Subthalamic Neurons Coordinate Basal Ganglia Function through Differential Neural Pathways

Introduction

The subthalamic nucleus (STN) is a component in the basal ganglia circuitry that mediates a variety of motor functions (Alexander and Crutcher, 1990; DeLong, 1990; Parent and Hazrati, 1995; Gerfen and Wilson, 1996). Accumulating evidence indicates that the STN plays important roles in movement control, behavioral performance, and motivational process (Hamada and DeLong, 1992a; Baunez et al., 1995, 2002; Baunez and Robbins, 1997; Dybdal and Gale, 2000). In addition, modulation of the STN activity is implicated in the motor symptoms in Parkinson's disease and in animal models with experimentally induced parkinsonism (Bergman et al., 1994; Hassani et al., 1996; Magnin et al., 2000). Lesion or high-frequency stimulation of the STN improves these motor symptoms (Bergman et al., 1990; Benazzouz et al., 1993; Witchmann et al., 1994; Limousin et al., 1995; Levy et al., 2000). The STN is thus a promising target for the surgical therapy of Parkinson's disease.

The output structures of the basal ganglia, including the substantia nigra pars reticulata (SNr) and entopeduncular nucleus (EPN), act as the interface of motor functions through their GABAergic inhibition of target brain regions. The STN neurons receive glutamatergic inputs from many cortical areas and GABAergic inputs from the globus pallidus (GP) (Smith et al., 1990; Bevan et al., 1997; Maurice et al., 1998; Nambu et al., 2000). These neurons send glutamatergic projections to the SNr and the EPN and also innervate the GP through their axonal collaterals (Kita et al., 1983; Kita and Kitai, 1987; Robledo and Féger, 1990; Nakanishi et al., 1991; Sato et al., 2000b). The GP neurons provide GABAergic inhibition to the SNr and the EPN (Bevan et al., 1997; Sato et al., 2000a). A standard model of the basal ganglia circuitry suggests that the direct projection from the STN to the output structures plays a major role in the regulation of motor functions (Alexander and Crutcher, 1990; DeLong, 1990; Parent and Hazrati, 1995; Gerfen and Wilson, 1996). In addition, recent evidence suggests that reciprocal connections between the STN and the GP generate network rhythmicity (Plenz and Kitai, 1999) and that changes in the rhythmicity may be involved in normal movement and pathological feature of movement disorders (Bevan et al., 2002). The mechanism by which the STN neurons regulate motor functions in the basal ganglia circuitry, however, is not fully understood.

In the present study, we perform conditional ablation of the STN neurons by using immunotoxin-mediated cell targeting (IMCT), which is a transgenic technique to eliminate specific cell types from a complex neural circuitry (Kobayashi et al., 1995; Watanabe et al., 1998; Sano et al., 2003). Then we examine the motor behavior dependent on dopamine (DA) transmission and analyze the firing activity of neurons that mediate the behavior in the basal ganglia circuitry. Our results provide evidence that the STN neurons act to coordinate basal ganglia function through differential neural pathways dependent on the state of DA transmission.

Materials and Methods

Generation of mutant mice. In the strategy of IMCT (Kobayashi et al., 1995), transgenic (Tg) mice are generated that express the human interleukin-2 receptor α-subunit (IL-2Rα) under the control of a cell type-specific promoter. These mice are then treated with a recombinant immunotoxin (IT), which is composed of the variable regions of the anti-IL-2Rα monoclonal antibody and a bacterial exotoxin fragment. The transgene construct contained a 10 kb DNA fragment encoding the 5′-flanking region of the mouse neuropsin (NP) gene (Hirata et al., 2001), the second intron of the rabbit β-globin gene (Kobayashi et al., 1992), the gene cassette encoding IL-2Rα fused to green fluorescent protein (IL-2Rα/GFP) (Watanabe et al., 1998), and the polyadenylation signals of the rabbit β-globin gene and simian virus 40 early gene (Kobayashi et al., 1992). The construct was microinjected into fertilized mouse eggs, which were then implanted into pseudopregnant females. Tg mice were identified by Southern blot hybridization or PCR with genomic DNA prepared from tail clips. We produced six independent Tg lines, in which the high-level expression of the transgene was observed in the STN neurons. In the present study, we used one Tg line termed NP-IL-2Rα/GFP3-5. Animal care and handling procedures were in accordance with the guidelines established by the Experimental Animal Center of Fukushima Medical University.

Intracranial injection. Mice (12-16 weeks old) were anesthetized with sodium pentobarbital (50 mg/kg, i.p.) and subjected to unilateral or bilateral treatment. For IT injection, anti-Tac(Fv)-PE38 (Kreitman et al., 1994) was diluted into a final concentration of 10 μg/ml in PBS containing 0.1% mouse albumin. The IT solution or PBS (0.3 μl/site) was injected into the STN through a glass microinjection capillary, which was surgically introduced by using the following coordinates (in mm): anteroposterior, -2.10; mediolateral, 1.50; and dorsoventral, 4.60 from bregma and dura according to an atlas of the mouse brain (Paxinos and Franklin, 2001). Injection was performed at a constant velocity of 0.1 μl/min with a microinfusion pump. In all experiments conducted in this study, the mice were used on day 7 after the IT injection (day 0). After the experiments, the brains were taken, and their sections were stained with neutral red. This staining confirmed the loss of the STN neurons in almost all of the IT-injected Tg mice.

For intrapallidal infusion of 2,3-dioxo-6-nitro-1,2,3,4-tetrahydro [f]quinoxaline-7-sulfonamide (NBQX), solution (0.25 μl/site) was injected into the GP through a 33 gauge stainless steel cannula by using the following coordinates (in mm): anteroposterior, -0.34; mediolateral, 1.80; and dorsoventral, 3.10 from bregma and dura. Administration was performed with a microinfusion pump. After the experiments, the placement of the tips of the cannula was confirmed by histological examination with neutral red staining.

Histological analysis. Mice were anesthetized with sodium pentobarbital and perfused transcardially with 4% paraformaldehyde in 0.1 m phosphate buffer (PB), pH 7.4. Fixed brains were cut into sections (30 μm thick) through the coronal plane with a cryostat. Free-floating sections were stained with neutral red. For immunohistochemistry (Sano et al., 2003), sections were incubated with primary antibodies against neuron-specific nuclear protein (NeuN) (1:500 dilution; Chemicon, Temecula, CA), glial fibrillary acidic protein (GFAP) (1:500 dilution; Lipshaw-Immunon, Pittsburgh, PA), tyrosine hydroxylase (TH) (0.4 μg/ml; Roche, Basel, Switzerland), and c-fos (1:5000 dilution; Santa Cruz Biotechnology, Santa Cruz, CA). The sections were incubated with biotinylated secondary antibodies at 1:1000 dilution and visualized with a Vectastain Elite ABC kit (Vector Laboratories, Burlingame, CA).

For cell counting, sections through the STN at the anteroposterior coordinates from bregma between -1.70 and -2.18 mm were stained with neutral red. Sections through the GP at the anteroposterior coordinates from bregma between -1.70 and -2.18 mm were immunostained with anti-c-fos antibody. Fifteen to 20 sections from each animal were used. The number of stained cells was counted with a computer-assisted imaging program (NIH Image 1.62; National Institutes of Health, Bethesda, MD). The average of the values obtained from each animal was calculated.

For in situ hybridization (Sano et al., 2003), fresh-frozen brain sections (10 μm thick) were fixed in a solution of 4% paraformaldehyde in 0.1 m PB and treated with 0.1 m triethanolamine, pH 8.0, containing 0.25% acetic anhydride. The sections were hybridized with antisense RNA probes for mouse DA D₁ or D₂ receptors labeled by using in vitro transcription with digoxigenin-11-UTP (Roche). The hybridized signals were visualized with a nonradioactive detection system using anti-digoxigenin Fab fragments conjugated to alkaline phosphatase (Roche).

Neurochemical analysis. Contents of DA and its metabolites 3,4-dihydroxy-phenylacetic acid (DOPAC) and homovanillic acid (HVA) were measured by using a HPLC equipped with an electrochemical detection system (Kobayashi et al., 1994). Tissues were homogenized in 5 vol of 0.2 m perchloric acid containing 0.1 m EDTA and 100 ng/ml isoproterenol. The mobile phase was 0.1 m sodium citrate, 0.1 m citric acid, 0.5 mm sodium octanesulfonate, 0.15 mm EDTA, and 12% methanol, pH 3.5.

Behavioral analysis. Locomotor activity was measured with a movement analyzer equipped with photobeam sensors (Nishii et al., 1998). The number of beam breaks was counted for every 10 min period (movement score). The total number of beam breaks in a 30 min test period was calculated to evaluate spontaneous locomotor activity during the pretreatment (-40 to -10 min) and drug-induced locomotor activity after drug treatment (10-40 min).

For rotational behavior, mice were placed in a spherical bowl (25 cm diameter), and their behavior was monitored with a digital video camera. One rotation was defined as a complete 360° turn. The total number of rotations in a 30 min test period was counted to evaluate spontaneous rotations during the pretreatment (-40 to -10 min) and drug-induced rotations after drug treatment (20-50 min).

Electrophysiological recording. Extracellular single-unit recording was performed according to the method used for recording of rat brain under the awake head-restraint condition (Koyama et al., 2003) with some modifications. Mice were anesthetized with sodium pentobarbital and surgically implanted with a U-shaped plastic plate on their skull. Four days after the surgery, the mice were placed in a plastic chamber, and the U-shaped plastic plate was fixed to the stereotaxic apparatus. Single-unit activity was recorded through a glass pipette microelectrode (10-12 MΩ at 1 kHz), amplified, and monitored with a digital oscilloscope. During single-unit recording, electroencephalograms were monitored to confirm the awake status of the mice. The recording data were processed by a computer system equipped with the Spike2 software (Cambridge Electronic Design, Cambridge, UK). The recording coordinates from the bregma and dura were as follows (in mm): SNr, anteroposterior, -3.80 to -2.80; mediolateral, 1.00-1.75; and dorsoventral, 3.80-5.30; GP, anteroposterior, -0.70 to -0.10; mediolateral, 1.50-2.10; and dorsoventral, 3.30-4.80. The placement sites of microelectrodes were verified by Pontamine Sky Blue deposit counterstained with neutral red.

For off-line analysis of spontaneous firing activity, the interspike interval (ISI) histogram, autocorrelogram, and mean firing rate were calculated from at least 500 spikes or 2 min duration of units. Each unit was classified into different firing types according to the ISI histogram and autocorrelogram as described previously (Ni et al., 2000). In brief, the regular type was characterized by a symmetrical distribution of ISI and an auto-correlogram showing at least three identifiable peaks. The irregular type was characterized by an asymmetrical distribution of ISI and an autocorrelogram with less than two peaks. The bursty type was characterized by the appearance of bursts separated by the absence of discharge or by low-frequency tonic activity, showing an asymmetrical or random distribution of ISI and an autocorrelogram without peak or with a single initial peak. Classification of the bursty firing was confirmed by using the computer algorithms described by Kaneoke and Vitek (1996).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Conditional ablation of STN neurons in NP-IL-2Rα/GFP Tg mice. A, Fluorescence microscopic image of coronal sections through the STN in the Tg mice indicating expression of IL-2Rα/GFP fusion gene in these neurons. B, Schematic illustration of the coordinate for the intrasubthalamic injection (-2.10 mm from bregma in the anteroposterior plane). Filled circle indicates the injection site. C, Neutral red staining of sections through the STN. Unilaterally injected mice (n = 5 for non-Tg mice; n = 6 for Tg mice) were used for the staining. The number of cells in the STN was counted by using a computer-assisted imaging program. ^*p < 0.001, significant difference from the IT-injected non-Tg mice. Representative light microscopic images of the IT-injected side are indicated, and bottom images are 16-fold magnified views of the STN (boxes in the top images). D, Neutral red staining of sections through the striatum, GP, ventral midbrain, and thalamus. Unilaterally IT-injected mice were used for the staining, and light microscopic images of the injected side are indicated. SNc, Substantia nigra pars compacta; VA, ventral anterior thalamic nucleus; VL, ventrolateral thalamic nucleus. E, Immunostaining for NeuN and GFAP of sections through the STN prepared from unilaterally IT-injected mice. Scale bars: A, E, 100 μm; C, 400 μm; D, 200 μm.

Statistical analysis. The ANOVA and the post hoc Tukey's test were used for statistical comparisons. All values were expressed as the mean ± SEM of the data.

Results

Conditional ablation of STN neurons by IMCT

We produced Tg mice that carry the IL-2Rα/GFP gene cassette downstream of the 5′-flanking region of the gene encoding NP, which is a serine protease expressed in the brain (Hirata et al., 2001). The NP-IL-2Rα/GFP mice expressed the fusion protein in the majority of neurons in the STN (Fig. 1A). We unilaterally injected IT solution or PBS into the STN by using a stereotaxic approach (Fig. 1B). The injected mice were subjected to brain sectioning, and the sections through the STN were stained with neutral red (Fig. 1C). On the IT-injected side, the number of the STN neurons in the Tg mice was markedly decreased to 35% of that in the non-Tg mice (p < 0.001), whereas the cell number on the PBS-injected side was not significantly different between the two kinds of mice. The sections through the striatum, GP, ventral midbrain, and thalamus were also stained with neutral red, and there was no apparent damage to these brain regions on the IT-injected side of the Tg mice (Fig. 1D). In addition, we performed immunohistochemistry for the neuronal marker NeuN and the glial marker GFAP. The number of NeuN-positive cells in the STN of the Tg mice was remarkably decreased relative to that of the non-Tg mice (Fig. 1E). There were few GFAP-positive signals in the STN of the non-Tg and Tg mice (Fig. 1E). These data indicate a selective and efficient elimination of the STN neurons in the Tg mice by the IT injection.

To examine the influence of the STN ablation on the midbrain DA system, we performed immunostaining with sections through the midbrain or the striatum prepared from unilaterally injected mice. The staining pattern of TH-positive neurons and intensity of their immunoreactive signals in the ventral midbrain exhibited no visible changes on the IT-injected side in the Tg mice compared with those for the PBS-injected Tg mice or the IT- or PBS-injected non-Tg mice (Fig. 2A). The distribution of TH-positive fibers in the striatum on the IT-injected side of the Tg mice was also unaffected (Fig. 2A). In addition, we determined the contents of DA and its metabolites (DOPAC and HVA) in the striatum of unilaterally injected mice (Fig. 2B). Statistical analysis displayed no significant difference in the contents of DA, DOPAC, or HVA between the two injection sides or the two genotypes. Furthermore, we analyzed the gross distribution of DA receptors in the striatum by using in situ hybridization for DA D₁ or D₂ receptor mRNA. Patterns of hybridization signals indicated apparently normal distribution of these two receptors in the striatum on the IT-injected side of the Tg mice (Fig. 2C).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

STN ablation does not influence midbrain DA system and DA receptor distribution. A, TH immunostaining of sections through the midbrain or the striatum prepared from unilaterally injected mice. The substantia nigra pars compacta (SNc) and the ventral tegmental area (VTA) in the ventral midbrain are indicated. The dorsal striatum (St) and the nucleus accumbens (NAc) are indicated. B, Contents of DA, DOPAC, and HVA in the striatum dissected from unilaterally injected mice (n = 10). C, In situ hybridization analysis for DA D₁ and D₂ receptor mRNAs in the striatum. Sections prepared from unilaterally injected mice were used for in situ hybridization with digoxigenin-labeled antisense RNA probes. Scale bars, 200 μm.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

STN ablation causes abnormality in motor behavior. A, B, Time course of spontaneous and METH-induced locomotion. Mice (n = 6) were bilaterally injected with IT solution (A) or PBS (B) for the behavioral test. After monitoring of spontaneous movement, the mice were administered METH (4.0 mg/kg, s.c.). Movement score for every 10 min was counted. ^*p < 0.05, significant differences from non-Tg mice. C, D, Locomotor activity. The total number of beam breaks in a 30 min test period was calculated to evaluate spontaneous locomotor activity during the pretreatment (-40 to -10 min) and METH-induced locomotor activity after treatment (10-40 min). Locomotor activity of IT-injected (C) and PBS-injected (D) mice is shown. ^*p < 0.01, significant differences from non-Tg mice. E, F, Rotational behavior. Mice (n = 6) were unilaterally injected with IT solution for the behavioral test. After monitoring of spontaneous rotations, the mice were administered METH (4.0 mg/kg, s.c.) (E) or apomorphine (2.0 mg/kg, s.c.) (F). The total number of contralateral or ipsilateral rotations in a 30 min test period was calculated to evaluate spontaneous rotations during the pretreatment (-40 to -10 min) and drug-induced rotations after drug administration (20-50 min). ^*p < 0.05, significant differences from non-Tg mice.

STN ablation alters spontaneous firing of the output neurons

To examine the influence of STN ablation on the firing activity of the output neurons, we performed extracellular single-unit recording of the activity of the SNr neurons. Previous electrophysiological studies show that these neurons are divided into several types with different firing patterns (Burbaud et al., 1995; Murer et al., 1997). We unilaterally injected PBS or IT into the STN and recorded the firing activity of units by using microelectrodes placed within the SNr (Fig. 4A). To evaluate the pattern of spontaneous firing, we analyzed the ISI histogram and the autocorrelogram of each unit and categorized the units into three different types: regular, irregular, and bursty. Typical firing patterns obtained from the IT-injected side are shown in Figure 5. The proportion of units belonging to each type on the PBS-injected side was indistinguishable between the non-Tg and Tg mice, whereas the proportion on the IT-injected side was different between the two genotypes (p < 0.001) (Fig. 4B). On the IT-injected side of the Tg mice, the proportion of the regular type was reduced, whereas that of the bursty type was increased. In addition, the mean firing rate of each type on the PBS-injected side was indistinguishable between the two genotypes, whereas the rate of the regular type on the IT-injected side was markedly reduced in the Tg mice relative to that of the non-Tg mice (p < 0.05), and the rate of other types tended to decrease in the Tg mice (Fig. 4C). On the IT-injected side, the mean firing rate of the total units in the Tg mice was 62% of that in the non-Tg mice (p < 0.05) (Fig. 4C). Therefore, the STN ablation shifted the pattern of spontaneous firing of the SNr neurons and reduced the firing rate of these neurons, although the rate reduction was evident in the regular-type neurons.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

STN ablation alters spontaneous firing of the SNr neurons. Unilaterally injected mice were used for extracellular single-unit recording of firing activity. A, Schematic illustration of representative microelectrode placement sites within the SNr. Number indicates the distance (in millimeters) from bregma in the anteroposterior plane. B, Proportion of the three firing types. Number of units recorded from four to eight mice is shown in parentheses. C, Spontaneous mean firing rate of the three firing types and total units on the PBS- or IT-injected side. ^*p< 0.05, significant difference from non-Tg mice.

Pharmacological blockade of the GP activation inhibits DA response of the output neuron activity and motor behavior

The results obtained from single-unit recording of METH-induced firing activity of the SNr and GP neurons suggest that ablation of the STN neurons, through the blockade of drug-induced activation of the GP neurons, attenuated the suppression of the SNr activity. If this possibility is indeed the case, a pharmacological blockade of glutamatergic excitation of the GP neurons would be expected to inhibit METH-induced suppression of the SNr neurons. We thus tested the efficacy of intrapallidal treatment with an AMPA receptor antagonist, NBQX, on METH-induced firing activity of these neurons. C57BL/6J inbred mice received intrapallidal treatment with NBQX (125 ng/site) or PBS and administered METH before single-unit recording of the SNr neurons. As shown in Figure 9A, NBQX inhibited METH-induced suppression of the firing rate of the SNr neurons, and this inhibition was evident at the late phase (p < 0.05; p < 0.01). To confirm whether the intrapallidal treatment with NBQX altered GP activation in response to METH, we performed immunostaining for the immediate-early gene c-fos, which is a marker of cellular responsiveness (Cenci et al., 1992; Ruskin and Marshall, 1995). The inbred mice were treated intrapallidally with NBQX or PBS and administered METH before brain sectioning. The sections through the GP were immunostained with anti-c-fos antibody (Fig. 9B). Many c-fos-positive cells were observed in the GP in the PBS-treated group, whereas the number of immunopositive cells in the NBQX-treated group was markedly decreased relative to that in the PBS-treated group (58%; p < 0.05). The data indicate that the intrapallidal treatment with NBQX indeed inhibited METH-induced activation of the GP neurons. These results support the view that the STN ablation influences METH-induced suppression of the SNr activity by blockading GP activation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Firing pattern of the SNr neurons. The digital raster display, ISI histogram, and autocorrelogram of the typical firing types (regular, irregular, and bursty) obtained from the IT-injected side of non-Tg and Tg mice are shown. The mean firing rate (Fr; spikes per second) of the unit is presented in the ISI histogram.

In addition, we explored whether the blockade of glutamatergic GP activation could diminish METH-induced hyperactivity in movement. Mice were treated intrapallidally with different doses of NBQX and administered METH for locomotion analysis (Fig. 9C). METH-induced locomotor activity in the NBQX-treated group was dose dependently decreased relative to that in the PBS-treated group (p < 0.05), and decreases were 33% at the lower dose (25 ng/site) and 42% at the higher dose (125 ng/site) of the control value. However, locomotor activity in the absence of METH administration was not altered by intrapallidal NBQX treatment (data not shown). These results indicate that the pharmacological blockade of the GP activity reduced the METH-induced hyperactivity.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

STN ablation does not change spontaneous firing of the GP neurons. Unilaterally IT-injected mice were used for single-unit recording of firing activity. The recording was performed from the injected side. A, Schematic illustration of representative microelectrode placement sites within the GP. Number indicates the distance (in millimeters) from bregma in the anteroposterior plane. B, Proportion of the firing types. Number of units recorded from eight mice is shown in parentheses. C, Spontaneous mean firing rate of the firing types and total units.

Activation of the STN neurons in response to systemic DA stimulation Administration of drugs that stimulate DA receptors is known to increase the firing activity of the STN neurons (Kreiss et al., 1996, 1997). To ascertain whether systemic administration of METH induced activation of the STN activity, we examined METH-induced expression of c-fos in the STN. C57BL/6J inbred mice were administered PBS or METH (4.0 mg/kg, s.c.) and used for brain sectioning. The sections through the STN were immunostained with anti-c-fos antibody (Fig. 10). There were a number of c-fos-positive cells in the STN in the METH-administered group (106 ± 10 cells per section; n = 4). These data support the previous reports that systemic DA stimulation activates the STN neurons (Kreiss et al., 1996, 1997).

• Download figure
• Open in new tab
• Download powerpoint

Figure 7.

STN elimination attenuates METH-induced suppression of the SNr activity. A, METH-induced firing of the SNr neurons. After recording of spontaneous firing, the mice were administered METH (1.0 mg/kg, s.c.), and recording was then performed from the IT-injected side. The firing pattern and digital raster display of representative units during pretreatment (Pre) and 25 min after treatment are shown. The mean firing rate (Fr; spikes per second) of the unit is presented. B, Time course of METH-induced firing rate. The average of firing rate for every 5 min period was calculated, and the ratio of drug-induced firing rate relative to spontaneous firing rate was plotted (n = 11). ^*p < 0.05, ^** p < 0.01, significant differences from non-Tg mice.

Discussion

In the present study, we performed conditional ablation of the STN neurons from basal ganglia circuitry by using the IMCT technique. We then examined DA-mediated motor behavior and analyzed the firing activity of the SNr and GP neurons that normally receive the STN inputs. Ablation of the STN neurons caused an increase in spontaneous movement, whereas it reduced hyperactivity in response to METH stimulation. Ablation of these neurons also modulated the pattern and rate of spontaneous firing of the SNr neurons, although it did not substantially affect spontaneous firing of the GP neurons. The ablation attenuated METH-induced suppression of the firing rate of the SNr neurons and inhibited METH-induced elevation of the firing rate of the GP neurons. In addition, pharmacological blockade of GP activation in response to METH stimulation inhibited the suppression of SNr activity and the resultant motor activation.

IMCT of the STN neurons caused spontaneous hyperactivity, which was accompanied by a shift in the pattern of spontaneous firing of the SNr neurons and a reduction in the rate of spontaneous firing of these neurons, while preserving the activity of the GP neurons (supplemental Fig. 1A, available at www.jneurosci.org as supplemental material). These results indicate a key role of the STN neurons in spontaneous motor behavior and spontaneous firing activity of the SNr neurons. The results suggest that the STN neurons normally act to maintain spontaneous firing of the SNr neurons predominantly through the direct subthalamonigral pathway, thereby suppressing spontaneous movement (supplemental Fig. 1B, available at www.jneurosci.org as supplemental material). These data support the standard model of basal ganglia circuitry that proposes a major role of the subthalamonigral pathway in suppression of motor functions (Alexander and Crutcher, 1990; DeLong, 1990; Parent and Hazrati, 1995; Gerfen and Wilson, 1996).

• Download figure
• Open in new tab
• Download powerpoint

Figure 8.

STN elimination blocks METH-induced elevation of the GP activity. A, METH-induced firing of the GP neurons. After recording of spontaneous firing, the mice were administered METH (1.0 mg/kg, s.c.), and recording was then performed from the IT-injected side. The firing pattern and digital raster display of representative units during pretreatment (Pre) and 25 min after treatment are shown. The mean firing rate (Fr; spikes per second) of the unit is presented. B, Time course of METH-induced firing rate. The average of firing rate for every 5 min period was calculated, and the ratio of drug-induced firing rate relative to spontaneous firing rate was plotted (n = 5). ^*p < 0.05, significant differences from non-Tg mice.

The shift in the pattern of spontaneous firing of the SNr neurons by STN ablation was characterized by a reduction in the proportion of the regular firing type and an increase in the proportion of the bursty firing type. However, the STN ablation did not influence the shape and cell density of the SNr. These data suggest that STN neurons contribute to the generation of the regularity in spontaneous firing of the output neurons. Previous electrophysiological studies show that the STN neurons possess intrinsic tonic firing activity (Nakanishi et al., 1987; Bergman et al., 1994; Song et al., 2000) and that this activity is generated from different types of subthreshold sodium currents (Bevan and Wilson, 1999; Beurrier et al., 2000; Do and Bean, 2003). The property of spontaneous firing of the STN neurons may be important to determine the firing pattern of the output neurons and thereby may subserve normal movement control.

• Download figure
• Open in new tab
• Download powerpoint

Figure 9.

Pharmacological blockade of GP inhibits electrophysiological response of output neurons and behavioral susceptibility to METH. A, Single-unit recording of the SNr neurons. C57BL/6J inbred mice (n=6) were treated intrapallidally with PBS or NBQX at 125 ng/site, and, 15 min later, they were administered METH (1.0 mg/kg, s.c.) for unit recording. The average of firing rate for every 5 min period was calculated, and the ratio of METH-induced firing rate relative to spontaneous firing rate was plotted. ^*p < 0.05, ^**p < 0.01, significant differences from the PBS-treated group. B, Influence of NBQX on c-fos induction in the GP. The inbred mice (n = 4) were treated intrapallidally with PBS or NBQX (125 ng/site), administered METH (4.0 mg/kg, s.c.), and, 1 h later, used for brain sectioning. The sections through the GP were immunostained with anti-c-fos antibody, and the number of c-fos-positive cells in the GP was counted. ^*p < 0.05, significant difference from the PBS-treated group. Representative microscopic images are shown, and bottom images are fourfold magnified views of the GP (boxes in the top images). Scale bar, 200 μm. C, METH-induced locomotion. C57BL/6J inbred mice (n = 6) were treated intrapallidally with NBQX at different doses, and, 15 min later, they were administered METH (4.0 mg/kg, s.c.) for measurement of drug-induced locomotor activity. ^*p < 0.05, significant differences from the PBS-treated group.

• Download figure
• Open in new tab
• Download powerpoint

Figure 10.

c-Fos induction in the STN after METH administration. C57BL/6J inbred mice were administered PBS or METH (4.0 mg/kg, s.c.), and, 1 h later, they were used for brain sectioning. The sections through the STN were immunostained with anti-c-fos antibody. Scale bar, 100 μm.

A previous study reports that excitotoxic lesion of the STN in primates causes a reduction in firing activity in the external (GPe) and internal (GPi) segments of the GP (Hamada and DeLong, 1992b). Although in the present study we did not examine the influence of the STN ablation on the firing activity of the EPN neurons, our results suggest a role of the STN neurons in exciting spontaneous firing of the output nuclei in basal ganglia and support the results obtained from primates showing the reduced GPi activity by the STN lesion. In contrast, the STN ablation in this study did not cause any significant changes in the activity of the GP neurons. Another previous study with excitotoxic STN lesion in rodents also shows no alterations in the firing rate of the GP in the lesioned animals (Ni et al., 2000). Some anatomical and electrophysiological studies indicate that axons of the STN innervate both the GP and SNr through their collaterals in rodents (Kita et al., 1983; Kita and Kitai, 1987; Robledo and Féger, 1990; Nakanishi et al., 1991). In primates, a recent study reports that the majority of the STN axons projects to the GPe and the GPi/SNr with different sets of collaterals innervating some targets (Sato et al., 2000b). Therefore, the difference in the influence of STN lesion on the GP or GPe activity between the aforementioned studies does not appear to be derived from the distinction in the pattern of STN projections between the species. Conversely, primate GPe neurons are divided into two different firing types, termed the high-frequency discharge-pause type and the low-frequency discharge-burst type, and the high-frequency discharge-pause cells are predominant (Hamada and DeLong, 1992b). Our unit recording indicated the presence of the bursty and irregular types of GP neurons in mice, and the bursty type was the major proportion. The difference in the firing property of these neurons may affect dependency of their spontaneous firing activity on the STN input.

IMCT of the STN neurons caused a reduction in METH-induced hyperactivity, which was accompanied by attenuation of drug-induced suppression of the SNr firing and blockade of drug-induced elevation of the GP firing (supplemental Fig. 1C, available at www.jneurosci.org as supplemental material). These results indicate that the STN neurons mediate METH-induced motor behavior and METH responsiveness of the SNr and GP neurons. In addition, our histological examination indicated activation of the STN neurons in response to METH administration, being consistent with previous electrophysiological studies (Kreiss et al., 1996, 1997). One explanation for these observations is that the STN activation in response to METH facilitates GP activity through the subthalamopallidal pathway and that this facilitation induces suppression of SNr activity through the pallidonigral pathway, thereby leading to drug-induced motor activation (supplemental Fig. 1D, available at www.jneurosci.org as supplemental material). The facilitation of the pallidonigral pathway may dominate the opposite activity of the subthalamonigral pathway to suppress the net activity of SNr neurons. Partial suppression of SNr activity under the cell-ablation condition may be derived from facilitation of the striatonigral pathway. Again, the data obtained from the pharmacological blockade of GP activation suggest that the blockade attenuates METH-induced suppression of SNr activity and in turn diminishes the behavioral response (supplemental Fig. 1 E, available at www.jneurosci.org as supplemental material), supporting the view that the STN activation contributes to drug-induced regulation of the output activity and the motor behavior mainly through the subthalamo-pallido-nigral pathway.

One possible mechanism by which the GP-mediated pathways dominate the subthalamonigral pathway in DA-induced motor behavior involves the property and distribution of synaptic terminals originating from the GP or the STN on the output neurons. The GP terminals are known to form symmetrical synaptic contacts with large boutons that distribute densely in the perikarya and proximal dendrites of the output neurons (Smith and Bolam, 1991; Bolam and Smith, 1992; Smith et al., 1998), whereas the STN terminals appear to make asymmetrical synaptic contacts with relatively smaller boutons that show a scattered distribution over the perikarya and dendrites of these neurons (Bevan et al., 1994; Smith et al., 1998). The facilitation of the GP neurons, because of their synaptic property and distribution, may produce a greater effect of GABAergic inhibition on the output neurons, shunting the glutamatergic excitation through the subthalamonigral pathway.

There are two potential mechanisms by which systemic administration of METH induces hyperactivity of the STN neurons. One is the activation of the cortical neurons in response to DA, which then stimulates the STN neurons through the direct corticosubthalamic route. Administration of DA receptor agonists is known to activate many neurons in the cortical areas that are innervated by midbrain DA neurons (Wang et al., 1995). Indeed, electrical stimulation of the cortical neurons increases the discharge of the STN neurons (Nambu et al., 2000). Another mechanism is the direct activation of the STN neurons by DA. These neurons receive direct inputs from midbrain DA neurons (Hassani et al., 1997), and DA receptor agonists alter the discharge of the STN neurons (Mintz et al., 1986; Baufreton et al., 2003). It remains to be determined which mechanism is involved in the activation of the STN neurons in response to systemic DA stimulation.

A selective loss of the nigrostriatal DA system in Parkinson's disease is reported to cause hyperactivity of the STN neurons (Bergman et al., 1994; Hassani et al., 1996; Magnin et al., 2000). This hyperactivity is considered to be derived from the decreased GP activity because of the loss of DA function and appears to be linked to the symptoms of the disease. Our results obtained from the analysis of the neural circuitry underlying DA-induced motor behavior indicate that activation of the STN neurons results in the facilitation of GP activity. Comparison of these facts suggests that the STN activation affects motor behavior depending on the state of neural activity of other basal ganglia nuclei.

In conclusion, our findings demonstrate that the STN neurons are involved in dual regulation of basal ganglia function through differential neural pathways dependent on the state of DA transmission. The STN neurons function to suppress spontaneous motor behavior by facilitating the output neurons mainly through the subthalamonigral pathway. In response to DA stimulation, an increase in the STN activity contributes to the expression of motor behavior by inducing suppression of the output neurons, in particular through the subthalamo-pallid-onigral pathway. Detailed understanding of the cellular mechanisms for differential control of the neural circuitry through the STN will provide the physiological basis of various motor functions and the clinical evidence for therapeutic approaches to movement disorders.