in mM) 250 sucrose, 3.5 KCl, 1 or 0.5 CaCl2 , 3 or 5 MgCl2 , 26 NaHCO3 , and 10 D-glucose. Normal bathing solution contained (in mM) 125 NaCl, 3.5 KCl, 2.5 CaCl2 , 1.3 MgCl2 , 26 NaHCO3 , and 10 D-glucose. The bathing solution was bubbled continuously with 95% O2-5% CO2 to maintain pH at Ç7.4. Bathing solution was circulated with a peristaltic pump at rates up to 4 ml/min. Pharmacological agents reached the slices in °1 min. Quis (Cambridge Research Biochemicals) was prepared as a 10 mM stock solution in 10% ammonium hydroxide. 1S,3R-1aminocyclopentane-1,3-dicarboxylic acid (ACPD), 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), and D(-)2-amino-5-phosphonovaleric acid ( D-APV) were from Tocris Neuramin. Bicuculline methiodide (BIC) was from Sigma. These compounds were prepared as 10 or 20 mM stock solutions and added to the bathing solution during the experiment to reach the desired concentration.

Whole cell recording Layer I neurons were identified by location below the pial surface and morphological appearance, as previously described (Zhou and Hablitz 1996a,b). All recordings were made at room temperature (22 { 17C). Patch electrodes were pulled from Garner KG33 glass using a Narishige PP-83 puller. Electrodes were coated with Sylgard. Series resistance (Rs ) was estimated according to Rs Å 10 mV/I, where I was the peak of the transient current (filtered ¢10 kHz) evoked by the 10-mV testing pulse when the pipette capacitance was compensated fully. Series resistance during recording was 4–18 MV and was not compensated. Recordings were terminated whenever significant increases in series resistance occurred. The intracellular solution contained (in mM) 10 KCl, 125 CsCl or KCl, 10 N-2-hydroxyethylpiperazine-N *-2-ethanesulfonic acid, 2 Mg-ATP, 0.2 Na-GTP, 0.5 ethylene glycol-bis( b-aminoethyl ether)-N,N,N *,N *-tetraacetic acid. pH was adjusted to 7.3 with KOH. In many cells, QX-314 (3 mM) was included in the intracellular solution; no effect on IPSC properties was observed. Osmolarity was adjusted to 270–290 mOsm. Before forming a tight seal, offsets including liquid junction potential were nullified using the offset feature of the patch-clamp amplifier. The liquid junction potential that developed after going whole cell was not corrected. Tight seals ( ú2 GV before breaking into whole cell mode) were achieved without first cleaning the cell. Electrical signals were recorded using an Axopatch-200 amplifier (Axon Instruments). Signals were low-pass filtered at 5 kHz, digitized at 44 kHz on-

9k20

/ 9k1d$$oc27 J292-7

We included in the following analysis 128 layer I neurons. In 35, the series resistance was 4–10 MV, and recordings from these cells were used to obtain IPSC kinetic parameters. The remaining cells were recorded with higher series resistance and used only for analyses less sensitive to series resistance errors, such as pharmacological properties and interevent interval. Spontaneous synaptic events recorded in these individual cells numbered from õ100 to ú5,000 depending on the frequency and recording time. Averaged IPSCs were obtained after aligning individual events on their rising phase. Double exponential functions, in the form of Af 1 {[exp( 0t/ t f )] / As 1 [exp( 0t/ ts )]}, was fitted to the decay of synaptic currents using a nonlinear Levenberg-Marquadt algorithm, where Af and As and t f and ts were amplitude and time decay time constant for the fast and slow component, respectively. Results are expressed as means { SD. Correlation coefficients (r) were calculated using a least-squares linear regression analysis. Comparisons of frequency and amplitude of miniature IPSCs before and after drugs were made using the Kolmgorov-Smirnov (K-S) test. P ° 0.01 was considered significant. RESULTS

IPSCs in layer I neurons To determine whether layer I neurons receive inhibitory synaptic inputs, synaptic activities were recorded in the presence of 20–100 mM D-APV and 10–50 mM CNQX to block ionotropic glutamate receptors. Under these conditions, spontaneous synaptic currents were recorded in all layer I neurons tested. Examples of such activities are shown in Fig. 1A, top. Amplitudes of these events, recorded at 050 to 070 mV under symmetrical Cl 0 conditions, ranged from 10 to 1,000 pA. Some of these synaptic events, particularly large amplitude events, were abolished by adding 0.5 mM tetrodotoxin (TTX) to the bathing solution. Most small amplitude events were TTX resistant. As shown in the lower trace of Fig. 1A, bath application of the GABAA receptor antagonist BIC (10 mM) blocked all synaptic events. Under symmetrical Cl 0 conditions, events were inward at negative holding potentials and reversed polarity near 0 mV, the expected Cl 0 equilibrium potential (Fig. 1B). These results indicate that these spontaneous synaptic events recorded in layer I neurons were GABAA receptor-mediated inhibitory postsynaptic currents (IPSCs). Properties of mIPSCs Miniature IPSCs (mIPSCs) were recorded in the presence of 0.5 mM TTX. The frequency of mIPSCs at 070 mV was Ç1 Hz in the majority of the cells ranging from 0.1 to 8 Hz. The variability in mIPSC frequency largely may reflect differential preservation of synaptic connections during slic-

10-29-97 14:18:38

neupa

LP-Neurophys

MGLURS

IN INTERNEURONS

2289

FIG . 1. Layer I neurons receive g-aminobutyric acid-A (GABAA ) receptor-mediated inhibitory synaptic inputs. A: examples of spontaneous synaptic events recorded in the presence of 20 mM D ( 0 )2-amino-5-phosphonovaleric acid ( D-APV) and 10 mM 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) at a holding potential of-60 mV. This cell received frequent large amplitude, presumably action potential-dependent synaptic inputs. Bath application of 10 mM bicuculline methiodide (BIC) completely blocked spontaneous synaptic activity (bottom), indicating that these synaptic inputs were sIPSCs. B: in another layer I neuron, spontaneous inhibitory postsynaptic currents (sIPSCs) were recorded at /50 (top), 0 (middle), and 050 mV (bottom) in the presence of 20 mM D-APV and 10 mM CNQX. sIPSCs had an apparent reversal potential near 0 mV, indicating that these IPSCs were mediated through GABAA-gated Cl 0 channels.

ing. At a holding potential of 070 mV, amplitudes of most individual mIPSCs were between 15 and 100 pA (Fig. 2, A and B). In most cells, there were also a number of events with amplitudes ranging from 100 to 200 pA (Fig. 2B). In some cells, mIPSCs up to Ç400 pA were observed. The amplitude of averaged mIPSCs was 49 { 3 pA (n Å 28) at a holding potential of 070 mV. Amplitude distributions for mIPSCs were broad and skewed toward large amplitude events (Fig. 2B). These histograms could not be fit with single- or multiple Gaussian functions. At holding potentials of 060 or 070 mV, the 10–90% rise time of individual mIPSCs was Ç0.65 ms, ranging from

0.3 to 3 ms (Fig. 2C). Averaged mIPSCs had a rise time of 0.63 { 0.04 ms (n Å 28). The decay of mIPSCs was complex. The decay of ú85% of individual mIPSCs and therefore the averaged mIPSCs, was best fitted with a double exponential function (Fig. 2A). The time constant for the fast ( t f ) and slow ( ts ) component for individual mIPSCs in these cells varied from Ç1 to Ç8 ms and Ç15 to Ç45 ms for t f and ts , respectively. The contribution from the slow component to the total amplitude was Ç60%, ranging from Ç30 to 70%. A small number of individual spontaneous IPSCs (sIPSCs) had single exponential decays with time constants Ç25 ms. However, the decay of averaged mIPSCs was always double exponential with a t f of 3.7 { 0.4 ms and a ts of 22 { 2.4 ms (n Å 28). The contribution to the total amplitude from the slow component was 61 { 7%. The homogeneity in the rise time and decay of averaged mIPSCs indicates that kinetics of GABAA channels were similar among these cells. To evaluate dendritic filtering, we plotted, for individual mIPSCs from 12 layer I cells, amplitude as a function of 10–90% rise time. Figure 2C shows that 10–90% rise times and amplitudes were not or only slightly correlated ( r Å 0.16 { 0.05, n Å 12). The 90% decay times of individual mIPSCs were correlated modestly with the 10–90% rise times (r Å 0.44 { 0.1). The small r values may indicate that the dendritic cable filtering effects are nonlinear (Spruston et al. 1993). Alternatively other factors such as vesicle size variation and asynchronous transmitter release (Bekkers et al. 1990; Vautrin and Barker 1995) could be involved in shaping the waveform of mIPSCs such that plots like those above are of limited use in evaluating dendritic filtering in these cells. Kinetics of mIPSCs in pyramidal neurons

FIG .

2. Kinetics of GABAA receptor-mediated miniature IPSCs (mIPSCs) in layer I neurons. A, top: mIPSCs of variable amplitudes. A, bottom: an individual mIPSC (left) and the averaged mIPSC (right). Both are superimposed with double exponential fits. Parameters are listed. B: mIPSC amplitude histogram, binwidth was 5 pA. Distribution was skewed toward large events. C: plot of mIPSC rise time vs. amplitude. No correlation was seen. D: plot of mIPSC rise vs. decay time. A correlation of 0.53 was observed. Recordings were made at a holding potential of 070 mV. Bathing solution contained 20 mM D-APV, 10 mM CNQX, and 0.5 mM tetrodotoxin (TTX). Events with contaminated rise and/or decays were excluded in C and D.

9k20

/ 9k1d$$oc27 J292-7

To determine whether mIPSCs in pyramidal neurons have different kinetics than those in layer I neurons, we recorded mIPSCs from 10 pyramidal neurons. The frequency of mIPSCs in these pyramidal neurons was 5 { 2.6 Hz (n Å 10, Fig. 3A). The amplitude distribution of pyramidal cell mIPSCs also was skewed toward large events (Fig. 3B). The mean amplitude of averaged mIPSCs was 62 { 4.5 pA at 070 mV, significantly larger than that of layer I neurons. This larger amplitude also resulted in a longer 10–90% rise time (0.7 { 0.03 ms). The mechanism underlying the ampli-

10-29-97 14:18:38

neupa

LP-Neurophys

2290


FIG . 3. Properties of mIPSCs in pyramidal neurons. A: frequent mIPSCs are observed in pyramidal cells. Bottom: double exponential decay shown of an individual mIPSC (left) and the averaged mIPSC (right). Fits are superimposed in both traces and the parameters are listed. B: mIPSC amplitude distribution was skewed toward large events. Binwidth was 5 pA. Recording was made in a layer II/III pyramidal neuron voltage clamped at 070 mV. Bathing solution contained 40 mM D-APV, 20 mM CNQX, and 0.5 mM TTX.

tude difference is unknown. The decay was also double exponential: t f Å 3.4 { 0.3 ms and ts Å 21 { 2.3 ms, not different from those of layer I neurons. The contribution from the slow component to the total amplitude was 63 { 4%. When mIPSCs of similar amplitudes from layer I and pyramidal neurons were compared, no significant difference was found. ACPD and Quis increases the frequency of sIPSCs To test whether activation of mGluRs can modulate spontaneous inhibitory synaptic transmission in layer I neurons, ACPD was bath applied in the presence of the ionotropic glutamate receptor blockers D-APV (20–100 mM) and CNQX (10–40 mM). As shown in Fig. 4A, ACPD increased the frequency of sIPSCs. At 10 mM, the lowest concentration used, the frequency of sIPSCs was increased by Ç500 { 250% (5 cells, Fig. 4A). At 20–40 mM, sIPSC frequency was increased by Ç10 ( Ç5–20)-fold (8 cells). In the presence of ACPD, averaged sIPSCs amplitude was increased from 40–60 to 80–200 pA. The increased amplitude of averaged sIPSCs was due to an increase in the number of large, presumably action potential-dependent, events. These enhancing effects were sustained on prolonged application of 10–30 mM ACPD. The response to 50–80 mM ACPD (n Å 7) was more complex. As shown in Fig. 4B, ACPD initially induced an increase in IPSCs. The response quickly declined to a plateau level where large events seen during the initial phase of the response disappeared (Fig. 4B). The effects of ACPD were fully reversible on washing. To identify the subtype of mGluR involved, we tested the effects of Quis, a relatively selective agonist for group I mGluRs (mGluR1 and mGluR5). Bath application of 0.5 mM Quis produced an increase in sIPSC frequency (Fig. 5A) of Ç10-fold (range 5- to 30-fold, n Å 11). The response was sustained for the duration of application of 0.5 mM Quis in 6 of the 11 cells (Fig. 5A). In the remaining cells, the initial increase in sIPSC frequency was followed by a decline to a plateau level. Quis also increased the mean amplitude of sIPSCs from Ç50 to Ç140 pA at 070 mV. When ¢1 mM Quis was used (4 cells), the initial increase in sIPSCs

9k20

/ 9k1d$$oc27 J292-7

was larger than that induced by 0.5 mM Quis. However, this response tended to decline to a relatively low level and most large events disappeared, similar to what was seen with the application of 80 mM ACPD (Fig. 4B). The mechanism for this decline was addressed in current-clamp recording (see below). The rise time and decay of sIPSCs were not affected by Quis or ACPD when compared with control sIPSCs of similar amplitude. The effects of Quis were largely reversible on washing, although three of seven cells showed a persistent increase in sIPSC frequency. Cells responded to repeated applications of Quis. mGluR effects on mIPSCs The enhancement of sIPSCs by Quis and ACPD could be caused by a direct modulation of spontaneous vesicle release at presynaptic terminals or by evoking somatic action potentials. To distinguish between these two possibilities, we examined mGluR effects on mIPSCs recorded in the presence of TTX. Under these conditions, both Quis (0.5–2 mM, n Å 7, Fig. 5B) and ACPD (20–60 mM, n Å 12, Fig. 6) did not induce a statistically significant change in the frequency (Fig. 6A) or amplitude (Fig. 6B) of mIPSCs. As shown in Fig. 6B, inset, the kinetics of mIPSCs also were not affected by Quis or ACPD. These results suggest that the enhancement of sIPSCs by ACPD and Quis was mediated through an action potential-dependent mechanism. mGluR-induced action potential firing in interneurons To test the idea that Quis and ACPD were inducing repetitive firing in inhibitory interneurons, current-clamp recordings were made. In the absence of TTX, Quis (0.5 mM) induced a sustained depolarization ( Ç10 mV, 3 of the 5 cells) or action potential firing riding on the depolarization (2 of the 5 cells). At 1 mM, Quis induced action potential firing in every cell tested (n Å 7, Fig. 7A). With 1 mM Quis, the action potential firing was transient in four of the seven cells (Fig. 7A), possibly due to Na / channel inactivation induced by the sustained depolarization. On washing, the cells regained their firing capability. Input resistance, measured by injecting a small hyperpolarizing current pulse, was reduced by 43 { 4% by 1 mM Quis in five layer I cells tested (Fig. 7A, insets). Similar to but less potent than Quis, 40 mM ACPD was able to induce a depolarization of Ç10 mV (3 of 5 cells) and action potential firing (2 of the 5 cells). At 80 mM, ACPD induced action potential firing in every cells tested (5 cells, Fig. 7B) and also decreased input resistance by Ç37 { 6% in 4 layer I cells tested (Fig. 7B, insets). To verify that layer I neurons represent a valid model for neocortical GABAergic interneurons, recordings also were made from deeper layer interneurons. We tested Quis (1 mM) in three deeper layer interneurons, identified by their fast spiking properties. From resting potentials near 060 mV, depolarizations and action potential firing were induced in the three cells tested (Fig. 8). Quis- and ACPD-induced inward current The above results suggest that Quis and ACPD may induce an inward current in inhibitory neurons. Indeed, bath

10-29-97 14:18:38

neupa

LP-Neurophys

MGLURS

IN INTERNEURONS

2291

FIG . 4. Bath application of 1S,3R-1-aminocyclopentane-1,3-dicarboxylic acid (ACPD) increases the frequency of sIPSCs in layer I neurons. A, left: records under control conditions (bathing solution containing 20 mM D-APV, 10 mM CNQX, and no TTX; holding potential 050 mV). Right: sIPSC frequency was markedly increased in presence of ACPD. Large amplitude, presumably action potential-dependent events, also were increased. B: bath application of 80 mM ACPD increased the frequency of sIPSCs in another layer I neuron bathed in 40 mM D-APV, 20 mM CNQX, and no TTX (holding potential was 070 mV). Increase was most robust at the beginning of ACPD application. Inward current induced by 80 mM ACPD was composed of 2 components: a direct ACPD-induced inward current and a current resulting from high-frequency sIPSCs.

application of Quis (0.5 or 1 mM), in the presence of 0.5 mM TTX, 40–50 mM CNQX, 40–100 mM D-APV, and 10 mM BIC, induced an inward current in layer I neurons. At a holding potential of 070 mV, this current was 33 { 8 pA (n Å 7) in the presence of 0.5 mM Quis (Fig. 5B) and 69 { 23 pA (n Å 11) in 1 mM Quis. This current did not desensitize during prolonged ( °30 min) Quis applications. ACPD was much less effective in inducing inward currents. At 20 mM, which increased sIPSCs, ACPD did not produce a detectable inward current (5 cells). ACPD at 40 mM produced an inward current of 5–10 pA in 4 of 12 cells. At 80 mM, ACPD induced an inward current of 8–20 pA

in five of nine cells. However, at 200 mM ACPD was able to induce an inward current (10–45 pA) in each of the six layer I neurons tested. Because of the relatively small amplitude, we did not attempt to determine the reversal potential of the ACPD-induced current. To determine the ionic mechanism and voltage sensitivity of the Quis-induced inward current, we performed voltageramp experiments. Because layer I cells have strong voltagedependent K / currents (Zhou and Hablitz 1996c), we used a CsCl-based intracellular pipette solution to block partially voltage-dependent K / currents in these experiments. Currents were evoked by ramping cells (n Å 4) from 070 to

FIG . 5. Quisqualic acid (Quis) increases the frequency of spontaneous but not mIPSCs in layer I neurons. A: in the control bathing solution without TTX, sIPSCs occurred at a frequency of Ç1 Hz. Bath application of 0.5 mM Quis caused a sustained increase in sIPSC frequency. More large amplitude, presumably action potential-dependent events were seen. B: in the presence of TTX, 0.5 mM Quis induced a sustained inward current ( Ç50 pA in this cell) with no effect on mIPSCs. In both recordings, the cell recovered on washing. Both recordings (A and B) were from the same layer I neuron at a holding potential of 070 mV.

9k20

/ 9k1d$$oc27 J292-7

10-29-97 14:18:38

neupa

LP-Neurophys

2292


Amplitude distribution of mIPSCs in layer I neurons

FIG . 6. Activation of mGluRs has no effect on frequency or amplitude of mIPSCs. A: interevent interval distributions for mIPSCs recorded under control conditions and after bath application of 40 mM ACPD. No significant effect on interevent intervals was observed (P Å 0.99, K-S test). B: cumulative frequency plots of mIPSC amplitudes were not changed by ACPD (P Å 0.02, K-S test). mIPSC kinetics also were unaffected (inset). Recording was made from a layer I neuron at a holding potential of 070 mV in the presence of 0.5 mM TTX.

30 mV in the absence and presence of 1 mM Quis. The difference current, obtained by subtracting the control current from the current evoked in the presence of Quis, represents the current induced by Quis. The current-voltage relation of this current was basically linear over the voltage range 070 to 30 mV (Fig. 9). Linear regression analysis indicated a reversal potential of 0.1 { 0.5 mV (n Å 4). DISCUSSION

The present results indicate that layer I neurons receive a significant inhibitory input, suggesting that interneurons integrate a broad range of synaptic signals. IPSCs in both layer I neurons and pyramidal cells had rapid rise times and complex decays. Activation of mGluRs, presumably mGluR1 and/or mGluR5, resulted in an increased frequency of IPSCs, By directly recording from GABAergic neurons, we have shown that mGluR effects were mediated via an inward current in these interneurons.

Amplitude distributions of mIPSCs were broad, ranging from 10 to 400 pA, and skewed toward large events. Similar broad and skewed amplitude distributions have been described for a number of synapses and have been suggested to result from variations in the transmitter released from individual synaptic vesicles (Bekkers et al. 1990; Frerking et al. 1995; Zhou and Hablitz 1997). Variations in postsynaptic receptor density also may contribute to the broad amplitude distribution (Liu and Tisen 1995). Amplitude distributions with multiple, equally distant peaks were reported for mIPSCs from hippocampal granule cells (Edwards et al. 1990). Layer I neurons had only one clear peak at Ç35 pA or 500 pS. This is higher than the first and largest peaks previously reported ( °200 pS) (Edwards et al. 1990; Kraszewski and Grantyn 1992; Salin and Prince 1996; Soltesz et al. 1995). The mechanism underlying this difference is unknown. In layer I neurons, the mean amplitude of mIPSCs at 070 mV was Ç50 pA, indicating a conductance of Ç700 pS. If we assume a single channel conductance of Ç25 pS (Edwards et al. 1990; Mistry and Hablitz 1990), then Ç28 GABAA channels were open at the peak of mIPSCs. mIPSC kinetics Under the present experimental conditions, averaged mIPSCs had a 10–90% rise time of 0.65 ms and an average rise rate of Ç80 pA/ms. This rise is faster than that of mIPSCs from dentate gyrus and cerebellar granule cells (Edwards et al. 1990; Puia et al. 1994) and neocortical pyramidal neurons (Salin and Prince 1996). The present value is also similar to the rise time of GABA currents evoked in membrane patches from hippocampal and cerebellar neurons (Jones and Westbrook 1995; Maconochie et al. 1994). The decay of mIPSCs in layer I neurons was complex. Averaged mIPSCs had double exponential decays with t f of Ç4 ms and ts of Ç22 ms with the slow component

FIG . 7. Quis and ACPD induce action potential firing in layer I neurons. A: 1 mM Quis induced a depolarization and action potential firing. Input resistance, as monitored by constant hyperpolarizing current pulses (insets) was decreased during Quis application. Combined effect of depolarization and decreased input resistance resulted in a loss of excitability at the peak of the response. B: 80 mM ACPD induced a depolarization and action potential firing and a decrease in input resistance (insets). Excitability was lost during the period of peak depolarization and decreased input resistance. Cell recovered on washing. Both recordings (A and B) were from the same layer I neuron. Resting membrane potential was 060 mV. BIC (10 mM), D-APV (50 mM), and CNQX (40 mM) were included in the control bathing solution.

9k20

/ 9k1d$$oc27 J292-7

10-29-97 14:18:38

neupa

LP-Neurophys

MGLURS

IN INTERNEURONS

2293 FIG . 8. Quis induces action potential firing in deep layer fast-spiking interneurons. Bath application of 0.5 mM Quis induced a depolarization and sustained action potential firing. Recovery was seen after Ç4 min of washing with control solution. Recording was from a layer III nonpyramidal fast-spiking neuron. Response to a depolarizing current pulse is shown (inset). Resting membrane potential was 057 mV. BIC (10 mM), D-APV (50 mM), and CNQX (20 mM) were included in the control bathing solution.

contributing Ç60% to the total amplitude. A similar double exponential decay has been reported for mIPSCs in hippocampal dentate (Edwards et al. 1990) and for sIPSCs in cerebellar granule cells (Puia et al. 1994). However, mIPSCs with a single exponential decay have been reported (Otis and Mody 1992; Salin and Prince 1996). Soltesz and Mody (1995) suggested that mechanical injury to dendrites eventually can transform a single into a double exponential decay. Our efforts to protect neurons from excitotoxic damage during slicing using sucrose-based high Mg 2/ -low Ca 2/ solutions and with 40 mM D-APV in a few cases had no effect on the decay of layer I mIPSCs. Jones and Westbrook (1995) have provided evidence that in cultured hippocampal neurons the slowly decaying component of IPSCs was due to the reopening of GABAA channels after exiting from desensitized states. It remains to be determined whether the slow component of mIPSCs in layer I neurons is generated by a similar mechanism. Even though the decays of averaged mIPSCs were relatively homogenous across cells, the decay of individual mIPSCs, even for those with similar rise times, was highly variable in individual cells. This suggests that differences in the kinetics of individual mIPSCs could result partly from variations in intrinsic properties of GABAA receptors at individual synapses. Because mIPSCs had similar kinetic properties in layer I neurons and pyramidal cells, these properties

FIG . 9. Current-voltage relation for the Quis-induced current in a layer I neuron. Control bathing solution contained 0.5 mM TTX, 50 mM D-APV, 50 mM CNQX, and 10 mM BIC. Intracellular pipette solution contained 125 mM CsCl. Currents were evoked by ramping the cell from 070 to 30 mV in the absence and presence of 1 mM Quis. Difference current, obtained by subtracting the control current from the current evoked in the presence of Quis, represents the Quis-induced current. Current was not voltage dependent and had a linear I-V relation. Linear regression analysis indicated a reversal potential of 0.07 mV.

9k20

/ 9k1d$$oc27 J292-7

do not distinguish between receptor subtypes in the two cell types. mGluRs enhance GABAergic transmission by firing cortical interneurons ACPD is a nonselective mGluR agonist, whereas Quis is a selective agonist for group I mGluRs (Pin and Bockaert 1995 for review). Bath application of ACPD or Quis caused an increase in the frequency of sIPSCs, but the two agonists had no effect on mIPSCs recorded in the presence of TTX. The mean amplitude of sIPSCs also was increased substantially by ACPD and Quis due to the occurrence of more large amplitude events. These results indicate that mGluR enhancement of sIPSCs is mediated through an action potential-dependent mechanism, not through a direct modulation of the spontaneous synaptic vesicle release process in the presynaptic terminal. One obvious candidate mechanism is that activation of mGluRs elicits action potentials in inhibitory interneurons. Consistent with this idea, Quis and ACPD induced an inward current of °100 pA in layer I neurons. Such a current is strong enough to depolarize these cells beyond action potential threshold (Zhou and Hablitz 1996b). Indeed, action potentials were observed in layer I and deep layer interneurons under current-clamp conditions. Similarly, it has been reported that, in hippocampal neurons, activation of mGluRs increased the frequency of sIPSCs without altering mIPSCs (Fitzimonds and Dichter 1996; Gereau and Conn 1995; McBain et al. 1994; Sciancalepore et al. 1995). However, in contrast to the enhancing effect in hippocampus (McBain et al. 1994), activation of mGluRs has a modest depressing effect on spontaneous excitatory synaptic currents in layer I and deep layer neurons (unpublished observation). Our present results suggest that Quispreferring group 1 mGluRs are present on the somata but not the axonal terminals of cortical interneurons, even though these terminals may possess different mGluRs coupled to other functions. This is consistent with the findings of Baude et al. (1993) that immunostaining with an mGluR1a antibody revealed an enriched presence of mGluR1a on the somata and dendrites but not axonal terminals of hippocampal and neocortical interneurons. Poncer et al. (1995) also provided indirect evidence that group 1 mGluRs increase sIPSCs in the hippocampus by depolarizing interneurons. However, we did not observe the oscillatory responses reported by McBain et al. (1994) in hippocampal interneurons in our cells upon the application ( °30 min) ACPD and Quis, even though our pipette solution also included ATP and GTP.

10-29-97 14:18:38

neupa

LP-Neurophys

2294


Quis- and ACPD-induced inward current Quis is a potent agonist for group I mGluRs. In layer I neurons, Quis (0.5–1 mM), in the presence of °100 mM D-APV and °50 mM CNQX, was able to induce a nondesensitizing inward current accompanied by a decrease in input resistance and slight increase in membrane noise. This Quisinduced current was linear and reversed at 0 mV, indicating that activation of group I mGluRs may open a type of nonselective cation channel. This is consistent with findings in hippocampal CA3 pyramidal neurons showing that activation of group I mGluRs can open nonselective cation channels with a single channel conductance of 14 pS (Guerineau et al. 1995). In contrast, in rat neostriatal neurons, activation of group I mGluRs caused a depolarization and an increase in input resistance by suppressing a K / conductance (Takeshita et al. 1996). Further, ACPD caused a hyperpolarization and a decrease in input resistance by opening large conductance calcium-dependent (BK) type K / channels in rat basolateral amygdala neurons (Holmes et al. 1996). Differential expression of mGluR subtypes coupled to varying effector mechanisms presumably underlies this variability. A high concentration of ACPD (200 mM) was needed to induce reliably an inward current. The apparent failure for low concentrations of ACPD to induce inward currents in a large proportion of layer I neurons may result from several factors. First, ACPD is a weak agonist for group I mGluRs (Pin and Bockaert 1995). Second, the expression of mGluRs may not be homogeneous among individual neurons. Some layer I neurons might have a relatively lower density of mGluRs than others. Third, some cells might be damaged partly such that they had functionally lost a portion of their responsiveness. Finally, under our recording conditions, a swing within 5 pA in baseline holding current was not uncommon in our recording. ACPD-induced current õ5 pA was usually below the detection threshold. However, because many layer I neurons and possibly many deep layer inhibitory interneurons have very high-input resistance (Zhou and Hablitz 1996b), a few picoamps of depolarizing current may be strong enough to evoke action potential firing. Further, each cortical neuron may receive inhibitory synaptic inputs from many individual inhibitory neurons, and some of these inhibitory neurons may be depolarized strongly by ACPD. Therefore, ACPD at low concentrations still can increase sIPSCs. Finally, it needs to be pointed out that we cannot fully rule out the possibility that ACPD directly evokes action potentials in axon terminals in a subset of cortical interneurons similar to the situation for the preterminal nicotinic receptors on GABAergic axon terminals in the rat interpedunclar nucleus (Lena et al. 1993). Functional consideration of mGluRs on interneurons The depolarizing action of mGluRs on cortical interneurons may have important functional implications. In a neuronal circuit, if pyramidal neurons become excited and release glutamate onto pyramidal and GABAergic cells, the latter cell type is more likely to fire action potentials due to their high-input resistance (McCormick et al. 1985; Zhou and Hablitz 1996b), rapid excitatory postsynaptic currents (Zhou and Hablitz 1997). and mGluR-induced depolariza-

9k20

/ 9k1d$$oc27 J292-7

tion (this study). mGluRs on interneurons help to produce feedback inhibition onto pyramidal cells. Thus mGluR agonists that can enhance selectively GABAergic synaptic transmission may be useful presynaptic enhancers of GABAergic inhibition. However, the decrease in input resistance, coupled with the depolarization was able to lead to a complete loss of excitability in layer I neurons when high concentrations of Quis and ACPD were used. This suggests that overexcitation of mGluRs actually may reduce the output from cortical GABAergic cells. In conclusion, layer I neurons receive GABAA receptormediated inhibitory synaptic inputs. mIPSCs had a fast rise with a 10–90% rise time of Ç0.6 ms and a double exponential decay. The kinetic profile of mIPSCs in layer I neurons and pyramidal neurons was similar. Activation of mGluRs, possibly mGluR1 and -5, enhanced GABAergic transmission by inducing an inward current and firing in cortical interneurons. We thank A. Margolies for excellent technical assistance. This work was supported by National Institute of Neurological Disorders and Strokes Grant NS-18145. Address for reprint requests: J. J. Hablitz, Dept. of Neurobiology, University of Alabama at Birmingham, Birmingham, AL 35294. Received 10 April 1997; accepted in final form 20 June 1997. REFERENCES BAUDE, A., NUSSER, Z., ROBERTS, J.D.B., MULVIHILL, E., MC ILHINNEY, R.A.J., AND SOMOGYI, P. The metabotropic glutamate receptor (mGluR1a ) is concentrated at perisynaptic membrane of neuronal subpopulations as detected by immunogold reaction. Neuron 11: 771–787, 1993. BEKKERS, J. M., RICHERSON, G. B., AND STEVENS, C. F. Origin of variability in quantal size in cultured hippocampal neurons and hippocampal slices. Proc. Natl. Acad. Sci. USA 87: 5359–5362, 1990. BEKKERS, J. M. AND STEVENS, C. F. Cable properties of cultured hippocampal neurons determined from sucrose-evoked miniature EPSCs. J. Neurophysiol. 75: 1250–1255, 1996. BUHL, E. H., HALASY, K., AND SOMOGYI, P. Diverse sources of hippocampal unitary inhibitory postsynaptic potentials and the number of synaptic release sites. Nature 368: 823–828, 1995. CONNORS, B. W. AND AMITAI, Y. Generation of epileptiform discharge by local circuits of neocortex. In: Epilepsy: Models, Mechanisms, and Concepts, edited by P. A. Schwartzkroin. Cambridge, UK: Cambridge, 1993, p. 388–423. EDWARDS, F. A., KONNERTH, A., AND SAKMANN, B. Quantal analysis of inhibitory synaptic transmission in the dentate gyrus of rat hippocampal slices: a patch-clamp study. J. Physiol. (Lond.) 430: 213–249, 1990. FITZIMONDS, R. M. AND DICHTER, M. C. Heterologus modulation of inhibitory synaptic transmission by metabotropic glutamate receptors in cultured hippocampal neurons. J. Neurophysiol. 75: 885–893, 1996. FRERKING, M., BORGES, S., AND WILSON, M. Variation in GABA mini amplitude is the consequence of variation in transmitter concentration. Neuron 15: 885–895, 1995. GABBOTT, P.L.A. AND SOMOGYI, P. Quantitative distribution of GABAimmunoreactive neurons in the visual cortex (area 17) of the cat. Exp. Brain Res. 61: 323–331, 1986. GEREAU, R. W., IV AND CONN, P. J. Multiple presynaptic metabotropic glutamate receptors modulate excitatory and inhibitory synaptic transmission in hippocampal area CA1. J. Neurosci. 15: 8679–8689, 1995. GUERINEAU, N. C., BOSSU, J. L., GAHWILER, B. H., AND GERBER, U. Activation of nonselective cationic conductance by metabotropic glutamatergic and muscarinic agonists in CA3 pyramidal neurons of the rat hippocampus. J. Neurosci. 15: 4395–4407, 1995. HOLMES, K. H., KEELE, N. B., ARVANOV, V. L., AND SHINNICK-GALLAGHER, P. Metabotropic glutamate receptor agonist-induced hyperpolarization in rat basolateral amygdala neurons: receptor characterization and ion channels. J. Neurophysiol. 76: 3059–3069, 1996.

10-29-97 14:18:38

neupa

LP-Neurophys

MGLURS

IN INTERNEURONS

HOWE, J. R., SUTOR, B., AND ZIEGLGANSBERGER, W. Characteristics of longduration inhibitory postsynaptic potentials in rat cortical neurons in vitro. Cell. Mol. Neurobiol. 7: 1–18, 1987. JONES, M. V. AND WESTBROOK, G. L. Desensitized states prolong GABAA channel responses to brief agonist pulses. Neuron 15: 181–191, 1995. KAWAGUCHI, Y. AND KUBOTA, Y. Physiological and morphological identification of somatostatin- or vasoactive intestinal polypeptide-containing cells among GABAergic cell types in rat frontal cortex. J. Neurosci. 16: 2701–2715, 1996. KRASZEWSKI, K. AND GRANTYN, R. Unitary, quantal and miniature GABAactivated synaptic chloride currents in cultured neurons from the rat superior colliculus. Neuroscience 47: 555–570, 1992. LENA, C., CHANGEUX, J.-P., AND MULLE, C. Evidence for ‘‘preterminal’’ nicotinic recptors on GABAergic axons in the rat interpeduncular nucleus. J. Neurosci. 13: 2680–2688, 1993. LI, J. Y. AND SCHWARK, H. D. Distribution and proportions of GABAimmunopositive neurons in cat primary somatosensory cortex. J. Comp. Neurol. 343: 353–361, 1994. LIU, G. AND TSIEN, R. W. Properties of synaptic transmission at single hippocampal synaptic boutons. Nature 375: 404–408, 1995. LLANO, I. AND MARTY, A. Presynaptic metabotropic glutamatergic regulation of inhibitory synapses in rat cerebellar slices. J. Physiol. (Lond.) 486: 163–176, 1995. MACONOCHIE, D. J., ZEMPEL, J. M., AND STEINBACH, J. H. How quickly can GABAA receptors open? Neuron 12: 61–71, 1994. MC BAIN, C. J., DICHIARA, T. J., AND KAUER, J. A. Activation of metabotropic glutamate receptors differentially affects two classes of hippocampal interneurons and potentiates excitatory synaptic transmission. J. Neurosci. 14: 4433–4445, 1994. MCCORMICK, D. A., CONNORS, B. W., LIGHTHALL, J. W., AND PRINCE, D. A. Comparative electrophysiology of pyramidal and sparsely spiny stellate neurons of the neocortex. J. Neurophysiol. 54: 782–806, 1985. MC KERNAN, R. M. AND WHITING, P. J. Which GABAA-receptor subtypes really occur in the brain? Trends Neurosci. 19: 139–143, 1996. MILES, R. Variation in strength of inhibitory synapses in the CA3 region of guinea-pig hippocampus in vitro. J. Physiol. (Lond.) 431: 659–676, 1990. MISTRY, D. K. AND HABLITZ, J. J. Activation of subconductance states by g-aminobutyric acid and its analogs in chick cerebral neurons. Pflu¨gers Arch. 416: 454–461, 1990. MODY, I., DE KONINCK, Y., OTIS, T. S., AND SOLTESZ, I. Bridging the cleft at GABA synapses in the brain. Trends Neurosci. 17: 517–525, 1994. MORIN, F., BEAULIEU, C., AND LACAILLE, J.-C. Membrane properties and synaptic currents evoked in interneuron subtypes in rat hippocampal slices. J. Neurophysiol. 76: 1–16, 1996. OTIS, T. S. AND MODY, I. Modulation of decay kinetics and frequency of GABAA receptor-mediated spontaneous inhibitory postsynaptic currents in hippocampal neurons. Neuroscience 49: 13–32, 1992. PAXINOS, G. AND WATSON, C. The Rat Brain in Stereotaxic Coordinates. (2nd ed.). Sydney, Australia: Academic, 1986. PIN, J.-P. AND BOCK AERT, J. Get receptive to metabotropic glutamate receptors. Curr. Opin. Neurobiol. 5: 342–349, 1995. PONCER, J.-C., SHINOZAKI, H., AND MILES, R. Dual modulation of synaptic inhibition by distinct metabotropic glutamate receptors in the rat hippocampus. J. Physiol. (Lond.) 485: 121–134, 1995.

9k20

/ 9k1d$$oc27 J292-7

2295

PUIA, G., COSTA, E., AND VICINI, S. Functional diversity of GABA-activated Cl0 currents in Purkinje versus granule neurons in rat cerebellar slices. Neuron 12: 117–126, 1994. REID, S.N.M., ROMANO, C., HUGHES, T., AND DAW, N. W. Immunohistochemical study of two phosphoinositide-linked metabotropic glutamate receptors (mGluR1a and mGluR5) in the cat visual cortex before, during, and after the peak of the critical period for eye-specific connections. J. Comp. Neurol. 355: 470–477, 1995. ROMANO, C., SESMA, M. A., MCDONALD, C. T., O’MALLEY, K., VAN DEN POL, A., AND OLNEY, J. W. Distribution of metabotropic glutamate receptor mGluR immunoreactivity in rat brain. J. Comp. Neurol. 355: 455– 469, 1995. SALIN, P. A. AND PRINCE, D. A. Spontaneous GABAA receptor-mediated inhibitory currents in adult rat somatosensory cortex. J. Neurophysiol. 75: 1573–1588, 1996. SCIANCALEPORE, M., STRATTA, F., FISHER, N. D., AND CHERUBINI, E. Activation of metabotropic glutamate receptors increase the frequency of spontaneous GABAergic currents through protein kinase A in neonatal rat hippocampal neurons. J. Neurophysiol. 74: 1118–1122, 1995. SHIGEMOTO, R., NAK ANISHI, S., AND MIZUNO, N. Distribution of the mRNA for a metabotropic glutamate receptor (mGluR1) in the central nervous system: an in situ hybridization study in adult and developing rat. J. Comp. Neurol. 322: 121–135, 1992. SOLTESZ, I. AND MODY, I. Ca 2/ -dependent plasticity of miniature inhibitory postsynaptic currents after amputation of dendrites in central neurons. J. Neurophysiol. 73: 1763–1773, 1995. SOLTESZ, I., SMETTERS, D. K., AND MODY, I. Tonic inhibition originates from synapses close to the soma. Neuron 14: 1273–1283, 1995. SPRUSTON, N., JAFFE, D. B., WILLIAMS, S. H., AND JOHNSTON, D. Voltageand space-clamp errors associated with the measurement of electrotonically remote synaptic events. J. Neurophysiol. 70: 781–802, 1993. TAKESHITA, Y., HARATA, N., AND AK AIKE, N. Suppression of K / conductance by metabotropic glutamate receptor in acutely dissociated large cholinergic neurons of rat caudate putamen. J. Neurophysiol. 76: 1545– 1558, 1996. THOMSON, A. M., WEST, D. C., HAHN, J., AND DEUCHARS, J. Single axon IPSPs elicited in pyramidal cells by three classes of interneurons in slices of rat neocortex. J. Physiol. (Lond.) 496: 81–102, 1996. VAUTRIN, J. AND BARKER, J. L. How can exocytosis account for the actual properties of miniature synaptic signals? Synapse 19: 144–149, 1995. ZHOU, F.-M. AND HABLITZ, J. J. Postnatal development of membrane properties of rat neocortical layer I neurons. J. Neurosci. 16: 1131–1139, 1996a. ZHOU, F.-M. AND HABLITZ, J. J. Rat neocortical layer I neurons. I. Action potential and repetitive firing properties. J. Neurophysiol. 76: 651–667, 1996b. ZHOU, F.-M. AND HABLITZ, J. J. Rat neocortical layer I neurons. II. Voltagedependent outward currents. J. Neurophysiol. 76: 668–682, 1996c. ZHOU, F.-M. AND HABLITZ, J. J. Morphological properties of intracellularly labeled rat neocortical layer I neurons. J. Comp. Neurol. 376: 198–213, 1996d. ZHOU, F.-M. AND HABLITZ, J. J. Spontaneous IPSCs in rat neocortical layer I neurons. Soc. Neurosci. Abstr. 22: 789, 1996e. ZHOU, F.-M. AND HABLITZ, J. J. Rapid kinetics and inward rectification of miniature EPSCs in layer I neurons of rat neocortex. J. Neurophysiol. 77: 2416–2426, 1997.

10-29-97 14:18:38

neupa

LP-Neurophys