Adenosine A2A receptor enhances GABAA-mediated IPSCs in the rat globus pallidus

Authors


Corresponding author
M. Ichimura: Pharmaceutical Research Institute, Kyowa Hakko Kogyo Co. Ltd, 1188 Shimotogari, Nagaizumi, Sunto, Shizuoka 411-8731, Japan.

Abstract

  • 1The actions of adenosine A2A receptor agonists were examined on GABAergic synaptic transmission in the globus pallidus (GP) in rat brain slices using whole-cell patch-clamp recording. GP neurones were characterized into two major groups, type I and type II, according to the degree of time-dependent hyperpolarization-activated inward rectification and the size of input resistance.
  • 2The A2A receptor agonist 2-[p-(2-carboxyethyl)phenethylamino]-5′-N-ethylcarboxamido- adenosine (CGS21680; 0.3-3 μm) enhanced IPSCs evoked by stimulation within the GP. The actions of CGS21680 were blocked by the A2A antagonists (E)-8-(3,4-dimethoxystyryl)-1,3-dipropyl-7-methylxanthine (KF17837) and 4-(2-[7-amino-2-(2-furyl)[1,2,4]triazolo[2,3-a][1,3,5]triazin-5-ylamino]ethyl)phenol (ZM241385).
  • 3The CGS21680-induced increase in IPSCs was associated with a reduction in paired-pulse facilitation. CGS21680 (0.3 μm) increased the frequency of miniature IPSCs (mIPSCs) without affecting mIPSC amplitude. These observations demonstrated that the enhancement of IPSCs in the GP was attributable to presynaptic, but not postsynaptic, A2A receptors.
  • 4The results suggest that A2A receptors in the GP serve to inhibit GP neuronal activity, thereby disinhibiting subthalamic nucleus neurone activity. Thus, the A2A receptor-mediated presynaptic regulation in the GP, together with the A2A receptor-mediated intrastriatal presynaptic control of GABAergic neurotransmission described previously, may play a crucial role in controlling the neuronal functions of basal ganglia. This A2A receptor-mediated presynaptic dual control in the striatopallidal pathway could also afford the mode of action of A2A antagonists for ameliorating the symptoms of Parkinson's disease in an animal model.

The globus pallidus (GP) is the principal component of the basal ganglia circuitry and is considered to perform several critical operations that influence the basal ganglia output (Albin et al. 1989; DeLong 1990; Smith et al. 1998). It receives and relays GABAergic signals from striatal medium spiny neurones (MSNs) to the subthalamic nucleus (STN) and other basal ganglia nuclei. Since selective damage to the GP causes slowness of voluntary movements and involuntary posture (Robertson et al. 1989; Mink & Thach, 1991), the GP may play a critical role in some pathological states. In Parkinson's disease, hypoactivity in the GP may contribute to excessive inhibition of basal ganglia targets leading to akinesia (Albin et al. 1989; Crossman 1989; Delong 1990).

One of the four known receptors for the neuromodulator adenosine, the adenosine A2A receptor, has been demonstrated to have a distinct distribution in the brain, restricted primarily to dopamine-innervated regions such as the striatum, nucleus accumbens and olfactory tubercles on the basis of in situ hybridization studies (Schiffmann et al. 1991; Fink et al. 1992; Augood & Emson, 1994). It has been further revealed that A2A receptor mRNA in the striatum is mainly expressed in striatopallidal MSNs but not in striatonigral MSNs (Fink et al. 1992; Augood & Emson, 1994). Interestingly A2A receptor expression has been detected also in the GP, as determined by ligand binding (Jarvis & Williams, 1989; Parkinson & Fredholm, 1990; Martinez-Mir et al. 1991) and immunohistochemistry (Rosin et al. 1998). However, little is known regarding the physiological and pathophysiological function of A2A receptors in the GP. Very recently, an in vivo microdialysis study has suggested that pallidal A2A receptors regulate the GABA release in the GP (Ochi et al. 2000). We therefore examined the A2A receptor function in the GP in terms of the regulation in the GABAergic synaptic transmission.

We previously reported that the striatal A2A receptor presynaptically inhibits GABAergic transmission in the striatal MSNs (Mori et al. 1996), an effect which could lead to increase striatopallidal output activity and thus inhibition of GP neurones. Here, we find that the A2A receptors in the GP presynaptically enhance GABAergic synaptic transmission. This finding implies that in the GP, there exists an A2A receptor-mediated suppression of the GP neurones, thereby disinhibiting the neuronal activity of STN. This may also give a new insight into the pathophysiological role of the A2A receptor in the GP. Current models of the pathophysiology of Parkinson's disease emphasize an increase in the overall activity in the striatopallidal indirect pathway (Wichmann & DeLong, 1996). Blockade of A2A receptors exert an anti-Parkinsonian activity in animal models of this disease (Kanda et al. 1994, 1998; Shimada et al. 1997; Grondin et al. 1999). We thus propose that the pallidal A2A receptor-mediated regulation, together with the intrastriatal A2A receptor-mediated disinhibitory control, is significant for anti-Parkinsonian activity of A2A receptor antagonists. This adenosinergic presynaptic dual regulation in the indirect pathway has led us to new insights into the role of A2A receptors in the basal ganglia.

METHODS

All experiments were carried out in accordance with the Guiding Principles for the Care and Use of Laboratory Animals as set out by the Japanese Pharmacological Society and approved by the Ethical Committee on Animal Experiments of the Pharmaceutical Research Institute, Kyowa Hakko Kogyo.

Slice preparation

The experiments were performed on male Sprague-Dawley rats (12-16 days postnatal). Animals were anaesthetized with ether and decapitated. The brain was removed rapidly. Slices containing the globus pallidus were cut on a DTK-1000 microslicer (Dosaka, Japan) at a thickness of 250 μm in an oblique plane, about 30 deg rostral-up to the horizontal (Kawaguchi et al. 1989). Slices were then incubated in oxygenated Ringer solution at a temperature of 29-30 °C for 1 h. The standard Ringer solution had the following composition (mm): NaCl, 124; KCl, 3.0; CaCl2, 2.4; MgCl2, 1.2; NaHCO3, 26.0; NaH2PO4, 1.0; and glucose, 10.0; continuously bubbled with a mixture of 95 % O2 and 5 % CO2. After incubation, a single slice was transferred to a recording chamber placed on the stage of an upright microscope, and was continuously perfused (3-4 ml min−1) with oxygenated Ringer solution at 30 °C. The remaining slices were kept in a holding chamber containing oxygenated Ringer solution at room temperature.

Whole-cell recording and data analysis

Neurones in the globus pallidus were visualized using infrared differential interference videomicroscopy (IR-DIC) with a × 40 water immersion objective (Nikon, Tokyo, Japan). Micropipettes for whole-cell recordings were pulled with a P-97 Flaming-Brown electrode puller (Sutter Instrument Company, Novato, CA, USA) from borosilicate glass (1.5 mm outer diameter, Clark Electromedical Instruments, Reading, UK). These had a final resistance of 2-5 MΩ when filled with intracellular solution. For whole-cell current-clamp recordings, micropipettes were filled with a solution containing (mm): potassium methanesulfonate, 115; KCl, 5.0; EGTA, 0.5; MgCl2, 1.7; ATP, 4.0; GTP, 0.3; and Hepes, 8.5; pH 7.2-7.4; osmolarity, 285 mosmol l−1. Current-clamp recordings were made in the bridge mode with an Axoclamp-2B amplifier (Axon Instruments, Foster City, CA, USA). Resting potentials were measured just after the patched membranes were ruptured by suction. Input resistance of cells was determined by passing hyperpolarizing current pulses (duration, 600-900 ms), which induced voltage shifts of 6-15 mV negative to rest.

For whole-cell voltage-clamp recordings, the pipette solution contained (mm): caesium methanesulfonate, 120; KCl, 5.0; EGTA, 10.0; CaCl2, 1.0; MgCl2, 2.0; ATP, 4.0; GTP, 0.3; Hepes 8.0; and QX314, 5.0; pH 7.2-7.4; osmolarity, 295 mosmol l−1. The quaternary lidocaine derivative QX314 was included to suppress fast sodium currents. Voltage-clamp recordings were made with an Axopatch-1D amplifier (Axon Instruments). Throughout voltage recordings, series resistance was monitored via a voltage step (-5 mV, 10 ms), and typical values for the series resistance were 10-20 MΩ. Experiments were discarded if changes over 20 % of the series resistance were seen. The experiments showing no recovery of IPSC amplitude when washout of the drug was omitted. Liquid junction potentials were not corrected.

Evoked synaptic currents were elicited by focal stimulation (0.1 Hz, 200 μs, 10-100 V) via a glass micropipette filled with the extracellular solution and placed within 500 μm of the recorded neurone.

Signals were filtered at 2 kHz and digitized at 5 kHz using an ITC-16 interface (Instrutech, Great Neck, NY, USA) connected to a Power Macintosh computer running Pulse (HEKA Elektronik, Lambrecht, Germany) software. Analysis for evoked IPSCs was performed using Pulse (HEKA) and IGOR Pro (WaveMetrics, Lake Oswego, OR, USA) software. For drug effects on evoked IPSCs, a normalized value for the evoked IPSC peak amplitude under drug application was obtained by averaging the peak amplitude of eight consecutive evoked IPSCs and by subsequently dividing this mean value by the average obtained in the control period.

Spontaneous synaptic currents were automatically analysed using the Mini Analysis Program (Synaptosoft Inc., Leonia, NJ, USA). Events were ranked by amplitude and inter-event interval for preparation of cumulative probability distributions within 200-300 s epochs for control and drug conditions. The cumulative probability distributions were compared by the Kolmogorov-Smirnov test; P < 0.05 was taken as indicating statistical significance.

All data are given as means ±s.e.m., unless stated otherwise. Responses to agonists are expressed as a percentage of the control obtained before the addition of the agonists. The effects of agonists, expressed as percentages, were compared to the effects of agonist-free solution using Steel's test. The effects of antagonists were compared using Wilcoxon's rank sum test. Student's paired t test was performed to compare the raw values of the control with the responses in the presence of agonists applied on the same cell for analysis of paired-pulse facilitation and spontaneous synaptic currents.

Histochemical procedures

In an attempt to identify the recorded cells morphologically, 20 mm biocytin was included in the pipette solution so that they were filled by diffusion (Horikawa & Armstrong, 1988). Slices containing biocytin-loaded cells were fixed by immersion in 4 % paraformaldehyde and 0.2 % picric acid in 0.1 m phosphate buffer (PB) overnight at 4 °C, rinsed in PB for 30 min, and incubated in PB containing 0.5 % H2O2 for 30 min to suppress endogenous peroxidase activity. They were then incubated in 10 and 20 % sucrose for 30 min and 1 h, respectively. The slices, without resectioning, were then washed with Tris-buffered saline (TBS) containing 0.5 % Triton X-100 and avidin-biotin-peroxidase complex (1:100; Vector Laboratories, Burlingame, CA, USA) for 4 h at room temperature. After rinsing, the slices were reacted with 3,3′-diaminobenzidine tetrahydrochloride (DAB; 0.05 %) and H2O2 (0.003 %) in TBS, and mounted on slides.

Drugs

Drugs were applied by replacing the solution superfusing the slice with one containing a set concentration. d-2-Amino-5-phosphonovaleric acid (APV), 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) and ZM241385 were obtained from Tocris Cookson (Bristol, UK); tetrodotoxin (TTX) was from Wako (Tokyo, Japan); CGS21680 was from Research Biochemicals International (Natick, MA, USA); and (-)-bicuculline methiodide was from Sigma (St Louis, MO, USA). (E)-8-(3,4-dimethoxystyryl)-1,3-dipropyl-7-methylxanthine (KF17837) was synthesized at the Medicinal Chemistry Department of the Pharmaceutical Research Laboratories, Kyowa Hakko Kogyo Co. Ltd (Shizuoka, Japan).

RESULTS

Characteristics of GP neurones

Neurones were sampled by whole-cell recording from visually identified neurones in the GP (Fig. 1A and B) and only cells that were located within the GP with resting potentials more negative than -40 mV and overshooting action potentials were used for further analysis. These GP neurones were characterized into two major groups, type I and type II, by the degree of time-dependent, hyperpolarization-activated inward rectification and the size of input resistance, in accordance with previous reports (Nambu & Llinas, 1994, 1997; Stanford & Cooper, 1999). More recent investigation has further confirmed the presence of two major subtypes of GP neurones electrophysiologically and morphologically (Cooper & Stanford, 2000).

Figure 1.

Visual and electrophysiological characterization of GP neurones

A, infrared photograph of a GP neurone in a rat slice preparation. B, micrograph of a biocytin-filled GP neurone that was visualized using histochemical procedures after being recorded in the whole-cell configuration. C, voltage responses and spontaneous discharges of type I neurone in the GP. Type I neurones showed little or slight sag during hyperpolarizing pulses (C1, average steady-state to peak sag ratio of 0.82 ± 0.03 in 8 neurones). Spontaneous action potentials fired at 9.6 Hz (C3). D, voltage responses and spontaneous discharges of type II neurone in the GP. Injection of a hyperpolarizing pulse produced a prominent sag in the membrane potential in type II neurones (D1, average steady-state to peak sag ratio of 0.66 ± 0.01 in 16 neurones, P < 0.005 vs. type I neurones by Wilcoxon's rank sum test). Both type of neurones fired regular spikes from hyperpolarized potentials (C2 and D2). Voltage, time and current calibrations in C1 also apply to C2, D1 and D2. Scale bars: A, 20 μm; B, 100 μm.

Type I neurones (8 of 24 cells) exhibited little or slight sag of membrane potential back toward the resting membrane potential during hyperpolarization (Fig. 1C1). Most of these neurones (7 of 8 cells) fired action potentials spontaneously at mostly regular rates which ranged from 7.4 to 15.3 Hz (mean 10.5 ± 0.9 Hz) at zero holding current (Fig. 1C3). The input resistance of these neurones was 208 ± 14 MΩn = 8. The resting potential of spontaneously firing neurones was difficult to establish because of the absence of a stable membrane potential. On the other hand, type II neurones (16 of 24 cells) displayed a prominent sag in the membrane potential (Fig. 1D1). Type II neurones had a higher input resistance (438 ± 60 MΩ; n= 16) than type I neurones. Thus, type II neurones could be separated from type I neurones by the presence of hyperpolarization-activated inward rectification and the size of input resistance (Stanford & Cooper, 1999). Twelve type II neurones regularly fired spontaneous action potentials at 0.9-14.4 Hz (mean 4.1 ± 1.0 Hz) when no holding current was applied (Fig. 1D3). Four type II neurones did not fire spontaneously and had a resting potential of -56 ± 3 mV n = 4. Both type of neurones fired regular spikes from hyperpolarized potentials (Fig. 1C2 and D2).

GABAergic IPSCs in the GP

Whole-cell voltage-clamp recordings were performed in GP neurones with a Cs+-filled patch pipette. Evoked outward-going currents were elicited by focal stimulation at a holding potential of 0 mV in a solution containing blockers of excitatory transmission (10 μm CNQX and 50 μm APV). These outward currents were completely and reversibly suppressed by the GABAA receptor antagonist bicuculline (10 μm) (Fig. 2A) and reversed polarity at about -50 mV (-50 ± 1 mV, n= 15), close to the equilibrium potential for chloride events calculated from the Nernst equation (ECl= -61 mV) (Fig. 2B). Therefore these evoked synaptic currents were GABAA receptor-mediated IPSCs. Since evoked IPSCs were often observed as a small amplitude when cells were held near their resting membrane potentials, all further data reported in this study were obtained from GP neurones recorded at a holding potential of 0 mV with Cs+-filled electrodes except where specifically indicated.

Figure 2.

Evoked GABAergic postsynaptic currents in GP neurones

Evoked IPSCs were induced by focal stimulation within the GP in GP neurones in a solution containing antagonists for non-NMDA and NMDA receptors (10 μm CNQX and 50 μm APV). A, time course of the amplitude of evoked IPSCs during the application of GABAA receptor antagonist (10 μm bicuculline) (left), and superimposed traces (right) of an average of consecutive evoked IPSCs (5 traces) before (control) and during application of bicuculline and after (wash), taken at the indicated time points. Bicuculline was applied to the superfusion medium for the period indicated by the bar. Membrane potential = 0 mV. B, reversal potentials of GABAergic IPSCs. It reversed polarity at -50 mV, close to the chloride equilibrium potential.

The adenosine A2A receptor-mediated enhancement of evoked IPSCs in the GP

We examined the effect of the adenosine A2A receptor agonist CGS21680 on evoked IPSCs. Evoked IPSCs were recorded at a holding potential of 0 mV in the presence of 10 μm CNQX and 50 μm APV. Bath application of CGS21680 (1 μm) significantly increased the amplitude of evoked IPSCs within 5 min of application of the drug (Fig. 3A). After drug removal, the IPSC amplitude slowly recovered. CGS21680 increased the evoked IPSC amplitude in a concentration-dependent manner (Fig. 4). Moreover, to assess whether this action of CGS21680 was dependent on the type of recorded neurone, we examined the effect of CGS21680 on evoked IPSCs after identifying the membrane properties of the recorded neurones in current-clamp recordings. CGS21680 (1 μm) increased the evoked IPSC amplitude in both type I and type II neurones (2 type I and 4 type II neurones, 133 ± 7 % of control, n= 6).

Figure 3.

CGS21680-induced enhancement of evoked IPSCs in the GP

A, enhancement of evoked IPSCs by an adenosine A2A receptor agonist, CGS21680. Left, the time course of the amplitude of evoked IPSCs during the application of CGS21680 (1 μm); right, superimposed traces of an average of consecutive evoked IPSCs (8 traces) before (control) and during application of CGS21680 (1 μm), taken at the indicated time points. B, effect of the adenosine A2A receptor antagonist KF17837 (0.5 μm) on the CGS21680-induced enhancement of evoked IPSCs. KF17837 was applied to the superfusion medium for the period indicated by the bar. Evoked IPSCs were recorded in the presence of CNQX (10 μm) and APV (50 μm) at a holding potential of 0 mV.

Figure 4.

Summary of pharmacological characterization of the adenosine A2A receptor-mediated GABAergic synaptic transmission

Data are normalized as a percentage of control values. The error bars represent s.e.m. The numbers of cells examined are given in parentheses. *P < 0.05 vs. 0 μm CGS21680 by Steel's test; †P < 0.05, ††P < 0.005 vs. 1 μm CGS21680 by Wilcoxon's rank sum test.

The CGS21680-induced enhancement of evoked IPSCs was suppressed by application of KF17837 (0.5 μm), an adenosine A2A receptor-selective antagonist (113 ± 5 % of control, n= 5, P < 0.05, Figs 3B and 4). ZM241385 (0.5 μm), another antagonist of the adenosine A2A receptor, also blocked the CGS21680-induced enhancement of evoked IPSCs (93 ± 1 % of control, n= 4, Fig. 4). These results indicate the involvement of an adenosine A2A receptor in the regulation of the GABAergic synaptic transmission in the GP. In addition, KF17837 (0.5 μm) alone had no effect on the amplitude of evoked IPSCs (95 ± 3 % of control, n= 5).

Effects of an adenosine A2A agonist on paired-pulse facilitation

To examine whether the regulation of the GABAergic synaptic transmission in the GP was mediated by post- and/or presynaptic A2A receptor, we first examined the effects of CGS21680 on paired-pulse facilitation (PPF), a presynaptic phenomenon. A pair of synaptic responses was elicited with an interval of 50 ms, and the magnitude of increase of the second response relative to the first one was monitored. In the absence of CGS21680, the ratio of the amplitude of the second IPSC to the first (PPF ratio) was 1.97 ± 0.13 n = 8. As expected, application of CGS21680 (1 μm) increased the first IPSCs by 148 ± 7 % of control (n= 8, Fig. 5A). However, CGS21680 significantly reduced the PPF ratio to 1.55 ± 0.07 in the same neurones (P < 0.001, paired t test, Fig. 5B).

Figure 5.

Effects of CGS21680 on PPF

A, reduction of PPF by CGS21680. The upper traces show typical superimposed traces of an average of consecutive IPSCs (8 traces) evoked by paired stimulation (50 ms interval) before (control) and during application of CGS21680 (1 μm). The lower traces are the same as those above, except that the amplitude of the first IPSC recorded in control conditions has been normalized to the first IPSC recorded during CGS21680 application. B, mean PPF ratios in 8 different neurones under control conditions and in the presence of CGS21680 (1 μm). PPF is decreased in the presence of CGS21680. The error bars represent s.e.m.*P < 0.001 vs. control by paired t test.

Effects of an adenosine A2A agonist on mIPSCs and spontaneous IPSCs

To investigate further the locus of action of CGS21680, we analysed the spontaneous mIPSCs in the GP slices. Outward-going spontaneous mIPSCs were observed in the presence of APV (50 μm), CNQX (10 μm) and TTX (0.5 μm, a concentration that blocked sodium channel-dependent action potentials and evoked-postsynaptic currents) at a holding potentials of 0 mV. Under these conditions, mIPSCs in GP neurones were observed to occur at frequencies ranging between 3 and 15 Hz. These events ranged in amplitude from 5 to > 100 pA. Because these outward currents were reversibly and completely suppressed by bicuculline (10 μm), they were mediated by GABAA receptors (Fig. 6).

Figure 6.

Spontaneous miniature GABAergic IPSCs in GP neurones

Spontaneous mIPSCs were recorded in a solution containing CNQX (10 μm), APV (50 μm) and TTX (0.5 μm). A, a continuous trace showing time course of mIPSCs during application of GABAA receptor antagonist (10 μm bicuculline). B, example traces with an expanded time scale from A, taken at the indicated time points. These mIPSCs were recorded at a holding potential of 0 mV.

Typical experimental data in which the effect of CGS21680 (0.3 μm) on mIPSC frequency and amplitude was examined are shown in Fig. 7. The representative consecutive current traces taken before and during CGS21680 application show that CGS21680 caused an increase of synaptic activity (Fig. 7A). The frequency of mIPSCs increased within 5 min after addition of CGS21680 to the bath solution and recovery upon washout was slow, similar to that observed for evoked responses (Fig. 7B). Application of CGS21680 did not change the cumulative distribution of mIPSC amplitudes (Fig. 7C, left), but shifted the cumulative distribution of intervals between successive mIPSCs toward shorter intervals (Fig. 7C, right). On average, CGS21680 increased the mean mIPSC frequency from 7.4 ± 0.9 to 9.8 ± 1.4 Hz (130 ± 4 % of control, P < 0.01 by paired t test; n= 9), whereas the mean amplitude was not changed (mean amplitude in the presence of CGS21680: 29.76 ± 2.36 pA, 104 ± 1 % of control, n= 9) (Fig. 8).

Figure 7.

Analysis of spontaneous miniature IPSCs in the GP neurones

The data are taken from the same neurone. A, two consecutive traces of mIPSCs before (control) and during application of CGS21680 (0.3 μm). B, time course of the frequency of mIPSCs during application of CGS21680 (0.3 μm). C, cumulative probability distribution of mIPSC amplitude (left) and inter-event interval (right) before (1255 events) and during application of CGS21680 (1802 events). CGS21680 had no effect on the amplitude distribution (P > 0.2, Kolmogorov-Smirnov test for control vs. CGS21680), but shifted the frequency distribution to shorter inter-event intervals (P < 0.001, Kolmogorov-Smirnov test for control vs. CGS21680). The frequencies and amplitudes (mean ±s.d.) of mIPSCs were 4.59 Hz and 38.75 ± 24.08 pA in control and 6.01 Hz and 37.72 ± 23.14 pA during CGS21680 application, respectively.

Figure 8.

Summary graph of the experiments which tested the effect of CGS21680 on mIPSC frequency and amplitude

Pooled data of 9 neurones show that CGS21680 (0.3 μm) increased the mean frequency without affecting the mean amplitude of mIPSCs. Mean frequency ( inline image) and mean amplitude (□) of mIPSC were 7.4 ± 0.9 Hz and 29.76 ± 2.36 pA before (control) and 9.8 ± 1.4 Hz and 30.75 ± 2.39 pA during application of CGS21680, respectively. CGS21680 increased the frequency of mIPSCs by 130 ± 4 %. The error bars represent s.e.m.*P < 0.01 vs. control by paired t test.

CGS21680 (0.3 μm) caused a moderate increase in the frequency of spontaneous IPSCs from 17.6 ± 4.1 to 19.5 ± 4.1 Hz (116 ± 7 % of control, P < 0.05 by paired t test; n= 7), whereas the mean amplitude was not changed (mean amplitude in the presence of CGS21680: 31.01 ± 3.45 pA, 101 ± 4 % of control, n= 7) in the presence of APV (50 μm) and CNQX (10 μm). The potentiating effect of CGS21680 on the frequency of spontaneous IPSCs was significantly less than that of mIPSCs (P < 0.05 by Wilcoxon's rank sum test).

DISCUSSION

Presynaptic regulation via adenosine A2A receptors for GABAergic synaptic transmission in the GP

The CGS21680-induced enhancement of GABAergic synaptic transmission in the GP was caused via adenosine A2A receptors since CGS21680 is an adenosine A2A receptor-selective agonist (Jarvis et al. 1989) and the enhancement was blocked by the adenosine A2A receptor-selective antagonists KF17837 (Nonaka et al. 1994) and ZM241385 (Poucher et al. 1995) (Figs 3 and 4). We examined the CGS21680-induced enhancement of IPSCs using rat GP slices (12-16 days). The physiological responses to current injection and synaptic stimulation and the morphology of such juvenile rat GP neurones are very similar to those of adult rat GP (Ogura & Kita, 2000) and the binding of [3H]CGS21680 is saturated in rat GP at 10-15 days postnatal (Johansson et al. 1997). To investigate the locus of the A2A receptor-mediated regulation, we have examined the effects of CGS21680 on the PPF, a presynaptic process (Isaacson & Walmsley 1995), and analysed the mIPSCs. The enhancement of evoked IPSCs induced by CGS21680 was accompanied by a reduction in the PPF ratio (Fig. 5). CGS21680 also increased the frequency of mIPSCs without affecting their amplitude distribution (Figs 7 and 8). These observations revealed that the CGS21680-mediated enhancement of GABAergic synaptic transmission was due to an increased probability of presynaptic transmitter release and was not due to an increase in postsynaptic receptor sensitivity. Although facilitatory effects of adenosine via A2A receptors are well known in excitatory synaptic transmission (Sebastiao & Ribeiro, 1996, for a review), the enhancement of inhibitory GABAergic synaptic transmission was directly demonstrated for the first time in this study. In the GP, neurochemical studies have shown that A2A receptors increase electrically stimulated GABA release in rat GP slices (Mayfield et al. 1993). More recently, microdialysis data have demonstrated that intrapallidal infusion of CGS21680 increases GABA levels in the rat GP in vivo (Ochi et al. 2000). These results are consistent with the present finding that presynaptic A2A receptors serve to enhance GABAergic synaptic transmission in the GP. In addition, the selective A2A receptor antagonist KW-6002, a derivative of KF17837, is shown to cause a decrease of extracellular GABA levels in the GP of nigrostriatal dopamine-lesioned rats, but has little effect on that of non-lesioned rats in vivo (Ochi et al. 2000). In the present study, KF17837 had little effect by itself on IPSCs. One possibility is that the endogenous adenosine concentration at the A2A receptor sites in normal rat GP might be insufficient to be detectable to an A2A receptor antagonist action. Although the study using synaptosome preparations of the rat GP showed a different effect of A2A receptors from our present data on GABA release (Kurokawa et al. 1994), it may be attributed to differences in methodologies and/or experimental preparations. Further investigation is necessary to resolve this point.

Electrophysiological (Kita et al. 1983; Nakanishi et al. 1987) and immunohistochemical (Oertel & Mugnaini, 1984; Smith et al. 1987) studies indicated that GP projection neurones are GABAergic. The GP neurones give rise to widespread intranuclear axon collaterals (Kita & Kitai, 1994; Nambu & Llinas, 1997). The GP also receives massive GABAergic inputs from striatum. Therefore, the recorded IPSCs in the present study are attributable to activation of GABAergic inputs from striatum and/or activation of local axon collaterals of GP neurones, i.e. the A2A receptor-mediated presynaptic modulation could be caused by an action on striatopallidal terminals and/or terminals of axon collaterals of GP neurones. A recent immunohistochemical study indicated that A2A receptors exist in the GP (Rosin et al. 1998), although A2A receptor mRNA has been detected in striatopallidal neurones, but not in the GP (Schiffmann et al. 1991; Fink et al. 1992; Augood & Emson, 1994). Thus, A2A receptors in the GP are most likely to be located on the axonal arborizations of striatopallidal neurones. In addition our results showed that CGS21680 was more effective in increasing the frequency of mIPSCs than that of spontaneous IPSCs. The former are elicited by spontaneous release from any GABAergic terminals (including those of striatopallidal neurones), whereas the latter depend on action potential firing by spontaneously active GABA-containing neurones. Most GP neurones were found to be spontaneously active (19 of 24 neurones, 79 %) in this slice preparation, whereas striatopallidal neurones are generally silent in such preparations (Wilson, 1993). These results might further support the idea that most of the A2A receptor-mediated effects are exerted on the striatopallidal terminals, although there is no direct physiological evidence about the source of the GABAergic inputs to pallidal neurones that are being modulated. Because striatopallidal axons terminate mainly on the dendritic shafts of GP neurones (Falls et al. 1983; Okoyama et al. 1987) while the collateral axons terminate on the somata and proximal dendrites of GP neurones (Kita 1994), not only locally evoked IPSCs but also spontaneous IPSCs and mIPSCs may originate from the axon collaterals of GP neurones (Ogura & Kita, 2000). We need to resolve this issue in further studies.

Physiological implications of adenosine A2A receptor-mediated modulation in the indirect pathway

In the striatum, presynaptic adenosine A2A receptors have been demonstrated to suppress GABAergic inhibitory synaptic transmission in the striatal MSNs on the basis of electrophysiological (Mori et al. 1996; Chergui et al. 2000) and neurochemical data (Kirk & Richardson, 1994; Kurokawa et al. 1994). This striatal A2A receptor-mediated disinhibition could cause an overall increase in the activity of the striatopallidal projection neurones, resulting in suppression of GP neurone activity (Richardson et al. 1997). This has recently been demonstrated also by in vivo microdialysis experiments in which the adenosine receptor stimulation in the striatum increases GABA release in the GP (Ochi et al. 2000). The pallidal presynaptic A2A receptor-mediated enhancement of GABA release in the present findings could cause further suppression of GP neurones. The suppression of GP neurone activity via A2A receptors would consequently lead to disinhibition of the STN neurone activity. This presynaptic regulation mechanism may work very efficiently in terms of a high degree of anatomical convergence or funnelling in the striatum-GP system. Thus, adenosine may significantly contribute to the control of the basal ganglia system through the A2A receptor-mediated presynaptic dual regulation in the indirect pathway.

With regard to the A2A receptor-mediated regulation in the striatum and GP, it would be very interesting if the presynaptic regulation by the A2A receptor within the same striatopallidal neurones were inhibitory in the striatum and stimulatory in the GP. It remains to be identified which neurones within the striatum and within the GP really contribute to the A2A receptor-mediated regulation of GABA release.

It has been suggested that an increase in GP GABAergic synaptic transmission is a critical consequence of dopamine depletion (Chesselet & Delfs, 1996, for a review). The suppression of the neuronal activity of GP through the A2A receptor-mediated presynaptic dual control could afford an action mechanism of A2A antagonists in Parkinsonian states. Indeed the A2A receptor-selective antagonist KW-6002 (Shimada et al. 1997), a derivative of KF17837 (Nonaka et al. 1994), causes a marked decrease of extracellular GABA levels in the GP of nigrostriatal dopamine-lesioned rats, while having little effect on that of non-leisioned rats (Ochi et al. 2000). This could therefore explain the ability of the A2A receptor antagonists KF17837 or KW-6002 to exhibit a prominent anti-Parkinsonian activity in mice (Kanda et al. 1994; Shimada et al. 1997) and 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-treated monkeys (Kanda et al. 1998; Grondin et al. 1999). Thus, the action of A2A receptor antagonists in the GP, together with its blockade of the intrastriatal disinhibition by adenosine, affords a novel mode of action of A2A receptor antagonists in ameliorating the symptoms of Parkinson's disease. Interestingly Plenz & Kitai (1999) recently proposed that STN and GP constitute a central pacemaker of the basal ganglia, and that synchronized oscillatory activity of this STN-GP circuitry could be linked to basal ganglia dysfunction resulting from striatal dopamine deficiency. Since adenosine turns out to give a dual control on the STN-GP circuitry through the A2A receptors in the striatum and the GP, the adenosine A2A receptor in the indirect pathway may be important in the generation of synchronized oscillatory bursting in pathological conditions such as Parkinson's disease.

It was recently reported that an A2A receptor-selective antagonist showed the neuroprotective effect against kainate-induced excitability in the hippocampus (Jones et al. 1998). In this context, it is worth exploring whether the A2A antagonist action in the indirect pathway might originate a neuroprotective activity in the basal ganglia.

In conclusion, the data in this study demonstrate that the A2A receptor enhances GABAergic synaptic transmission in the GP. Adenosine could control the basal ganglia function by acting on the indirect pathway via both striatal and pallidal presynaptic A2A receptors. The action of A2A receptor antagonists in the GP, together with its blockade of the intrastriatal disinhibition by adenosine, affords a novel mode of action of A2A receptor antagonists in producing efficient anti-Parkinsonian effects.

Acknowledgements

We thank P. J. Richardson for critical comments on this manuscript.

Ancillary