Dopamine–Adenosine Interactions in the Striatum and the Globus Pallidus: Inhibition of Striatopallidal Neurons through Either D₂ or A_2A Receptors Enhances D₁Receptor-Mediated Effects on c-fos Expression

Abstract

D₁ receptors located on striatonigral neurons and D₂ receptors located, together with A_2Areceptors, on striatopallidal neurons are known to interact functionally. Using in situ hybridization, we examined the effects of D₁ and D₂ agonists and of an A_2A antagonist on c-fos mRNA in identified striatal neurons and in globus pallidus. The full D₁agonist, SKF 82958 (1 mg/kg), induced a homogenous increase ofc-fos mRNA in the striatum. This increase occurred to a similar extent in D₁ and D₂ receptor-containing striatal neurons. Conversely, the D₂ agonist, quinelorane (2 mg/kg), decreased c-fos mRNA in these populations but increased it in globus pallidus. The adenosine A_2A receptor antagonist, SCH 58261 (5 mg/kg), also decreased c-fosmRNA in D₂ receptor-containing neurons in striatum but did not affect pallidal c-fos mRNA. Concomitant administration of either D₁ plus D₂ agonists or D₁ agonist plus A_2A antagonist caused a potentiation of c-fos mRNA in striatal neurons expressing the D₁ receptor and in globus pallidus. However, only the combination of D₁ and D₂ agonists modified the c-fos mRNA expression to a “patchy” distribution. Our data show that (1) c-fos expression can be activated through D₁ and inhibitedthrough A_2A or D₂ receptors in both striatal output pathways in normal rats, and (2) D₂ receptor stimulation as well as A_2A receptor blockade can interact with D₁ receptor activation to potentiatec-fos expression in the striatum and the globus pallidus. The data also suggest that the topological alteration ofc-fos expression after coadministration of D₁ and D₂ agonists involves D₂receptors located on interneurons or presynaptically on dopaminergic nerve terminals.

The basal ganglia are involved in the integration of sensorimotor, associative, and limbic information to produce motor behaviors. The central component of these structures, the striatum, integrates excitatory glutamatergic inputs from cortex, thalamus, and limbic areas, with dopaminergic inputs from mesencephalon. It is composed of a large proportion of medium-sized spiny output neurons (95%) and of interneurons (5%). Striatal output neurons are GABAergic and project to either substantia nigra (pars reticulata) or globus pallidus and differ in their neuropeptide content: the striatonigral pathway contains substance P/dynorphin and the striatopallidal enkephalin (for review, see Graybiel, 1990; Gerfen and Wilson, 1996).

Dopamine regulates striatal neurotransmission via two types of receptor families, D₁-type (D₁ and D₅) and D₂-type (D₂, D₃, D₄) receptors, which have distinct pharmacological profiles and mechanisms of transduction (Creese et al., 1983; Jaber et al., 1996). It has been suggested that dopamine differentially regulates the two striatal output pathways and that a balanced control is essential for the proper function of the extrapyramidal motor system (for review, see Alexander and Crutcher, 1990; Gerfen, 1992). Accordingly, several anatomical studies have demonstrated a segregation of D₁ and D₂receptors, respectively, in striatonigral/substance P and striatopallidal/enkephalin neurons (Gerfen et al., 1990; Le Moine et al., 1990a, 1991; Hersch et al., 1995; Le Moine and Bloch, 1995, 1996;Yung et al., 1996). However, many physiological data indicate synergistic effects after coactivation of D₁- and D₂-type receptors (for review, see Waddington and Daly, 1993; White and Hu, 1993).

In the basal ganglia A_2A receptors are restricted to striatopallidal/D₂-containing neurons and, in contrast to D₂ receptors, are not present on dopaminergic nerve terminals and are virtually absent from cholinergic interneurons (Schiffmann et al., 1991; Fink et al., 1992; Augood and Emson, 1994;Svenningsson et al., 1997). An alternative way to investigate how D₁/D₂ interactions occur is to study how adenosine modulates neurotransmission via adenosine A_2Areceptors and how they can be involved in interactions with D₁ receptor-mediated effects. Indeed, it has been shown that dopamine acting on D₂ receptors and adenosine acting on A_2A receptors have opposing actions on neurotransmitter release, gene expression, and several motor behaviors (for review, seeFerré et al., 1992; Ongini and Fredholm, 1996). Accordingly, selective A_2A antagonists share with D₂agonists the ability to potentiate motor effects induced by D₁ receptor agonists as well as D₁-inducedc-fos expression in dopamine-depleted striatum (Jiang et al., 1993; Pinna et al., 1996; Pollack and Fink, 1996).

In this context, detailed analysis of the modulation of D₁ or D₂ agonist-mediated effects by an A_2A antagonist may help to elucidate the D₁/D₂ interactions in the basal ganglia. We therefore used sensitive in situ hybridization with riboprobes to examine how pharmacological treatments involving dopamine or adenosine receptors might up- or downregulate the expression ofc-fos in the basal ganglia. In particular, c-fosexpression was studied in phenotypically identified striatal neurons, with double-labeling, after challenges with selective compounds acting at D₁, D₂, and A_2Areceptors given alone or in combination.

MATERIALS AND METHODS

Pharmacological manipulations and tissue preparation.All experiments have been performed in accordance with the guidelines of the French Agriculture and Forestry Ministry (decree 87849, license 01499) and with the Centre National de la Recherche Scientifique approval. Adult male Sprague Dawley rats (200–280 gm) (Iffa Credo, France) were maintained in standard housing conditions several days before the experiments. Animals were treated with systemic injections of saline (NaCl 0.9%); ±SKF-82958 (Research Biochemicals, Natick, MA), a full dopamine receptor agonist that has a 200-fold selectivity for D₁ over D₂ receptors (Andersen and Jansen, 1990); quinelorane or LY-163,502 (Research Biochemicals), a dopamine receptor agonist that conversely shows at least a 50-fold selectivity for D₂ over D₁ receptors (Bymaster et al., 1986; Andersen and Jansen, 1990); or SCH-58261 (Schering-Plough, Milan, Italy), an adenosine receptor antagonist that is 60-fold selective for A_2A over A₁ receptors (Zocchi et al., 1996). All rats had been handled the day before the injection and had received two injections. The different treatment groups were as follows: saline plus saline (n = 5), quinelorane 2 mg/kg plus saline (n = 4), SKF-82958 0.5 mg/kg plus saline (n = 3), SKF-82958 1 mg/kg plus saline (n = 5), SKF-82958 2 mg/kg plus saline (n = 2), SCH-58261 5 mg/kg plus saline (n = 4), SKF-82958 1 mg/kg plus quinelorane 2 mg/kg (n = 5), SKF-82958 1 mg/kg plus SCH-58261 5 mg/kg (n = 4), or quinelorane 2 mg/kg plus SCH-58261 5 mg/kg (n = 5). SKF-82958 and quinelorane were dissolved in saline, whereas SCH-58261 was dissolved in saline/5% Tween 80 after careful sonication. Drugs were injected intraperitoneally, 0.5 ml per injection, and the rats were decapitated 1 hr after the injections. The brains were dissected out, frozen over liquid nitrogen, and then sectioned into 10 μm sections, collected on gelatin-coated slides, and stored at −80°C until used.

Probe synthesis. ³⁵S-labeled cRNA probes were prepared by in vitro transcription from cDNA clones corresponding to fragments of the rat c-fos cDNA (Curran et al., 1987) (a gift from Dr. T. Curran, Roche Institute of Molecular Biology, Nutley, NJ), rat D₁ and D₂ dopamine receptor cDNAs (Monsma et al., 1989, 1990) (a gift from Dr. D. Sibley, National Institute of Health, NINDS, Bethesda, MD), and rat μ-opioid receptor cDNA (Thompson et al., 1993) (a gift from Dr. S. J. Watson, University of Michigan, Ann Arbor, MI). Transcriptions were performed from 50 ng of linearized plasmid, using either ³⁵S-UTP (>1000 Ci/mmol; DuPont de Nemours, Les Ulis, France) or digoxigenin-11-UTP (Boehringer Mannheim, Meylan, France) and SP6, T3, or T7 RNA polymerases as described by Le Moine and Bloch (1995). After alkaline hydrolysis to obtain 250 bp cRNA fragments, the³⁵S-labeled probes were purified on G50-Sephadex. The³⁵S-labeled probes and the digoxigenin-labeled probes were precipitated in 3 m sodium acetate/absolute ethanol (0.1:2.5, v/v), pH 5.

Single detection of c-fos mRNA on cryostat sections.Sections were hybridized as described by Le Moine and Bloch (1995,1996) with minor modifications. Cryostat sections were post-fixed in 4% paraformaldehyde (PFA) for 5 min at room temperature, rinsed twice in 4× SSC, and placed into 0.25% acetic anhydride in 0.1m triethanolamine/4× SSC, pH 8, for 10 min at room temperature. After dehydration, the sections were hybridized overnight at 55°C with 10⁶ cpm of ³⁵S-labeled cRNA probe in 50 μl of hybridization solution (20 mmTris-HCl, 1 mm EDTA, 300 mm NaCl, 50% formamide, 10% dextran sulfate, 1× Denhardt’s, 250 μg/ml yeast tRNA, 100 μg/ml salmon sperm DNA, 100 mm DTT, 0.1% SDS, and 0.1% sodium thiosulfate). After 20 min of RNase A treatment (20 mg/ml), the sections were washed with 2× SSC (5 min, twice), 1× SSC (5 min), 0.5× SSC (5 min) at room temperature, and rinsed in 0.1× SSC at 65°C (30 min, twice) before dehydration (the latter SSC washes contained 1 mm DTT). Sections either were exposed on x-ray films (Kodak BIOMAX, Rochester, NY) for 3–6 d or dipped into Ilford K5 emulsion, exposed for 7 weeks, developed, and stained with toluidine blue.

Simultaneous detection of c-fos mRNA with D₁ or D₂ mRNAs on cryostat sections. Two combinations of probes were used for the simultaneous detection of two mRNAs on a single section: a ³⁵S-labeled c-fosprobe in combination with digoxigenin-labeled D₁ or D₂ probes. Cryostat sections were pretreated as mentioned above. After dehydration the sections were hybridized overnight at 55°C with a combination of ³⁵S- and digoxigenin-labeled probes (10⁶ cpm of ³⁵S-labeled probe and 10–20 ng of digoxigenin-labeled probe in 50 μl of hybridization solution). After 20 min of RNase A treatment at 37°C (20 μg/ml), the slides were washed in various concentrations of SSC as mentioned above, but without DTT. After washing, the sections were put in 0.1× SSC at room temperature and then processed directly for detection of the digoxigenin signal. The sections were rinsed twice for 5 min in buffer A (1 m NaCl, 0.1 m Tris, and 2 mm MgCl₂, pH 7.5) and then for 30 min in buffer A containing 3% normal goat serum and 0.3% Triton X-100. After 5 hr of incubation at room temperature with alkaline phosphatase-conjugated anti-digoxigenin antiserum (Boehringer Mannheim; 1:1000 in buffer A, 3% normal goat serum, and 0.3% Triton X-100), the sections were rinsed in buffer A (5 min, twice) and then for 10 min twice in STM buffer (1 m NaCl, 0.1 m Tris, and 5 mm MgCl₂, pH 9.5), and finally for 10 min twice in 0.1 m STM buffer, pH 9.5 (0.1 mNaCl, 0.1 m Tris, and 5 mmMgCl₂, pH 9.5). Then the sections were incubated overnight in the dark at room temperature in 0.1 m STM buffer, pH 9.5, containing 0.34 mg/ml nitroblue tetrazolium and 0.18 mg/ml bromo-chloro-indolylphosphate. They were rinsed in 0.1m STM buffer, pH 9.5, and then in 1× SSC, dried, and dipped into Ilford K5 emulsion (diluted 1:3 in 1× SSC). After being exposed for 10 weeks in the dark, the sections were developed and mounted without counterstaining.

Counting of labeled neurons. Labeled neurons both from single-labeling and double-labeling experiments (exposure times: 7 weeks for single in situ hybridization and 10 weeks for double in situ hybridization) were counted as previously described on similar material (Le Moine and Bloch, 1995). Accordingly, a labeled neuron corresponded to a density of silver grains at least twofold higher than background. One section per animal was analyzed for counting in single in situ hybridization, and one section per animal was counted for the double labeling. The densities ofc-fos mRNA-containing neurons were studied in the striatum (+1 mm from bregma) and globus pallidus (−0.8 mm from bregma) according to Swanson (1992). The areas examined were 2–4 mm² for the caudate putamen and 1.5–2 mm² for the globus pallidus. The labeled neurons were counted using an image analyzer system for cartography (HISTO 200, Biocom, Les Ulis, France). For double in situ hybridization, quantification was performed only on the sections with simultaneous detection of c-fos and D₂ mRNAs, and thec-fos mRNA-labeled neurons were divided into two populations: the D₂ mRNA-positive (+) and D₂mRNA-negative (−) neurons. The densities ofc-fos-expressing neurons (number of c-fosmRNA-positive neurons per mm²) were pooled and averaged for each group, and statistical analysis was performed by a two-way ANOVA, followed by post hoc t tests corrected for the experiment-wise α level by the Bonferroni correction.

RESULTS

Effects of D₁ and D₂ agonists onc-fos expression in the striatum and in the globus pallidus

Under control conditions (i.e., saline-treated rats), neurons containing c-fos mRNA were observed in several cortical areas, especially the endopiriform and piriform cortices, in the septum and in the caudate putamen and nucleus accumbens (Fig.1). The densities ofc-fos-positive neurons (mean ± SEM) were 35.25 ± 4.34 per mm² in the caudate putamen and 24.2 ± 3.4 per mm² in the globus pallidus (Table1).

Fig. 1.

D₁/D₂ and D₁/A_2A receptor interactions onc-fos expression. Dark-field photomicrographs afterin situ hybridization with a ³⁵S-labeled riboprobe show the localization of c-fos mRNA-containing neurons in the striatum after saline (A), D₁ agonist SKF-82958 (B), D₁ agonist SKF-82958 + A_2A antagonist SCH-58261 (C), and D₁ agonist SKF-82958 + D₂ agonist quinelorane (D). Under basal conditions (A) c-fosmRNA-containing neurons are few and scattered in the caudate putamen (cp) and the nucleus accumbens (acb).c-fos is induced after the D₁ agonist both in the caudate putamen and the nucleus accumbens (B). As compared with D₁ agonist + A_2A antagonist (C), the combined treatment with D₁ + D₂ agonists potentiates the D₁-induced expression of c-fos with a heterogeneous “patchy” pattern (arrowheads inD). Cortical expression of c-fos in layer VIb is seen clearly after D₁ agonist alone or in combination with either A_2A antagonist or plus D₂ agonists (arrows in B–D). Magnification, 11×.

Table 1.

Density of neurons containing c-fos mRNA after D₁ or/and D₂ agonists and A_2Aantagonist, alone or in combination

Treatment group	n	Caudate putamen	Globus pallidus
Saline (a)	6	35.25  ± 4.34	24.20  ± 3.40
Quinelorane (b)	5	8.55  ± 1.70*a	42.20  ± 4.00*a
SKF-82958 (c)	4	132.75  ± 15.30*a	12.50  ± 1.60
SCH-58261 (d)	4	16.70  ± 1.50*a	18.25  ± 4.20
SKF-82958 + quinelorane	5	156.20  ± 6.50*b,ns,c	122.30  ± 17.60*b,c
SKF-82958 + SCH-58261	4	183.90  ± 6.30*c,d	69.80  ± 13.90*c,d
Quinelorane + SCH-58261	5	17.35  ± 1.39*b,ns,d	60.30  ± 8.10*d,ns,b

Rats were treated with saline (NaCl 0.9%), with the D₂ agonist quinelorane (2 mg/kg), with the D₁agonist SKF-82958 (1 mg/kg), with the A_2A antagonist SCH 58261 (5 mg/kg), or various combinations: SKF-82958 (1 mg/kg) + quinelorane (2 mg/kg), SCH 58261 (5 mg/kg) + SKF-82958 (1 mg/kg), and quinelorane (2 mg/kg) + SCH 58261 (5 mg/kg). c-fos mRNA was detected with single in situ hybridization (exposure times, 7 weeks). Values represent the mean ± SEM of the number of c-fos mRNA-containing neurons per mm². Two-way ANOVA, followed by post hoc t tests corrected for the experiment-wise alpha level (Bonferroni correction). The results of the global ANOVA were for quinelorane/SKF-82958 interaction:F_(1,16) = 11.48, p < 0.005 for caudate putamen (CP) and F_(1,16) = 23.87,p < 0.001 for globus pallidus (GP); for SKF-82958/SCH-58261 interaction: F_(1,14) = 19.02, p < 0.001 for CP andF_(1,14) = 19.84, p < 0.001 for GP; for quinelorane/SCH-58261 interaction:F_(1,16) = 21.27, p < 0.001 for CP and F_(1,16) = 5.163, p < 0.05 for GP. For the multiple post hoc t tests Bonferroni correction, an asterisk indicates relevant significant differences between indicated groups (p < 0.05).

One hour after administration of the D₁ agonist SKF-82958 at the dose of 1 mg/kg, the number of c-fos mRNA-containing neurons dramatically increased in the caudate putamen (+277%) and the nucleus accumbens (Figs. 1, 2, Table 1). An increase also was found in the cortex (with a particularly high concentration in layer VIb) and in the septum (Fig. 1). By contrast, the number of c-fos mRNA-containing neurons tended to decrease (by 48%, p = 0.08) in the globus pallidus (Fig. 3, Table 1). In all of the examined areas, the effects of SKF-82958 were similar over the dose range tested (0.5–2 mg/kg; data not shown).

Fig. 2.

D₁- and D₂-mediated regulation of c-fos expression in the caudate putamen. Dark-field photomicrographs from single in situhybridization with a ³⁵S-labeled riboprobe showc-fos mRNA after treatments with D₁ and D₂ agonists alone or in combination. The D₁agonist SKF-82958 increases the number of c-fos-positive neurons (B), whereas the D₂ agonist quinelorane decreases it (C), as compared with saline-treated rats (A). Association of D₁ and D₂ agonists changes the D₁-induced c-fos expression into a heterogeneous “patchy” pattern (arrowheads inD). Quantitative data are listed in Table 1. Magnification, 40×.

Fig. 3.

D₁- and D₂-mediated regulation of c-fos expression in the globus pallidus. Dark-field photomicrographs from single in situhybridization with a ³⁵S-labeled riboprobe showc-fos mRNA after treatments with D₁ and D₂ agonists alone or in combination. The level ofc-fos mRNA observed under basal conditions inA is increased after the D₂ agonist (C), whereas it tends to decrease with the D₁ agonist (B). Combined treatment with both D₁ and D₂ agonists potentiated the D₂-mediated induction of c-fos in the globus pallidus (D). Stars indicate the internal capsule. Quantitative data are listed in Table 1. Magnification, 40×.

Conversely, the D₂ agonist quinelorane, at the dose of 2 mg/kg, caused a decrease in the number of c-fosmRNA-containing neurons in the caudate putamen (−75%, Table 1). Detection of such a decrease is directly related to our ability to consistently detect and quantify c-fos mRNA in basal conditions by using sensitive riboprobes (Fig. 2). In contrast, the density of labeled neurons in the globus pallidus was increased after treatment with the D₂ agonist (+74%) (Fig. 3, Table1).

When quinelorane (2 mg/kg) was coadministered with SKF-82958 (1 mg/kg), the density of c-fos-labeled neurons in the caudate putamen and the nucleus accumbens was increased to the same extent as after SKF-82958 alone (Table 1). However, as shown in Figures 1 and4, the homogenous distribution of thec-fos mRNA-containing neurons after SKF-82958 treatment was heterogeneous (“patchy”) after coadministration of the two drugs. Comparison on adjacent sections shows that the distribution ofc-fos mRNA after D₁ plus D₂ agonists was parallel to the distribution of μ-opioid receptor mRNA (Fig. 4). At the same time, in the globus pallidus, the coadministration of both SKF-82958 (1 mg/kg) and quinelorane (2 mg/kg) increased by 190% the density of c-fos-labeled neurons as compared with quinelorane alone (Fig. 3, Table 1).

Fig. 4.

Striatal c-fos expression in patches after combined treatment with D₁ and D₂agonists. Dark-field photomicrographs after in situhybridization with ³⁵S-labeled riboprobes on adjacent sections show that the “patches” of c-fosmRNA-containing neurons (arrowheads in A) correspond to patches of μ-opioid receptor mRNA expression in the striatum (arrowheads in B). Also note the concomitant expression of c-fos and μ-mRNA in the subcallosal patch. cc, Corpus callosum. Magnification, 23×.

Effects of an A_2A antagonist alone or in combination with a D₁ agonist on c-fos expression in the striatum and in the globus pallidus

The adenosine A_2A antagonist SCH-58261 had similar effects to the D₂ agonist quinelorane in the striatum. Treatment with SCH-58261 at a dose of 5 mg/kg induced adecrease in the density of c-fos-labeled neurons in the caudate putamen (−53%). In contrast to quinelorane, it had no effect on the density of labeled neurons in the globus pallidus (Fig.5, Table 1). The coadministration of SKF-82958 (1 mg/kg) and SCH-58261 (5 mg/kg) induced a further increase in the density of c-fos mRNA-containing neurons in the caudate putamen (+38%) as compared with SKF-82958 alone (Fig. 5, Table1). The distribution pattern of the c-fos-labeled neurons after the coadministration was homogeneous in the striatum and not patchy, as seen after D₁ plus D₂ agonists (Figs. 1, 2, 4, 5).

Fig. 5.

Effect of the A_2A antagonist alone or in combination with the D₁ agonist on c-fosexpression in the caudate putamen (A–C) and the globus pallidus (D–F). Dark-field photomicrographs from single in situ hybridization with a ³⁵S-labeled riboprobe show the basal levels ofc-fos mRNA in the caudate putamen (A) and in the globus pallidus (D). The A_2A antagonist SCH-58261 alone decreases the number ofc-fos-positive neurons in the caudate putamen (B) but has no effect on the globus pallidus (E). Coadministration of the D₁agonist, together with the A_2A antagonist, inducesc-fos both in the caudate putamen (C) and in the globus pallidus (F) with a synergistic effect, as compared with the D₁ agonist alone (see also Table 1).Stars indicate the internal capsule. Quantitative data are listed in Table 1. Magnification, 40×.

In the globus pallidus the coadministration of SKF-82958 with SCH-58261 induced a dramatic increase in the density of labeled neurons as compared with the saline-treated rats (+188%) but also as compared with SKF-82958 alone (+458%) (Fig. 5, Table 1).

Effects of an A_2A antagonist alone or in combination with a D₂ agonist on c-fos expression in the striatum and in the globus pallidus

As mentioned above, the D₂ receptor agonist quinelorane (2 mg/kg) decreased the density of c-fosmRNA-containing neurons in the caudate putamen and increased it in the globus pallidus, whereas the A_2A receptor antagonist SCH 58261 (5 mg/kg) affected c-fos mRNA expression only in the caudate putamen, where it caused a decrease in the density of labeled neurons (Table 1). The coadministration of D₂agonist and A_2A antagonist significantly counteracted the decrease induced by quinelorane in the caudate putamen (from −75 to −52%). No synergistic effect of the two drugs on c-fosexpression was found in the globus pallidus as compared with quinelorane alone (Table 1).

Phenotypical identification of the c-fosmRNA-containing neurons in the caudate putamen after D₁ and D₂ agonists, given alone or in combination

To examine in which type of striatal neurons the above-mentioned changes in c-fos expression occurred, we used double-labeling experiments with probes for either D₁ or D₂ receptor mRNA, together with a probe forc-fos mRNA. Because the results, analyzed in two separate experiments (as illustrated in Fig. 6), were identical, quantitative data were generated only fromc-fos plus D₂ mRNAs simultaneous detection (Table 2). Therefore, in the following, D₂ mRNA-negative (−) neurons are referred to as D₁ mRNA-positive (+) neurons on the basis of both experiments and previously published data (Le Moine and Bloch, 1995,1996).

Fig. 6.

Phenotypical characterization of the striatal neurons expressing c-fos after D₁ and D₂ agonists, alone or in combination. Double in situ hybridization detects D₁ or D₂receptor mRNA with digoxigenin-labeled riboprobe (stained cells), together with c-fos mRNA, with a ³⁵S-labeled riboprobe (silver grains). A and D show that c-fos mRNA is present both in D₁mRNA-containing (A) and D₂mRNA-containing (D) neurons under basal conditions. The D₁ agonist SKF-82958 increasesc-fos expression both in D₁ mRNA-containing neurons (arrowheads in B) and in D₂ mRNA-containing neurons (arrowhead inE). As compared with the D₁ agonist alone, coadministration of D₁ and D₂ agonists potentiates the increase of c-fos expression in D₁ mRNA-containing neurons (arrowheads inC) and decreases it in D₂ mRNA-containing neurons. Quantitative data are listed in Table 2. Magnification, 640×.

Table 2.

Density of D₁− or D₂ striatal neurons expressing c-fos mRNA after D₁ or/and D₂ agonists and A_2A antagonist, alone or in combination

Treatment group	n	Fos+/D2− neurons	Fos+/D2+ neurons
Saline (a)	5	34.2  ± 4.6	26.7  ± 3.7
Quinelorane (b)	4	5.3  ± 1.7*a	0.75  ± 0.4*a
SKF-82958 (c)	5	84.2  ± 14.8*a	73.0  ± 10.4*a
SCH-58261 (d)	4	22.5  ± 3.8ns,a	10.8  ± 3.1*a
SKF-82958 + quinelorane	5	233.5  ± 25.3*b,c	22.4  ± 2.5*b,c
SKF-82958 + SCH-58261	4	166.6  ± 18.0*c,d	65.6  ± 7.7*d,ns,c
Quinelorane + SCH-58261	5	10.8  ± 2.3*b,ns,d	1.5  ± 0.5*ns,b,d

Rats were treated with saline (NaCl 0.9%), with the D₂ agonist quinelorane (2 mg/kg), with the D₁agonist SKF-82958 (1 mg/kg), with the A_2A antagonist SCH 58261 (5 mg/kg), or various combinations: SKF-82958 (1 mg/kg) + quinelorane (2 mg/kg), SCH 58261 (5 mg/kg) + SKF-82958 (1 mg/kg), and quinelorane (2 mg/kg) + SCH 58261 (5 mg/kg). c-fos mRNA was detected with double in situ hybridization (exposure times, 10 weeks). Values represent the mean ± SEM of the number ofc-fos mRNA-containing neurons per mm². Two-way ANOVA, followed by post hoc t tests corrected for the experiment-wise alpha level (Bonferroni correction). The results of the global ANOVA were for quinelorane/SKF-82958 interaction:F_(1,15) = 31.55, p < 0.001 for D₂-negative neurons and F_(1,15) = 4.155 for D₂-positive neurons; for SKF-82958/SCH-58261 interaction: F_(1,14) = 15.45, p< 0.001 for D₂-negative neurons andF_(1,14) = 0.35 for D₂-positive neurons. For the multiple post hoc t tests Bonferroni correction, an asterisk indicates relevant significant differences between indicated groups (p < 0.05).

Figure 6 shows that administration of the D₁ agonist SKF-82958 (1 mg/kg) increased the number of both D₁ and D₂ mRNA-containing neurons that express c-fosmRNA (Table 2). Conversely, the D₂ agonist quinelorane (2 mg/kg) decreased the density of c-fos-labeled neurons both for D₁ and D₂ mRNA-containing neurons (Table2). The coadministration of D₁ and D₂ agonists had opposite effects on c-fos expression in these two populations because it induced an increase in the density ofc-fos-labeled neurons containing D₁ mRNA and a decrease in the density of c-fos-labeled neurons containing D₂ mRNA, as compared with the D₁ agonist alone (Fig. 6, Table 2). Indeed, in SKF-82958 treated rats 53% of thec-fos expressing neurons were D₁ mRNA-positive, whereas in rats treated by SKF-82958 plus quinelorane, the proportion of these neurons reached 91% (Table 2). Note here and below that the relative changes observed in the density of c-fos-labeled neurons in the caudate putamen are comparable to what was observed in the single-labeling experiments and summarized in Table 1.

Phenotypical identification of the c-fosmRNA-containing neurons in the caudate putamen after A_2A antagonist and D₁ agonist, given alone or in combination

Similar experiments, performed with the A_2A antagonist SCH 58261 (5 mg/kg), showed a decrease in the density ofc-fos-labeled neurons and in D₂ mRNA-containing neurons, but not in D₁ mRNA-containing neurons (Table 2). As mentioned above, the D₁ agonist SKF-82958 increased the density of c-fos-labeled neurons both in D₁ and D₂ mRNA-positive neurons (Table 2). The coadministration of the D₁ agonist and the A_2A antagonist potentiated the increase in the density of c-fos-labeled neurons that were positive for D₁ mRNA but had no effect on the density of c-fos labeled in D₂mRNA-containing neurons, as compared with the D₁ agonist alone (Table 2).

Phenotypical identification of the c-fosmRNA-containing neurons in the caudate putamen after A_2A antagonist and D₂ agonist, given alone or in combination

The density of D₂ mRNA-positive neurons that expressc-fos mRNA was lower in SCH-58261 (−60%) and quinelorane-treated animals (−97%) as compared with saline (Table 2). At the same time, quinelorane—and not SCH-58261—induced a reduction of c-fos in neurons positive for D₁ mRNA (−84.5%). When SCH-58261 and quinelorane were coadministered, there was no synergistic effect on c-fos expression in the D₁-containing nor in the D₂-containing neurons (Table 2).

DISCUSSION

Individual and synergistic effects of dopamine D₁ and D₂ receptor agonists and of an adenosine A_2Areceptor antagonist on c-fos expression were analyzed in the striatum and globus pallidus. Our data, summarized in Figure7, show that (1) c-fosexpression can be either activated through D₁ andinhibited through A_2A or D₂receptors in the two striatal output pathways in normal rats, and (2) D₂ receptor stimulation as well as A_2A receptor blockade can interact with D₁, but not D₂, receptor activation to potentiatec-fos expression in both the striatum and the globus pallidus.

Fig. 7.

Schematic representation of the interactions in the basal ganglia after treatments with D₁ and D₂ agonists or combined treatment with D₁ + D₂ agonist or D₁ agonist + A_2Aantagonist. The variations of expression of c-fos mRNA as compared with basal conditions are indicated inside the structure or the neuronal populations that we have studied. Dark arrows represent excitatory pathways, and white arrows represent inhibitory pathways. Thethickness of the arrows changes according to the supposed neuronal activity in the different pathways.ST, Striatum; GP, globus pallidus;SNc, substantia nigra pars compacta;SNr/EPN, substantia nigra pars reticulata/entopeduncular nucleus; STN, subthalamic nucleus; fos,c-fos mRNA.

Effect of D₂ and D₁ agonists given alone onc-fos expression in the striatum

Selective activation of D₂ receptors by the D₂ agonist produced a significant decrease in the number of striatal neurons expressing c-fos in the caudate putamen. The decrease was found in both D₁- and D₂-positive neurons. In D₂-containing neurons this decrease may be explained by the fact that dopamine is likely to have an inhibitory action on striatopallidal neurons via postsynaptic D₂ receptors (Gerfen et al., 1990). Conversely, the D₂ agonist effect on c-fos in D₁-containing neurons might be related to activation of presynaptic D₂ autoreceptors located on dopaminergic terminals, because this strongly decreases striatal dopamine release (Imperato et al., 1988; Suaud-Chagny et al., 1991) and thereby the D₁-mediated activity in striatonigral neurons. Decreases of mRNA coding for the immediate early gene NGFI-A (zif 268) have been described after treatment with drugs acting on D₂or A_2A receptors (Gerfen et al., 1995; Svenningsson et al., 1995), but we describe here for the first time the D₂-mediated inhibition of c-fos expression in the two striatal output neurons.

The full D₁ agonist SKF-82958 increased c-fosexpression in the striatum in normal rats, as previously reported byWang and McGinty (1996). A strong induction of c-fosexpression in the D₁ rich cortical layer VIb (Gaspar et al., 1995) also was found. Interestingly, c-fos mRNA increased to a similar extent in D₁- and D₂-containing neurons in the striatum. The stimulation ofc-fos expression in D₁-positive neurons was expected, because many studies have demonstrated that the dopamine-mediated induction of striatal Fos is dependent on D₁ activation [see Hughes and Dragunow (1995) and references therein]. The increased number of D₂-positive neurons expressing c-fos after SKF-82958 was unexpected. In previous studies, using the partial D₁ agonist SKF-38393, researchers observed c-fos induction only in the D₁ receptor-containing striatonigral neurons (Robertson et al., 1990; Gerfen et al., 1995). However, these studies were performed in animals with nigrostriatal lesions, and we therefore suggest thatc-fos induction by the D₁ agonist in striatopallidal neurons requires intact nigrostriatal neurons. We hypothesize that the D₁ agonist, when injected systemically, acts on D₁ receptors located on striatonigral terminals (Caillé et al., 1996) and stimulates GABA release (Cameron and Williams, 1993), which in turn inhibits nigrostriatal neurons and decreases the extracellular striatal dopamine level (Suaud-Chagny et al., 1992). This effect would be indirectly responsible for an increase of c-fos in striatopallidal neurons. Nevertheless, cholinergic interneurons expressing D₅ (C. Le Moine, unpublished results) in addition to D₂ receptors (Le Moine et al., 1990b) and corticostriatal glutamatergic neurons (Gaspar et al., 1995) also may be involved in this D₁-dependent c-fos activation in the D₂-containing neurons (Berretta et al., 1992).

Effect of combined D₁ and D₂ agonists onc-fos expression in the striatum

Thus, the effects of D₁ or D₂ agonists probably can be attributed to both direct postsynaptic effects and indirect effects mediated by the mesencephalic dopamine neurons. However, when these drugs are combined, the effects of endogenous dopamine are likely to be masked. Indeed, in the striatum, combined treatment with D₁ and D₂ agonists potentiatedc-fos expression in D₁-containing neurons but inhibited it in D₂-containing neurons. The fact that the combined treatment induces c-fos at 92% in D₁-containing neurons is consistent with data obtained in conditions that enhance extracellular dopamine concentration (Graybiel et al., 1990; Young et al., 1991; Moratalla et al., 1993; Jaber et al., 1995; Wang et al., 1995; Chergui et al., 1996).

Effect of an A_2A antagonist alone or in combination with D₁ or D₂ agonists in the striatum

A_2A and D₂ receptors regulate pallidal GABA release in an opposite manner (Ferré et al., 1993; Mayfield et al., 1993, 1996) and are colocalized in striatopallidal neurons, but not in interneurons nor on nigrostriatal terminals (Schiffmann et al., 1991; Fink et al., 1992; Augood and Emson, 1994; Svenningsson et al., 1997). Therefore, studying the effects of A_2A receptors on striatal neurotransmission may be of interest not only to better understand adenosinergic modulation but also to delineate effects specifically related to an altered activity of striatopallidal neurons. We show here that the A_2A antagonist SCH-58261 shared with the D₂ agonist the ability to decrease c-fosexpression in the striatum. This decrease occurred only in D₂-containing neurons, suggesting that this effect is mainly postsynaptic. Indeed, unlike the D₂ agonist, the A_2A antagonist does not affect dopamine release (Ferré et al., 1993). This supports the idea that endogenous adenosine acting at A_2A receptors regulates the constitutive expression of immediate early genes in the striatum (Svenningsson et al., 1995).

Coadministration of the A_2A antagonist with the D₁ agonist potentiated the D₁-induced increase in c-fos expression in D₁-containing neurons, like treatment with D₁ and D₂ agonists. However, this combination, unlike the D₁ plus D₂ combination, caused no inhibition of D₁-mediated c-fos induction in D₂-containing neurons. This suggests that regulation ofc-fos by dopamine is more potent than A_2A-mediated effects on these neurons in our conditions. Whereas the D₁/D₂ combined treatment produced a change of the initial homogeneous striatal expression ofc-fos into a “patchy” pattern, as previously described (Paul et al., 1992; Wang and McGinty, 1996), the pattern ofc-fos expression after the D₁/A_2A combination was homogeneous in the striatum. These results suggest that D₂ receptors located postsynaptically on striatopallidal neurons, like the A_2A receptors, are involved in the quantitative enhancement of c-fos mRNA in striatal neurons, whereas D₂receptors located presynaptically or on interneurons might be involved more specifically in differential dopaminergic regulations between the patch/matrix compartments.

D₁/D₂ and D₁/A_2A interactions in the globus pallidus

In accordance with previous immunohistochemical studies (Robertson et al., 1992; Marshall et al., 1993), we show here an increase ofc-fos expression in the globus pallidus after administration of the D₂ agonist. A strong tendency for a decrease ofc-fos expression was found after D₁ agonist treatment, although not significant in our statistical conditions. This tendency might be attributable to the D₁-mediatedc-fos expression in striatopallidal neurons. Taken together, these data suggest that stimulation of D₁ and D₂ receptors has opposite effects on pallidal neurons also.

The combined treatment with D₁ and D₂ agonists potentiated the increase in c-fos expression induced by the D₂ agonist alone, as previously shown (Paul et al., 1992,1995; Marshall et al., 1993). This agrees with electrophysiological data showing that the D₁ plus D₂ coactivation is required for the maximal excitatory effect, demonstrating a potentiated effect mediated by D₁ receptors on D₂ receptor-activated responses (Walters et al., 1987). There was also a strong induction of c-fos expression after combined treatment, using the A_2A antagonist together with the D₁ agonist. Interestingly, coadministration of A_2A antagonist together with the D₂ agonist had no synergistic effects on c-fos expression in the globus pallidus. This implies that, despite their coexpression and their well established interactions (Ferré et al., 1992, 1993), the D₂ and A_2A receptors are not solely the key for adenosine/dopamine interactions in the basal ganglia. Instead, our findings suggest that the most important functional interactions may be between drugs that affect A_2A and dopamine receptors in distinct neuronal populations. This conclusion also has implications for our understanding of the D₁/D₂interactions.

Disinhibition of striatopallidal neurons is one of the mechanisms whereby c-fos is induced in globus pallidus. However, ifc-fos expression can correlate with the activity of striatopallidal neurons, these neurons are likely to be stimulated rather than inhibited by combined treatments with D₁ plus D₂ agonists or D₁ agonist plus A_2Aantagonist. Thus, the increase in pallidal c-fos expression may be attributable to the involvement of additional inputs to the globus pallidus. This may be attributable to an increased activity in the excitatory input from subthalamic nucleus. It has been found that NMDA receptor antagonists inhibit the induction of pallidal Fos immunoreactivity after combined administration of D₁ and D₂ agonists (Paul et al., 1992, 1995). Thus, it might turn out that concomitant stimulation of an excitatory input and inhibition of striatopallidal neurons act in synergy to increase c-fosin globus pallidus.

Conclusion

Although c-fos generally is used as a neuronal activation marker, we demonstrate here that basal c-fosexpression is upregulated by a D₁ agonist butdownregulated by a D₂ agonist or an A_2A antagonist. This suggests that c-fos mRNA levels may be used as an indicator of inhibition as well as activation of a neuronal pathway. Synergistic effects have been observed in the striatal output pathways after coadministration of D₁ plus D₂ agonists or D₁ agonist plus A_2A antagonist, providing evidence for important interactions between these parallel pathways. This work gives a basis for further investigations to elucidate the mechanisms whereby these synergistic effects occur, especially in the globus pallidus.