Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The basal ganglia functional model predicts that dopamine, by stimulating the striatonigral pathway via D₁ receptors and inhibiting the striatopallidal pathway via D₂ receptors, should inhibit output from the substantia nigra pars reticulata and internal pallidal segment (SNpr/GPi). Inhibition of output from the SNr/GPi should in turn disinhibit the thalamus to facilitate movement (for review, see Alexander and Crutcher, 1990). Many of the predictions of the model have been supported by electrophysiological studies in humans with Parkinson's disease and animal models of this disorder (Filion and Tremblay, 1991; Bergman et al., 1994; Papa et al., 1999; Levy et al., 2001). However, the question of whether these predictions are valid and supported in normal animals without dopaminergic lesions has been less rigorously addressed.

A definitive test of this hypothesis has been made difficult by the complexity of the circuitry, in particular by the fact that D₁ and D₂ receptors are expressed not only within the striatum but also in other nuclei of the basal ganglia that directly or indirectly influence the SNpr/GPi (Smith and Kieval, 2000). Dopaminergic drugs, when administered systemically, may therefore act at sites besides the striatum to modulate the basal ganglia output nuclei, thereby preventing a clear assessment of the specific role of striatal dopamine receptors in the responses. Another problem arises when attempting to distinguish the individual roles of D₁ and D₂ receptors because a synergistic interaction between them is normally required for generating a "full" behavioral or electrophysiological response to dopamine (Arnt et al., 1987; Walters et al., 1987; White et al., 1988; Waddington and Daly, 1993; White and Hu, 1993). D₁/D₂ synergism complicates defining the roles of each receptor even when using D₁ and D₂ receptor-selective agonists because the unintended receptor can still be stimulated by endogenous dopamine to yield a mixed D₁/D₂ response. For these reasons, in studies where dopamine agonists have been given systemically to animals with an intact dopamine system, it has not been possible to attribute a particular behavioral or electrophysiological response to activation of a specific receptor type at a specific location in the circuitry. Moreover, few attempts have been made to correlate the behavioral consequences of striatal D₁ and/or D₂ receptor stimulation with the electrophysiological consequences of that same manipulation.

The goal of our studies has been to fill these gaps by assessing how discrete stimulation of striatal D₁ and D₂ dopamine receptors, individually and concurrently, influences behavioral and electrophysiological output from the basal ganglia in normal rats via the SNpr/GPi. To circumvent the problems posed by systemic routes of administration, we and others have adopted the use of microinjections of dopaminergic drugs directly into the corpus striatum. In agreement with previous reports (Kelley et al., 1988; Wang and Rebec, 1993; Dickson et al., 1994), we found that bilateral (but not unilateral) infusions into the rat ventral-lateral striatum (VLS) of the dopamine releaser d-amphetamine (20 µg/µl/side) cause a robust behavioral activation consisting of primarily oral movements (biting, licking, tongue protrusions, and jaw movements) and sniffing. However, contrary to a key prediction of the basal ganglia model, we showed in parallel electrophysiological studies that these same infusions do not cause an inhibition of firing of SNpr neurons. In fact, during the period of peak behavioral activation, we noted extremely variable changes in the firing of SNpr neurons, but the average response was a significant increase in SNpr activity to approximately 120% of basal firing rates (Martin et al., 1997; Waszczak et al., 2001). The variable nature of the responses and the overall excitation of SNpr cell firing were not the result of anesthesia because similar results were obtained in both anesthetized rats and animals that were awake, locally anesthetized, and paralyzed.

To determine the relative contributions of D₁ and D₂ receptors to these responses, we have now used bilateral infusions into VLS of drug combinations intended to selectively stimulate D₁, D₂, or both receptors concurrently. Several strategies were considered to prevent endogenous dopamine from stimulating the unintended receptor. 6-Hydroxydopamine lesions of the nigrostriatal dopamine neurons, or pretreatment of the animals with reserpine or -methyl-p-tyrosine (AMPT), have been used in previous studies to destroy or deplete endogenous dopamine stores. However, chronic dopamine depletion by either of these treatments seriously disrupts the normal interactions between neurons within the circuitry (Huang and Walters, 1994), altering their expression of dopamine receptors and leading to an uncoupling D₁/D₂ synergism (Hu et al., 1990; LaHoste et al., 1993; Marshall et al., 1993). Such conditions would be inappropriate for assessing the functional roles of striatal D₁ or D₂ receptors in the normosensitive system. Acute (2-4 h) depletion of dopamine by AMPT, which avoids the disruptions of long-term depletion, has been valuable in demonstrating the phenomenon of D₁/D₂ synergism (Walters et al., 1987; White et al., 1988; Hu et al., 1990) and was a reasonable approach for our studies. However, AMPT causes a generalized central and peripheral (autonomic) catecholamine depletion that might interfere in the behavioral responses to striatal dopamine agonist infusions. Consequently, we chose a strategy where selective stimulation of each receptor was accomplished by coadministration of an agonist at one receptor together with an antagonist at the unintended receptor. The antagonist would block the effects of both endogenous dopamine and any nonselective effects of each agonist at the opposing dopamine receptor, thereby allowing stimulation of only a single receptor in the area of the infusion.

To selectively activate D₁ receptors, we infused a solution containing the full-efficacy D₁ agonist SKF 82958 [(±) 6-chloro-APB hydrobromide] (O'Boyle et al., 1989) together with the potent and selective D₂ antagonist YM 09151-2 (Terai et al., 1983; Niznik et al., 1985). In this case, although SKF 82958 displays only about 10-fold selectivity for D₁ versus D₂ receptors (Murray and Waddington, 1989; Gnanalingham et al., 1995), any D₂ agonist activity should have been prevented by coinfusion of the D₂ blocker. To selectively stimulate D₂ receptors, we infused the D₂ agonist quinpirole (Tsuruta et al., 1981) plus the selective D₁ antagonist SCH 23390 (Hyttel, 1983; Iorio et al., 1983). For concurrent stimulation of both D₁ and D₂ receptors, we infused either the mixed D₁/D₂ agonist apomorphine, or SKF 82958 and quinpirole together. As in our earlier studies with amphetamine, we monitored the effects of individual and concurrent stimulation of the two receptors on behavior and electrophysiological output from the SNpr in separate but parallel studies following identical drug infusion paradigms.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Striatal Drug Infusions. Male rats (250-275 g) were implanted bilaterally with 23-gauge stainless steel guide cannulae into the VLS (coordinates: A, 9.8 mm to ; L, 3.8 mm; V, 3.3 mm) 1 week before behavioral testing or just before electrophysiological experiments. For the placement of chronic guide cannulae (behavioral studies only), three screws were set in the skull near the cannula, and the assembly was secured to the skull by using dental acrylic. For striatal drug or saline infusions, a 30-gauge injection cannula was inserted through the guide, and extended 3 mm beyond it, so that its tip reached the target site in the VLS. Infusions were made from a pair of 10-µl Hamilton syringes by using a Harvard Apparatus dual infusion pump. For behavioral studies, the pump was remotely activated and controlled by a computer to avoid behavioral effects due to handling. After completion of experiments, rats were sacrificed and cannulae placements in the VLS were confirmed histologically.

Behavioral Monitoring Techniques. Behavior was monitored in a test chamber with clear plastic walls, a glass floor (50 × 40 cm), transverse infrared photobeams 10 cm from the end walls, and a video camera mounted below the cage floor. The chamber was housed in a small, darkened room separated from the main laboratory. The rat was handled briefly at the beginning of each test session to remove the dummy cannulae and insert the bilateral injection cannulae through the guides. The injection cannulae were attached via thin polypropylene tubing to a liquid swivel held by a spring arm over the chamber, and the swivel was in turn connected to syringes controlled by the infusion pump. The entire infusion assembly was out of reach of the rat, but allowed the animal complete freedom of movement within the chamber. After a 20-min acclimation period, the infusion pump was remotely activated by the experimenter from outside the testing room, and a drug combination or saline (1 µl/side) was infused bilaterally over 2 min into the VLS. Behavior was monitored for 45 min by videotaping through the cage floor, and by computer acquisition of consecutive (A/B) photobeam breaks. All rats were observed during two baseline sessions (no infusions) and were then randomly given two infusions of each drug combination and saline with at least a 2-day washout between trials. In one group of rats, drug combinations were SKF 82958 (10 µg/µl/side) + YM 09151-2 (2 µg/µl/side) and SKF 82958 (10 µg/µl/side) + quinpirole (30 µg/µl/side). In a second group of rats, drug combinations were quinpirole (30 µg/µl/side) + SCH 23390 (10 µg/µl/side) and quinpirole (30 µg/µl/side) + SKF 82958 (10 µg/µl/side).

Extracellular Single Unit Recording Techniques. Activities of SNr neurons were recorded in rats anesthetized with chloral hydrate (400 mg/kg i.p.). All surgical procedures were carried out in strict compliance with the National Research Council's Guide for the Care and Use of Laboratory Animals, 1996. We have previously shown that chloral hydrate anesthesia does not blunt or otherwise alter the nature of responses of SNpr neurons to striatal infusions of d-amphetamine, a dopamine releaser (Martin et al., 1997; Waszczak et al., 2001). SNr neurons were located within the pars reticulata region of the substantia nigra, within the following stereotaxic coordinates: L, 2.0 to 2.4 mm; A, 2.8 to 3.2 mm; V, >7.0 mm. These neurons are distinguished electrophysiologically by their sharp, biphasic extracellular waveforms, duration (<1 ms), firing rates (10-40 spikes/s), and location just ventral to the pars compacta dopamine neurons, according to criteria described previously (Waszczak et al., 1980).

Drugs. All drugs used in these studies were obtained from Sigma Chemical (St. Louis, MO) and were prepared fresh before use.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effects of Striatal Infusion of D₁/D₂ Receptor Agonist Apomorphine on Behavior and SNpr Cell Firing. Bilateral infusions of the mixed D₁/D₂ receptor agonist apomorphine (5 and 10 µg/µl/side) into VLS caused a robust, dose-dependent behavioral activation with an onset ranging from immediate to 20-min postinfusion and persisting for at least 40 min after the infusions (Fig. 1). At the 10-µg/µl/side dose, animals tended to be stationary initially, and engaged primarily in intense oral movements and sniffing. By 20-min postinfusion, these behaviors began to decrease somewhat and locomotor activity began to increase. Both the 5- and 10-µg/µl/side doses of apomorphine caused significant increases in oral movements (two-way ANOVA: p < 0.0001, F = 19.84, df = 2,135 at 5 µg/µl/side and p < 0.0001, F = 15.36, df = 2,135 at 10 µg/µl/side), sniffing (two-way ANOVA: p < 0.0001, F = 36.99, df = 2,135 at 5 µg/µl/side and p < 0.0001, F = 19.44, df = 2,135 at 10 µg/µl/side), and locomotor activity (data not shown; two-way ANOVA: p < 0.0001, F = 27.75, df = 2,20 at 5 µg/µl/side and p < 0.0001, F = 57.42, df = 2,20 at 10 µg/µl/side) relative to bilateral infusions of saline, or baseline behavior with no infusion. Grooming and rearing were not significantly altered at either dose (data not shown), and there was no significant effect of postinfusion time on any of the behaviors.

View larger version (24K): [in this window] [in a new window]   Fig. 1.   Behavioral responses of six rats after bilateral infusions of apomorphine (5 or 10 µg/µl/side) or saline into the ventral-lateral striatum. Each animal was tested twice with no drug infusion (baseline), and twice after randomized infusions of saline and each of the two doses of apomorphine. Oral movements and sniffing were scored by the experimenter upon watching videotapes of behavioral activity (see Materials and Methods). , baseline; , saline; , APO 5 µg; , APO 10 µg.

View larger version (33K): [in this window] [in a new window]   Fig. 2.   Responses of substantia nigra pars reticulata neurons after bilateral infusions of apomorphine (10 µg/µl/side) into the ventral-lateral striatum of chloral hydrate-anesthetized rats. The drug was infused over a 2-min interval indicated by a bar over each graph. Top, responses of individual neurons as a percentage of their preinfusion baseline firing rates. Bottom, average responses (±S.E.M.) of all neurons depicted in the top graph.

Effects of Selective Activation of Striatal D₂ Dopamine Receptors on Behavior and SNpr Cell Firing. To clarify the role of D₂ receptors in mediating the effects of the mixed D₁/D₂ agonist apomorphine, we used bilateral infusions into VLS of the D₂ agonist quinpirole (30 µg/µl/side) + the D₁ antagonist SCH 23390 (10 µg/µl/side). In contrast to the effects of apomorphine (see above) and amphetamine in previous studies, selective stimulation of D₂ receptors failed to elicit a behavioral response (Fig. 3). In fact, oral movements, sniffing, and locomotor activity (data not shown) tended to be suppressed at some time points after quinpirole + SCH 23390 infusions, relative to responses observed after bilateral infusions of saline, or baseline behavior with no infusion. Overall, there was no significant change in any of the behavioral measures after selective stimulation of striatal D₂ dopamine receptors.

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Behavioral responses of six rats after bilateral infusions of quinpirole (QUIN; 30 µg/µl/side) + SCH 23390 (SCH; 10 µg/µl/side), or quinpirole (30 µg/µl/side) + SKF 82958 (SKF; 10 µg/µl/side), or saline into the ventral-lateral striatum. Each animal was tested twice with no drug infusion (baseline), and twice after randomized infusions of saline and each of the two drug pairs. Oral movements and sniffing were scored by the experimenter upon watching videotapes of behavioral activity (see Materials and Methods). , baseline; , saline; , QUIN + SCH; , QUIN + SKF.

View larger version (29K): [in this window] [in a new window]   Fig. 4.   Responses of substantia nigra pars reticulata neurons after bilateral infusions into the ventral-lateral striatum of quinpirole (QUIN; 30 µg/µl/side) + SCH 23390 (SCH; 10 µg/µl/side) in chloral hydrate-anesthetized rats. The drug was infused over a 2-min interval indicated by a bar over each graph. Top, responses of individual neurons as a percentage of their preinfusion baseline firing rates. Bottom, average responses (±S.E.M.) of all neurons depicted in the top graph.

Effects of Selective Activation of Striatal D₁ Dopamine Receptors on Behavior and SNpr Cell Firing. To assess the effects of selective stimulation of striatal D₁ receptors on motor and electrophysiological output from the SNpr, we coinfused the D₁ agonist SKF 82958 (10 µg/µl/side) and the D₂ antagonist YM 09151-2 (2 µg/µl/side) into VLS. Bilateral infusions of this drug combination caused a specific behavioral response, i.e., a modest but highly significant increase in oral movements that was sustained during the 30 min after the infusion (two-way ANOVA: p < 0.0001, F = 18.87, df = 2,195 versus baseline or saline infusion; Fig. 5). The increase in oral movements, although significant, was only about one-third the magnitude of that observed after coinfusions of both the D₁ and D₂ agonists (see below). The behavioral activation was also discrete in that sniffing and locomotor activity were not increased relative to saline or baseline controls.

View larger version (20K): [in this window] [in a new window]   Fig. 5.   Behavioral responses of six rats after bilateral infusions of SKF 82958 (SKF; 10 µg/µl/side) + YM 09151-2 (YM; 2 µg/µl/side, or quinpirole (QUIN; 30 µg/µl/side) + SKF 82958 (SKF; 10 µg/µl/side), or saline into the ventral-lateral striatum. Each animal was tested twice with no drug infusion (baseline), and twice after randomized infusions of saline and each of the two drug pairs. Oral movements and sniffing were scored by the experimenter upon watching videotapes of behavioral activity (see Materials and Methods). , baseline; , saline; , SKF + YM; , QUIN + SKF.

View larger version (37K): [in this window] [in a new window]   Fig. 6.   Responses of substantia nigra pars reticulata neurons after bilateral infusions of SKF 82958 (SKF; 10 µg/µl/side) + YM 09151-2 (YM; 2 µg/µl/side) into the ventral-lateral striatum of chloral hydrate-anesthetized rats. The drug was infused over a 2-min interval indicated by a bar over each graph. Top, responses of individual neurons as a percentage of their preinfusion baseline firing rates. Bottom, average responses (±S.E.M.) of all neurons depicted in the top graph.

Effects of Concurrent Activation of D₁ + D₂ Receptors on Behavior and SNpr Cell Firing. In both groups of animals that were evaluated for behavioral responses to selective D₁ or D₂ receptor stimulation, we also examined the effects of concurrent stimulation of both receptors by coinfusions into VLS of quinpirole (30 µg/µl/side) + SKF 82958 (10 µg/µl/side). Bilateral infusions of both the D₁ + D₂ agonist produced a profound behavioral activation similar to that seen previously with apomorphine, and well in excess of responses to either D₁ or D₂ receptor activation alone (Figs. 3 and 5). Oral movements, sniffing, and locomotor activity (data not shown) were all significantly increased in both groups of animals given the combined agonist infusions [two-way ANOVA: for oral movements, p < 0.001, F = 174.25, df = 2,195 and p < 0.0001, F = 29.33, df = 2,135; for sniffing, p < 0.0001, F = 117.44, df = 2,195 and p < 0.001, F = 109.23, df = 2,135; and for locomotor activity, p < 0.0001, F = 12.30, df = 2,64 and p < 0.0436, F = 3.68, df = 2,20].

View larger version (39K): [in this window] [in a new window]   Fig. 7.   Responses of substantia nigra pars reticulata neurons after bilateral infusions of SKF 82958 (SKF; 10 µg/µl/side) + quinpirole (QUIN; 30 µg/µl/side) into the ventral-lateral striatum of chloral hydrate-anesthetized rats. The drug was infused over a 2-min interval indicated by a bar over each graph. Top, responses of individual neurons as a percentage of their preinfusion baseline firing rates. Bottom, average responses (±S.E.M.) of all neurons depicted in the top graph.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The basal ganglia functional model put forth more than a decade ago by Albin et al. (1989) and Gerfen et al. (1990) has provided a valuable framework for understanding how disruptions of striatal dopamine transmission give rise to the motor symptoms of basal ganglia disorders, most notably Parkinson's disease. It is now well accepted that striatal release of dopamine regulates motor function via its influence over striatal efferent pathways that ultimately control output from the SNpr/GPi. Specifically, the model postulates that D₁ and D₂ receptors are predominantly segregated to striatonigral and striatopallidal neurons, respectively, so that each receptor independently regulates the activity of one striatal efferent pathway. These basic elements of the model give rise to three testable hypotheses that we have attempted to address in these studies. First, selective stimulation of striatal D₁ receptors is expected to activate the striatonigral pathway to directly inhibit the firing of SNpr neurons. This response presumably involves D₁ receptor-mediated potentiation of cortical glutamatergic inputs to striatal neurons via N-methyl-D-aspartate receptors (Hernández-Lopez et al., 1997; Cepeda et al., 1998). Second, selective stimulation of striatal D₂ receptors on striatopallidal neurons should have no direct effect on SNpr neurons, but is expected to indirectly reduce their firing by withdrawal of excitation from the subthalamic nucleus inputs to the SNpr. Finally, if D₁ and D₂ receptors are segregated on the two striatal efferents, as proposed, then the phenomenon of D₁/D₂ synergism must be mediated by interactions between neurons possessing the different receptors, rather than by interactions between the two receptors on the same striatal neurons.

To test these basic predictions of the model, it is necessary to restrict stimulation to a single dopamine receptor subtype within the striatum, and only the striatum, thus ruling out the use of systemically administered dopamine agonists. We chose an acute means of preventing dopamine's activation of the unintended receptor within the field of the agonist infusion to avoid any disruptions to the circuitry. Finally, to determine whether SNpr output is suppressed when movement is facilitated, we monitored the effects of striatal dopamine receptor stimulation on SNpr cell firing under conditions that were known to elicit motor activation in behaving animals. We chose the VLS striatum as the target for our infusions because bilateral amphetamine infusions at this site are known to cause intense oral movements (Kelly et al., 1988; Delfs and Kelley, 1990; Dickson et al., 1994; Waszczak et al., 2001), a behavior postulated to specifically involve output from the SNpr (Childs and Gale, 1983; DeLong et al., 1983; Redgrave et al., 1984).

The results of our studies corroborate and extend those of our earlier work with bilateral infusions of d-amphetamine into the VLS (Martin et al., 1997; Waszczak et al., 2001). We now conclude that concurrent bilateral stimulation of both D₁ and D₂ dopamine receptors by the dopamine releaser d-amphetamine, the mixed D₁/D₂ agonist apomorphine, or coinfusion of a D₁ + D₂ agonist produces an intense motor activation consisting of repetitive mouth movements, tongue protrusions, and vacuous chewing movements. Sniffing and locomotor activity were also increased by each treatment. These behavioral changes were correlated temporally with changes in the activity of neurons in the SNpr. However, the changes in firing were extremely variable in direction and magnitude, and included both dramatic increases and decreases in SNpr unit activity. These fluctuations in firing were particularly pronounced after coinfusions of the D₁ agonist SKF 82958 and the D₂ agonist quinpirole (Fig. 7), and occurred during the time when behavior was maximally stimulated (Fig. 3 and 5). When the responses of all neurons were averaged, the net response to concurrent D₁ and D₂ receptor stimulation by each drug regimen was an increase in SNpr cell firing (to 117-133% of baseline firing rates). It is possible that the electrophysiological responses we observed might not mimic changes that occur in awake, behaving animals because cortical activation of striatal neurons may be dampened in anesthetized rats. It seems unlikely, however, that consistent inhibitory effects would have occurred even in conscious rats because similarly variable changes in firing and a net increase in SNpr cell activity were also seen in awake, locally anesthetized rats given bilateral infusions of amphetamine into the VLS (Waszczak et al., 2001).

Our electrophysiological results are in conflict with a fundamental prediction of the basal ganglia functional model that stimulation of both striatal D₁ and D₂ receptors should lead, independently and concurrently, to an inhibition of SNpr output. We are not the first to have observed this discrepancy. Other investigators who have examined changes in SNpr activity during spontaneous movement (Gulley et al., 1999), or in response to unilateral striatal infusions of apomorphine (Murer et al., 1997a,b), amphetamine (Olds, 1988; Timmerman et al., 1998; Gulley et al., 1999), or a combination of SKF 38393 and quinpirole (Murer et al., 1997a) have found a mixed pattern of responses. Although drug doses and striatal infusion sites differed, and each of these studies used unilateral rather than bilateral infusions, the results consistently showed that concurrent stimulation of striatal D₁ and D₂ receptors leads to variable changes in SNpr cell firing. Collectively, these findings and our own reveal that the output signal from the SNpr coincident with movement is normally a complex, highly variable one consisting of increases, decreases, and biphasic changes in activity.

To determine whether the complexity of the electrophysiological response was caused by opposing and competing influences of D₁ and D₂ receptor mechanisms on striatal output, behavioral and electrophysiological responses were evaluated after coinfusions of selective agonist-antagonist pairs, thereby restricting stimulation to a single dopamine receptor subtype. With the same doses of the D₁ and D₂ agonist that caused profound behavioral effects when given together, neither agonist alone caused a behavioral response of the magnitude observed with concurrent D₁/D₂ receptor stimulation. Indeed, selective D₂ receptor stimulation actually suppressed all forms of behavior, whereas selective D₁ receptor stimulation caused a modest increase in oral movements to roughly one-third of that observed after combined D₁/D₂ receptor stimulation. Oral movements may therefore be unique among the array of dopaminergic behaviors in that they can be elicited to a modest extent by D₁ receptor stimulation alone. These findings are in good agreement with the early literature concerning the behavioral effects of systemically administered D₁ and D₂ receptor-selective agonists. For instance, i.p. administration of the partial D₁ agonist SKF 38393 has been shown to cause a mild increase in oral movements (Johansson et al., 1987) similar to that which we observed after striatal coinfusion of SKF 82958 + YM 09151-2. Conversely, i.p. administration of low doses of quinpirole were found to inhibit all movement, including oral movements (Johansson et al., 1987; Delfs and Kelley, 1990; Canales and Iversen, 1998), similar to the suppression of behavior that we observed after coinfusions of quinpirole plus SCH 23390. Numerous studies have corroborated the modest behavioral effects of systemically administered D₁- and D₂-selective agonists when given alone (for reviews, see Clark and White, 1987; Waddington and Daly, 1993). The additional information contributed by our studies is that these behaviors, or the lack thereof, can be entirely attributable to striatal actions of the drugs, rather than mediated in part by their actions at extrastriatal D₁ and D₂ receptors.

Consistent with the lack of a behavioral response with selective D₂ receptor stimulation, we observed only minor fluctuations in SNpr cell firing and no net change from baseline rates. Conversely, the mild oral behavior seen after selective D₁ receptor stimulation correlated with moderate and variable changes in SNpr cell firing, and a slight increase in activity similar to that observed with concurrent D₁/D₂ receptor stimulation. It is conceivable that the modest effects of SKF 82958 we observed were not entirely due to D₁ receptor stimulation because this agonist exhibits only 10-fold D₁:D₂ receptor selectivity (Murray and Waddington, 1989; Gnanalingham et al., 1995). In fact, SKF 82958 has been shown to inhibit substantia nigra dopamine cell firing, a response attributed to D₂ autoreceptor stimulation and reversed by D₂ antagonists (Ruskin et al., 1998). It has also been found to inhibit slowly inactivating K⁺ currents in rat striatal neurons by a non-D₁ receptor mechanism (Gabel and Nisenbaum, 1998; Nisenbaum et al., 1998). In our studies, however, it is unlikely that D₂ receptor stimulation contributed to the responses observed because such effects should have been prevented by coinfusion of the D₂ antagonist YM 09151-2. We cannot rule out the possibility, however, that the agonist and antagonist might have differed in their spread in tissue, thereby exposing some striatal neurons to the agonist without concurrent blockade of the opposing receptor. Similarly, we cannot rule out the possibility that in some animals a small amount of the drug solution might have spread caudally into the adjacent pallidum, a nucleus of the "indirect" pathway.

Previous investigators who have examined the effects of unilateral infusions of selective D₁ and D₂ receptor agonists on SNpr cell firing have made observations similar to ours. Timmerman et al. (1998) reported that striatal infusions of neither the D₁ agonist CY 208243 nor the D₂ agonist quinpirole (LY171555) significantly altered the firing rates of SNpr neurons. Akkal et al. (1996) and Murer et al. (1997a) both found that a subset of SNpr neurons did exhibit changes in firing after intrastriatal infusions of SKF 38393 or quinpirole, but the responses consisted of both excitations and inhibitions and appeared to be generally quite modest. The low incidence of responding neurons and the minor changes in firing in these reports may have been due in part to much lower doses of the agonists than we used, and/or to the use of unilateral rather than bilateral infusions. However, as in our studies, infusions of the D₁ and D₂ agonists concurrently at the same doses given alone invariably caused greater responses than occurred with either drug individually. Moreover, the different magnitude of responses seen with unilateral and bilateral infusions suggests that the contralateral side can play a role in regulating output from the basal ganglia via the SNpr, a point made in our earlier report (Waszczak et al., 2001).

Several interpretations can be drawn from these data. First, stimulation of neither striatal D₁ nor D₂ receptors leads to a consistent inhibition of SNpr cell firing, as is predicted by the basal ganglia model. Indeed, when behavior is activated, SNpr neurons are more frequently excited than inhibited. Second, the complex and variable effects of concurrent D₁/D₂ receptor stimulation are not due to separate and competing effects on striatal output mediated by the individual receptors, because their individual effects on SNpr output are also variable, modest, and mildly excitatory. These findings draw into question fundamental predictions of the basal ganglia model, at least in its most simplistic form. Nevertheless, it is important to note that aspects of the model may still be valid under conditions different from those of our study. For instance, it is possible that the ventral lateral striatum (where our infusions were targeted) may differ from the dorsal striatum in the manner by which D₁ and D₂ receptors regulate striatal output, and in turn SNpr output. It is also conceivable that neurons of the SNpr and GPi, the two output nuclei of the basal ganglia, are not equivalent in their striatal inputs and therefore differ in their responses to striatal dopamine receptor stimulation. Because we did not evaluate GPi neurons, it is unclear whether these neurons were inhibited by our striatal drug infusions.

A third interpretation of our results is that electrophysiological and behavioral output from the basal ganglia requires coactivation of both striatal dopamine receptors for full expression, in accordance with the findings of numerous previous investigators (Arnt et al., 1987; Walters et al., 1987; White et al., 1988; Waddington and Daly, 1993; White and Hu, 1993). The synergism between the receptors draws into question another tenet of the basal ganglia model, i.e., that D₁ and D₂ receptors are segregated in their expression and independently control the activities of the two striatal efferent populations. For instance, if D₁ receptors predominate in regulating the activity of the "direct" striatonigral -aminobutyric acidergic pathway, as predicted by the model then our data suggest that D₁ receptor activation inhibits striatonigral efferents to a greater extent than it activates them, a conclusion contrary to predictions of the model. Conversely, if D₂ receptors exert predominant (inhibitory) control over the activity of striatopallidal neurons, as predicted by the model, then our data imply that striatal D₂ receptors and the indirect pathway must have little influence over the tonic activity of the SNpr, also contrary to predictions of the model. Indeed, neither the proposed functional roles of striatal D₁ and D₂ receptors nor the segregated pattern of their expression offers a tenable explanation for our findings, and both appear to be overly simplistic.

A more parsimonious explanation, supported by a growing body of evidence, is that D₁ and D₂ receptors coexist on a substantial proportion of striatonigral and striatopallidal neurons, and interact synergistically on these neurons to generate a complex, highly variable output to SNpr/GPi. Indeed, a significant colocalization of the two receptors on striatal efferents now seems clear. Surmeier et al. (1992, 1993, 1996) showed that both efferent groups respond electrophysiologically to both D₁ and D₂ receptor agonists, and many contain mRNA for both classes of receptor. Meador-Woodruff et al. (1991) and Lester et al. (1993) found that D₁ and D₂ mRNAs were coexpressed by 27 to 33% of striatal efferents, whereas Ariano et al. (1992) observed D₂ receptor immunoreactivity on a minimum of 60% of striatonigral neurons. More strikingly, Aizman et al. (2000) showed complete colocalization of the two receptors on neurons from embryonic rat striatum. The means of interaction between the coexpressed receptors giving rise to the dramatic and fluctuating changes in striatal output transmitted to the SNpr is less clear. Possible mechanisms include modulation of neuronal excitability by D₁/D₂ synergistic regulation of Na⁺ fluxes through tetrodotoxin-sensitive Na⁺ channels and the Na⁺/K⁺ ATPase (Bertorello et al., 1990; Aizman et al., 2000). By whatever the mechanism, our results suggest that stimulation of striatal D₁ and D₂ receptors gives rise to a dynamic and complex signal that is transmitted, in turn, to the thalamus and other premotor nuclei to facilitate movement. It is increasingly apparent that movement is signaled by discrete changes in the firing of subpopulations of SNpr/GPi neurons, rather than by a uniform inhibitory response of the entire population. In view of these findings, the basal ganglia functional model needs to be revised to accommodate the fact that, at least in normal animals with an intact circuitry, dopamine does not produce simple, unidirectional changes in the activities of striatal efferents or their targets in the SNpr/GPi.