Materials and Methods

Determination of Agonist Efficacy at Recombinant h5-HT_1A, h5-HT_1B, and h5-HT_1DReceptors by [³⁵S]GTPγS Binding.

Efficacy at cloned h5-HT_1A, h5-HT_1B, and h5-HT_1D receptors was determined by measuring stimulation of [³⁵S]GTPγS binding, as described previously (Newman-Tancredi et al., 1999). Briefly, membranes prepared from Chinese hamster ovary (CHO) cells stably expressing h5-HT_1A, h5-HT_1B, or h5-HT_1D receptors were incubated at 22°C for 20 min (h5-HT_1A) or 30 min (h5-HT_1B and h5-HT_1D) with drugs or 5-HT in the following buffer: 20 mM HEPES pH 7.4, 10 mM NaCl, 3 μM GDP, 3 mM MgCl₂, and 0.1 nM [³⁵S]GTPγS. Agonist efficacy is expressed relative to that of 5-HT (defined as 100%), which was tested at a maximally effective concentration (10 μM) in each experiment. Experiments were terminated by rapid filtration through GF/B filters (Whatman, Maidstone, UK) using a Packard Instrument Company, Inc. (Downers Grove, IL) 96-well cell harvester and radioactivity determined by liquid scintillation counting. Binding densities (B_max values) at h5-HT_1A, h5-HT_1B, and h5-HT_1D receptors were 3.6, 8.5, and 1.6 pmol/mg, respectively.

Determination of Agonist Efficacy at h5-HT_2A, h5-HT_2B, and h5-HT_2C Receptors by [³H]PI Depletion.

The functional activity of antiparkinson compounds at h5-HT_2A, h5-HT_2B, and h5-HT_2Creceptors (VSV isoform) was determined as described previously (Cussac et al., 2002). Briefly, cells were labeled with 2 μCi/ml of [³H]myoinositol (10–20 Ci/mmol) for 24 h. Cells were washed and then incubated at 37°C for 30 min with drugs in Krebs-LiCl buffer: 15.6 mM NaH₂PO₄ pH 7, 120 mM NaCl, 4.8 mM KCl, 1.2 mM MgSO₄, 1.2 mM CaCl₂, 0.6% (w/v) glucose, 0.04% (w/v) bovine serum albumin, and 10 mM LiCl. At h5-HT_2A, h5-HT_2B, and h5-HT_2Creceptors in each case, in the absence of agonists, ∼40,000 dpm was typically detected, compared with ∼25,000 dpm in the presence of a maximally effective concentration of 5-HT (10 μM). Agonist efficacy is expressed relative to that of 5-HT (defined as 100%), which was tested at a maximally effective concentration (10 μM) in each experiment. For antagonist studies, cells were preincubated for 15 min with drug before the addition of 5-HT: 1 μM for h5-HT_2A receptors and 0.03 μM for h5-HT_2B and h5-HT_2Creceptors. In additional studies, the influence of the selective 5-HT_2A receptor antagonist MDL100,907 and of the selective 5-HT_2C antagonist SB242,084 against actions of pergolide, cabergoline, and bromocriptine were determined. Membranes were recovered by rapid filtration, and [³H]PI content was determined by scintillation counting. B_max values at h5-HT_2A, h5-HT_2B, and h5-HT_2C receptors were 2.0, 3.0, and 18 pmol/mg, respectively.

Data Analyses.

Binding isotherms were analyzed by nonlinear regression using the program PRISM (GraphPad Software, San Diego, CA). In antagonist studies, K_B values were calculated, as described previously (Newman-Tancredi et al., 1999), according to the equation K_B = IC₅₀/{[(2 + (agonist/EC₅₀)^n_H)^n_H−1] − 1}, where IC₅₀ is the inhibitory concentration₅₀ of the antagonist, agonist is agonist concentration, EC₅₀ is the effective concentration₅₀ of 5-HT alone, andn_H is Hill coefficient of the agonist stimulation isotherm. The EC₅₀ values of 5-HT at h5-HT_2A, h5-HT_2B, and h5-HT_2C receptors used in these calculations were 32, 3.1, and 1.3 nM, respectively. Protein concentrations were determined by use of a bicinchoninic acid kit (Sigma, St. Quentin Fallavier, France).

Drugs.

Lisuride maleate and terguride were donated by Schering (Berlin, Germany). Bromocriptine and pergolide methanesulfonate were purchased from Sigma/RBI (Natick, MA). Apomorphine hydrochloride was purchased from Sigma. Roxindole methanesulfonate was donated by Merck (Darmstadt, Germany), and talipexole was provided by Boehringer Ingelheim GmbH (Ingelheim, Germany). Cabergoline was obtained from Farmitalia Carlo Erba (Rueil-Malmaison, France). Pramipexole dihydrochloride, ropinirole, piribedil hydrochloride,R-(+)-α-(2,3-dimethoxyphenyl)-1-[2-(4-fluorophenylethyl)]-4-piperidine-methanol base [(+)-MDL100,907], ondansetron hydrochloride, and (6-chloro-5-methyl-1-[6-(2-methylpyridin-3-yloxy) pyridin-3-yl carbamoyl] indoline) hydrochloride (SB242,084) were synthesized by Institut de Recherches Servier (Paris, France) chemists.

Results

Drugs Evaluated.

The antiparkinson agents examined herein were those demonstrated to possess significant affinity (pK_i ≥ 6.0) at sites studied in competition binding assays documented in the accompanying article (Millan et al., 2002). Thus, quinpirole, quinelorane, talipexole, and TL99 were not included in the present article and both pramipexole and ropinirole were evaluated only at h5-HT_1Areceptors.

Drug Actions at h5-HT_1A Receptors.

At a maximally effective concentration (10 μM), 5-HT enhanced [³⁵S]GTPγS binding at h5-HT_1A receptors by ∼1.5-fold relative to basal values; it displayed a pEC₅₀ value of 7.7 (Fig. 1; Table1). All ligands stimulated [³⁵S]GTPγS binding at h5-HT_1A receptors, with efficacies ranging from partial for apomorphine (E_max = 35%) to full for cabergoline (93%) and lisuride (98%). Potencies for stimulation of [³⁵S]GTPγS binding varied considerably from low (piribedil, pEC₅₀ = 5.2) to pronounced (lisuride, 8.90). Pramipexole exhibited partial agonist properties (62.6 ± 8.5%) at high concentrations (pEC₅₀, 4.9). Ropinirole was also a weak (pEC₅₀, 5.3) partial agonist (73.0 ± 4.0%). The correlation coefficient, r (Pearson product-moment), between pEC₅₀ values and pK_i values determined in competition experiments (Millan et al., 2002) was 0.91 (P < 0.05).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Actions of antiparkinson agents at h5-HT_1A receptors expressed in CHO cells. [³⁵S]GTPγS binding was carried out as described underMaterials and Methods. Binding is expressed as a percentage of that observed with a maximally effective concentration (10 μM) of 5-HT (defined as 100%). Values shown are from representative experiments performed in triplicate and repeated on at least three occasions.

Drug Actions at h5-HT_1B Receptors.

At a maximally effective concentration (10 μM), 5-HT stimulated [³⁵S]GTPγS binding at h5-HT_1B receptors by ∼1.5-fold relative to basal values; it displayed a pEC₅₀ value of 8.1 (Fig. 2; Table 1). All ligands displayed agonist properties except for apomorphine (no stimulation, 10 μM) and piribedil (not tested, pK_i < 5.0;Millan et al., 2002). Certain drugs showed low intrinsic activity (such as terguride, 37%), whereas others were highly efficacious (such as pergolide, 90%, and cabergoline, 102%). Drug potencies (pEC₅₀) were modest, ranging from 5.7 (cabergoline) to lisuride (7.6).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Actions of antiparkinson agents at h5-HT_1B receptors expressed in CHO cells. [³⁵S]GTPγS binding was carried out as described underMaterials and Methods. Binding is expressed as a percentage of that observed with a maximally effective concentration (10 μM) of 5-HT (defined as 100%). Values shown are from representative experiments performed in triplicate and repeated on at least three occasions.

Drug Actions at h5-HT_1D Receptors.

At a maximally effective concentration (10 μM), 5-HT increased [³⁵S]GTPγS binding at h5-HT_1D receptors by ∼1.5-fold relative to basal values; it displayed a pEC₅₀ value of 8.9 (Fig. 3; Table 1). Apomorphine (no stimulation at 10 μM) was inactive and piribedil (pK_i < 5.0; Millan et al., 2002) was not evaluated. Potencies of other drugs for stimulation of [³⁵S]GTPγS binding were greater (pEC₅₀ values 1 to 2 log units higher) than at h5-HT_1B receptors. Relative efficacies also differed between h5-HT_1D versus h5-HT_1B receptors. For example, bromocriptine was more efficacious, whereas cabergoline was less efficacious.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Actions of antiparkinson agents at h5-HT_1D receptors expressed in CHO cells. [³⁵S]GTPγS binding was carried out as described under under Materials and Methods. Binding is expressed as a percentage of that observed with a maximally effective concentration (10 μM) of 5-HT (defined as 100%). Values shown are from representative experiments performed in triplicate and repeated on at least three occasions.

Drug Actions at h5-HT_2A Receptors.

5-HT depleted [³H]PI with a pEC₅₀ value of 7.5 (Fig. 4; Table2). Terguride, lisuride, bromocriptine, cabergoline, and pergolide also exhibited, in order of increasing efficacy, agonist properties. These actions were expressed with high potencies (pEC₅₀ values) ranging from 7.6 (terguride) to 8.8 (pergolide). In contrast, apomorphine and roxindole blocked stimulation of [³H]PI depletion by 5-HT (10 μM) with pK_bvalues of 6.4 and 7.7, respectively. Piribedil was inactive. The agonist properties of pergolide, bromocriptine, and cabergoline (1.0 μM) were concentration-dependently abolished by the selective 5-HT_2A receptor antagonist MDL100,907 with pK_b values close to its pK_b for blockade of [³H]PI depletion elicited by 5-HT, that is, pK_bvalues were 10.09 ± 0.13, 10.01 ± 0.09, and 10.15 ± 0.27 compared with 9.97 ± 0.11 for 5-HT.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Actions of pergolide, cabergoline, and bromocriptine at h5-HT_2A and h5-HT_2C receptors. Phospholipase C activity was assessed by determination of [³H]phosphatidylinositol depletion in membranes of CHO cells expressing h5-HT_2A (A) or h5-HT_2Creceptors (B). Depletion is expressed as a percentage of that observed with a maximally effective concentration (10 μM) of 5-HT (defined as 100%). Values shown are from representative experiments performed in triplicate and repeated on at least three occasions.

Drug Actions at h5-HT_2B Receptors.

5-HT depleted [³H]PI with a pEC₅₀ value of 8.52 (Table 2). Only cabergoline and pergolide acted as agonists in potently (pEC₅₀ values of 8.59 and 8.22, respectively) stimulating [³H]PI depletion with high efficacy (123 and 113%, respectively). All other ligands acted as antagonists with potencies ranging from weak (piribedil, 5.99) to pronounced (bromocriptine, 8.89).

Drug Actions at h5-HT_2C Receptors.

5-HT depleted [³H]PI with a pEC₅₀ value of 8.9 (Fig. 4; Table 2). Cabergoline and pergolide behaved as efficacious (96 and 87%) agonists, albeit with potencies lower than those at h5-HT_2A and h5-HT_2B receptors. Both bromocriptine and lisuride also revealed high efficacy (79 and 75%, respectively) in enhancing [³H]PI depletion. On the other hand, terguride, apomorphine, and roxindole manifested antagonist properties. Finally, in line with its low affinity (pK_i < 5.0; Millan et al., 2002), piribedil was inactive. The agonist properties of pergolide, bromocriptine, and cabergoline (1.0 μM) were concentration-dependently blocked by the selective 5-HT_2C receptor antagonist SB242,084, with pK_b values close to its pK_b for blockade of [³H]PI depletion by 5-HT, that is, pK_b values were 9.73 ± 0.04, 9.81 ± 0.06, and 9.35 ± 0.13 compared with 9.53 ± 0.14 for 5-HT.

Discussion

The present study demonstrates that antiparkinson agents display contrasting profiles of agonist and antagonist activity at multiple subtypes of 5-HT receptor implicated in the etiology and management of Parkinson's disease.

h5-HT_1A Receptors.

Using [³⁵S]GTPγS binding, a measure of coupling to G proteins, the clinically active antiparkinson agent lisuride displayed pronounced potency and efficacy at h5-HT_1A receptors (Newman-Tancredi et al., 1999). These observations are consistent with agonist properties at 1) 5-HT_1A receptors coupled to adenylyl cyclase in rat hippocampus, 2) postsynaptic 5-HT_1A receptors controlling behavioral parameters, and 3) 5-HT autoreceptors inhibitory to serotonergic neurons (Barnes and Sharp, 1999; Millan et al., 2000). A further ergot, terguride, likewise stimulated h5-HT_1A receptors, although data from in vivo models are lacking. Jackson et al. (1995) reported that bromocriptine possesses high affinity for native 5-HT_1A sites and assumed that its increase of 5-HT turnover reflected antagonist properties at 5-HT_1A autoreceptors. This explanation seems unlikely in light of its marked efficacy at h5-HT_1A receptors. Thus, a more likely interpretation for the enhancement of serotonergic transmission by bromocriptine is its antagonist actions at inhibitory α₂-AR heteroceptors (Millan et al., 2002;Newman-Tancredi et al., 2002). Cabergoline and pergolide, which display modest affinity for h5-HT_1A receptors (Millan et al., 2002), markedly enhanced [³⁵S]GTPγS binding. Although in vivo correlates of their actions remain to be documented, the potent agonist properties of roxindole at h5-HT_1A sites coincide well with its suppressive influence upon central serotonergic transmission (Newman-Tancredi et al., 1999).

Actions of antiparkinsonian agents at 5-HT_1Areceptors are of interest because their engagement abrogates the induction of dyskinesias by l-DOPA in 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-treated primates (Bibbiani et al., 2001). Furthermore, in pilot studies with the 5-HT_1A partial agonist, buspirone, an attenuation of spontaneous and l-DOPA-induced dyskinesias was seen in Parkinson's disease patients (Bonifati et al., 1994). However, there is no compelling evidence from experimental models (Ahlenius and Salmi 1995) that selective engagement of 5-HT_1Areceptors exerts clinically meaningful antiparkinson actions or that 5-HT_1A receptors are involved in the antiparkinson profiles of drugs evaluated herein. Furthermore, 5-HT_1A agonists exert a complex influence upon locomotor behavior and blunt DA release in the striatum (Barnes and Sharp, 1999; De La Garza and Cunningham, 2000; Millan, 2000): they also attenuate l-DOPA-induced DA release in the striatum from serotonergic neurons bearing 5-HT_1A autoreceptors (Arai et al., 1996; Kannari et al., 2001). These observations, together with the low activity at 5-HT_1A receptors of clinically effective antiparkinson agents, such as pramipexole, indicate that their stimulation is neither necessary nor sufficient for treatment of Parkinson's disease. Nevertheless, in association with agonist actions at dopaminergic receptors, modest agonist properties at 5-HT_1A receptors may improve motor dyskinesias (vide supra), mood, and cognitive performance (Barnes and Sharp, 1999;Meneses, 1999). This hypothesis justifies clinical evaluation.

h5-HT_1B and h5-HT_1D Receptors.

Lisuride and bromocriptine interact with native 5-HT_1B sites, and notwithstanding species differences (Barnes and Sharp, 1999; Audinot et al., 2001), they exerted agonist actions at h5-HT_1B receptors herein with high and modest potency, respectively. Furthermore, the structurally related ergot derivatives terguride, pergolide, and cabergoline likewise revealed high efficacies, whereas the chemically distinct roxindole is a weak partial agonist (Newman-Tancredi et al., 1999). Except for modest affinities of bromocriptine and lisuride at native 5-HT_1D receptors (Barnes and Sharp, 1999), and the low efficacy of roxindole at h5-HT_1Dsites (Newman-Tancredi et al., 1999), actions of antiparkinson drugs at h5-HT_1D sites have not been reported. The demonstration then that bromocriptine, lisuride, and the other ergots terguride, cabergolide, and pergolide (but neither piribedil nor apomorphine) are agonists at h5-HT_1D receptors was unanticipated. Moreover, drug potencies were ∼1 to 2 log units higher than at h5-HT_1B sites. In view of the implication of 5-HT_1B receptors in the control of dopaminergic transmission, motor behavior, and mood, and of the high concentration of 5-HT_1D receptors in human basal ganglia (see Introduction), evaluation of their potential significance in the beneficial and deleterious actions of antiparkinson agents would be justified. However, the lack of activity of apomorphine, pramipexole, and other agents at these sites indicates that their stimulation is not required for therapeutic activity.

h5-HT_2A Receptors.

A preliminary report documented partial agonist actions of bromocriptine at SH-SY5Y cells expressing h5-HT_2A receptors coupled to cytosolic inositol phosphates (Mitchell et al., 1998). Using the complementary approach of depletion of membrane-bound [³H]PI (Cussac et al., 2002b), bromocriptine similarly activated phospholipase C at h5-HT_2A receptors, consistent with its high affinity for h5-HT_2A (Millan et al., 2002). Lisuride likewise behaved as a potent partial agonist, corroborating its modest efficacies at h5-HT_2A receptors in human embryonic kidney-293 cells (inositol phosphates), NIH-3T3 cells (agonist/antagonist binding ratios), and CHO and SH-SY5Y cells (intracellular Ca²⁺ levels) (Porter et al., 1999;Egan et al., 2000; Jerman et al., 2001). Like lisuride, a further ergot possessing high affinity for 5-HT_2A sites, pergolide (Hagen et al., 1994; Millan et al., 2002), was an efficacious agonist at h5-HT_2A receptors, actions mimicked by terguride and cabergoline. In contrast, the structurally distinct roxindole, a potent ligand of h5-HT_2A sites (Millan et al., 2002), and apomorphine behaved as antagonists at h5-HT_2A receptors, whereas piribedil was inactive. The pronounced agonist properties of pergolide at h5-HT_2A sites coincide with the finding that 5-HT_2A receptors participate in its induction of hyperlocomotion in rats (Moore et al., 1999).

Although the latter observation is consistent with the facilitatory influence of 5-HT_2A sites upon locomotor behavior in general (Millan et al., 1999), its relevance to the motor dysfunction of Parkinson's disease is unclear, and (selective) activation of 5-HT_2A sites is not known to be associated with antiparkinson properties. Moreover, antagonist actions at 5-HT_2A receptors, probably on corticostriatal-pallidostriatal afferents, improve extrapyramidal motor dysfunction (Gresch and Walker, 1999; Bubser et al., 2001; Oh et al., 2001). Furthermore, stimulation of 5-HT_2Areceptors and DA release in limbic structures may be related to the psychiatric (hallucinogenesis) side effects of antiparkinson agents (Roth and Meltzer, 1995; Millan et al., 1999; see Introduction). Nevertheless, antiparkinson agents devoid of affinity for 5-HT_2A sites (such as pramipexole) also display psychiatric effects as well as therapeutic efficacy. Thus, experimental and clinical studies comparing antiparkinson drugs with differing efficacies/potencies at h5-HT_2A receptors would help elucidate their significance to beneficial and undesirable actions of antiparkinson agents.

h5-HT_2B Receptors.

In contrast to h5-HT_2A receptors, lisuride is a potent antagonist at h5-HT_2B receptors (Porter et al., 1999; Jerman et al., 2001; Cussac et al., 2002), and all other antiparkinson agents weakly (piribedil) to potently (bromocriptine) blocked h5-HT_2B receptors except cabergolide and pergolide, which were potent agonists. The significance of central h5-HT_2B receptors to antiparkinson agents remains unclear because the only functional role ascribed to their activation is a reduction in anxiety, an observation awaiting confirmation (Duxon et al., 1997). On the other hand, activation of peripheral populations may influence respiratory and gastrointestinal function (Barnes and Sharp, 1999).

h5-HT_2C Receptors.

Herein, lisuride displayed significant efficacy at h5-HT_2C receptors (VSV isoform), in analogy to its partial agonist properties at the VNV isoform, whereas it exhibited low efficacy at SH-SY5Y cells expressing the wild-type (INI) isoform (Egan et al., 2000; Jerman et al., 2001). The contrasting lack of agonist activity of lisuride in a previous study of CHO cells expressing the VSV isoform (Porter et al., 1999) presumably relates to the low receptor density (0.2 pmol/mg) compared with this study (18 pmol/mg). Indeed, the high receptor reserve for induction of [³H]PI depletion in our cellular model favors low levels of drug efficacy (Cussac et al., 2002a,b). Thus, the antagonist properties of apomorphine, roxindole, and terguride at h5-HT_2C receptors herein are of special note. This high sensitivity should also be borne in mind as regards the marked efficacy of cabergoline and pergolide, the only drugs showing potent and high efficacy at all 5-HT₂ receptor subtypes. Bromocriptine has high affinity for h5-HT_2C receptors (Millan et al., 2002) and mimicked the high efficacy of lisuride at h5-HT_2C sites herein. Although 5-HT_2C antagonist properties may improve the influence of antiparkinson agents upon mood and motor function (see Introduction), the risk of weight gain and proepileptic actions should not be neglected (Barnes and Sharp, 1999; Fox and Brotchie, 1999). On the other hand, 5-HT_2C agonist properties may oppose the favorable influence of antiparkinson agents upon motor function and mood.

General Discussion.

Together with the two accompanying articles (Millan et al., 2002; Newman-Tancredi et al., 2002), the present observations evoke several general comments. First, the most striking conclusion of this comprehensive evaluation of antiparkinson agents is that they cannot be regarded as a homogeneous group of “dopaminergic agonists”. Like other classes of drug, such as antipsychotics, they present contrasting patterns of interactions with multiple subtypes of monoaminergic receptor. Imaging studies in Parkinson's disease patients would be instructive in identifying the receptors with which they interact at clinically used doses. Second, the use of cloned, heterologously expressed populations of human receptors under uniform conditions offered important advantages for our comparative studies. However, drug potencies, efficacies, and in vivo actions depend upon a multitude of factors, including colocalization of different receptor types permitting intracellular interactions and formation of heterodimers, coupling to various subtypes of G protein that can be differentially recruited by specific agonists, receptor density, and both isoform and species differences (Backstrom et al., 1999; Barnes and Sharp, 1999; Devi, 2001; Jerman et al., 2001; Cussac et al., 2002a,b). Moreover, although comprehensive, the present studies could be extended to other receptor types, such as 5-HT₃, 5-HT₄ and 5-HT₆ receptors, which are also of potential relevance to the actions of antiparkinson agents (Barnes and Sharp, 1999). Third, as emphasized throughout these articles, the relevance of monoaminergic properties of antiparkinson agents is not restricted to their influence upon motor performance but is equally pertinent to mood and cognitive function. Furthermore, sensory disturbances, including pain, are an important feature of Parkinson's disease (Chulder and Dong, 1995), and monoaminergic mechanisms exert a pronounced influence upon the perceptive and affective-cognitive dimensions of pain via actions at central and peripheral loci (Millan, 2002). Fifth, the involvement of monoaminergic receptors in potentially deleterious actions of antiparkinson agents should not be neglected. Finally, the present observations provide a springboard for additional studies of the role of individual monoaminergic receptors in the pathogenesis and management of Parkinson's disease. They suggest, further, the need for a reexamination of previous experimental and clinical findings of differences and similarities in the functional profiles of antiparkinson agents that have not integrated the notion of contrasting actions at diverse monoaminergic receptors. Additional comparative studies of the actions of antiparkinson drugs in animal models of Parkinson's disease and in Parkinsonian patients would be instructive. Such functional studies, complementary to the present cellular approach, should permit reasonable predictions concerning the therapeutic potential of future antiparkinson agents.

Conclusions

Antiparkinson drugs display a diverse pattern of agonist and antagonist actions at multiple classes of 5-HT receptor. These observations complement those of the accompanying articles (Millan et al., 2002; Newman-Tancredi et al., 2002) in underlining the heterogeneity of antiparkinson agents. Although modest agonist activity at 5-HT_1A sites may be favorable, marked stimulation of 5-HT₂ receptor subtypes may well be disadvantageous. Thus, although serotonergic mechanisms are clearly not essential for therapeutic activity in Parkinson's disease, further elucidation of the functional significance of the contrasting actions of antiparkinson agents at multiple subtypes of 5-HT receptor would be of considerable interest.