Journal of Neurophysiology

Presynaptic Inactivation of Action Potentials and Postsynaptic Inhibition of GABAA Currents Contribute to KA-Induced Disinhibition in CA1 Pyramidal Neurons

Ning Kang, Li Jiang, Wei He, Jun Xu, Maiken Nedergaard, Jian Kang

Abstract

Kainate-type glutamate ionotropic receptors (KAR) mediate either depression or potentiation of inhibitory transmission. The mechanisms underlying the depressant effect of KAR agonists have been controversial. Under dual patch-clamp recording techniques in synaptically coupled pairs of CA1 interneurons and pyramidal neurons in hippocampal slices, micromolar concentrations of KAR agonists, kainic acid (KA, 10 μM) and ATPA (10 μM), induced inactivation of action potentials (APs) in 58 and 50% of presynaptic interneurons, respectively. Inactivation of interneuronal APs might have significantly contributed to KA-induced decreases in evoked inhibitory postsynaptic currents (eIPSCs) that are obtained by stimulating the stratum radiatum. With controlled interneuronal APs, KAR agonists induced a decrease in the potency (mean amplitude of successful events) and mean amplitude (including failures) of unitary inhibitory postsynaptic currents (uIPSCs) without significantly changing the success rate (Ps) at perisomatic high-Ps synapses. In contrast, KAR agonists induced a decrease in both the Ps and potency of uIPSCs at dendritic high-Ps synapses. KAR agonists induced an inhibition of GABAA currents by activating postsynaptic KARs in pyramidal neurons; this was more prominent at dendrites than at soma. Both the exogenous GABA-induced current and the amplitude of miniature IPSCs (mIPSCs) were attenuated by KAR agonists. Thus the postsynaptic KAR-mediated inhibition of GABAA currents may contribute to the KAR agonist-induced decrease in the potency of uIPSCs and KA-induced disinhibition.

INTRODUCTION

Glutamate ionotropic kainate receptors (KARs) have been reported to mediate excitatory synaptic currents in a certain area of neurons (Castillo et al. 1997; Cossart et al. 1998; DeVries 2000; DeVries and Schwartz 1999; Frerking et al. 1998; Kidd and Isaac 1999; Li et al. 1999; Vignes and Collingridge 1997) and to modulate both excitatory (Chittajallu et al. 1996; Kamiya and Ozawa 2000; Schmitz et al. 2000) and inhibitory synaptic transmission (Ali et al. 2001; Binns et al. 2003; Clarke et al. 1997; Cossart et al. 2001b; Jiang et al. 2001; Mulle et al. 2000; Rodriguez-Moreno et al. 1997). Application of KAR agonists depresses fiber stimulation-evoked inhibitory postsynaptic currents (eIPSCs) in the CA1 area of the hippocampus (reviewed by Lerma et al. 2001). This disinhibition has attracted considerable attention because the phenomenon is comparable to the convulsant and excitotoxic properties of kainate (Coyle 1983; Fisher and Alger 1984; Westbrook and Lothman 1983). The precise mechanisms by which kainate depresses eIPSCs are unclear, although a direct presynaptic G-coupled protein kinase C (PKC) pathway has been proposed (Rodriguez-Moreno and Lerma 1998; Rodriguez-Moreno et al. 2000). Additionally, another group reported indirect effects of increased spontaneous IPSCs including activation of presynaptic GABAB receptors and dendritic shunt due to sustained activation of postsynaptic GABAA receptors (Frerking and Nicoll 2000; Frerking et al. 1999). One observation that argues for a direct presynaptic depressant role of KARs on GABA release is that the frequency of miniature IPSCs (mIPSCs) is reduced by kainate (Rodriguez-Moreno and Lerma 1998). However, later studies have shown that kainate has little or no effect on mIPSCs (Ben-Ari and Cossart 2000; Bureau et al. 1999; Frerking and Nicoll 2000; Semyanov and Kullmann 2001). In addition to the controversial arguments about mechanisms underlying kainate-induced depression of eIPSCs, facilitating effects of low (submicromolar) concentrations of KAR agonists on GABA release have recently been reported. First submicromolar concentrations of kainate trigger interneurons to fire action potentials (APs) spontaneously either by increasing excitatory inputs to interneurons (Frerking and Nicoll 2000; Frerking et al. 1998) or by depolarizing axons and lowering the AP threshold in interneurons (Semyanov and Kullmann 2001). Second, submicromolar concentrations of kainate enhance both the frequency of mIPSCs and the amplitude of eIPSCs at inhibitory-inhibitory connections in interneurons in the hippocampal CA1 area (Cossart et al. 2001b; Mulle et al. 2000). Third, pair recordings from synaptically coupled CA1 interneurons and pyramidal neurons reveal that submicromolar KAR agonists or synapse-released glutamate increases the efficacy of low release probability GABAergic synapses (Jiang et al. 2001).

GABAergic synapses are heterogeneous in their release probability (Jiang et al. 2000), location (Buhl et al. 1994; Maccaferri et al. 2000; Miles et al. 1996), and function (Cossart et al. 2001a; Lambert et al. 1991). In neuronal circuits, a perisomatic GABAergic synapse controls action potential firing (Melinek and Muller 1996; Miles and Wong 1987; Moore et al. 1983) while a dendritic GABAergic synapse modulates synaptic strength (Davies et al. 1991; Miles et al. 1996). KARs may differentially modulate these perisomatic and dendritic GABAergic synapses to play different roles in modulating dendritic inputs and controlling AP production. To investigate the mechanisms underlying KA-induced disinhibition, we tested the effects of KAR agonists KA (10 μM) and ATPA (10 μM) on interneuronal firing, mIPSCs, and uIPSCs at perisomatic or dendritic GABAergic synapses in hippocampal slices. We found that 10 μM KA and 10 μM ATPA caused inactivation of APs in 58 and 50% of interneurons, respectively. With control interneuronal APs, ATPA and KA induced a decrease in potency of uIPSCs at perisomatic high-Ps synapses but induced a more prominent decrease in both potency and Ps at dendritic high-Ps GABAergic synapses. Postsynaptic KAR-mediated inhibition of GABAA currents contributed to this decrease in the potency of uIPSCs.

METHODS

Preparation of slices

Brain slices were prepared as described previously (Jiang et al. 2001; Kang et al. 1998). Briefly, 14- to 20-day-old (P14–P20) Sprague-Dawley rats were anesthetized with pentobarbitone sodium (55 mg/kg) and decapitated. Brains were removed rapidly and glued with the posterior surfaces down. Transverse brain slices of 300 μm were cut with a vibratome (TPI, St Louis, MO) in a cutting solution containing (in mM) 2.5 KCl, 1.25 NaH2PO4, 10 MgSO4, 0.5 CaCl2, 10 glucose, 26 NaHCO3, and 230 sucrose. Slices containing the hippocampus were incubated in the slice solution gassed with 5% CO2-95% O2 for 1–7 h and then transferred to a recording chamber (1.5 ml) that was perfused with the slice solution gassed with 5% CO2-95% O2 at room temperature (23–24°C) for recording. The standard slice solution contained (in mM) 126 NaCl, 2.5 KCl, 1.25 NaH2PO4, 2 MgCl2, 2 CaCl2, 10 glucose, and 26 NaHCO3 (pH at 7.4 when gassed with 95% O2-5% CO2).

Electrophysiology

The recording chamber was placed on the stage of an Olympus BX51 upright microscope (Olympus Optical) equipped with DIC (differential inference contrast) optics, and cells were visualized with a ×63 water-immersion lens. Two electrical manipulators (TS Products, Arleta, CA) were mounted on the stage in opposing positions and moved along a plane 18° to the horizontal. Patch electrodes with a resistance of 4–7 MΩ were pulled from KG-33 glass capillaries (1.0 mm ID, 1.5 mm OD, Garner Glass, Claremont, CA) using a P-97 electrode puller (Sutter Instrument, Novato, CA). A seal resistance <5 GΩ was rejected. Dual patch-clamp recordings were performed in pairs of interneurons whose somata were located in the stratum radiatum and pyramidal neurons in the CA1 pyramidal layer. Interneurons were patched in the whole cell current-clamp configuration (Hamill et al. 1981) using a MultiClamp 700 amplifier (Axon Instruments, Burlingame, CA). The pipette solution contained (in mM) 120 K-gluconate, 10 KCl, 1 MgCl2, 10 HEPES, 0.1 EGTA, 0.025 CaCl2, 1 ATP, 0.2 GTP, and 4 glucose (pH adjusted to 7.2 with KOH). Pyramidal neurons were patched in the voltage-clamp configuration and recorded at a holding potential of −60 mV using an Axopatch 200B amplifier (Axon Instruments). The pipette solution contained (in mM) 120 CsCl, 10 KCl, 1 MgCl2, 10 HEPES, 0.1 EGTA, 0.025 CaCl2, 1 ATP, 0.2 GTP, 5 QX-314, and 4 glucose (pH adjusted to 7.2 with KOH). Experiments with a holding current > −200 pA were rejected. It is important for comparing the amplitude of uIPSCs to obtain relative constant access resistance during experiments. We started experiments after access resistance getting relative constant (5–10 min), and recordings with changes in series resistance >10% of control in pyramidal neurons were rejected.

Identification of synaptically connected pairs

Cell pairs that included an interneuron in the s. radiatum and a pyramidal neuron in the pyramidal layer in the hippocampal CA1 region were identified under DIC microscopy and patched in the whole cell configuration (interneuron: I clamp and pyramidal neuron: V clamp). Synaptically connected pairs were identified by a train of three depolarization pulses that evoked three interneuronal APs with an interval of 20 ms and a duration of 15 ms. A successful uIPSC event was defined as a downward current with decay >20 ms and amplitude >5 pA starting within the mean latency (for a given pair) ±2 SD (fast: 0.2 ms, slow: 1.1 ms).

Data analysis

Recording signals were filtered through an 8-pole Bessel low-pass filter with a 2-kHz cut-off frequency and sampled by a PCLAMP 8.1 program (Axon Instruments) with an interval of 50 or 200 μs. Data were further processed with Origin 5.1 (Microcal Software, Northampton, MA) and CorelDraw 9.0 (Corel) programs. Statistical data are presented as means ± SE if not indicated. Data from experiments with GYKI 53655 and SYM 2206 were pooled for analysis.

Local application of the GABAA agonists

In the presence of 1 μM TTX, 50 μM SYM 2206, and 50 μM APV, whole cell GABAA currents were induced by pressure application of GABA or muscimol (50 μM) through another pipette, which contained the slice solution supplemented with GABA or muscimol. Under DIC optics, the pipette was placed on the top of or 100–150 μm away from the recorded pyramidal neurons. Three to 5 psi pressure was applied to the puff pipette through a picospritzer III (Parker Hannifin General Valve Operation).

Biocytin-staining

Neurons in pairs were patched with a biocytin-containing pipette (0.3%). After finishing experiments, slices were perfused with 4% paraformaldehyde +0.2% picric acid +0.1% glutaraldehyde (pH = 7.4) for 10 min. Slices were postfixed overnight at 4°C. After several washed with 0.1 M phosphate-buffered saline (PBS), slices were treated with 0.3% H2O2 for 10 min at room temperature. Then slices were incubated in 0.2% Triton X-100 and 0.2% albumin for 1 h at 4°C, and stained with the ABC kit (1:500, Pierce 32054) for 1 h at room temperature. Last, slices were incubated in 0.025% DAB plus 0.003% H2O2 Tris buffer solution for 15 min and mounted in Permount. Morphological subtypes of interneurons were identified with a criteria previously reported by Buhl et al. (1994).

RESULTS

Micromolar KAR agonists inactivate interneuronal APs

In our previous study, we found that submicromolar KAR agonists (KA or ATPA, 300 nM) increased the strength of low-Ps GABAergic synapses but did not induce a significant change in the Ps and amplitude of uIPSCs at high-Ps synapses (Jiang et al. 2001). In this study, we examined the effects of micromolar concentrations (5 or 10 μM) of KAR agonists on uIPSCs. Evoked IPSCs (eIPSCs) in CA1 pyramidal neurons are obtained by stimulating the s. radiatum. However, micromolar KA may induce changes in the firing ability of stimulated fibers and affect recording of eIPSCs. Therefore we first examined the effects of KAR agonists on interneuronal firing using pair recordings. KA (10 μM) or ATPA (10 μM) was applied to the recording chamber in the presence of the N-methyl-d-aspartate (NMDA) receptor antagonist APV (50 μM) and the AMPA receptor antagonist GYKI 53655 (50 μM) or SYM 2206 (50 μM). During application of KA, the patched interneuron, which lacked holding currents, depolarized (ΔV = 28.5 ± 3.3 mV) and fired a high-frequency of spontaneous APs in a short time period (Fig. 1A, Int, sAP). Meanwhile, the pyramidal neuron showed an inward current with a high frequency of spontaneous IPSCs (Fig. 1A, Pyr). The amplitude of current injection-evoked APs (Fig. 1A, Int, eAP) decreased gradually. After 2 min, evoked interneuronal APs disappeared, and only the current injection depolarization remained (Fig. 1, B and C, 4 and 5), indicating that the interneurons lost the ability to fire APs. Fifty-eight (11/19) and 50% (5/10) of interneurons temporarily lost firing ability during application of KA and ATPA, respectively. A transient decrease in the input resistance was also noted (Fig. 1B, 2), reflecting increased channel conductance in the interneuron. Thus inactivation of interneuronal APs significantly contributes to KA-induced depression of eIPSCs. When interneuronal APs were fully inactivated, a strong depolarization pulse (600–800 pA) could not successfully trigger uIPSCs (16 pairs), suggesting that AP regeneration is required for triggering synaptic release of GABA. Furthermore, the attenuated APs evoked uIPSCs with a similar efficiency as control APs (Fig. 1, B and C, 3 and 6), suggesting that attenuated APs could still regenerate along the axon. To assess AP width during KA application, we measured the AP half-widths (Fig. 1, B and C, 2 and 3, arrows). The mean half-width of APs after KA application (Fig. 1D, KA) was significantly longer than that during control (Fig. 1D, CON, P < 0.01). Increased AP width might increase AP-driven GABA release due to increased opening time of voltage-gated calcium channels (VGCC), which could counteract the depressant effects of decreased AP amplitude on GABA release.

FIG. 1.

Kainate receptor (KAR) agonist-induced inactivation of interneuronal action potentials (APs). A: continuing recording traces showing membrane currents in a pyramidal neuron (V-clamp) and membrane potential in the interneuron (I-clamp) before and after KA application in the presence of APV (50 μM) and GYKI 53655 (GYKI, 50 μM). Spontaneous inhibitory postsynaptic currents (IPSCs) increase simultaneously with increased spontaneous spikes (sAP) during KA application. Evoked spikes (eAP) were triggered every 10 s by a depolarization pulse. B: enlarged time scale from period indicated in A, showing that amplitude of interneuronal APs (bottom line, arrows) gradually decreases (2, 3) during KA application, fully inactivates (4, 5) after 2 min, and recovers after washing (6, 7). Top traces: unitary IPSCs (uIPSCs) recorded in the patched pyramidal neuron. Controls (1) include 10 continuous uIPSCs. Data were collected from a representative pair. C: the amplitude of interneuronal APs (•) and uIPSCs (○) is plotted against time for the data in A. The amplitude was measured from the starting point of AP to the peak of AP. D: the KA-induced increase in the width of spikes. The half-width of APs was measured during control (CON), 2 min after KA application (KA), and after washout (Wash). *, P < 0.01, compared with control, paired t-test, n = 7 pairs.

FIG. 2.

KAR agonists induce a decrease in the potency of uIPSCs at perisomatic high-Ps synapses. A: biocytin staining for paired pyramidal neuron and interneuron that produced large and fast uIPSCs in B. A branch of the interneuronal axon (←) makes a contact (▴) on the soma of the pyramidal cell. The bar represents 10 μm. B: in the presence of APV and SYM 2206 (SYM), 10 μM ATPA attenuates the potency and mean amplitude of perisomatic high-Ps synapses. Top: uIPSCs evoked by 10 continuous spikes during control (CON), during ATPA application (ATPA), and after washout (Wash). C: the amplitude of uIPSCs plotted against time. D: the Ps, potency (Pot), and amplitude (Amp) of uIPSCs during control (□), during ATPA application (▪), and after washout (▒). No significant decrease was found in the Ps (P = 0.135), but the potency and amplitude were significantly reduced. *, P < 0.05, compared with control, paired t-test, n = 5 high-Ps pairs. E: the Ps, potency (Pot), and amplitude (Amp) of uIPSCs during control (□), during KA application (▪), and after washout (▒). *, P < 0.05, compared with control, paired t-test, n = 4 high-Ps pairs.

FIG. 3.

ATPA decreases both the Ps and amplitude of uIPSCs at dendritic high-Ps synapses. A: biocytin-staining paired pyramidal neuron and interneuron that produced small and slow uIPSCs in B. The interneuronal axon (↓) makes a dendritic contact (▴) on the pyramidal cell. The bar represents 10 μm. B: traces of uIPSCs evoked by 10 continuous spikes during control (CON), during ATPA application (ATPA), or after washout (Wash) from a representative dendritic high-Ps pair. C: a plot of the amplitude of uIPSCs against time from data in A. D: the Ps, potency (Pot), and amplitude (Amp) of uIPSCs during control (□), during ATPA application (▪), and after washout (▒). *, P < 0.01 and **, P < 0.001, compared with control, paired t-test, n = 6 high-Ps pairs. E: the Ps, potency (Pot), and amplitude (Amp) of uIPSCs during control (□), during KA application (▪), and after washout (▒). *, P < 0.01 and **, P < 0.001, compared with control, paired t-test, n = 4 high-Ps pairs.

KAR agonists induce a decrease in the potency and amplitude of uIPSCs at perisomatic high-Ps synapses

To study KAR-mediated depression of uIPSCs at perisomatic and dendritic GABAergic synapses, we located GABAergic synapses in pyramidal neurons by analyzing the rising time and amplitude of uIPSCs (Jiang et al. 2000; Maccaferri et al. 2000). To confirm the locations of the GABAergic synapses, some pairs were recorded with patch pipettes containing 0.3% biocytin followed by immunoassay of biocytin. Similar to findings by Maccaferri et al. (2000), biocytin-staining showed that axons of the presynatic interneurons (basket cells) in three pairs with large, fast uIPSCs (potency range: 67–292 pA; potency, mean amplitude of nonfailure events) were distributed around the pyramidal layer and made contacts on somata of postsynaptic cells (Fig. 2A, ▴). All interneurons in five pairs with small, slow uIPSCs (potency range: 9.9–32.7 pA) had axonal distribution in the s. radiatum and had axonal contacts on dendrites of pyramidal neurons (Fig. 3A, ▴).

GABAergic synapses are heterogeneous in their release probability and can be divided into high-Ps (Ps ≥ 0.5) and low-Ps (Ps < 0.5) synapses (Jiang et al. 2001). The activities of high-Ps synapses contribute to the majority of eIPSCs, whereas low-Ps synapses contribute much less to eIPSCs under normal conditions due to their very low release probability. Therefore we examined KA- or ATPA-induced changes in uIPSCs at high-Ps synapses in the presence of 50 μM SYM 2206 and 50 μM APV. To prevent interneurons from over-depolarization-induced inactivation of APs, we maintained interneuronal membrane potential at less than −55 mV (a value that did not affect spontaneous firing) by injecting hyperpolarizing currents during application of ATPA (10 μM) or KA (10 μM).

At perisomatic high-Ps synapses (potency >50 pA, T1/2 <2 ms), ATPA did not reduce the number of successful events (Fig. 2, B and C, ATPA) and did not induce a significant decrease in the Ps of uIPSCs (Fig. 2D, Ps, P = 0.135). The potency and amplitude of uIPSCs were significantly attenuated by ATPA (Fig. 2D, Pot: CON, 103.6 ± 20.1 pA; ATPA, 69.2 ± 15.9 pA; Amp: CON, 82.0 ± 21.6 pA; ATPA, 53.4 ± 17.1 pA, P < 0.05). A decrease of 34.7 ± 3.0% in the control potency of uIPSCs was induced by ATPA application. KA also induced a significant decrease in the potency and amplitude of uIPSCs (Fig. 2E, Pot: CON, 93.0 ± 4.1 pA; KA, 46.0 ± 14.6 pA; Amp: CON, 84.3 ± 9.1 pA; ATPA, 39.7 ± 15.1 pA, P < 0.05) but induced only a marginal decrease in the Ps (Fig. 2E, Ps, P = 0.07). KA induced a decrease of 51.5 ± 15.3% in the control potency of uIPSCs. These results suggest that at perisomatic high-Ps GABAergic synapses, micromolar amounts of KAR agonists induce a significant decrease in the potency of uIPSCs and a minor effect on the release probability. The decrease in the potency of perisomatic uIPSCs implies a postsynaptic mechanism.

KAR agonists induce a decrease in both the Ps and the potency of dendritic high-Ps uIPSCs

Because the amplitude of dendritic uIPSCs was small, a successful uIPSC event was defined as a downward current with decay >20 ms and amplitude >5 pA starting within the mean latency (for a given pair) ± 2 SD (1.1 ms). In dendritic synapses (potency <40 pA, T1/2 ≥ 2 ms), ATPA reduced successful events of uIPSCs at dendritic high-Ps synapses (Fig. 3, B and C, ATPA). The Ps of uIPSCs during application of ATPA (Fig. 3D, Ps, ATPA, P < 0.01) or KA (Fig. 3E, Ps, KA, P < 0.01) was significantly lower than during the control, suggesting that activation of KARs induces a decrease in the release probability at dendritic GABAergic synapses. The potency and amplitude of uIPSCs was also significantly decreased during application of ATPA (Fig. 3D, Pot: CON, 25.4 ± 4.7 pA; ATPA, 18.0 ± 3.4 pA; Amp: CON, 16.9 ± 3.0 pA; ATPA, 5.1 ± 1.3 pA, P < 0.01) or KA (Fig. 3E, Pot: CON, 20.5 ± 4.7 pA; KA, 13.0 ± 4.5 pA; Amp: CON, 14.9 ± 3.2 pA; ATPA, 2.4 ± 0.9 pA, P < 0.01), again implying the involvement of postsynaptic mechanisms.

Micromolar ATPA facilitates low-Ps GABAergic synapses

In previous studies, we found that submicromolar KAR agonists increase the efficacy of low-Ps GABAergic synapses (Jiang et al. 2001). To test whether this facilitating effect is limited to submicromolar concentrations of KAR agonists, we tested the effects of 5 μM ATPA on uIPSCs in low-Ps pairs. At perisomatic low-Ps synapses with large fast uIPSCs, ATPA induced a significant increase in the Ps of uIPSCs (Fig. 4, A and B, ATPA, Ps, P < 0.05). The mean amplitude of uIPSCs were also increased by ATPA (Fig. 4, A and B, ATPA, P < 0.05 for both Pot and Amp), suggesting that micromolar concentrations of KAR agonists have facilitating effects on perisomatic GABAergic synapses that are similar to those of submicromolar quantities (Jiang et al. 2001). In three of six dendritic low-Ps pairs, ATPA induced biphasic responses: a transient increase (∼1 min) in successful events of uIPSCs followed by a subsequent decrease. Statistical data from all six dendritic low-Ps pairs showed no significant changes in the Ps, potency, or mean amplitude (Fig. 4C, ATPA: P = 0.35, 0.78, and 0.41, respectively).

FIG. 4.

Effects of ATPA on uIPSCs at low-Ps synapses. A: ATPA potentiates perisomatic low-Ps GABAergic synapses. Top: traces of uIPSCs evoked by 10 continuous spikes during control (CON), during ATPA application (ATPA), and after washout (Wash) from a representative perisomatic low-Ps pair. Bottom: a plot of the amplitude of uIPSCs against time. B: the Ps, potency (Pot), and mean amplitude (Amp) of uIPSCs during control (□) and ATPA application (▪). *, P < 0.05, compared with control, n = 4 perisomatic low-Ps pairs. C: the Ps, potency (Pot), and mean amplitude (Amp) of uIPSCs at dendritic low-Ps synapses during control (□) and ATPA application (▪). No significant changes in the Ps (P = 0.35), potency (P = 0.78), or mean amplitude (P = 0.41) were found, paired t-test, n = 6 low-Ps pairs.

KAR agonists induce a postsynaptic inhibition of GABAA currents

Application of ATPA or KA induced a decrease in the potency of perisomatic (Fig. 2, D and E, Pot) and dendritic (Fig. 3, D and E, Pot) uIPSCs, possibly by a postsynaptic mechanism. To define the postsynaptic mechanism, we recorded the GABAA-mediated Cl current in pyramidal neurons by locally applying the GABAA receptor agonists GABA (50 μM) or muscimol (50 μM). Hippocampal slices were perfused with TTX (1 μM), SYM 2206 (50 μM), and APV (50 μM) during application of KAR agonists. Pyramidal neurons were first tested by whole cell patch-clamp recording at the soma, and GABAA agonists were applied to the area around the soma. Application of ATPA (10 μM) induced a slow inward current (IKAR) and a decrease in the amplitude of GABAA currents (Fig. 5B, P < 0.01, 1-way ANOVA). The mean amplitude of IKAR in the presence of TTX (Fig. 5, B and C, TTX) was similar to that in the absence of TTX (Fig. 5, A and C, TTX). The decrease in the amplitude of GABAA currents was also similar in the presence or absence of TTX (Fig. 5D, P = 0.89, 2-way ANOVA). These results suggest that activation of postsynaptic KARs attenuates GABAA currents, contributing to KAR agonist-induced decreases in the potency of perisomatic uIPSCs.

FIG. 5.

ATPA induces a somatic IKAR and inhibition of somatic GABAA currents. A: ATPA induces a slow inward current (IKAR) in pyramidal neurons during pair recording. B: ATPA induces IKAR and inhibition of somatic GABAA currents in the presence of TTX (1 μM). C: ATPA-induced IKAR in the presence (▪) and absence (□) of TTX. No statistical difference was found (P = 0.83, unpaired t-test, n = 11 and 16 cells for TTX and CON, respectively). D: normalized amplitude of somatic GABAA currents plotted against time in the presence (•) or absence of TTX (○). n = 7 and 6 cells for TTX and Con, respectively.

To investigate the postsynaptic mechanisms underlying ATPA- or KA-induced depression of dendritic uIPSCs, we performed whole cell patch-clamp recording at a dendritic position 60–80 μm from the soma of pyramidal neurons (Fig. 6C, arrow). GABA (50 μM) was locally applied to the recording position through a puff pipette (Fig. 6C, arrowhead). Application of ATPA (10 μM) induced IKAR and a simultaneous decrease in the amplitude of dendritic GABAA currents (Fig. 6, A and E). KA (10 μM) induced a larger IKAR (Fig. 6, B and D, KA, 206.6 ± 41.9 pA) than ATPA (Fig. 6, A and D, ATPA, 109.4 ± 39.7 pA, P < 0.05, paired t-test). In 2 of 10 tested cells, the ATPA-induced IKAR was undetectable, but the KA-induced IKAR was noticeable. The reversal potential of IKAR was +43.3 ± 5.1 mV (n = 6 cells), implying that KAR channels in pyramidal neurons may be Ca2+-permeable. KA also induced a greater inhibition of dendritic GABAA currents (Fig. 6, B and E, KA) than ATPA (Fig. 6, A and E, ATPA). The maximal inhibition by KA (59 ± 7% of control) was significantly larger than that by ATPA (44 ± 9% of control, P < 0.05, paired t-test). The IKAR and inhibition of GABAA currents were blocked by the AMPA/KAR antagonist CNQX (20 μM), confirming that IKAR current and inhibition are mediated by KAR activation (Fig. 6, D and E, KA/CNQX). The preceding results suggest that postsynaptic KAR-mediated inhibition of dendritic GABAA currents contributes to the ATPA- or KA-induced decrease in the potency of dendritic uIPSCs (Fig. 3).

FIG. 6.

ATPA induces dendritic IKAR and inhibition of dendritic GABAA currents. A, top: ATPA induces dendritic IKAR and inhibition of dendritic GABAA currents in the presence of TTX (1 μM). Bottom: enlarged GABAA currents from indicated period. B: KA-induced dendritic IKAR and inhibition of dendritic GABAA currents. C: a DIC image showing a patch pipette (↓) on an apical dendrite of a pyramidal neuron and a puff pipette (▴) used to apply GABA locally. Bar represents 10 μm. D: the average amplitude of the IKAR induced by ATPA and KA. KA-induced IKAR is blocked by CNQX (KA/CNQX). E: normalized amplitude of dendritic GABAA currents (IGABA) plotted against time. Neither ATPA (○) nor KA (•) significantly decreases the amplitude of dendritic GABAA currents (P < 0.01 for both, 1-way ANOVA). The decrease is blocked by CNQX (KA/CNQX). F: relationship between the change in reversal potential of GABAA currents (VRev) and the decrease in dendritic GABAA currents. R = 0.74, P < 0.05.

To test whether the GABAA current activated by synaptic release of GABA is similarly inhibited by KAR agonists, we examined the effects of ATPA or KA on dendritic mIPSCs while monitoring exogenous GABA-induced currents. Frequency and amplitude of mIPSCs during a 50-s period were calculated for control, ATPA, or KA application and washout. The amplitude of mIPSCs was significantly attenuated by ATPA (Fig. 7, A and B, Amp), but the frequency was not significantly altered (P = 0.50). The decrease in the amplitude of mIPSCs occurred simultaneously with and correlated with inhibition of the exogenous GABA-induced current (Fig. 7C, IGABA, R = 0.84, P < 0.01), suggesting that the decrease in the mIPSC amplitude is due to the postsynaptic KAR-mediated inhibition of the GABAA current. The effect of KA on mIPSCs was similar to that of ATPA. The amplitude of mIPSCs was attenuated by KA (Fig. 7D, Amp, P < 0.01) while the frequency was reduced without statistical significance (P = 0.11). The results suggest that dendritic KARs in pyramidal neurons mediate an inhibition of the GABAA current.

FIG. 7.

A decrease in the amplitude of miniature IPSCs (mIPSCs) simultaneously occurs with the KAR-mediated inhibition of dendritic GABAA currents. A: recording traces from a representative pyramidal neuron showing that the amplitude of mIPSCs is attenuated by ATPA. B: pooled data showing that the amplitude of mIPSCs is decreased by ATPA (Amp, ▪). **, P < 0.01, compared with control, paired t-test, n = 10 cells. C: the change in the amplitude of mIPSCs in each cell is linearly related to the inhibition of GABAA currents. R = 0.84, P < 0.01. D: KA-induced dendritic inhibition of mIPSCs. The mean amplitude and frequency of mIPSCs during control (□) and after KA application (▪). **, P < 0.01, paired t-test, n = 7 cells.

DISCUSSION

Many studies on KA-induced depression of inhibitory transmission have focused on eIPSCs. However, a serious disadvantage is that the firing ability of stimulated fibers is not well controlled. In this study, we have shown that 58% of patched interneurons were unable to fire APs during application of 10 μM KA (Fig. 1). This inactivation of interneuronal APs may be due to depolarization-induced inactivation of voltage-gated Na+ channels and may contribute significantly to KA-induced depression of eIPSCs. Thus firing ability of stimulated fibers should be well controlled when eIPSCs are used to study the depressant effects of KAR agonists. During KA application, attenuated spikes could still induce uIPSCs (Figs. 1B, 2, 3, and 6) while strong depolarization pulses could not trigger uIPSCs successfully when interneuronal spikes were fully inactivated. This suggests that attenuated spikes could still regenerate along the axon. The attenuation of spikes is due to partial inactivation of Na+ channels. The initiation of action potential may occur at axons 30–60 μm from soma (Colbert and Johnston 1996; Colbert and Pan 2002), and the distance between the axonal initiation and terminals may be of the same scale as that between the axonal initiation and somata in interneurons. Thus when we recorded an attenuated spike from the soma, a comparable AP should also propagate to the terminals. When no spike was recorded from soma, the AP at the initiation position was either nonexistent or was too small to regenerate.

With controlled interneuronal APs, we found that micromolar KAR agonists differentially modulated individual GABA synapses. ATPA and KA induced a significant decrease in the potency and amplitude of uIPSCs without significantly changing the Ps at perisomatic high-Ps synapses. Thus ATPA and KA may not significantly affect the presynaptic release of GABA at perisomatic high-Ps synapses. In contrast, at dendritic high-Ps synapses, ATPA and KA induced a decrease in both the Ps and potency of uIPSCs (Fig. 3). The results suggest that KAR-mediated presynaptic inhibition of GABA release at perisomatic GABAergic synapses is weaker than that at dendritic GABAergic synapses. The significant decrease in potency of perisomatic high-Ps uIPSCs during application of ATPA and KA (Fig. 3) indicates a postsynaptic mechanism. ATPA induced a decrease in somatic GABAA currents (Fig. 5, A, B, and D) both in the presence or absence of TTX (Fig. 5D), suggesting that the decrease is not related to the ATPA-induced increase in sIPSCs and that it contributes to a part of the ATPA-induced decrease in the potency of uIPSCs.

At dendritic high-Ps synapses, ATPA and KA induced a decrease in both the Ps and potency of uIPSCs (Fig. 3). Because the amplitude of dendritic uIPSCs is small during control, the further decrease in the amplitude of dendritic uIPSCs could cause postsynaptic failures (i.e., synaptic events become smaller than background levels). Besides previously reported activation of presynaptic GABAB autoreceptors (Frerking et al. 1999; Frerking and Nicoll 2000) or direct inhibition of presynaptic release through the PKC pathway (Rodriguez-Moreno and Lerma 1998; Rodriguez-Moreno et al. 2000), postsynaptic failures due to inhibition of dendritic IPSCs might also have contributed to the KAR agonist-induced decrease in the Ps of dendritic uIPSCs. Moreover, KAR agonist-induced postsynaptic depolarization could trigger release of other transmitters or modulators such as endocannabinoids (Carlson et al. 2002; Hoffman et al. 2003; Wilson et al. 2001) or adenosine (Oliet and Poulain 1999; Scanziani et al. 1992). These transmitters/modulators may inhibit presynaptic GABA release and should be examined in further studies. One postsynaptic mechanisms underlying the KAR agonist-induced decrease in the potency of uIPSCs (Fig. 3C) is the KAR agonist-induced decrease in postsynaptic dendritic GABAA currents, which we demonstrated using dendritic whole cell recordings (Fig. 6). The time course of ATPA-induced decreases in GABAA currents (Fig. 6, A and B, 3–4 min) is similar to that in the potency of uIPSCs (3–4 min). The experimental results in this study suggest that postsynaptic KAR-mediated decreases in GABAA currents contribute to the ATPA- and KA-induced decreases in the potency of dendritic uIPSCs.

The mechanisms underlying KAR-mediated inhibition of GABAA currents are not fully understood. Several mechanisms may contribute to this inhibition. First, KAR activation may trigger a shift of the reversal potential for GABAA currents by modifying intracellular Cl and HCO3 concentrations (Alger and Nicoll 1979, 1982; Andersen et al. 1980; Connors et al. 1988; Grover et al. 1993; Gulledge and Stuart 2003; Kaila et al. 1993, 1997; Lambert et al. 1991; Michelson and Wong 1991; Newberry and Nicoll 1985; Perkins 1999; Staley and Proctor 1999; Staley and Smith 2001; Staley et al. 1995; Taira et al. 1997; Van den Pol 1996; Wong and Watkins 1982). Intracellular Cl concentrations in CNS neurons can be bidirectionally regulated by the K+-Cl co-transporter KCC2 and by the Na+-K+-2Cl cotransporter NKCC1 (Beck et al. 2003; DeFazio et al. 2000; Rivera et al. 1999; Sung et al. 2000; Vu et al. 2000). Activity-dependent modulation of KCC2 and NKCC1 has been reported by several groups (Ganguly et al. 2001; Schomberg et al. 2003; Woodin et al. 2003). Second, ATPA and KA application resulted in a large current at somata and dendrites (Figs. 5 and 6). The ATPA- and KA-induced opening of KAR channels decreased the membrane resistance and caused a shunt effect on GABAA currents. The average peak current of ATPA- and KA-induced IKAR at a holding potential (Vh) of −60 mV was 109 and 207 pA, respectively (Fig. 6D). The reversal potential (Vrev) of IKAR is +43.3 ± 5.1 mV. The increase in membrane conductance (S) can be calculated by the equation: S = I/(VhVrev). The maximal increase in membrane conductance is near 1 and 2 nS for ATPA and KA, respectively. The average membrane conductance in our control dendritic recordings is 5.1 ± 0.3 nS (n = 15 cells), and the maximal membrane conductance during ATPA and KA application is near 6 and 7 nS (control conductance + conductance increase), respectively. The maximal percent change in membrane resistance is near 17 and 30% for ATPA and KA, respectively. Thus at least part of ATPA- and KA-induced decreases in GABAA currents in Fig. 6E (17 and 30% for ATPA and KA, respectively) is due to the postsynaptic KAR channel opening-induced shunt effect. Because uIPSCs were recorded without TTX, increased sIPSC-induced shunt effects (Frerking et al. 1999) and desensitization of GABAA channels may also contribute to the KA-induced changes in the potency of uIPSCs.

The experimental data from recordings of dendritic mIPSCs (Fig. 7) reveal the effects of ATPA and KA on dendritic mIPSCs. The amplitude of dendritic mIPSCs decreased during application of ATPA (Fig. 7, A and B) and KA (D). This decrease is due to the KAR-mediated inhibition of GABAA currents because the decrease in the amplitude of mIPSCs correlates with the inhibition of GABAA currents (Fig. 7C). Moreover, the decrease in mIPSCs occurred simultaneously with the inhibition of GABAA currents. Therefore the ATPA- and KA-induced decrease of dendritic mIPSCs results from a postsynaptic inhibition of GABAA currents rather than from presynaptic inhibition of GABA release (Rodriguez-Moreno et al. 1997, 1998).

In this study, we demonstrated that micromolar concentrations of the KAR agonists KA and ATPA over-depolarized presynaptic interneurons and inactivated spikes, which may have significantly contributed to KA-induced depression of eIPSCs. With controlled interneuronal AP, ATPA (10 μM) and KA (10 μM) decreased the potency of both perisomatic and dendritic high-Ps uIPSCs but only reduced the release probability at dendritic high-Ps GABAergic synapses. Postsynaptic KAR receptors mediate an inhibition of GABAA currents that contributes to the decrease in the potency of uIPSCs.

GRANTS

The work was supported by National Institute of Neurological Disorders and Stroke Grants IR29NS-37349 and RO1NS-39997 and an AES Predoctoral Fellowship.

Footnotes

  • * N. Kang and L. Jiang contributed equally to the article.

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

REFERENCES

View Abstract