Introduction

Top Abstract Introduction Methods Results Discussion References

Receptors for the neurotransmitter DA fall into two types, the D₁ and D₂ receptor subfamilies,¹ distinguished biochemically by their opposing actions on adenylate cyclase (Kebabian and Calne, 1979). These two receptor types are also distinguished by many pharmacological agents. The prototypical selective D₁ agonist, the substituted benzazepine SKF 38393 (Setler et al., 1978) and its selective antagonist derivative SCH 23390 (Hyttel, 1983; Iorio et al., 1983) have allowed the specific pharmacological investigation of D₁ receptor functions (Clark and White, 1987). SKF 38393 is selective for D₁ over D₂ receptors in in vitro binding assays (Andersen et al., 1985; Lovenberg et al., 1989; Sibley et al., 1982) and has been widely used in vivo to examine D₁ receptor influences on behavior (Braun and Chase, 1986; Gershanik et al., 1983; Molloy and Waddington, 1984; Setler et al., 1978), as well as influences on the physiology of the basal ganglia, several nuclei of which express D₁ receptors (Boyson et al., 1986; Dearry et al., 1990; Yung et al., 1995). SKF 38393 has been shown to inhibit striatal neurons when applied iontophoretically (Hu and Wang, 1988), and systemic administration excites subthalamic nucleus neurons (Kreiss et al., 1996) and potentiates the excitatory effect of D₂ agonists on GP neurons, although alone it is without consistent effect on these latter cells (Carlson et al., 1987a; Walters et al., 1987). Notably, SKF 38393 does not have consistent or robust effects on DAergic neurons of the SNPC in normal animals (Carlson et al., 1987b; Wachtel et al., 1989), in accord with the D₂ receptor subfamily classification of the autoreceptors on these cells (Mansour et al., 1990; Sokoloff et al., 1990; Weiner et al., 1991).

Although selective for D₁ receptors, SKF 38383 has the disadvantage of being a partial agonist in terms of in vitro adenylate cyclase stimulation, with an intrinsic activity of 45% to 70% of dopamine (Andersen and Jansen, 1990; Arnt et al., 1992; O'Boyle et al., 1989; Setler et al., 1978). Also, it has been suggested that SKF 38393 is poorly absorbed into the brain, based primarily on the greater potency of its more lipophilic 3-N-substituted analogs in causing behavioral activation after peripheral administration (Arnt et al., 1992; Murray and Waddington, 1989). Therefore, the pattern of effects of SKF 38393 on basal ganglia physiology, in particular those cases in which a significant effect is absent, could be due to some extent to the partial agonist nature of this drug or, in situations with systemic treatment, due to poor central absorption.

Since the introduction of SKF 38393, several new compounds have been proposed to be full, selective D₁ agonists. These include the benzazepines SKF 82958 (chloro-APB) and SKF 81297 (6-chloro-PB; Murray and Waddington, 1989; O'Boyle et al., 1989; Pfeiffer et al., 1982), the benzo[a]phenanthridine DHX (Lovenberg et al., 1989) and the isochroman A-77636 (Kebabian et al., 1992). These drugs are typically reported as having intrinsic activities for adenylate cyclase stimulation similar to DA (Andersen and Jansen, 1990; Arnt et al., 1992; Gilmore et al., 1995; Izenwasser and Katz, 1993), although some studies have described SKF 82958, DHX and A-77636 as actually having higher activity than DA (Kebabian et al., 1992; Lovenberg et al., 1989; Mottola et al., 1992; O'Boyle et al., 1989). Although still D₁ receptor preferring, these latter three compounds have less D₁/D₂ receptor selectivity in vitro than SKF 38393 (Kebabian et al., 1992; Lovenberg et al., 1989; Mottola et al., 1992); SKF 81297 appears to have a selectivity comparable to that of SKF 38393 (Andersen and Jansen, 1990; Arnt et al., 1992).

These full D₁ agonists have been used to further investigate D₁ receptor actions in a variety of experimental preparations. However, few studies have investigated the effects of systemic administration of any of these D₁ compounds on basal ganglia electrophysiology (Heidenreich et al., 1995; Kreiss et al., 1996; Nichols et al., 1992). Hence, the present study has extended the characterization of full D₁ agonist systemic treatment on basal ganglia firing rates, with two major goals. First, the D₁/D₂ receptor selectivity of the full D₁ agonists has been primarily examined in vitro, so it remains unclear whether the systemic doses used for in vivo studies are D₁ receptor selective. Therefore, we have screened these four full D₁ agonists for D₂ activity (in a dose range shown to be maximal for D₁ receptor effects on behavior; Arnt et al., 1992; Darney et al., 1991; Kebabian et al., 1992) by examining effects on GP and SNPC DA neuron firing rates. Drugs with D₂ receptor activity act on postsynaptic D₂ receptors to increase GP firing rates, and they do so in a manner that is dependent on and potentiated by D₁ receptor activity (Carlson et al., 1986; Carlson et al., 1987a). Because presynaptic D₂ autoreceptors are more sensitive to agonists than postsynaptic D₂ receptors (Bergstrom et al., 1986; Carlson et al., 1987a; Skirboll et al., 1979), testing these D₁ agonists for inhibition of SNPC DA neurons offers an even more stringent test for D₂ receptor activity. The second goal was to compare the four full D₁ agonists with the partial agonist SKF 38393 by testing these full agonists in a protocol in which SKF 38393 pretreatment has been demonstrated to potentiate the rate-increasing effects of a subsequently administered D₂ agonist on GP unit activity (Carlson et al., 1987a; Walters et al., 1987).

	Methods

Top Abstract Introduction Methods Results Discussion References

Extracellular single-unit recordings were performed in neurologically intact male Sprague-Dawley rats (Taconic Farms, Germantown, NY), weighing 350 to 450 g, as previously described (Bergstrom and Walters, 1981; Bunney et al., 1973). Dopaminergic cells of the SNPC were recorded in rats anesthetized with chloral hydrate (400 mg/kg i.p.). Supplemental chloral hydrate was given as needed during the experiments. Because general anesthesia has been shown to greatly attenuate the effects of DA agonists on the GP (Bergstrom et al., 1984), GP neurons were recorded in locally anesthetized, paralyzed rats. Under halothane anesthesia, rats were tracheotomized, the trachea was intubated with a cannula and incision sites and pressure points were thoroughly infiltrated with the long-acting local anesthetic mepivacaine HCl. Corneal drying was prevented with the application of Lacri-Lube (Allergan Pharmaceuticals, Irvine, CA). After being placed in a stereotaxic instrument, halothane anesthesia was discontinued, and rats were paralyzed with the injection of gallamine triethiodide (16 mg/kg) through a lateral tail vein. Rats then were artificially ventilated on room air at a rate adjusted to maintain expired CO₂ levels between 3.4% and 4.5%. Supplements of gallamine were given as needed during the experiments. Body temperature of both paralyzed and chloral hydrate-anesthetized rats was maintained at 36° to 38°C with a heating pad. All surgical procedures were in accordance with the National Institutes of Health Guide for Care and Use of Laboratory Animals (Cohen et al., 1985).

Glass microelectrodes were filled with 2 M NaCl containing 1% Pontamine Sky Blue, and microelectrode tips were broken back under microscopic control until in vitro tip impedance measured 2.6 to 6.0 M (at 135 Hz). Microelectrodes were stereotactically guided through drilled skull holes to the following coordinates: for the SNPC, 2.8 to 3.2 mm anterior to lambda, 1.8 to 2.2 mm lateral to the midline and 6.5 to 7.5 mm ventral to dura; for the GP, 0.8 to 1.2 mm posterior to bregma, 2.6 to 3.0 mm lateral to the midline and 5.0-7.0 mm ventral to dura. Electrical signals were passed through an Axoclamp 2A amplifier (Axon Instruments, Burlingame, CA) in bridge mode, and amplified signals were monitored with an oscilloscope and audio monitor. Single-unit activity was isolated with a window discriminator, and firing rate data were collected on a computer with Spike2 software (version 2.18, Cambridge Electronic Design, Cambridge, UK). Dopaminergic neurons in the SNPC were identified on the basis of their characteristic long-duration, biphasic (+/) or triphasic (+/) action potential waveforms. All selected GP units had biphasic type II (+/) waveforms and had basal firing rates above 10 Hz.

Drugs used include SKF 82958 [(±)-chloro-APB; (±)-N-allyl-6-chloro-2,3,4,5-tetrahydro-7,8-dihydroxy-1-phenyl-1H-3-benzazepine] HBr, SKF 81297 [(±)-6-chloro-PB; (±)-6-chloro-2,3,4,5-tetrahydro-7,8-dihydroxy-1-phenyl-1H-3-benzazepine] HBr, (R)-(+)-SCH 23390 [(R)-(+)-7-chloro-8-hydroxy-3-methyl-1-phenyl-2,3,4,5-tetrahydro-1H-3-benzazepine] HCl, (S)-()-eticlopride HCl and ()-quinpirole HCl were obtained from Research Biochemicals (Natick, MA). Haloperidol was obtained from McNeil Pharmaceuticals (Spring House, PA). DHX [trans-10,11-dihydroxy-5,6,6a,7,8,12b-hexahydrobenzo[a]phenanthridine] HCl was obtained from Interneuron Pharmaceuticals (Lexington, MA). A-77636 [(1R,3S)3-(1'-adamantyl)-1-aminomethyl-3,4-dihydro-5,6-dihydroxy-1H-2-benzopyran] HCl was obtained from Abbott Laboratories (Pomezia, Italy). All drugs were dissolved in distilled water, except SCH 23390, which was dissolved as a 10× stock solution in 0.001 N HCl before dilution to final volume with distilled water.

Drugs were administered after basal data had been collected for 2 to 5 min for SNPC cells and 5 to 10 min for GP neurons. All drugs (and vehicle) were administered i.v. through a lateral tail vein at 0.5 ml/kg, and doses refer to the weight of the salts. Agonist drugs were given at typically given at 1.0 or 5.0 mg/kg, except in dose-response experiments. SCH 23390, eticlopride and haloperidol were given at 0.5, 0.2 and 0.2 mg/kg, respectively. One unit was recorded per rat. At the end of recording, Pontamine Sky Blue was iontophoresed at 15 µA for 20 to 30 min to mark the recording site. Brains were removed and frozen and later sectioned for histological confirmation of the recording site.

Typically, for SNPC cells, the average rate from 10 to 60 sec postdrug was expressed as a percentage of the basal rate within the last minute before drug. For GP neurons, the average rate from 5 to 10 min after drug was expressed as a percentage of the 5 min before drug. Overall effects were analyzed with one- or two-way ANOVA, post hoc comparisons between vehicle and drug treatments were done using Dunnett's t tests and post hoc examinations of the effect of quinpirole and DA receptor antagonists on GP rates were performed with paired t tests. All post hoc tests were two-sided. Relationships between basal rate and drug effects were studied with Pearson correlations. Dose-response curves were evaluated with Allfit (version 2.7; NICHD, NIH).

	Results

Top Abstract Introduction Methods Results Discussion References

DAergic neurons of the SNPC. The basal firing rates of the DAergic units studied ranged from 1.0 to 6.3 Hz, with a mean ± SEM of 3.5 ± 0.2 Hz. The effect of 1 mg/kg i.v. bolus D₁ agonist administration on firing rate of DAergic neurons is illustrated in figure 1. Basal firing rates of the neurons in these treatment groups were not significantly different (F_4,39 = .24, N.S.). ANOVA of drug effects on firing rate 10 to 60 sec after injection revealed a significant effect of treatment group (F_4,39 = 26.1, P < .001). Post-hoc comparisons demonstrated that SKF 82958 (1.0 mg/kg) significantly inhibited DAergic units compared with vehicle treatment, whereas the other compounds were not significantly different from vehicle. Of 11 units, 10 were inhibited >30% by SKF 82958; 5 were inhibited by 90%. The average inhibition was 69 ± 8%. Although antagonist drugs were typically given 1 to 2 minutes after SKF 82958, some units were held without antagonist. Two of these units are shown in figure 2, A and B. These cells demonstrated a slow recovery of firing rate after the initial robust inhibition. The units in figure 2, A and B, are also examples of the inverse relationship between the effect of SKF 82958 and basal firing rate (i.e., cells with slower basal firing rates were typically more robustly inhibited by SKF 82958; r = .74, P < .02, data not shown). Analysis of cumulative dose-response curves to SKF 82958 indicated an ED₅₀ value of 0.70 ± 0.14 mg/kg (fig. 3).

View larger version (64K): [in this window] [in a new window]   Fig. 1.   The effects of D1 agonists on SNPC DAergic unit firing rate. All drugs were given intravenously at 1.0 mg/kg. The rate 10 to 60 sec post-drug is plotted as percent of basal rate. Vehicle, SKF 81297, DHX and A-77636 have little effect on firing rate. However, SKF 82958 causes a robust inhibition in most cells. Bars represent mean ± S.E.M.; points indicate single units. Number of cells per group is indicated at the bottom of the bars. ***, P < .001 Dunnett's post-hoc comparison with vehicle treatment.

View larger version (27K): [in this window] [in a new window]   Fig. 2.   Rate histograms of SNPC DAergic and GP units. A and B, Two units inhibited by SKF 82958. Both cells slowly recover; in addition, the slower unit (B) is more robustly inhibited. C, D1 receptor blockade by SCH 23390 (SCH) does not prevent the SKF 82958 inhibitory effect, which is reversed by the D2 antagonist haloperidol (Hal). D, A cumulative dose-response curve of DHX (0.64-10.2 mg/kg at 1-min intervals) has little effect on this unit's firing rate within the 10 min after injection but a cumulative dose-response curve to quinpirole (Quin; 1.25-40 µg/kg at 1-min intervals) robustly decreases firing rate. This effect is reversed by eticlopride (Etic). E, This GP neuron is mildly excited after SKF 82958 but has a large rate increase after quinpirole. The example is one of the largest rate increases seen in this study with a D1 agonist/quinpirole combination. SCH 23390 has little effect on this excitation; however, a later injection of eticlopride reverses it. F, This cell demonstrates a large rate increase to quinpirole after SKF 81297 pretreatment, without a rate effect of the pretreatment alone. SCH 23390 effectively reversed the excitation. G, This cell has a large rate increase due to quinpirole after A-77636; this effect is reversed by SCH 23390. In all cases, rate is shown in spikes per 10 sec, and the units of time are minutes. SKF 82958, DHX, SKF 81297, A-77636 and quinpirole are all given at 1.0 mg/kg, except where indicated. SCH 23390 is given at 0.5 mg/kg. Haloperidol and eticlopride are given at 0.2 mg/kg. All injections are intravenous.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Cumulative dose-response curves for D1 agonist effects on SNPC DAergic neurons. Doses were doubled with successive injections at 1-min intervals up to the maximal dose or until the unit was completely inhibited. , SKF 82958; , DHX; , SKF 81297; , A-77636. Open symbols indicate no pretreatment, filled circles indicate 0.5 mg/kg SCH 23390 injection 10 min before dose-response curve. SCH 23390 pretreatment does not block the SKF 82958 effect, nor does it significantly shift the ED50 value, which is 0.70 ± 0.14 mg/kg for SKF 82958 alone and 0.49 ± 0.07 mg/kg for SCH 23390 plus SKF 82958. Other D1 agonists have minimal effect up to 10.2 mg/kg. Rates 10 to 60 sec postinjection are plotted as percent of basal rate. Points represent mean ± SEM. n = 4 or 5 for all groups except SKF 82958 alone (n = 7). Both fitted curves were assigned to 100% response for X = 0.

GP neurons. GP-type II unit basal firing rates ranged from 12 to 69 Hz, with a mean ± S.E.M. of 32 ± 1.7 Hz. The effects of 1.0 mg/kg D₁ agonist injection and subsequent D₂ agonist (quinpirole) injection are illustrated in figures 2 and 4. Basal firing rates of the neurons in these treatment groups were not significantly different (F_4,33 = 1.4, N.S.). A two-way ANOVA on firing rates 5 to 10 min postdrug showed no significant effect of D₁ agonist treatment (between-subjects factor; F_4,33 = 1.4, N.S.) but did reveal a significant effect of quinpirole treatment (within-subjects factor; F_1,36 = 57.7, P < .001). The interaction of factors (D₁ agonist treatment × quinpirole treatment) did not reach significance (F_9,28 = 2.0, P = .11, N.S.). Figure 4 demonstrates that vehicle or the D₁ agonists (1.0 mg/kg) alone had minor effects on rate in the large majority of units, although occasional cells had rate increases >30% (3 of 8 for SKF 81297; 2 of 8 for DHX) or rate decreases >30% (3 of 8 for SKF 81297). These varied rate responses to D₁ agonist alone were not significantly correlated to basal rate (r = .29, N.S.). Although quinpirole (1.0 mg/kg), given 10 min after vehicle, increased GP firing rate 32% on average, this effect was not significant in a post-hoc test. However, post-hoc comparisons indicated that quinpirole significantly increased GP firing rate when given 10 min after any of the D₁ agonists at 1.0 mg/kg (figs. 2, E-G, and 4). Furthermore, the mean rate increases in these groups (98-121%) were much larger than after vehicle/quinpirole and were qualitatively similar for all D₁ agonist/quinpirole combinations. Rate effects due to D₁ agonist/quinpirole combinations were inversely correlated with basal firing rate, such that units with slower basal rates tended to have greater percent increases in rate due to D₁/D₂ receptor activation (r = .41, P < .05, data not shown), a relationship that has also been noted for the GP rate effects of amphetamine (Bergstrom and Walters, 1981) and apomorphine (Ruskin DN and Walters JR, unpublished data). DHX/quinpirole was distinguished from other treatments in that some units had only marginal rate changes after this combination (3 of 8 units had <20% rate change; fig. 4).

View larger version (41K): [in this window] [in a new window]   Fig. 4.   The effects of D1 agonists at 1.0 mg/kg and subsequent quinpirole treatment on GP unit firing rate. The rate 5 to 10 min postdrug is plotted as percent of basal rate. Quinpirole (1.0 mg/kg) was given 10 min after vehicle or D1 agonist in all cases. Vehicle and the D1 agonists have no consistent effects on firing rate alone. However, quinpirole causes a large excitation of GP cells when given after any of the D1 agonists, although not after vehicle. Bars represent mean ± S.E.M.; points indicate single units. Numbers of cells per group are indicated at the bottom of the bars. *, P < .05, **, P < .01, ***, P < .001, paired t tests for rate before and after quinpirole.

View larger version (33K): [in this window] [in a new window]   Fig. 5.   Effects of SKF 82958 and DHX at 5.0 mg/kg on GP unit firing rate. Drugs were given intravenously. The rate 5 to 10 min postdrug is plotted as percent of basal rate. At this higher dose, SKF 82958 causes rate increases in four of six units, an effect significantly different from vehicle injection. DHX has no significant effect alone at this higher dose, but quinpirole (1.0 mg/kg, given 10 min afterward) does significantly increase firing rates. Bars represent mean ± S.E.M.; points indicate single units. Numbers of cells per group are indicated at the bottom of the bars. , P < .05 Dunnett's post-hoc comparison with vehicle treatment; *, P < .05 paired t test for rate before and after quinpirole.

	Discussion

Top Abstract Introduction Methods Results Discussion References

Most previous studies have used the partial D₁ agonist SKF 38393 to characterize D₁ receptor influences on the electrophysiology of GP neurons and the DAergic neurons of the SNPC. The data presented here using full D₁ agonists strongly support this characterization. Systemic SKF 38393 has little effect on the rate of SNPC DAergic units (Carlson et al., 1987b; Wachtel et al., 1989). SKF 81297, DHX and A-77636 are similarly without effect in a dose range that is maximal for D₁ receptor stimulation as measured behaviorally (Arnt et al., 1992; Darney et al., 1991; Kebabian et al., 1992), and although SKF 82958 strongly inhibited these cells, this action is D₂ receptor mediated (see below). D₁ receptor agonists also fail to change SNPC firing rate when applied iontophoretically (Carlson et al., 1987b; Wachtel et al., 1989) and do not modulate DA synthesis in vivo (Brooderson et al., 1990; Brown et al., 1985; Wachtel et al., 1989) or DA release from mesostriatal terminals in vitro (Lehmann et al., 1983). Modulatory effects of D₁ agonists on SNPC neuron rate have been shown in some preparations (Kelland et al., 1988; Momiyama et al., 1993), but robust D₁ receptor-mediated effects on firing rate occur only after DA depletion in awake, paralyzed rats (Huang and Walters, 1992; Sun et al., 1993), and this action is not mediated by local nigral receptors (Sun et al., 1996). In the GP, SKF 38393 has no consistent influence on type II unit firing rate but will potentiate the excitatory actions of a D₂ agonist (Carlson et al., 1987a; Walters et al., 1987), effects also seen with the full D₁ agonists. This same pattern has been demonstrated for GP immediate-early gene expression (Marshall et al., 1993; Ruskin and Marshall, 1995). The present data indicate that the lack of effect of SKF 38393 (when given alone) on pallidal and SNPC DA neuron firing rate is not simply a consequence of the partial D₁ receptor activity of this compound. A similar conclusion applies to the synergism of SKF 38393 with D₂ agonists in causing rate increases in the GP. Therefore, the current results show that the electrophysiological influences of D₁ receptor activation, as studied with the systemic administration of SKF 38393, are not idiosyncratic to this drug or to substituted benzazepines but are characteristic of the activation of D₁ receptors generally.

It should be noted that peripherally administered D₁ agonists can have significant effects alone on some basal ganglia nuclei, other than the GP. Specifically, D₁ agonists increase firing rate in both the ventral pallidum and subthalamic nucleus (Kreiss et al., 1996; Maslowski and Napier, 1991). Hence, the actions of D₁ receptors modulating activity in these nuclei are not dependent on simultaneous agonist-induced activation of D₂ receptors, although they clearly can be modulated by D₂ receptors (Heidenreich et al., 1995; Kreiss and Walters, 1997). In another basal ganglia nucleus, the substantia nigra pars reticulata, the pattern of D₁/D₂ receptor influence in neurologically intact rats more resembles the pattern found for GP type II units (i.e., little effect of D₁ agonists alone but significant firing rate changes due to concomitant D₁/D₂ receptor activation), although these latter rate changes include both increases and decreases (Walters et al., 1992; Waszczak et al., 1984).

Although SKF 81297, DHX and A-77636 were apparently free of D₂ agonist activity within the tested dose range, peripherally administered SKF 82958 had a novel inhibitory action on the firing rate of SNPC DAergic neurons. Although it is possible that this effect is mediated via the full efficacy activation of D₁ receptors by this drug, such an explanation is unlikely because (1) the effect was neither prevented nor reversed by the D₁ antagonist drug SCH 23390, (2) it was completely reversed in virtually every tested unit (17 of 18) by either of the D₂ antagonists haloperidol or eticlopride and (3) it was not reproduced by the other full D₁ agonists examined. These results strongly suggest that the inhibitory effect is due to an agonist action of SKF 82958 on D₂ subfamily receptors. In addition, the effect has characteristics typical of autoreceptor activation, with the response being rapid, inhibitory and desensitizing, and the size of the response being inversely related to basal firing rate (Aghajanian and Bunney, 1973; Staunton et al., 1980; White and Wang, 1984). Therefore, intravenous SKF 82958, at the dose range presently studied, activates inhibitory D₂ autoreceptors. It remains to be determined whether these doses also have this D₂ receptor effect when injected with less potent routes of administration, such as subcutaneously or intraperitoneally.

SKF 82985 also demonstrated apparent postsynaptic D₂ receptor activation. At 5.0 mg/kg, but not 1.0 mg/kg, SKF 82958 significantly increased GP unit firing rates, and where tested, these increases were reversed by eticlopride. However, even at 5.0 mg/kg, only two thirds of the units were excited, suggesting that this dose of SKF 82958 does not fully activate postsynaptic D₂ receptors. Given that the autoreceptor effect of this drug had an ED₅₀ value of 0.70 mg/kg, SKF 82958 clearly has a lower potency at postsynaptic D₂ receptors, in accord with the lower sensitivity of this receptor population (Bergstrom et al., 1986; Carlson et al., 1987a; Skirboll et al., 1979). The other full agonists examined showed no evidence of postsynaptic D₂ agonist activity at the present doses. The lack of postsynaptic D₂ receptor activity of these compounds is emphasized when it is considered that because D₁ receptor activation greatly potentiates the GP rate-increasing effects of D₂ agonists, the full D₁ receptor activity of these drugs would amplify any postsynaptic D₂ receptor agonism and so increase the probability of causing an increase in firing rate. Because SKF 81297 and SKF 38393 (each with no 3-N substitution) have good D₁/D₂ selectivity, the comparatively lower D₁/D₂ selectivity of SKF 82958 presumably involves the 3-N-allyl substituent of this compound.

SKF 82958 also differs from A-77636 and SKF 81297 in that the pallidal firing rate increases due to this drug (when combined with a D₂ agonist) are only partially reversed with 0.5 mg/kg SCH 23390. A similar pattern is seen in the subthalamic nucleus, wherein this dose of SCH 23390 only partially reverses SKF 82958-induced rate increases but completely reverses those due to SKF 38393 (Kreiss et al., 1996). Because studies have typically reported lower K_i values for SKF 82958 displacement of SCH 23390 binding compared with those for A-77636, SKF 81297 and SKF 38393 (Andersen and Jansen, 1990; Mottola et al., 1996), 0.5 mg/kg may be a dose of SCH 23390 that only partially displaces 1.0 mg/kg SKF 82958 but more completely displaces 1.0 mg/kg of the other agonists. However, this dose of SCH 23390 also only partially reverses the rate effects of a lower dose of SKF 82958 (0.43 mg/kg; Kreiss et al., 1996). In molar terms, under these conditions SCH 23390 HCl is present at ~1.5 times the dose of SKF 82958 HBr. Therefore, the data may reflect a "stickiness" of SKF 82958, such that other compounds poorly compete with this drug for the D₁ receptor once it has bound, although this binding is clearly not irreversible. Further experiments will be needed to distinguish between these possibilities.

Some results suggest that although D₁ receptor preferring, DHX does have significant activity at D₂ receptors. This drug competes with high affinity for [³H]spiroperidol binding (Mottola et al., 1992); it inhibits stimulated adenylate cyclase activity in a manner antagonized by the D₂ antagonist sulpiride (Mottola et al., 1992); and it blocks D₂ receptor-mediated inhibition of stimulated DA release in cultured cells (Kilts et al., 1996). Yet DHX is free of apparent D₂ receptor effects in the present study, since it did not inhibit SNPC DAergic units, block quinpirole-induced inhibition of these neurons or excite pallidal units when given alone. Although the effects of combined quinpirole and DHX (at both 1.0 and 5.0 mg/kg) on GP firing rates differed from the effects of other D₁/D₂ agonist combinations, the differences were subtle. The lack of presynaptic D₂ agonist and antagonist activity here confirms a preliminary electrophysiological study of intravenous DHX effects on SNPC DAergic units (Nichols et al., 1992). It has been hypothesized that DHX acts as an antagonist at D₂ receptors linked to K⁺ channels (characterized in pituitary lactotrophs), and as an agonist at those linked to adenylate cyclase (Smith et al., 1996). Increased K⁺ current appears to mediate most electrophysiological effects of autoreceptor activation on midbrain DAergic neurons (Kim et al., 1995; Lacey et al., 1987; Lacey et al., 1988). Therefore, it is surprising that DHX does not antagonize the inhibitory effect of quinpirole on these neurons. It is possible that the specific nature of the interaction among DHX, D₂ receptors and K⁺ channels depends on the particular D₂ receptor isoforms or G protein alpha subunit subtypes expressed in the cell type being studied.

A substantial body of work indicates that partial and full D₁ agonists can have equal efficacies with respect to physiological measures besides adenylate cyclase activation. Iontophoretic application of partial and full D₁ agonists results in similar inhibitions of nucleus accumbens septi neurons in vivo (Johansen et al., 1991), and systemic administration of partial and full agonists results in comparable rate increases in the ventral pallidum (Heidenreich et al., 1995) and also in the subthalamic nucleus (Kreiss et al., 1996). Similarly, the rank order for potency and/or efficacy of D₁ agonists for inducing behavioral activation is markedly different from the rank order of efficacy for adenylate cyclase activation (Arnt and Hyttel, 1988; Arnt et al., 1992; Arnt and Perregaard, 1987; Murray and Waddington, 1989). In some of these studies, agonists substantially weaker than SKF 38393 (e.g., SKF 75670 and SKF 83959) are as effective behaviorally as full agonists. The findings presented here support this dissociation, since the full D₁ agonists (when combined with a D₂ agonist) had rate-increasing actions on pallidal neurons that were similar in magnitude to that seen with the partial agonist SKF 38393 (Carlson et al., 1987a; Walters et al., 1987). Furthermore, this magnitude of rate increase is also seen after release of the endogenous agonist (DA) by amphetamine (Bergstrom and Walters, 1981; Carlson et al., 1986). Because full and partial agonists can have equivalent maximal effects in systems with spare receptors (for a review, see Ruffolo, 1982), it may be that the D₁ receptors mediating many of these behavioral and electrophysiological actions include a substantial receptor reserve. Indeed, there is evidence for spare adenylate cyclase-linked D₁ receptors in the striatum (Battaglia et al., 1986; Hess et al., 1987). There is also electrophysiological evidence for a D₁ receptor reserve. Selective reduction of D₁ receptor number by 80% (caused by systemic EEDQ with eticlopride pretreatment) does not affect the ability of the mixed D₁/D₂ agonist apomorphine to increase GP firing rates, whereas a 60% to 70% reduction in D₂ receptor number (by EEDQ with SCH 23390 pretreatment) abolished this effect of apomorphine (Walters et al., 1988).

Alternatively, the D₁ receptors in question may be acting through an effector system other than adenylate cyclase. The existence of D₁ receptors not coupled to adenylate cyclase was proposed soon after the introduction of D₁ receptor-specific ligands (Andersen et al., 1985; Mailman et al., 1986), and some electrophysiological effects of D₁ agonists are independent of the adenylate cyclase system (Harvey and Lacey, 1996; Johansen et al., 1991; Martin and Waszczak, 1994), although others clearly are not (Cameron and Williams, 1993; Hernandez-Lopez et al., 1997; Schiffmann et al., 1995; Surmeier et al., 1995). Recent evidence demonstrates a link between D₁ subfamily receptors and phosphatidylinositol hydrolysis (Friedman et al., 1997; Mahan et al., 1990; Pacheco and Jope, 1997; Wang et al., 1995), although studies have not yet investigated the effects of full D₁ agonists on this effector system, nor have they investigated the involvement of this system in the electrophysiological effects of D₁ subfamily receptors. It will be of particular interest to characterize the potencies and efficacies of various D₁ agonists for the stimulation of phosphatidylinositol hydrolysis.

The present findings support the role of D₁ receptors in basal ganglia electrophysiology as previously characterized with SKF 38393. However, the full D₁ receptor agonists tested here are distinguishable pharmacologically. SKF 82958 has D₂ agonist activity, most clearly demonstrated at rate-inhibiting autoreceptors. Because of this effect, and because of evidence demonstrating direct interactions of this drug with voltage-activated K⁺ channels (Nisenbaum et al., in press), SKF 82958 effects should be interpreted as selectively D₁ receptor-mediated only with appropriate pharmacological controls. Although apparently free of D₂ receptor effects in the present study, DHX has been shown to have a complex interaction with this receptor subfamily in addition to its D₁ receptor activity (Mottola et al., 1992; Smith et al., 1996). In contrast, the isochroman A-77636 and the benzazepine SKF 81297 are apparently highly selective D₁ receptor ligands in vivo, based on their lack of inhibition of SNPC DAergic cells and (when given alone) lack of excitation of GP neurons. Recently, these four full DAergic agonists have been tested for therapeutic potential in unilateral (Johnson et al., 1995; Vermeulen et al., 1993) and bilateral (Blanchet et al., 1996; Gnanalingham et al., 1995; Kebabian et al., 1992; Pearce et al., 1995; Taylor et al., 1991) experimental models of Parkinson's disease in primates, with each drug demonstrating differing antiparkinsonian effects, and well as differing adverse side effects (dyskinesias, tachyphylaxis). It is likely that the particular spectrum of dopamine receptor activity of each of these drugs contributes to these differences and will mark their usefulness as palliative agents in clinical Parkinson's disease.