Discussion

Receptor profiles of S 16924 as compared to clozapine and haloperidol in relationship to their functional actions.

As described in the companion paper, S 16924 possesses a pattern of interaction at multiple dopaminergic, serotonergic and adrenergic receptors that differs markedly to that of haloperidol and that closely resembles that of clozapine (table 8). In particular, modest and equilibrated affinity for D₂ and D₁ receptors and more pronounced affinity at D₄, α₁-AR, 5-HT_2A and 5-HT_2Cvs. D₂ receptors. In addition, S 16924 may be distinguished from both haloperidol and clozapine by its potent, partial agonist properties at 5-HT_1A receptors. The following discussion illustrates the importance of the distinctive receptorial profile of S 16924 in determining its functional actions in vivo.

Potential antipositive properties of S 16924: significance of D₂ receptor blockade.

In line with previous studies, and reflecting its potent blockade of postsynaptic D₂receptors in limbic structures, haloperidol was highly active in each of the tests predictive of potential antipositive activity (Brunelloet al., 1995; Kahn and Davis, 1995). Similarly in line with previous work, clozapine was active in each paradigm, although only at substantially higher doses, in correspondence with its lower affinity at D₂ receptors (Brunello et al., 1995; Meltzer, 1995). The activity of haloperidol and clozapine in these models is consistent with their clinical efficacy over, respectively, low and high dose ranges in controlling the positive symptoms of schizophrenia (Kane and Freeman, 1994; Meltzer, 1995; Pickar, 1995). In accordance with its intermediate affinity for D₂ receptors, S 16924 also manifested robust and dose-dependent activity in each of these models over a dose-range lying between those of haloperidol and clozapine (Maurel-Remy et al., 1995a and b). These data support, then, the hypothesis that S 16924 should be effective in controlling the positive symptoms of schizophrenia via antagonist actions at D₂ receptors. The robust inhibition by S 16924 of hyperlocomotion induced by the NMDA receptor channel blocker, dizocilpine, is of particular interest inasmuch as a reduction in glutamatergic transmission may be involved in the pathogenesis of schizophrenia (Bartha et al., 1997; Evins et al.,1997). Further, clozapine reduces the psychosis elicited by a further open channel blockade at NMDA receptors, ketamine, in man (Malhotraet al., 1997). Notwithstanding the importance of D₂ receptor blockade, selective antagonists at D₁ receptors, such as SCH 39166, are also potently active in the models used herein. Further, D₁ receptor antagonism may enhance the antipsychotic effects of D₂ receptor blockade although, clinically, selective blockade of D₁receptors does not appear to be sufficient for antipsychotic activity (Josselin et al., 1997; Martin et al., 1994;Pickar, 1995). Indeed, as indicated by the rotation model employed herein (table 1), in contrast to haloperidol, S 16924 and clozapine both exert antagonist properties at D₁ receptors at doses close to those blocking D₂ receptors. Further, the selective α₁-AR antagonist, prazosin, similarly attenuates the hyperlocomotion provoked by dizocilpine and other psychostimulants (Blanc et al., 1994; Mathé et al., 1996). These observations suggest that the antagonist properties of S 16924 and clozapine at D₁ and α₁-AR receptors may also contribute to their potential antipositive actions. Indeed, it might be conjectured that combined D₂, D₁ and α₁-AR receptor blockade, a property shared by clozapine and S 16924, may improve efficacy in patients who respond poorly to selective D₂receptor blockade with haloperidol (Prinssen et al., 1994). Finally, it has been suggested that stimulation of 5-HT_1Areceptors, although poorly effective alone, may enhance the antipositive effects of D₂ receptor blockade (Evenden, 1992). This observation is evidently of pertinence to S 16924, a potent 5-HT_1A partial agonist, although we have not, as yet, acquired evidence that this property contributes to its actions in the models of antipositive activity summarized in table 1.

Preferential blockade of PCP-induced hyperlocomotion and DOI-induced HTWs by S 16924: importance of 5-HT_2Aantagonist properties.

In contrast to amphetamine, administration of PCP to normal subjects mimics the negative-cognitive as well as the positive symptoms of schizophrenia (Gorelick and Balster, 1995;Steinpreis, 1996). We have hypothesized that preferential blockade of PCP- vs. amphetamine-induced locomotion may suggest a clozapine-like ability to control negative symptoms (Maurel-Remyet al., 1995a and b). Indeed, our study confirms the greater sensitivity of PCP- as compared to amphetamine-induced (and spontaneous) locomotion to clozapine, but not haloperidol, and a similar profile was presented by S 16924. This differential responsiveness of PCP- as compared to amphetamine-, cocaine- and dizocilpine-induced locomotion to clozapine and S 16924 may reflect the contrasting mechanisms underlying their induction of locomotion. Thus, amphetamine-, cocaine- and dizocilpine-induced locomotion is mediated (independently of 5-HT_2A receptors) by nucleus accumbens-localized D₂ receptors (Jackson et al., 1994; Maurel-Remy et al., 1995a and b; Moore and Kenyon, 1994). In distinction, although dependent on dopaminergic mechanisms for its expression, PCP-evoked hyperlocomotion involves serotonergic mechanisms. Indeed, it may involve the activation of nucleus accumbens-localized 5-HT_2A receptors by an increase in extracellular 5-HT levels: this results from an interference by PCP of 5-HT reuptake processes (Jackson et al., 1994; Kehneet al., 1996; Maurel-Remy et al. 1995a and b;Moore and Kenyon, 1994). Correspondingly, the more potent 5-HT_2Avs. D₂ antagonist properties of S 16924 (table 8) and clozapine likely underlies their preferential blockade of the actions of PCP as compared to amphetamine (Maurel-Remyet al., 1995a). S 16924 and clozapine were also highly potent in blocking induction of HTWs by the hallucinogen, DOI and, in analogy to PCP-induced locomotion, 5-HT_2A receptors mediate the induction of HTWs by DOI, although this population of 5-HT_2A sites is localized in the FCX (Schreiber et al., 1995; Willins and Meltzer, 1997). Thus, a common feature of PCP-induced hyperlocomotion and DOI-induced HTWs is their mediation by 5-HT_2A receptors and sensitivity to the potent 5-HT_2A antagonist properties of S 16924 and clozapine. These data are, thus, consistent with the suggestion that, like clozapine, and by virtue of its potent 5-HT_2A antagonist properties, S 16924 may control the negative symptoms of schizophrenia in which a dysfunctionment of serotonergic as well as dopaminergic transmission has been implicated (Martin et al., 1997;Meltzer, 1995; Svensson et al., 1995).

DS properties of S 16924.

As mentioned above, DOI behaves as an agonist at 5-HT_2A receptors, stimulation of which underlies its DS properties in rats (Schreiber et al.,1994). Correspondingly, in line with their potent 5-HT_2Aantagonist properties, S 16924 and clozapine, but not haloperidol, abolished the DS properties of DOI. A direct comparison of the DS properties of S 16924 and clozapine revealed that, in clozapine-trained animals, S 16924 showed complete (≥80%) generalization suggesting that animals share DS properties with those of clozapine. This interpretation is reinforced by the complete (≥80%) generalization of clozapine in animals trained to recognize S 16924 itself. In distinction, in line with previous studies, haloperidol achieved no more than partial (≤50%) generalization to clozapine or to S 16924 (Carey and Bergman, 1997; Nielsen, 1988). Thus, S 16924 and clozapine clearly share similar DS properties, whereas those of haloperidol are different. It has proven difficult to identify the receptorial interactions underlying the DS properties of clozapine, although a role of muscarinic receptors (in rats) and 5-HT_2C receptors (in pigeons) has been proposed (Hoenicke et al., 1992; Nielsen, 1988). The present data are consistent with the latter possibility inasmuch as S 16924 is a potent ligand at 5-HT_2C—but not muscarinic receptors (companion paper and tables 7 and 8). In addition, complementary studies (Dekeyne A and Millan MJ, unpublished observations) have indicated that antagonist properties at 5-HT_2A and 5-HT_2C receptors may be involved in the DS properties of S 16924 and clozapine in rats. Nevertheless, multiple receptorial interactions likely underlie the complex DS properties of clozapine (Carey and Bergman, 1997) and S 16924. Notably, the actions of S 16924 and clozapine in these DS paradigms were expressed at doses lower than those disrupting response rates, in line with their lack of cataleptic activity (see below).

EEG profile of S 16924.

Although it is not possible to specify which cerebral structures and neuronal circuits are implicated in the DS properties of S 16924 and clozapine, it is not unreasonable to assume that both limbic and cortical regions are involved in view of their rich monoaminergic innervation and importance in the actions of antipsychotic agents. Thus, it is of interest that S 16924 and clozapine exerted a similar influence on EEG power spectra derived from the FCX, whereas the action of haloperidol was clearly different. It is difficult to attribute particular EEG complexes to interactions at specific receptor types. However, in complementary studies (Sebban C,et al., unpublished observations) the D₂/D₃ antagonist, raclopride, reproduced the profile of haloperidol, consistent with a role of antagonist properties at D₂ sites, whereas the profile of S 16924 and clozapine was mimicked by 5-HT_2A antagonists, consistent with a role of antagonist actions at 5-HT_2A receptors. The similarity in the EEG power spectra of S 16924 and clozapine, as compared to haloperidol, obtained from the FCX is of particular significance inasmuch as a functional perturbation (“hypofrontality”) of this region is implicated in the negative and cognitive-attentional symptoms of schizophrenia (Andreasen et al., 1992; Knable and Weinberger, 1997).

Modulation of cognitive-attentional function: induction of LI by S 16924.

To more directly explore the possibility that S 16924 may improve cognitive-attentional performance, and increase the “filtering” of irrelevant sensory stimuli, a function compromised in schizophrenic patients, we examined its influence in a paradigm of LI: that is, the decrease in the conditioning power of a stimulus after its previous presentation (Weiner and Feldon, 1997). In the CER model used, SRs were slightly higher for animals preexposed (10 times) to an auditory stimulus during a drinking session as compared to those which had not been preexposed (Weiner and Feldon, 1997). In line with recent studies (Moran et al., 1996), clozapine enhanced the difference between these groups via a selective action on SRs in the latter, an action mimicked by S 16924. That is, in their presence, there was a significant LI, indicating that both S 16924 and clozapine may enhance the ability to ignore irrelevant sensory stimuli. In distinction, haloperidol was ineffective herein, although under certain conditions LI can also be induced by haloperidol (Moran et al., 1996; see Ruob et al., 1997 and Weiner and Feldon, 1997). The mechanisms underlying the potential ability of S 16924 and clozapine to improve cognitive-attentional function are not, as yet, clear. However, activation and blockade of 5-HT_2A receptors can disrupt and induce LI, respectively (Hitchcock et al.,1997; Moser and Moran, 1994). Thus, 5-HT_2A antagonist properties of S 16924 and clozapine may contribute, although 5-HT_1A agonist properties of S 16924 are unlikely to be involved because 8-OH-DPAT does not induce LI (Gray et al.,1995; Millan MJ and Schreiber R, unpublished observations). Thalamic populations of α₁-AR receptors inhibit the filtering of information en route to the cortex and prazosin blocks PCP-induced disruption in a pre-pulse inhibition model (Bakshi and Geyer, 1997;McCormick and Pape, 1990). Thus, α₁-AR receptor blockade might also be involved in the actions of S 16924 and clozapine. Several arguments also favor a implication of D₄ receptors blockade in the LI-inducing actions of S 16924 and clozapine. Thus, the hippocampus is implicated in mechanisms underlying LI (Gray et al., 1995) and D₄ receptors are concentrated both therein as well as in the FCX and other structures controlling cognition (Primus et al., 1997). Further, selective D₄ receptor antagonists induce LI in rats (Millan et al., in preparation). Finally, the ability of S 16924 and clozapine to reinforce mesocortical vs. mesolimbic dopaminergic transmission (companion paper) may be of significance inasmuch as a disruption of LI is associated with a decrease and increase, respectively, in the activity of these pathways (Broersenet al., 1996; Gray et al., 1995). Indeed, disruption of LI by amphetamine likely reflects an enhancement of mesolimbic dopaminergic transmission thereby activating D₂receptors in the nucleus accumbens (Gray et al., 1995;Weiner and Feldon, 1997). Correspondingly, the ability of haloperidol and higher doses of S 16924 and clozapine to prevent disruption of LI by amphetamine presumably reflects antagonism of mesolimbic D₂ receptors (Gray et al., 1995; Moran et al., 1996; Weiner and Feldon, 1997). It would be of interest to further explore the, potentially multiple, mechanisms involved in the control of LI and other cognitive functions by S 16924 and to determine its potential influence on LI and cognition in schizophrenic patients (Gray et al., 1995).

Low extrapyramidal potential of S 16924: importance of 5-HT_1A agonist properties.

Haloperidol potently elicited catalepsy in rats, a response reflecting interruption of activity at D₂ receptors in the striatum and predictive of an extrapyramidal motor syndrome in man (Cunningham-Owens, 1996;Hoffman and Donovan, 1995). In contrast, S 16924 did not elicit catalepsy suggesting that, in analogy to clozapine (Cunningham-Owens, 1996; Keks, 1996), its extrapyramidal potential is low. Indeed, clozapine and, more potently, S 16924 inhibited the induction of catalepsy by haloperidol and the D₁ receptor antagonist, SCH 39166 (Bartoszyk et al., 1996; Cunningham-Owens, 1996;Kalkman et al., 1997). There are several mechanisms that may be relevant to this low extrapyramidal potential of S 16924 and clozapine. First, selective D₁ or D₂antagonists may elicit catalepsy by provoking a dysequilibrium in activity at striatal D₁ and D₂ sites (Brunelloet al., 1995; Josselin et al., 1997) and both S 16924 and clozapine, in contrast to haloperidol and SCH 39166, show balanced antagonist activity at D₁ and D₂receptors (table 1 and companion paper). Second, the anticholinergic properties of clozapine have been implicated in its low extrapyramidal potential although, in view of its low affinity for muscarinic receptors, this mechanism is not relevant to S 16924 (Brunello et al., 1995; Olianas et al., 1997; Zorn et al., 1994) (table 7). Third, a marked preference of antipsychotics for 5-HT_2Avs. D₂ receptors has been convincingly associated with a low extrapyramidal potential, a property share by both S 16924 and clozapine (Kehne et al., 1996;Roth and Meltzer, 1995; although see Wadenberg, 1996). Fourth, 5-HT_1A agonists may counter the extrapyramidal properties of neuroleptic agents, likely reflecting activation of inhibitory 5-HT_1A autoreceptors in raphe nuclei (Invernizzi et al., 1988). Indeed, S 16924 is a 5-HT_1A receptor partial agonist and, in contrast to haloperidol, inhibits striatal turnover of 5-HT although only weakly augmenting striatal DA synthesis (companion paper). Such a pattern of actions has been correlated with a low extrapyramidal potential, although it is controversial as to whether a reduction in striatal serotonergic transmission plays a causal role in limiting the influence of D₂ receptor antagonism upon induction of catalepsy (Andersen and Kilpatrick, 1996;Lucas et al., 1997). In any case, the selective 5-HT_1A antagonist, WAY 100,635, dose-dependently abolished the anticataleptic actions of S 16924 against haloperidol demonstrating that this is its principal mechanism of action. This assertion was supported by a study of methylphenidate-induced gnawing, a behavior also involving activation of striatal population of D₂receptors. The ability of S 16924 to block this response was also abolished by WAY 100,635, in line with data obtained with other 5-HT_1A agonists (Kleven et al., 1996). Interestingly, WAY 100,635 was less effective in blocking the anticataleptic actions of clozapine and did not modify its antagonism of methylphenidate. This suggests that agonist actions at 5-HT_1A receptors are less important to the low extrapyramidal potential of clozapine than that of S 16924 (Bartoszyket al., 1996; Kalkman et al., 1997; Klevenet al., 1996). Clozapine moderates extrapyramidal movement disorders (tardive dyskinesias) elicited by prior treatment with neuroleptic drugs (Cunningham-Owens, 1996; Lieberman et al.,1991). This clinical, antidyskinetic action of clozapine may correspond to its ability to block haloperidol-induced catalepsy. Thus, S 16924 might also be of utility in the control of dyskinesias due to antecedent neuroleptic treatment. Notably, even after chronic treatment, S 16924 (and clozapine) were still capable of blocking the cataleptic actions of haloperidol suggesting that their ability to reduce extrapyramidal effects may not adapt.

Limited impact of S 16924 upon PRL secretion.

In contrast to the increase in PRL levels provoked by haloperidol, S 16924 and clozapine elicited only a modest augmentation in PRL levels, even at high doses. This is of significance since hyperprolactinaemia is associated with ovulatory dysfunction, ammenorrhea and galactorrhea (Cunningham-Owens, 1996). The differential induction of PRL secretion by S 16924 and clozapine vs. haloperidol may likely be explained by the relatively weak affinity of the former at lactotrophic D₂ receptors tonically controlling PRL release (Cunningham-Owens, 1996; Millan et al., 1995a). Interestingly, at least in a transfected cell line, D₂receptors controlling PRL release may be “constitutively active” (Nilsson et al., 1996). That is, the induction of PRL secretion by haloperidol may reflect negative intrinsic activity at D₂ sites rather than interruption of a tonic inhibitory control exerted by DA. Although this is an intriguing hypothesis, it is unlikely to fully account for the influence of drugs upon PRL secretionin vivo inasmuch as both haloperidol and clozapine behave as inverse agonists at cloned hD₂ receptors (Hall and Strange, 1997). Currently, it has not been established whether S 16924 possesses negative efficacy at D₂ sites, an issue justifying further investigation inasmuch as inverse agonist actions at various populations of D₂ receptors might, in theory, also explain the induction of catalepsy and DA synthesis. Irrespective of the underlying mechanisms, these data collectively suggest that, as with clozapine, S 16924 may be distinguished from haloperidol as concerns its low potential for extrapyramidal motor and endocrine side-effects.

Absence of a serotonergic syndrome.

8-OH-DPAT and other high efficacy agonists at postsynaptic 5-HT_1A receptors provoke a constellation of behaviors termed the “serotonin syndrome,” and including FBP (Tricklebank, 1985). They also elicit STFs, a behavior reflecting interference with sensory transmission in the dorsal horn of the spinal cord (Millan et al., 1994). S 16924 did not, however, elicit either FBP or STFs consistent with its partial agonist properties in a cellular model of coupling at postsynaptic h5-HT_1A receptors (companion paper) and other in vivo models, such as induction of hypothermia, where the magnitude of the actions of S 16924 are less pronounced than those of full agonists (Millan MJ and Dekeyne, unpublished observations). Indeed, S 16924 fully and potently blocked the induction of STFs and FBP by 8-OH-DPAT, in analogy to other ligands possessing partial agonist properties at postsynaptic 5-HT_1A receptors (Millanet al., 1994). However, as described previously (Millanet al., 1994, 1995b; Tricklebank, 1985), these behaviors are dependent on dopaminergic and adrenergic mechanisms for their expression, and the actions of S 16924 at dopamine D₂ and α₁-AR receptors may, thus, also contribute to its inhibition of STFs. This mechanism also likely accounts for the blockade of this behavior by clozapine and haloperidol. Indeed, the mutual involvement of 5-HT_1A receptors, dopaminergic and adrenergic receptors in the modulation of various motor, endocrine and other functional parameters renders the “isolation” and definition of in vivo actions of S 16924 and other multireceptorial agents at 5-HT_1A, D₂ and other receptor types difficult (Millan et al., 1992, 1995b). A complementary and more direct approach to estimating relative potency at 5-HT_1Avs. D₂ receptors was undertaken in the companion paper. Thus, in line with itsca. 10-fold higher affinity at native, rat 5-HT_1Avs. D₂ receptors, S 16924 (WAY 100,635-reversibly) inhibited the electrical activity of (raphe-localized) serotonergic neurones at doses (ID₅₀ = 0.02 mg/kg, i.v.) ca. 10-fold lower than those preventing the inhibition of the electrical activity of (ventrotegmental area) dopaminergic neurones by apomorphine (0.2 mg/kg, i.v.). This question of the relative occupation and expression of activity at 5-HT_2A, D₂ and other receptor types is, thus, a complex issue, which is currently being clinically addressed for S 16924 using the technique of Position Emission Tomography.

Side-effect profile of S 16924 vs. clozapine: low muscarinic and histaminic affinity.

S 16924 exhibited only modest affinity for H₁ receptors, a finding of significance inasmuch as the potent interaction of clozapine at H₁ sites underlies several, undesirable actions affecting patient compliance. That is, sedation, drowsiness, fatigue and weight gain (Cunningham-Owens, 1996). S 16924 displayed negligible affinity for multiple, muscarinic (M₁–M₄) receptors, an observation of importance inasmuch as the potent interaction of clozapine with specific muscarinic receptor types is associated with a disturbance of vision, gastrointestinal transit, urinary flow and cardiovascular function, including tachycardia due to vagal (M₂) blockade (Brunello et al., 1995; Cunningham-Owens, 1996). Interestingly, in contrast to other antipsychotics possessing high muscarinic affinity, clozapine elicits hypersalivation (sialorrhoea) rather than a dry mouth, possibly due to an agonist action at M₄ receptors (Zorn et al., 1994; but see Olianas et al., 1997). As concerns other side-effects of clozapine, seizure activity occurs in around 5% of patients, probably due to an interference with inhibitory or excitatory amino acid transmission in the hippocampus, temporal cortex and other regions (Geaney, 1995). In this light, it is of interest that S 16924 is devoid of affinity for GABA_A, GABA_B, benzodiazepine, NMDA and AMPA receptors. The most disquieting, adverse effect of clozapine is, nevertheless, agranulocytosis, which occurs inca. 1 to 2% of patients and which can be fatal (Cunningham-Owens, 1996). The particular chemical structure of clozapine, allowing for the generation of a nitrenium ion, has been implicated (Liu and Uetrecht, 1995). The contrasting structure of S 16924 as compared to clozapine should be underlined in this context.

A comparison to other antipsychotic agents.

Several other, novel antipsychotics have been described that possess more marked activity at serotonergic (notably, 5-HT_2A) and adrenergic (α₁-AR) as compared to dopaminergic receptors, including olanzapine and quetiapine, which are structurally related to clozapine, the benzoizoxazole, risperidone, and the indole derivative, sertindole (Pickar, 1995). Although these agents do not possess significant activity at 5-HT_1A receptors, the benzothiazole, ziprasidone, displays high affinity for 5-HT_1A receptors, at which it behaves as a partial agonistin vitro (Seeger et al., 1995). However, in contrast to S 16924, its affinity for 5-HT_1A sites is not superior to that at D₂ receptors, and agonist actions at 5-HT_1A receptors are not a pronounced feature of itsin vivo profile of activity (Seeger et al., 1995; Millan MJ and Brocco M, unpublished observations). More recently, several novel antipsychotics have been documented that show significant activity at 5-HT_1A receptors in vitro andin vivo. Of these, both 1192U90 and SM 9018 are of note inasmuch as, as with S 16924, they also display antagonist activity at 5-HT_2A receptors (Jones-Humble et al., 1996;Kato et al., 1990). Nevertheless, their chemical structures and precise functional profiles clearly differ to those of S 16924. It will be of interest to establish to what extent activity at 5-HT_1A receptors contributes to the clinical actions of S 16924 and other drugs in the management of schizophrenia.