While neuroleptics, such as haloperidol and chlorpromazine, interact preferentially with D₂ receptors, serotonergic mechanisms may participate in the superior clinical profile of clozapine (Brunello et al. 1995; Meltzer 1995; Roth and Meltzer 1995; Cunningham-Owens 1996; Lieberman et al. 1998). In line with this hypothesis, the recently-characterized antipsychotics, risperidone, olanzapine and ORG5222, display more pronounced activity at 5-HT_2A vs. D₂ receptors and control positive symptoms in the relative absence of extrapyramidal side-effects (Meltzer 1995; Arnt and Skarsfeldt 1998; Ichikawa and Meltzer 1999; Remington and Kapur 2000). A further, clinically-effective antipsychotic, ziprasidone, as well as the novel benzopyrrolidine, S16924, share this preference for 5-HT_2A vs. D₂ sites and likewise reveal "atypical" profiles in functional studies (Seeger et al. 1995; Millan et al. 1998; see Millan 2000). In addition¾in analogy to clozapine¾they show agonist actions at 5-HT_1A autoreceptors, activation of which is implicated in their functional actions: notably an improvement of mood, cognitive and negative symptoms and an attenuation of extrapyramidal, motor side-effects (Ichikawa and Meltzer 1999 and Millan 2000).

Although they have received comparatively little attention, 5-HT_1B receptors (Moret and Briley 1993; Boess and Martin 1994; Gerhardt and van Heerikhuizen 1997; Barnes and Sharp 1999) are also of potential interest in the pathogenesis and management of schizophrenia.

First, postsynaptic populations are concentrated in regions implicated in psychotic disorders and their treatment: notably, the hippocampus, frontal cortex, amygdala and striatum. These populations principally comprise heteroceptors on the terminals of non-serotonergic neurones (Bruinvels et al. 1993; Saudou and Hen 1994; Bonaventure et al. 1997, 1998; Barnes and Sharp 1999; Sari et al. 1999). Further, inhibitory 5-HT_1B receptors are situated presynaptically on serotoninergic neurones themselves (Saudou and Hen 1994; Barnes and Sharp 1999). Of these, a minority may be co-localized with 5-HT_1A sites on serotoninergic cell bodies, but they are predominantly found on their terminals (Trillat et al. 1997; Gaster et al. 1998; Evrard et al. 1999; Sarhan and Fillion 1999; Millan et al. 2000a). In this regard, bearing in mind the broad implication of serotonergic mechanisms in schizophrenia (Brunello et al. 1995; Leiberman et al. 1998; Millan 2000), it is of interest to note that terminal-localized 5-HT_1B receptors interact with uptake sites for 5-HT (Daws et al. 2000). Indeed, there are indications that polymorphisms of the 5-HT transporter may be associated with alterations in psychosis scores (Malhotra et al. 1998) and that 5-HT transporters are altered in schizophrenics (Joyce et al. 1993; Dean et al. 1995). Further, 5-HT reuptake inhibitors may modify the subjective effects of hallucinogenic agents and antipsychotics in schizophrenic patients (Goff et al. 1990; Bonson et al. 1996; Evins and Goff 1996; Silver et al. 2000; Sills et al. 2000).

Second, 5-HT_1B receptors modulate the activity of ascending dopaminergic pathways via actions integrated both at terminals and cell bodies in the ventrotegmental area (mesolimbic and frontocortical pathways) and substantia nigra (nigrostriatal projection) (Johnson et al. 1992; Iyer and Bradberry 1996; Gobert et al. 1997; Barnes and Sharp 1999; Millan et al. 2000a). Activation of 5-HT_1B receptors enhances DA release in the nucleus accumbens (Hållbus et al. 1997; Boulenguez et al. 1998) and frontal cortex (Iyer and Bradberry 1996; Matsumoto et al. 1999), yet suppresses DA release in the striatum (Sarhan et al. 1999). Although the relationship of such local actions to the overall influence of systemically-administered selective 5-HT_1B ligands remains to be clarified (Galloway et al. 1993; Gobert et al. 1997; Fletcher and Korth 1999a,b; Millan et al. 1999), stimulation of 5-HT_1B receptors facilitates cocaine-induced release of mesolimbic DA in rats (Parsons et al. 1999). This action probably reflects disinhibition of mesolimbic dopaminergic pathways via inhibition of GABAergic interneurones (Johnson et al. 1992) and likely underlies potentiation of cocaine-induced reinforcement by engagement of 5-HT_1B receptors (Rocha et al. 1998; Parsons et al. 1998; Belzung et al. 2000;¾though see Fletcher and Korth 1999a, b). It is also related to the positive influence of 5-HT_1B receptors upon motor behavior (Rempel et al. 1993; Saudou et al. 1994; Skingle et al. 1996; Chaouloff et al. 1999; O'Neill et al. 2000). Moreover, 5-HT_1B receptors reduce hippocampal release of acetylcholine suggesting an inhibitory influence upon cognitive-attentional processes (Maura et al. 1989; Buhot 1997; Barnes and Sharp 1999; Malleret et al. 1999; Meneses 1999; Sarhan and Fillion 1999). Notably, 5-HT_1B receptor activation disrupts "pre-pulse inhibition" and "latent inhibition" (Cassaday et al. 1993; Boulenguez et al. 1998; Dulawa et al. 2000a,b)¾sensory processes defective in schizophrenics.

Third, activation of 5-HT_1B receptors inhibits aggressive behavior and facilitates anxious states, both of which are prominent in psychotic patients (Saudou et al. 1994; Buhot 1997). Fourth, inasmuch as clozapine and other antipsychotic agents provoke weight gain, the inhibitory and facilitory influence of 5-HT_1B receptor agonists and antagonists, respectively, upon food intake deserves mention (Trail et al. 1996; Lucas et al. 1998; De Vry and Schreiber 1999).

One reason underlying the dearth of information concerning actions of antipsychotics at 5-HT_1B receptors is the marked difference of rodent vs. homologous, human 5-HT_1B receptors (Price et al. 1996; Bonaventure et al. 1999; Barnes and Sharp 1999). Several drugs, despite high affinities at h5-HT_1B sites, fail to recognize their rodent counterparts. While guinea pig 5-HT_1B sites resemble those of man, this species is difficult to exploit for functional studies of antipsychotics (Sipes and Geyer 1996; Bonaventure et al. 1999; Barnes and Sharp 1999).

In addition to 5-HT_1B sites, closely-related 5-HT_1D receptors also operate as inhibitory autoreceptors upon serotoninergic perikarya (Davidson and Stamford 1995a,b; Pineyro et al. 1995; Bonaventure et al. 1998; Millan et al. 1999; Xie et al. 1999) and may, in theory, likewise be implicated in psychotic states and their treatment. However, owing to a lack of selective antagonists and knock-out mice, their functional significance remains unclear (Bruinvels et al. 1993; Barnes and Sharp 1999).

In light of the above, the present study examined the actions of haloperidol, clozapine and several other antipsychotics at native, rat and guinea pig 5-HT_1B sites and, in particular, at cloned h5-HT_1B and h5-HT_1D receptors. For determination of affinities, the novel ligand, [³H]-GR 125,743 (Audinot et al. 1997; Bonaventure et al. 1999) was employed and, for evaluation of efficacies, [³⁵S]-GTPS binding was exploited. This technique quantifies the initial step of h5-HT_1B and h5-HT_1D receptor coupling to G-proteins (Thomas et al. 1995; Selkirk et al. 1998; Millan et al. 1999; Mize and Alper 1999; Newman-Tancredi et al. 2000; Audinot et al. 2001). Inasmuch as 5-HT_1B and 5-HT_1D receptors show constitutive activity (Thomas et al. 1995; Pauwels et al. 1996; Selkirk et al. 1998; Middlemiss et al. 1999; Millan et al. 1999; Audinot et al. 2001), actions of antipsychotics were compared both to those of 5-HT and to those of the prototypical inverse agonist, methiothepin. In addition, interaction studies with the selective, neutral 5-HT_1B/1D antagonist, S18127 (Millan et al. 1999) were performed in order to confirm specificity of drug actions.

The procedures employed were as described in Audinot et al. (1997). Briefly, for rat cortex and guinea pig caudate, [³H]-GR125,743 (0.8nM; 70Ci/mmol, Amersham Pharmacia Biotech, Orsay, France) binding assays were carried out for 60 min at 22°C in a buffer containing 50mM Tris-HCl (pH 7.7 at 22°C), 4mM CaCl₂, 0.1% ascorbic acid and 10M pargyline. Cell membranes expressing the human 5-HT_1B and 5-HT_1D receptors respectively denominated CHO-h5-HT_1B and CHO-h5-HT_1D, were incubated for 60 min at 22°C in the same buffer without pargyline, with [³H]-GR125,743 (1nM; Audinot et al. 1997; Domenech et al. 1997). 5-HT (10M) was used to define non- specific binding. Incubations were terminated by rapid filtration using a cell harvester through Whatman GF/B filters pretreated with polyethylenimine (0.1% v/v). Data were analyzed by non-linear regression using the program PRISM (GraphPad Software Inc., San Diego, CA), to yield IC₅₀ (Inhibitory Concentration₅₀) values. Inhibition constants (K_i) were calculated according to the Cheng-Prusoff equation: K_i = IC₅₀/[1+(L/K_D)], where L is the concentration of the radioligand and K_D the dissociation constant.

The procedures employed were as described in Millan et al. (1998). Briefly, for [³H]-raclopride (2 nM, NEN, Les Ulis, France) binding to rat striatal membranes, incubations lasted 30 min at 22°C in a buffer containing 50mM Tris-HCl (pH 7.7 at 22°C), 4mM CaCl₂, 0.1% ascorbic acid, 5mM KCl, 1mM MgCl₂, 120mM NaCl and 10M pargyline. Non-specific binding was defined with spiperone (10M). For [¹²⁵I]-iodosulpride (0.1nM, Amersham Pharmacia Biotech, Orsay, France) binding to recombinant human D₂ receptors expressed in CHO cells, incubations lasted 30 min at 30°C in the same buffer without pargyline. Non-specific binding was defined with raclopride (10M).

Effects of receptor ligands on [³⁵S]-GTPS binding at CHO-h5-HT_1B and CHO-h5-HT_1D membranes

Receptor-linked G protein activation at h5-HT_1B and h5-HT_1D receptors was determined by measuring stimulation of [³⁵S]-GTPS (1000 Ci/mmol, Amersham Pharmacia Biotech, Orsay, France) binding. Briefly, as previously (Audinot et al. 2001) membranes (15-25g) were incubated (30 min at 22°C) with ligand in a final volume of 250l of buffer B (20 mM HEPES, pH 7.4 at 22°C, 3 M GDP, 3 mM MgCl₂, 100 mM NaCl and 0.1 nM [³⁵S]-GTPS). Non-specific binding was defined with GTPS (10M). Incubations were terminated by rapid filtration through Whatman GF/B filters. Data were analyzed by non-linear regression using the program PRISM, to yield EC₅₀ (Effective Concentration₅₀) or IC₅₀ values (Inhibitory Concentration₅₀). For antagonism experiments, K_B values of S18127 were calculated according to Lazareno and Birdsall (1993).

CHO-h5-HT_1B and CHO-h5-HT_1D cell membranes expressing 8 and 1.6 pmol/mg receptors respectively were purchased from Euroscreen (Brussels, Belgium). 5-HT sulphate, haloperidol, clozapine and chlorpromazine were purchased from SIGMA (Saint Quentin Fallavier, France) and, methiothepin maleate from Tocris-Fisher Bioblock (Illkirch, France). Olanzapine was obtained from Eli Lilly (Indianapolis, USA), ORG5222 from Organon (Oss, the Netherlands) and ocaperidone from Janssen (Beerse, Belgium). S16924 HCl and ziprasidone HCl were synthesized by G. Lavielle, Servier and S18127 HCl and risperidone by J.L. Peglion, Servier. GR125,5743 is (N-[4-methoxy-3-(4-methylpiperazin-1-yl)phenyl}-3-methyl- 4-(4-pyridyl)benzamide); S16924 is (R)-2-1-{1-[2-(2,3-dihydro-benzo[1,4] dioxin-5-yloxy)-ethyl]-pyrrolodin-3yl}-1-(4-fluorophenyl) ethanone; ORG-5222 is {trans-5-chloro-2-methyl-2,3,3a,12b-tetrahydro-1H dibenz[2,3:6,7]oxepino[4,5-c]pyrrole} and S18127 is N-[1-(1,4-benzodioxan-5yl)piperidin 4-yl] N-(indan 2-yl) amine.

As shown in Table 1, each antipsychotic displayed low affinity at native, rat 5-HT_1B receptors relative to native rat D₂ receptors (Millan et al. 1998, 2000b). The affinity of haloperidol at native, guinea pig 5-HT_1B receptors was similarly low as compared to native, rat D₂ receptors. Chlorpromazine and olanzapine shared the relatively low affinity of haloperidol at guinea pig 5-HT_1B vs. rat D₂ receptors. On the other hand, clozapine, risperidone and ocaperidone showed affinities at guinea pig 5-HT_1B sites which were only 10 to 20-fold lower than at native rat D₂ receptors. Further, for S16924, ORG5222 and ziprasidone, affinities at guinea pig 5-HT_1B sites were similar to rat D₂ receptors.

Drug affinities at h5-HT_1B receptors were similar to those determined at native, guinea pig 5-HT_1B receptors and markedly higher than affinities at native, rat 5-HT_1B receptors (Table 1). The affinities of haloperidol, olanzapine and chlorpromazine at h5-HT_1B receptors were substantially inferior to their affinities at cloned hD₂ receptors. Clozapine, risperidone and ocaperidone displayed affinities some 10-fold lower than those at hD₂ sites, while S16924, zisprasidone and ORG5222 revealed affinities at h5-HT_1B sites similar to their affinities at hD₂ receptors. With the exception of haloperidol, which displayed (like 5-HT) a 10-fold preference for h5-HT_1D vs. h5-HT_1B receptors, antipsychotics manifested similar affinity for h5-HT_1D as compared to h5-HT_1B sites.

Influence upon [³⁵S]-GTPS binding at h5-HT_1B and h5-HT_1D receptors

5-HT dose-dependently stimulated [³⁵S]-GTPS binding with a maximal effect (relative to basal values = 100 %) of +252% and +133% at h5-HT_1B and h5-HT_1D receptors, respectively, (Figure 1), Table 2). These data are in line with our previous studies of these sites (Audinot et al. 2001, Newman-Tancredi et al. 2000). In contrast, methiothepin dose-dependently inhibited basal [³⁵S]-GTPS binding at h5-HT_1B and h5-HT_1D receptors with a similar magnitude at both subtypes (Figure 1). S18127 did not significantly modify [³⁵S]-GTPS binding at either h5-HT_1B or h5-HT_1D receptors over a substantial concentration range (Figure 1, Table 2). At h5-HT_1B receptors, haloperidol, chlorpromazine, olanzapine, clozapine, S16924, risperidone, ocaperidone and ORG 5222, all suppressed basal [³⁵S]-GTPS binding in analogy to methiothepin and suggestive of inverse agonist properties (Figure 2, Table 2). S16924 and olanzapine showed the least and most pronounced "negative" efficacy in this respect. In contrast to these agents, ziprasidone behave as a weak partial agonist relative to 5-HT (Table 2). Antipsychotic potencies (pIC₅₀s) in exerting inverse agonist actions correlated significantly (r = 0.74, p < .05) with their affinities (pK_is) at these sites. At h5-HT_1D receptors, all antipsychotics manifested inverse agonist actions. Risperidone, zipradone and haloperidol were the most "efficacious" in this regard, while clozapine and S16924 showed relatively modest negative efficacy (Figure 2, Table 2). Again, antipsychotic potencies in exerting inverse agonist actions correlated well (r = 0.90, p < .005) with their affinities at these sites. Negative efficacies were significantly (p < .05) more pronounced at h5-HT_1D vs. h5-HT_1B sites for haloperidol, ziprasidone and risperidone¾as well as ziprasidone¾whereas they were similar for other drugs.

For antagonist studies with S18127, six key drugs were selected, two from each group of antipsychotics; those with relatively low h5-HT_1B/1D vs. hD₂ affinity (haloperidol and olanzapine); those with modest affinity (clozapine and risperidone) and those with relatively high h5-HT_1B/1D activity (S16924 and ziprasidone). S18127, which did not itself markedly affect [³⁵S]-GTPS binding at h5-HT_1B/1D receptors (Figure 3) concentration-dependently and completely reversed the inverse agonist effects of haloperidol, olanzapine, clozapine, risperidone and S16924 at both h5-HT_1B and 5-HT_1D receptors (Figures 3 and 4). This was also the case for the inverse agonist effect of ziprasidone at h5-HT_1D receptors (Figure 4). While its partial agonist effect at h5-HT_1B receptors was also reversed, this effect was too weak to obtain dose-dependent inhibition (not shown). S18127 exerted its antagonistic action with similar potencies in all cases, although pK_B values were somewhat lower for S16924 and methiothepin at h5-HT_1B receptors, and ziprasidone and methiothepin at h5-HT_1D receptors (Table 3).

[³⁵S]-GTPS binding at h5-HT_1B and h5-HT_1D receptors: agonist, antagonist and inverse agonist actions

In corroboration of prior investigations both by this laboratory (Millan et al. 1999; Newman-Tancredi et al. 2000; Audinot et al. 2001) and by others (Thomas et al. 1995; Pauwels et al. 1996; Gaster et al. 1998; Selkirk et al. 1998), [³⁵S]-GTPS binding revealed agonist properties of 5-HT at both h5-HT_1B and h5-HT_1D receptors at concentrations corresponding to its affinity for these sites. The stimulation by 5-HT of [³⁵S]-GTPS binding was less pronounced at h5-HT_1D than h5-HT_1B sites. This difference is possibly due to the lower density of h5-HT_1D vs. h5-HT_1B receptor expression inasmuch as receptor reserve, together with factors such as receptor-G protein stoichiometry, may influence agonist efficacy at these sites (Adham et al. 1992; Thomas et al. 1995; Zgombick et al. 1996; Selkirk et al. 1998; Milligan 2000; Audinot et al. 2001; Newman-Tancredi et al. 2000). Both h5-HT_1B and h5-HT_1D receptors show constitutive activity (Gaster et al. 1998; Millan et al. 1999; Newman-Tancredi et al. 2000; Audinot et al. 2001) and, in confirmation of previous work, a robust inhibition of basal [³⁵S]-GTPS binding was elicited by the inverse agonist, methiothepin (Thomas et al. 1995; Pauwels et al. 1996; Zgombick et al. 1996). In distinction, the novel, selective h5-HT_1B/h5-HT_1D ligand, S18127 (Millan et al. 1999), behaved as a "neutral" antagonist. The maximal reduction of [³⁵S]-GTPS binding was more marked at h5-HT_1D than at h5-HT_1B receptors and, in analogy to agonists, such differences likely reflect diverse factors including receptor-G-protein stoichiometry and the proportion of spontaneously G-protein-coupled vs. non-coupled receptors: this issue is discussed in detail elsewhere (Audinot et al. 2001; Adham et al. 1992; Thomas et al. 1995; Dickenson and Hill 1998; Newman-Tancredi et al. 1998; Millan et al. 1999; Milligan 2000). Irrespective of such considerations, the present conditions are, evidently, appropriate for the characterization and differentiation of potential agonist vs. inverse agonist actions of antipsychotics at h5-HT_1B and h5-HT_1D receptors.

In corroboration of previous investigations, antipsychotic agents displayed low to negligible affinity for native, rat 5-HT_1B receptors (Leysen et al. 1988; Moore et al. 1993; Seeger et al. 1995) suggesting that interactions at these sites are unlikely to play a major role in functional actions in rodents. On the other hand, at both guinea pig 5-HT_1B receptors and homologous h5-HT_1B sites, they showed a broad range of affinities. In this regard, ziprasidone was a potent agent, extending previous report of its high affinity at bovine, caudate 5-HT_1B receptors labeled with the non-selective agent, [³H]5-HT, in the presence of mesulergine and 8-OH-DPAT to mask binding to other sites (Seeger et al. 1995). Further, likewise employing an agonist, [³H]-alniditan, ziprasidone displayed high affinity at h5-HT_1B receptors expressed in a L929 cell line (Schotte et al. 1996). Indeed, the latter study, while employing experimental conditions different from the present investigation, observed affinities for various antipsychotics similar to those acquired herein, underpinning the coherence of these findings. In addition, employing a C6 glioma cell line, Schotte et al. 1996 reported affinities for several antipsychotics at h5-HT_1D receptors remarkably close to those obtained in the present study and, as herein, haloperidol revealed a (modest) preference for h5-HT_1D vs. h5-HT_1B receptors. S16924 (Millan et al. 1998), which has not been previously evaluated, mimicked the low affinity of other antipsychotics at rat 5-HT_1B sites yet manifested marked affinity for both h5-HT_1B and h5-HT_1D as well as guinea pig 5-HT_1B receptors.

For all known antipsychotic agents, blockade of (mesolimbic) dopamine D₂ receptors is a crucial element in the control of core, positive symptoms (Seeman and Tallerico 1999; Remington and Kapur 2000). It is, thus, important to express and interpret antipsychotic affinities at h5-HT_1B and h5-HT_1D receptors relative to those at D₂ receptors, an approach previously adopted for other 5-HT receptor types (Roth and Meltzer 1995; Lieberman et al. 1998; Newman-Tancredi et al. 1998; Millan et al. 2000b). Within this perspective, haloperidol, chlorpromazine and olanzapine interacted with h5-HT_1B and h5-HT_1D sites only at concentrations far higher than hD₂ sites, whereas the degree of separation was modest for clozapine, risperidone and ocaperidone. For the latter agents, h5-HT_1B and h5-HT_1D sites may, therefore, be of functional pertinence in vivo. Moreover, this assertion is clearly valid for S16924, ziprasidone and ORG5222, which display comparable affinities for h5-HT_1B, h5-HT_1D and hD₂ receptors.

With the exception of a very recent report showing inverse agonist action of ocaperidone at 5-HT_1D receptors expressed in a bacillovirus expression system (Brys et al. 2000), this is the first study of functional actions of antipsychotic agents at cloned h5-HT_1B and h5-HT_1D receptors. Importantly, we used a mammalian expression system, and the key finding was that antipsychotics all behaved as inverse agonists at both h5-HT_1D receptors and, with the exception of ziprasidone, at h5-HT_1B receptors. In fact, the magnitude of h5-HT_1B receptor stimulation elicited by ziprasidone was low. In the presence of 5-HT, then, ziprasidone will exert principally antagonist properties at 5-HT_1B receptors. This contention is consistent with the report of Seeger et al. 1995 that ziprasidone antagonizes the influence of 5-HT upon adenylyl cyclase in guinea pig substantia nigra, a structure enriched in 5-HT_1B receptors. Nevertheless, under conditions of low serotonergic tone, partial agonist properties of ziprasidone at 5-HT_1B receptors may be revealed.

In terms of inverse agonist efficacy, haloperidol, oraperidone and ziprasidone revealed a more marked influence at h5-HT_1D vs. h5-HT_1B sites, although the significance of differential degrees of inverse agonist efficacy remains unclear (see Millan et al. 1999 and below). Notably, pIC₅₀ values for inhibition of basal [³⁵S]-GTPS binding at h5-HT_1B or h5-HT_1D receptors were significantly correlated with drug affinities at these sites. This observation supports the specificity of drug actions at h5-HT_1B and h5-HT_1D receptors in the mediation of their inverse agonist properties. This contention is powerfully supported by studies with the selective, neutral antagonist, S18127 (Millan et al. 1999), which abolished the influence of antipsychotics upon [³⁵S]-GTPS binding¾without itself exerting activity.

As indicated in Table 3, pK_b values for blockade of the actions of methiothepin at both h5-HT_1B and h5-HT_1D sites by S18127 were lower than those for 5-HT. This difference might reflect blockade of inverse agonist as compared to agonist actions. However, this appears improbable inasmuch as pK_b values for blockade of the actions of several other inverse agonists (such as haloperidol) by S18127 were comparable to those of 5-HT. In fact, it has previously been suggested that a low pK_b for S18127 against methiothepin at h5-HT_1B sites reflects a distinctive influence of methiothepin upon receptor-G protein coupling or stability (Audinot et al. 2001). Similar arguments have been advanced as concerns the unusual interaction of methiothepin at h5-HT_1A receptors (McLoughlin and Strange 2000). Such differences may similarly underlie contrasting pK_b values for S18127 against other ligands, and notably low pK_bs vs. S16924. Information concerning pK_b values of selective neutral antagonists in blocking the actions of inverse agonists are virtually non-existent. Indeed, while K_b values of neutral antagonists should be constant against various agonists, it remains unclear as to whether this likewise applies to inverse agonists. This question would justify, thus, additional investigation.

The above observations demonstrate that¾relative to 5-HT¾all antipsychotics behaved as antagonists at h5-HT_1B and h5-HT_1D receptors and, with the exception of ziprasidone at the former, all showed inverse agonist properties. Based on observations outlined at the beginning of this article, antagonism (or inverse agonism) at h5-HT_1B receptors in schizophrenic patients may have important functional consequences for ziprasidone, ORG5222 and S16924 as well as, possibly, clozapine, ocaperidone and risperidone¾although it is improbable that haloperidol, chlorpromazine and olanzapine attain concentrations sufficient to occupy these sites. In particular, actions at 5-HT_1B sites may modulate mesolimbic, frontocortical or nigrostriatal DA release, cognitive function and sensory gating and mood.

In fact, it has been suggested that methiothepin elevates resting 5-HT release in the hypothalamus by inverse agonist actions at terminal 5-HT_1B receptors (Moret and Briley 1993). However, subsequent studies of methiothepin and selective 5-HT_1B antagonists, employing diverse measures of receptor-coupling, 5-HT release and other functional models, have not provided additional evidence for such inverse agonist actions in corticolimbic structures: indeed, it is difficult to differentiate potential inverse agonist actions from antagonism of the inhibitory effects of endogenous 5-HT upon serotonergic neurons (Roberts et al. 1997; Alper and Nelson 1998; Selkirk et al. 1998; Mize and Alper 1999; Stenfors et al. 2000). Moreover, although an increase in 5-HT release by antipsychotics could, theoretically, alleviate depressive states co-morbid with psychotic symptoms, agonist properties of clozapine, S16924 and ziprasidone at inhibitory 5-HT_1A autoreceptors would mask any potential increase in 5-HT release due to their inverse agonist/ antagonist actions at 5-HT_1B receptors (Sprouse et al. 1997; Millan et al. 1998; Millan 2000). Interestingly, actions at 5-HT_1B sites may be involved in the elevation by risperidone of frontocortical 5-HT release in rats (Hertel et al. 1998, 1999), although this action also involves blockade of ₂-adrenoceptors (Hertel et al. 1997). Finally, though studies in guinea pig could, in principle, reveal inverse agonist actions of antipsychotic agents, at 5-HT_1B/1D receptors, their numerous interactions at other receptors (Brunello et al. 1995) would render identification of such effects challenging. Further, it remains under discussion as to whether inverse agonist properties are genuinely expressed at 5-HT_1B, 5-HT_1D or other 5-HT receptor types in vivo (see Millan et al. 1999).

In conclusion, antipsychotic agents examined herein show low affinity for rat 5-HT_1B receptors questioning the significance of actions at these sites to their functional properties in rodents. However, at both guinea pig and homologous, human 5-HT_1B (and 5-HT_1D) receptors, they show a spectrum of affinities relative to hD₂ receptors ranging from weak (haloperidol, chlorpromazine and olanzapine) through intermediate (clozapine, olanzapine and risperidone) to pronounced (S16924, ORG5222 and ziprasidone). All antipsychotics manifested low efficacy at h5-HT_1B and h5-HT_1D sites indicating essentially antagonist properties with respect to 5-HT. Indeed, with the exception of ziprasidone at h5-HT_1B sites, all behaved as inverse agonists. The potential significance of inverse agonist properties at cerebral populations of h5-HT_1B and h5-HT_1D receptors remains unclear. However, as outlined at the beginning of this article, the broad implication of h5-HT_1B receptors in the modulation of functions pertinent to schizophrenia and its treatment (see Introduction) suggests that, for certain antipsychotic agents actions at h5-HT_1B and/or h5-HT_1D receptors may be of functional significance.

Acknowledgements

We thank C. Moreira, V. Pasteau, M. Touzard, N. Fabry and C. Lahaye for skillful technical assistance.