Introduction

Top Abstract Introduction Methods Results Discussion References

DA is implicated in the modulation of several physiological functions, including endocrine secretion, motor behavior, thermoregulation and mood (Creese and Fraser, 1987). Its actions are expressed via at least five different receptor types. On the basis of their molecular structure and pharmacological properties, these receptors have been divided into two families: D₁-like, including D₁ and D₅ receptors, and D₂-like, including D₂, D₃ and D₄ receptors (Sokoloff and Schwartz, 1995; Strange, 1993). As concerns the D₂ receptor family, the low level of expression of mRNA encoding D₄ receptors in rodent and human tissue, and the hitherto lack of selective radioligands for D₄ sites, have complicated the elucidation of their functional significance, although selective D₄ antagonists are now becoming available (Boyfield et al., 1996; Bristow et al., 1997; Kulagowski et al., 1996; Merchant et al., 1996; Millan et al., 1996). In contrast, both in situ hybridization and polymerase chain reaction amplification studies have allowed for the localization of mRNA encoding D₃ receptors in rodent and human brain, and the use of antibodies directed against D₃ receptors has permitted localization of the corresponding protein (Bouthenet et al., 1991; Landwehrmeyer et al., 1993; Lévesque et al., 1992). These approaches, together with studies of the preferential D₃ agonists [³H]-(+)-PD 128,907 and [³H]-(+)-7-OH-DPAT (Hall et al., 1996; Herroelen et al., 1994), suggest that D₃ sites are predominantly localized in limbic regions such as the olfactory tubercles, the nucleus accumbens and the islands of Calleja. This restricted D₃ receptor distribution contrasts with the broad distribution of D₂ receptors (Gehlert et al., 1992; Larson and Ariano, 1995; Levant et al., 1993; Murray et al., 1994; Richtand et al., 1995). In addition, a minor population of D₃ autoreceptors may exist with D₂ autoreceptors on dopaminergic cell bodies of the ventral tegmental area and substantia nigra, pars compacta (Diaz et al., 1995; Gobert et al., 1995; Koeltzow et al., 1998; Meador-Woodruff et al., 1996).

Despite the utility of gene knockout and antisense strategies (Accili et al., 1996; Baik et al., 1995; Carta et al., 1996, Tepper et al., 1997), ligands that interact selectively with D₃ vs. D₂ receptors are essential for an improved knowledge of their localization, modulation and functional significance. In the absence of appropriate tissue preparations, the characterization of novel ligands has generally been performed by using mammalian cell lines transfected with human or rodent D₂ or D₃ receptor subtypes and radioligands such as [¹²⁵I]-iodosulpride or [³H]-spiperone that do not differentiate these sites (Chio et al., 1993; Millan et al., 1995; Sokoloff et al., 1992; Tang et al., 1994). On the basis of such studies, the aminotetralins (+)-AJ 76 and (+)-UH 232 were found to display a modest (about 2-5-fold) preference for D₃ over D₂ sites (Lévesque, 1996; Sokoloff et al., 1992; Van Vliet et al., 1996). Subsequently, several antagonists with improved selectivity at D₃ vs. D₂ sites have been identified: the aminoindane U 99194 (Waters et al., 1993), the benzofurane (±)-S 11566 and its active (+)-isomer (+)-S 14297 (Gobert et al., 1995, 1996; Millan et al., 1994; 1995; Morris et al., 1997), the arylpiperazine GR 103,691 (Murray et al., 1995) and the benzamide derivative nafadotride (Sautel et al., 1996). Of these, (+)-S 14297 has been extensively employed in the functional evaluation of the potential significance of D₃ receptors in vivo. It has been shown that (+)-S 14297 inhibits the hypothermia induced by dopaminergic agonists such as (+)-7-OH-DPAT and quinpirole in rodents, which suggests an involvement of D₃ receptors in the control of CT (Millan et al., 1994; 1995; Rivet et al., 1996). At comparable doses, (+)-S 14297 neither modifies PRL secretion nor induces catalepsy in rats, two responses reflecting blockade of D₂ receptors (Brocco et al., 1995; Millan et al., 1995; 1997). Interestingly, (+)-S 14297 does not modify the synthesis or release of DA in cerebral tissues, and it fails to modify the activity of ventral tegmental area-localized dopaminergic neurons (Gobert et al., 1995; Lejeune and Millan, 1995; Millan et al., 1995; Rivet et al., 1994; 1996). These observations suggest that D₂ rather than D₃ autoreceptors tonically inhibit the activity of central dopaminergic neurons. Nevertheless, (+)-S 14297 attenuates the inhibitory influence of (+)-7-OH-DPAT and (+)-PD 128,907 upon DA release and synthesis, which suggests that D₃ (auto)receptors may contribute to the "phasic" inhibition of dopaminergic pathways (Gobert et al., 1995; Lejeune and Millan, 1995; Rivet et al., 1994).

Evidently, there now exist several antagonists and agonists of potential utility for the in vitro and in vivo functional characterization of actions mediated by D₃ receptors. However, there has not yet been any comparative evaluation of the properties of these ligands. Thus, using the radioligands [³H]-(+)-S 14297 and [³H]-(+)-PD 128,907, both of which have recently been radiolabeled (Akunne et al., 1995; Newman-Tancredi et al., 1995), as well as the mixed D₂/D₃ antagonist [¹²⁵I]-iodosulpride, the present study undertook a comparative evaluation of the binding profiles of (+)-S 14297, nafadotride, U 99194, GR 103,691, (+)-AJ 76, (+)-UH 232 and haloperidol at recombinant hD₂ and hD₃ receptors. In addition, we examined their in vivo actions in models of putative activity at D₃ receptors (hypothermia) and D₂ receptors (catalepsy, PRL secretion and DA synthesis). These analyses allowed for the classification of drugs in terms of 1) their rank order of potency at hD₃ receptors and 2) their rank order of selectivity for hD₃ vs. hD₂ receptors. In addition, a direct comparison of drug doses in putative models of D₃ as compared with D₂ receptor-mediated activity was possible.

	Material and Methods

Top Abstract Introduction Methods Results Discussion References

Binding at cloned, human D₂ and D₃ receptors. Binding studies were carried out at recombinant hD₂ receptors (Receptor Biology, Beltsville, MA) and hD₃ receptors (J.-C. Schwartz, INSERM, Paris, France) expressed in CHO cell lines as described previously (Millan et al., 1995). Briefly, affinities at hD₂ and hD₃ receptors were determined using [¹²⁵I]-iodosulpride (0.1 and 0.2 nM for D₂ and D₃, respectively), [³H]-(+)-S 14297 (7 nM for D₃) and [³H]-(+)-PD 128,907 (2 nM for D₃). Nonspecific binding was defined using 10 µM raclopride. Binding conditions for [¹²⁵I]-iodosulpride and [³H]-(+)-S 14297 were as previously described (Newman-Tancredi et al., 1995; Sokoloff et al., 1992), and they were the same for [³H]-(+)-PD 128,907 except that incubations were carried out for 90 min.

Binding at other sites. Binding conditions at other receptor sites were as described in Millan et al. (1995). For each receptor, the tissue, radioligand and ligands used to define nonspecific binding were as follows: hD₄ receptors (4-repeat variant) expressed in CHO cells; [³H]-spiperone (0.2 nM); haloperidol 10 µM; rat striatal D₂ receptors; [³H]-spiperone (0.1 nM), raclopride 10 µM; rat hippocampal 5-HT_1A receptors and cloned, human serotonin (h5-HT_1A) receptors expressed in CHO cells; [³H]-(±)-8-OH-DPAT (0.4 nM); 5-HT 10 µM; cloned, human muscarinic (hM₁) receptors expressed in CHO cells; [³H]-N-methylscopolamine (0.5 nM); atropine 10 µM; rat frontal cortex alpha-1 adrenoceptors; [³H]-prazosin (0.2 nM); phentolamine 10 µM.

Influence on CT. For evaluation of the influence of drugs alone on CT, basal CT was determined, drug or vehicle was administered, and CT was determined again 30 min later. This time corresponds to that at which (+)-PD 128,9078 and (+)-7-OH-DPAT exert their maximal influence on CT (Salmi et al., 1993; Dekeyne, A., unpublished observation). The difference between basal and postdrug values was calculated. For antagonist studies, the procedure was as follows. Basal CT was measured, drug or vehicle was administered and (+)-7-OH-DPAT, (+)-PD 128,907 or vehicle was administered 30 min later. A further 30 min later, CT was reevaluated and the difference from basal values calculated. The percent of drug inhibition of the actions of (+)-7-OH-DPAT and (+)-PD 128,907 was computed as 1-[(antagonist + agonist)  vehicle alone)/(vehicle + agonist)  vehicle alone)] × 100. Drug potency was expressed as ID₅₀ 95% CL.

Evaluation of catalepsy. Catalepsy was determined in rats using a previously described procedure (Waldmeier and Delini-Stula, 1979). The animals were placed in a position whereby the left and right hind paws were crossed over the ipsilateral front paws; the time during which this position was maintained was determined up to a maximum of 30 sec. The mean value of three consecutive tests, separated by intervals of 1 min, was determined. Cataleptogenic potency was expressed in terms of AD₅₀ (95% CL)that is, the dose required for the induction of a half-maximal response (equivalent to 15 sec).

Evaluation of PRL secretion. As described previously (Gobert et al., 1995), we measured PRL 30 min after drug administration in plasma by radioimmunoassay, using a highly specific (<0.5% cross-reactivity to all other hormones examined) antibody against rat PRL (Amersham, Buckingham, England). The detection limit was 70 pg/tube, and the intravariation and interassay variation were 5% and 15%, respectively. Because there is no theoretical limit to PRL levels in plasma, the MED was defined as the lowest dose significantly different (P < .05) from vehicle control values in Dunnett's test after ANOVA.

Evaluation of DA turnover. The method we used was described previously (Gobert et al., 1995). The effect of drugs alone was determined 30 min after their s.c. injection. The striatum, nucleus accumbens, olfactory tubercle and frontal cortex were examined (the inclusion of a portion of cingulate cortex in "frontal" cortex should be noted).

Electrical activity of serotonergic neurons in the DRN. As described in detail previously (Lejeune et al., 1994), rats were anesthetized with chloral hydrate (500 mg/kg s.c.). The femoral vein was cannulated and the rat placed in a stereotaxic apparatus. A tungsten electrode was lowered into the DRN, and the firing activity was measured by means of a Spike 2 analysis system obtained from Cambridge Electronic Design (Cambridge, U.K.). Drug effects were quantified over 60-sec bins at their time of maximal effect (1-2 min) after their injection i.v. Administered alone, GR 103,691 was injected in cumulative doses every 2 min. It was also injected at a single dose 2 min after administration of the 5-HT_1A receptor agonist (±)-8-OH-DPAT (0.005 mg/kg i.v.). Firing rates were expressed as a percentage of basal, preinjection values (defined as 100%). The mean, basal firing rate was 1.1 Hz.

Data analysis and statistics. All binding data were analysed with nonlinear regression ("Prism," GraphPad, San Diego, CA). K_i values were calculated according to the Cheng-Prussof equation, K_i = IC₅₀/(1 + L/K_d), where IC₅₀ is the concentration of compound that gives 50% inhibition of radioligand binding, L is the concentration of radioligand and K_d is the dissociation constant of the radioligand. In vivo dose-response curves were initially analyzed by ANOVA followed by Dunnett's test, for which the level of significance was set at P < .05. As indicated above, MEDs, ID₅₀ values, AD₅₀ values or AD₁₅₀ values were determined to estimate drug potencies. MOEs of drugs were also calculated.

Drugs and chemicals. [¹²⁵I]-Iodosulpride (2000 Ci/mmol) and [³H]-(+)-PD 128,907 (90-130 Ci/mmol) were purchased from Amersham, and [³H]-(+)-S 14297 (145 Ci/mmol) was radiolabeled by C.E.A (Gif-sur-Yvette, France). For in vitro binding studies, drugs were dissolved in water or in dimethylsulfoxide. For in vivo studies, drugs were dissolved in sterile water, plus a few drops of lactic acid if necessary, and pH was adjusted to as close to neutrality (>5.0) as possible. Drugs were injected s.c., and doses are expressed in terms of the base. Drug sources, salts and structures were as follows: (+)-AJ 76 and (+)-UH 232 (Tocris Cookson, Southampton, England); (+)-7-OH-DPAT HCl (J. Besselièvre, CNRS, Paris, France); (+)-PD 128,907 HCl and (±)-8-OH-DPAT HBr (Research Biochemicals, Inc., Natick, MA) and haloperidol base (Sigma, Chesnes, France). (±)-S 11566 HCl, (+)-S 14297 dibenzoyltartrate, ()-S 17777 dibenzoyltartrate, U 99194 HCl, GR 103,691 and nafadotride were synthesized by Servier chemists (J.-L. Peglion and G. Lavielle).

	Results

Top Abstract Introduction Methods Results Discussion References

In vitro ligand binding at hD₃ and hD₂ receptors. In an initial series of experiments, we investigated the binding characteristics of [³H]-(+)-PD 128,907 to hD₃ receptors. [³H]-(+)-PD 128,907 associated rapidly and monophasically with a t_1/2 of 21 ± 3 min. Dissociation of [³H]-(+)-PD 128,907 from hD₃ receptors was biphasic with a t_1/2 value for the first component of the isotherm of 58 ± 8 min (k¹ = 0.013 ± 0.002 min¹). The second component of the isotherm dissociated slowly, with 27% ± 1% of binding still remaining after 6 hr of incubation. An estimate of the K_d derived from the association and dissociation (rapid component) constants yielded a value of 0.6 nM. In saturation binding experiments, [³H]-(+)-PD 128,907 yielded a K_d value at hD₃ receptors of 1.61 ± 0.31 nM. The B_max value (6.06 ± 0.59 pmol/mg) was similar to that determined with [¹²⁵I]-iodosulpride in the same membrane preparation (not shown). Antagonist competition binding isotherms at hD₃ receptors were derived for the preferential D₃ receptor agonist [³H]-(+)-PD 128,907, the D₃ antagonist [³H]-(+)-S 14297 and the D₂/D₃ antagonist [¹²⁵I]-iodosulpride (fig. 1). K_i values at hD₃ receptors were similar regardless of the radioligand used (table 1). The rank order of antagonist potency against [³H]-(+)-PD 128,907 binding at hD₃ sites was GR 103,691 > nafadotride > haloperidol > (+)-UH 232 > (+)-S 14297 > (±)-S 11566 > (+)-AJ 76 > U 99194 > ()-S 17777. The agonists (+)-7-OH-DPAT and (+)-PD 128,907 yielded biphasic isotherms at hD₂ receptors (table 1, legend). The selectivity ratios (K_i, hD₂/K_i, hD₃) for the three radioligands are shown in table 1. (±)-S 11566, its isomer (+)-S 14297 and GR 103,691 exhibited the highest selectivity for hD₃ receptors. Affinities at hD₃ receptors correlated positively between radioligands with coefficients (r values) of +0.98 to +0.99. The affinity of the antagonists at other key receptor subtypes was determined. K_i values are shown for hD₄, h5-HT_1A and hM₁ receptors and for rat alpha-1 adrenoceptors (table 2). (+)-AJ 76 was only 8-fold more selective for hD₃ than for hD₄ receptors, and GR 103,691 was only 11-fold more selective for hD₃ than for h5-HT_1A receptors. (+)-S 14297 exhibited modest affinity at hM₁ receptors, for which its affinity was 30-fold less than at hD₃ receptors labeled by [³H]-PD 128,907 (Millan et al., 1995).

View larger version (27K): [in this window] [in a new window]   Fig. 1.   Competition binding of dopaminergic antagonists at cloned, hD3 and hD2 receptors expressed in CHO cells. [3H]-(+)-PD 128,907, hD3 (panel A), [3H]-(+)-S 14297, hD3 (panel B) and [125I]-iodosulpride, hD3 (panel C). Competition binding experiments at hD2 receptors were carried out against [125I]-iodosulpride (panel D). Points shown are means of triplicate determinations from a single, representative experiment performed at least three times.

Electrical activity of DRN-localized serotonergic neurons. Administered at a dose of 5 µg/kg i.v., (±)-8-OH-DPAT completely abolished the firing rate of DRN serotonergic neurons (100.0 ± 0.0% relative to vehicle values, defined as 0%). Further, the alpha-1 adrenoceptor antagonist prazosin (150 µg/kg i.v.) reduced firing levels to 77.9 ± 6.4% of vehicle values. Both effects were significant in Student's two-tailed t test (P < .05). By contrast, administered in incremental doses up to 500 µg/kg i.v., GR 103,691 failed (±0.0 ± 5.5%) to modify the firing rate of these neurons. Similarly, at a dose of 250 µg/kg i.v., GR 103,691 did not modify the influence of 8-OH-DPAT on DRN firing (100.0 ± 0.0%).

Blockade of (+)-PD 128,907- and (+)-7-OH-DPAT-induced hypothermia. Both (+)-PD 128,907 and (+)-7-OH-DPAT dose-dependently elicited hypothermia (fig. 2). Their actions were prevented by (±)-S 11566, as well as by its active eutomer (+)-S 14297, whereas the inactive distomer ()-S 17777 was ineffective (table 3; fig. 3). (+)-AJ 76, (+)-UH 232, nafadotride and U 99194 also antagonized (+)-PD 128,907- and (+)-7-OH-DPAT-induced hypothermia, whereas GR 103,691 was inactive (table 3; fig. 3). Haloperidol was active at low doses. Antagonist potencies in blocking PD 128,907- and (+)-7-OH-DPAT-induced hypothermia were highly correlated (r = 0.97, P < .01).

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Dose-dependent induction of hypothermia by the preferential agonists at hD3 receptors, (+)-PD 128,907 and (+)-7-OH-DPAT. The data represent the difference between basal and post-treatment values and are means ± S.E.M. (n  5 per value). ANOVA results: (+)-PD 128,907, F(6,31) = 55.8, P < .01 and (+)-7-OH-DPAT, F(4,38) = 51.1, P < .01. * Significantly different from vehicle values in Dunnett's test after ANOVA; P < .05.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Dose-dependent blockade of (+)-PD 128,907-induced hypothermia by dopaminergic antagonists. Antagonists were given 30 min before (+)-PD 128,907 (0.63 mg/kg s.c.) (closed symbols) or vehicle (open symbols). Data are means ± S.E.M. (n  4 per value). ANOVA results follow. Panel A: (±)-S 11566/vehicle, F(3,25) = 4.4, P < .05; (±)-S 11506/(+)-PD 128,907, F(3,15) = 24.5, P < .001; (+)-S 14297/vehicle, F(3,33) = 8.5, P < .001; (+)-S 14297/(+)-PD 128,907, F(3,15) = 12.4, P < .001; ()-S 17777/vehicle, F(2,12) = 3.1, P > .05; ()-S 17777/(+)-PD 128,907, F(3,16) = 3.7, P < .05. Panel B: (+)-UH 232/vehicle, F(3,10) = 0.3, P > .05; (+)-UH 232/(+)-PD 128,907, F(3,17) = 15.7, P < .01; (+)-AJ 76/vehicle, F(3,30) = 4.0, P < .05; (+)-AJ 76/(+)-PD 128,907, F(3,14) = 12.5, P < .01; nafadotride/vehicle, F(3,13) = 1.6, P > .05. Panel C: nafadotride/(+)-PD 128,907, F(4,18) = 19.4, P < .01; U 99,194/vehicle, F(3,11) = 0.2, P > .05; U 99,194/(+)-PD 128,907, F(4,26) = 11.1, P < .01 and GR 103,691/(+)-PD 128,907, F(3,13) = 0.8, P > .05. * Significantly different from vehicle/(+)-PD 128,907 values in Dunnett's test after ANOVA: P < .05.

Induction of catalepsy, PRL secretion and DA synthesis by dopaminergic antagonists. Haloperidol potently elicited catalepsy, increased plasma PRL levels and elevated DA synthesis (tables 4 and 5; fig. 4). (+)-AJ 76, (+)-UH 232 and nafadotride likewise induced both PRL secretion and DA synthesis (tables 4 and 5; fig. 4). U 99194 failed to elicit catalepsy and increased PRL release and DA turnover only at very high doses (40.0 mg/kg). (+)-S 14297 and GR 103,691 were devoid of activity in each of these models (tables 4 and 5; fig. 4).

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Dose-dependent induction by dopaminergic antagonists of catalepsy (panel A), PRL secretion (panel B) and striatal DA synthesis (panel C). Data are means ± S.E.M. (n  4 per value). ANOVA results follow. Catalepsy: (+)-AJ 76, F(4,26) = 4.8, P < .05; GR 103,691, F(3,18) = 0.6, P > .05; haloperidol, F(4,43) = 58.3, P < .05; nafadotride, F(3,26) = 11.8, P < .05; (+)-UH 232, F(3,19) = 11.7, P < .05; U 99194, F(2,8) = 0.8, P > .05 and (+)-S 14297, F(2,12) = 1.5, P > .05. Prolactin secretion: (+)-AJ 76, F(4,32) = 9.7, P < .05; GR 103,691, F(3,30) = 0.8, P > .05; haloperidol, F(5,47) = 18.7, P < .05; nafadotride, F(5,36) = 24.3, P < .05; (+)-UH 232, F(3,13) = 5.8, P < .05; U 99194, F(4,15) = 17.7, P < .05 and (+)-S 14297, F(4,28) = 0.5, P > .05. Striatal DA synthesis: GR 103,691, F(1,6) = 3.5, P > .05; nafadotride, F(4,19) = 90.5, P < .01; (+)-UH 232, F(3,12) = 34.5, P < .01; U 99194, F(3,17) = 108.2, P < .01 and S 14297, F(3,31) = 0.7, P > .05. (+)-AJ 76 and haloperidol data are from Gobert et al., (1995). * Significantly different from vehicle values in Dunnett's test after ANOVA; * P < .05.

	Discussion

Top Abstract Introduction Methods Results Discussion References

Competition binding studies. All antagonists displayed similar affinities at hD₃ receptors irrespective of the radioligand used, though slightly higher affinities were generally observed with [³H]-(+)-PD 128,907. This observation may reflect the capacity of agonist radioligands to stabilize hD₃ receptors in a high-affinity conformation, an interpretation consistent with its slow dissociation ("Results") and implied by a ternary complex model of receptor coupling (Kenakin, 1996). In any case, a similar order of hD₂/hD₃ selectivity was observed regardless of the D₃ radioligand. With subnanomolar affinities, GR 103,691 and nafadotride were potent ligands at hD₃ receptors (Murray et al., 1995; Sautel et al., 1996). The hD₂/hD₃ selectivity ratio of nafadotride was, however, only modest, whereas GR 103,691 showed 60-fold selectivity. Because of a higher affinity at D₂ sites in our hands, this value is about 2-fold less than that previously reported (table 1; Murray et al., 1995) but is still pronounced. (±)-S 11566 and its active eutomer (+)-S 14297 also exhibited marked selectivity ratios at D₃ vs. D₂ sites, whereas a modest hD₂/hD₃ selectivity ratio was obtained for the less active distomer ()-S 17777, as well as for U 99194 (Millan et al., 1994; 1995; Waters et al., 1993). Overall, the lowest D₂/D₃ ratios were observed for (+)-AJ 76 and (+)-UH 232 and the highest for (+)-S 14297 and GR 103,691 (table 1). However, GR 103,691 displayed high affinity at 5-HT_1A receptors and alpha-1 adrenoceptors (table 2), in accordance with the findings of Murray et al. (1994). (+)-S 14297 also showed modest (about 30 fold-lower) activity at muscarinic hM₁ receptors (Millan et al., 1995). Nevertheless, among the ligands tested, (+)-S 14297 presents a reasonable combination of marked hD₃ affinity, pronounced hD₃/hD₂ selectivity and modest interactions at other receptor sites. Indeed, (+)-S 14297 shows 100-fold lower affinity at hD₁ and hD₅ receptors, multiple serotonergic and adrenergic receptors and all other sites (>50) as yet examined, with the exception of ₁ sites, for which it shows modest affinity (196 nM) (Millan et al., 1995).

Induction and blockade of hypothermia. In a recent study, the comparative importance of D₂ and D₃ sites in mediating hypothermia was addressed in detail (Millan et al., 1994; 1995; Rivet et al., 1996). In a result consistent with a putative role of D₃ receptors, (+)-PD 128,907 elicited hypothermia herein (fig. 2). Notwithstanding its superior affinity at D₃ (and D₂) receptors as compared with (+)-7-OH-DPAT (table 1; Akunne et al., 1995; Bristow et al., 1996), (+)-PD 128,907 was less potent in decreasing CT. We have also noted an inferior potency of (+)-PD 128,907 relative to (+)-7-OH-DPAT in a diversity of in vivo models in which pre- and/or postsynaptic D₃ and/or D₂ receptors are implicated, including a reduction of locomotion, inhibition of the firing rate of ventral tegmental area-localized dopaminergic neurons and suppression of PRL secretion (Brocco et al., 1995; Gobert et al., 1995; Lejeune, F., unpublished observation). Collectively, these observations suggest that (+)-PD 128,907 is a less potent ligand than (+)-7-OH-DPAT in vivo, which probably reflects differential absorption, metabolism or access to the CNS.

PRL secretion, DA synthesis and catalepsy: blockade of tonically active D₂ receptors. Tuberoinfundibular dopaminergic pathways exert a tonic, inhibitory control on the secretion of PRL (Ben-Jonathan et al., 1989; McDonald et al., 1984). Although multiple dopaminergic receptor types may indirectly modulate PRL secretion via these tuberoinfundibular neurons (Berry and Gudelsky, 1990), dopaminergic ligands modulate PRL secretion primarily via actions at dopaminergic receptors on lactotrophs in the adenohypophysis (McChesney et al., 1991). Neuroanatomical studies have indicated that these are of the D₂ type (Levant et al., 1993; Lévesque, 1996). Pharmacological support for a predominant role of D₂ receptors has also been obtained. Thus the relative potencies of dopaminergic agonists and antagonists in decreasing and increasing PRL secretion, respectively, correlate with their affinity at D₂ (but not D₃) sites (Millan et al., 1995; Rivet et al., 1996). Indeed, compared with haloperidol, (+)-S 14297 modified PRL secretion very little (fig. 4). Further, U 99194, in line with its low D₂ affinity, exerted only a weak influence on PRL levels. A previous study has also reported that (+)-AJ 76 weakly modifies PRL secretion in vivo (Eriksson et al., 1986). Although GR 103,691 was also inactive, this observation must be interpreted in the light of the following comments about its poor activity in vivo. Indeed, the affinity of GR 103,691 at hD₂ receptors is virtually identical to that of (+)-UH 232, which potently elicited PRL secretion.

Mechanisms underlying induction of catalepsy, PRL secretion and DA turnover. The patterns of data obtained in the models of PRL secretion, induction of cerebral DA turnover and catalepsy were broadly similar. In contrast to the potent actions of haloperidol, (+)-S 14297 was inactive and U 99194 displayed modest activity. Nafadotride, (+)-UH 232 and (+)-AJ 76, which manifest only a mild preference for D₃ receptors, displayed an intermediate pattern of behavior. Thus these in vivo findings correspond to the relative preference of these antagonists for D₃ vs. D₂ sites in vitro. Interestingly, differences among the antagonists were apparent not only as concerns their relative potencies but also as regards their maximal effects. Thus haloperidol consistently produced the most marked actions, (+)-S 14297 was inactive, U 99194 was weakly active and the other antagonists provoked intermediate responses (tables 4 and 5; fig. 4). The question arises why maximal effects should differ. If antagonist actions simply reflected interruption of the activity of spontaneously released DA at D₂ receptors, they should, in theory, all elicit the same maximal effect. There are several possible explanations. First, the non-D₂ (or D₃) receptor interactions of drugs may modify their respective maximal effects. Second, the dose ranges of drugs employed may have been insufficient to occupy D₂ receptors substantially. This may be the case for (+)-S 14297, but dose-response curves for other drugs were pursued up to doses at which D₂ antagonist properties are clearly expressed (Brocco et al., 1995; Brocco, M., unpublished observation). Third, a functional interplay between postsynaptic D₂ and D₁ sites is well established (Creese and Fraser, 1987), and D₂ and D₃ receptors colocalized on individual neurons may mutually oppose each other's activities (Surmeier et al., 1992; Le Moine and Bloch, 1996). Indeed, as mentioned above, the enhancement of motor behavior by D₃ receptor blockade counters the motor-suppressive effects of D₂ receptor antagonism (Millan et al., 1995; 1997). Thus, for the ligands tested herein, a progressive reduction in cataleptogenic potential paralleled a progressive reinforcement of D₃ vs. D₂ antagonist preference. A fourth possibility is that haloperidol behaves as an inverse agonist at D₂ receptors (Hall and Strange, 1997; Nilsson et al., 1996). A differential degree of inverse agonist activity at D₂ sites could explain the contrasting maximal effects of antagonists on PRL secretion, etc. (fig. 4). However, the concept of inverse agonist actions cannot easily explain the occurrence of catalepsy in D₂ knockout mice (Baik et al., 1995). Further, like haloperidol, clozapine is an inverse agonist at D₂ receptors, but it does not evoke catalepsy (Hall and Strange, 1997). Irrespective of the reasons underlying the difference in maximal effects, the present data suggest that antagonists possessing a marked preference for D₃ versus D₂ receptors may have a low extrapyramidal syndrome potential.

The inactivity of GR 103,691. Pharmacokinetic factors also contribute to differential drug effects, and GR 103,691 was inactive, even at high doses relative to its in vitro affinities, in each of the in vivo paradigms employed herein. It was recently shown that, in contrast to haloperidol, GR 103,691 does not influence cerebral c-fos expression (Hurley et al., 1996). Indeed, the only in vivo effect documented to date for GR 103,691 is its ability to block the locomotion elicited by infusion of muscimol into the ventral tegmental area (Murray et al., 1995). In this model, GR 103,691 (0.3 mg/kg s.c.) was 6-fold less potent than haloperidol (0.05 mg/kg s.c.) despite its 5-fold higher affinity at D₃ receptors, an observation in line with the present data indicating a lower bioavailability than anticipated. However, its low affinity at D₃ receptors (K_i = 200 nM) notwithstanding, clozapine (0.005 mg/kg s.c.) was more potent than GR 103,691 in blocking the actions of muscimol. This observation suggests that activity in this model may reflect the involvement of receptors other than, or in addition to, D₃ sites. Thus, to date, there are no unambiguous data for functional actions of GR 103,691 at D₃ (or D₂) sites in vivo. We have, moreover, found that high doses of GR 103,691 (10 mg/kg s.c.) are inactive in several other behavioral models of D₂ receptor-mediated activity in rats, including inhibition of amphetamine-induced locomotion and reduction of conditioned avoidance responses (Brocco, M., unpublished observation) models, whereas haloperidol is active, with ID₅₀ values of 0.04 and 0.05 mg/kg s.c., respectively. In addition, haloperidol abolishes inhibition of the firing of dopaminergic neurons in the ventrotegmental area by (+)-PD 128,905 with an ID₅₀ of 0.003 mg/kg i.v. In contrast, even at an 80-fold higher dose of 0.25 mg/kg i.v., GR 103,691 only partially (about 40%) inhibits the action of (+)-PD 128,907 (Lejeune, F., unpublished observation). Herein, we also examined this issue by exploiting the marked affinity of GR 103,691 for 5-HT_1A receptors and alpha-1 adrenoceptors. Agonists at 5-HT_1A sites and antagonists at alpha-1 adrenoceptors both inhibit serotonergic neurons in the DRN (Hjorth and Sharp, 1990; Lejeune et al., 1994). However, GR 103,691, at doses up to 0.5 mg/kg i.v., neither modified basal firing rates of serotonergic neurons nor attenuated the inhibitory influence on these of the 5-HT_1A agonist (±)-8-OH-DPAT (see "Results"). These observations suggest that a lack of in vivo activity is a general property of GR 103,691 irrespective of the receptor type concerned. This electrophysiological study was performed by the i.v. route in order to minimize problems of metabolism. Further, GR 103,691 does not modify PRL secretion even though the relevant D₂ receptors are localized outside the blood-brain barrier. Thus it is not clear whether rapid metabolism, poor CNS penetration and/or other factors underlie the low activity of GR 103,691 in vivo.

General discussion. Of the ligands examined, (+)-S 14297 and GR 103,691 appear the most appropriate ligands for the characterization and differentiation of activity at D₃ as compared with D₂ receptors in vitro. In this regard, GR 103,691 possesses an advantage in terms of its higher affinity and overall selectivity for D₃ vs. D₂ sites. However, GR 103,691 displays the disadvantage of marked affinity at both 5-HT_1A receptors and alpha-1 adrenoceptors. Although these properties may not interfere with its use in transfected cell lines, they compromise its utility in functional studies of more complex systems. Furthermore, GR 103,691 appears to possess little bioavailability in vivo. In this respect, although its modest affinity at hM₁ and ₁ receptors must be mentioned (Millan et al., 1995), (+)-S 14297 appears to be a more suitable ligand for in vivo studies. In fact, few ligands have been reported that possess a marked D₃ vs. D₂ preference comparable to that of (+)-S 14297 and GR 103,691 in vitro. Moreover, in vivo data are generally not available. Nevertheless, the recently described, 2-substituted 2-aminotetralin GR 218,231 [2-(R,S)-(dipropylamino)-6-(4-methoxyphenylsulfonylmethyl)-1,2,3,4-tetrahydronaphtalene] possesses >100-fold selectivity for D₃ vs. D₂ sites, and its functional actions in vivo will be of interest to explore (Murray et al., 1996). Moreover, a modestly preferential (10-20-fold) antagonist, L 741,626, at D₂ receptors has been documented (Bowery et al., 1996). As concerns the D₂ receptor "family" in general, several potent and selective D₄ receptor antagonists have now become available, including S 18126 ({2-[4-(2,3-dihydro benzo [1,4]dioxin-6-yl) piperazin-1-yl methyl] indan-2-yl}) and L 745,870 (3-(4-[4-chlorophenyl]piperazin-1-yl)methyl-1H-pyrrolo[2,3b]pyridine). Notably, both S 18126 and L 745,870 were inactive in the functional models employed herein, which indicates a lack of involvement of D₄ receptors (Boyfield et al., 1996; Bristow et al., 1997; Kulagowski et al., 1996; Millan et al., 1996 and in press).

Conclusion. In conclusion, the present data underpin the hypothesis that the selective blockade of D₃ receptors does not provoke extrapyramidal side effects and fails to perturb dopaminergic transmission. On the other hand, whether an antagonist action at D₃ receptors contributes to the therapeutic efficacy of antipsychotic drugs remains to be determined (Levant, 1997; Sokoloff and Schwartz, 1995). More generally, the potential significance of D₃ receptors in depression, drug abuse and other psychiatric disorders requires further evaluation (Acri et al., 1995; Caine and Koob, 1993; Roberts and Ranaldi, 1995; Sokoloff and Schwartz, 1995; Spealman, 1996; Strange, 1993; Wallace et al., 1996; Willner, 1983). To develop an improved understanding of the physiological and therapeutic significance of D₃ (and D₂) receptors, we need additional, chemically diverse, potent and selective D₃ and D₂ receptor antagonists.