Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The dibenzepine derivative clozapine is effective in a subpopulation of patients refractory to neuroleptics, and exerts antipsychotic activity in the absence of extrapyramidal side effects and tardive dyskinesia (Kahn and Davis, 1995; Wirsching et al., 1995; Cunningham-Owens, 1996). In addition, negative symptoms of schizophrenia, which are generally resistant to neuroleptic treatment, may respond to treatment with clozapine (Kahn and Davis, 1995; Meltzer, 1995; Wirsching et al., 1995). In view of this "atypical" antipsychotic profile, considerable efforts have been devoted to the identification of novel, clozapine-like antipsychotic agents. Indeed, several drugs both chemically and pharmacologically similar-although not identical-to clozapine have been described, including olanzapine, quetiapine, and {trans-5-chloro-2-methyl-2,3,3a,12b-tetrahydro-1H dibenz[2,3:6,7]oxepino [4,5-c]pyrrole (ORG-5222) (Moore et al., 1992; Boer et al., 1993; Meltzer, 1995; Ellenbroek et al., 1996). Several others, although chemically distinct from clozapine, have also been suggested to possess an "atypical" or "clozapine-like" profile: for example, risperidone, ziprasidone, and sertindole (Sánchez et al., 1991; Meltzer, 1995; Seeger et al., 1995; Schotte et al., 1996).

Considerable evidence suggests that the interaction of clozapine with multiple monoaminergic receptors plays an important role in its functional and clinical actions (Brunello et al., 1995). Although the modest affinity of clozapine at D₂ receptors has repeatedly been emphasized, several other aspects of its distinctive receptorial profile as compared with that of haloperidol may also be evoked, notably, equilibrated affinity for D₁ and D₂ receptors (Josselin et al., 1997), antagonist properties at D₃ receptors (Levant, 1997), preferential affinity for D₄ versus D₂ receptors (Seeman et al., 1997), and potent antagonist properties at ₁-adrenoceptors (ARs; Baldessarini et al., 1992). Furthermore, the preferential blockade by clozapine of serotonin (5-HT)_2A versus D₂ receptors has been correlated with a relative lack of extrapyramidal activity (Roth and Meltzer, 1995), whereas the marked antagonist actions of clozapine at 5-HT_2C receptors may reinforce mesocortical dopaminergic transmission, a perturbation of which contributes to negative symptoms (Meltzer, 1995; Knable and Weinberger, 1997; Millan et al., 1998a). More recently, it has been suggested that the partial agonist actions of clozapine at 5-HT_1A receptors play a role in its benign extrapyramidal profile. Furthermore, they may contribute to its efficacy against negative symptoms (Brunello et al., 1995; Newman-Tancredi et al., 1996; Christoffersen and Meltzer, 1998; Millan et al., 1998d).

Although the above comments underline the distinctive monoaminergic properties of clozapine and its advantages relative to neuroleptics in the management of schizophrenia, clozapine is not an ideal drug (see Cunningham-Owens, 1996). Its potent antagonist actions at histaminic receptors are associated with sedation, drowsiness, and weight gain, whereas its marked affinity for multiple muscarinic receptors results in a disturbance of vision and intestinal function, as well as salivation and tachycardia (Keks, 1996). In addition, clozapine elicits seizures in a significant minority of patients. Finally, the most severe adverse response to clozapine is agranulocytosis, a potentially fatal blood disorder which occurs in 1 to 2% of patients.

The above discussion exemplifies the need for novel antipsychotic agents lacking the undesirable actions of clozapine yet possessing its improved antipsychotic profile. Within this framework, we recently described a novel benzodioxopyrrolidine, (R)-2-{1-[2-(2,3-dihydro-benzo[1,4]dioxin-5-yloxy)-ethyl]-pyrrolidin-3yl}-1-(4-fluorophenyl)-ethanone (S-16924), which, although chemically distinct from clozapine, shares its potent antagonist properties at 5-HT_2A, 5-HT_2C, ₁-AR and D₄ receptors and its modest affinity for D₁/D₅, D₂ and D₃ receptors. In addition, the moderate 5-HT_1A partial agonist properties of clozapine are considerably (50-fold) reinforced in S-16924 (Millan et al., 1998b,d). S-16924 may, however, be distinguished from clozapine by its markedly lower affinity for both histaminic and muscarinic receptors (Millan et al., 1998b,d).

In characterizing psychoactive drugs and determining their potential mechanisms of action, one interesting approach is to examine their discriminative stimulus (DS) properties. These reflect their "interoceptive" actions and may reasonably be considered to involve their overall modulation of corticolimbic activity and mood. Clozapine elicits robust stimulus control in several species, whereas neuroleptics, such as haloperidol, do not themselves reliably generate a DS, and fail to generalize to clozapine (Wiley and Porter, 1992; Hoenicke et al., 1992; Carey and Bergman, 1997; Goudie and Taylor, 1998). The present studies had, thus, the following aims: 1) using antagonists selective for specific receptor types implicated in the actions of S-16924 and clozapine, the receptorial mechanisms underlying their DS properties were characterized and compared and 2) the generalization patterns of S-16924 and clozapine to other chemically-diverse, "multireceptorial" antipsychotic agents were evaluated.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals. Male Wistar rats (Iffa-Credo, L'Arbresle, France), weighing 180 to 200 g upon arrival, were housed singly in sawdust-lined cages with free access to water and restricted access to food (approximately 10 g/day). Laboratory temperature was 21 ± 1.0°C and humidity was 60 ± 5%. There was a 12-h light/dark cycle with lights on at 7:30 AM.

Experimental Procedure. Animals were trained to discriminate S-16924 (2.5 mg/kg i.p.) from saline, or clozapine (5.0 mg/kg i.p.) from saline, using a standard two-lever Fixed-Ratio 10 food-reinforced operant procedure as described by Schreiber et al. (1994). Each daily session started 15 min (S-16924 group) or 30 min (clozapine group) after drug or saline injection and was terminated after 15 min. The dose of clozapine was based on previous studies of 5.0 mg/kg s.c. (Wiley and Porter, 1992; Goudie and Taylor, 1998), a dose that corresponds, further, to its activity in various models of antipsychotic activity (Millan et al., 1998d). In such paradigms, S-16924 is some 2- to 3-fold more potent than clozapine, so 2.5 mg/kg was an appropriate dose for the former (Millan et al., 1998d). A 30-min pretreatment time with clozapine was that routinely used in several previous drug discrimination studies with other drug classes (Schreiber et al., 1994). As concerns S-16924, its onset of action is particularly rapid (M.J.M., unpublished observations), so we reduced the pretreatment time to 15 min. Importantly, over the 15- to 30- and 30- to 45-min periods selected herein for study of S-16924 and clozapine, respectively, both of these drugs exert their functional actions not only in behavioral models but also in dialysis models of the modulation of monoaminergic transmission (Millan et al., 1998b, in press). Drug or saline sessions alternated randomly. The discrimination criterion consisted of 10 consecutive sessions with correct responding, i.e., no more than 13 responses on both the reinforced and the nonreinforced levers before the first reinforcement was obtained. Subsequently, test sessions were conducted every Wednesday and Friday, whereas training sessions continued on the other days (5 days/week). Rats responding incorrectly on the two most recent training days were submitted to an additional training session instead of a test session. During testing, responding on the selected lever, i.e., the lever on which 10 responses were recorded first, was reinforced for the remainder of the session. Test compounds were substituted for the training drug and administered at the corresponding time before the session. In the antagonist study with (N-{2-[4-(2-methoxyphenyl)-1-piperazinyl]ethyl}-N-(2-pyridinyl)cyclohexanecarboxamide (WAY-100,635), it was administered 30 min before the training dose of S-16924 or clozapine. Data recorded during a test session were lever selection and response rates (RR), i.e., the total number of presses on both levers. Lever selection data were expressed as the percentage of rats selecting the drug lever and were compared by a Fisher's exact probability test to control value (0% in generalization studies, 100% in the antagonist study). ED₅₀ values plus 95% CL were calculated by means of the method of Finney to estimate drug potency. RRs in the presence of drug were compared by a paired t test to RR obtained during the preceding saline (or drug, in the antagonist study) training session. The ID₅₀ values (95% CL) were calculated.

Drug Selection and Dosing. The choice of drugs (and doses) used herein was based upon extensive in vitro and in vivo studies performed both in our own and in other laboratories. Indeed, for the "multireceptorial" antipsychotics examined, the dose ranges used correspond to those at which they express their actions in vivo in well established models of antipsychotic activity (Millan et al., 1998c and M.J.M. and M. Brocco, unpublished observations). Similarly, for other drugs (the most selective available for their respective targets), doses were based on studies in which it was established that they occupy and express their functional actions at their respective receptors (see Discussion). In addition, for several drugs, such as ORG-5222 and 1-(1-methylindol)-5-yl)-3-(3-pyridyl)urea (SB 200,646), doses tested against clozapine were pursued up to a range markedly higher than those already generalizing to S-16924 (see Results). In certain cases, such as for 5-HT_2C or D₄ antagonists, which do not markedly influence motor behavior (Kennett et al., 1996; Millan et al., 1998c), it was not appropriate to pursue doses up to those at which a marked decrease in RR might be encountered. More generally, limitations in drug solubility or potential toxicity [such as for {(+)-[7-(N,N-dipropylamino)-5,6,7,8-tetrahydro-naphto(2,3b)dihydro,2,3-furane]} (S-14297)] may restrict increases in doses.

Correlation Analyses. The relationship between the potency of multireceptorial antipsychotics for generalization to clozapine or S-16924 and their affinities for several receptor types implicated in the management of schizophrenia (Table 6) was examined using an extensive correlation analysis. Only drugs for which precise ED₅₀ values for generalization to clozapine and/or S-16924 were obtained were incorporated into these analyses. For clozapine, this limited the analysis to four multireceptorial antipsychotics. However, for S-16924, a total of six could be incorporated. Correlation analyses for hM₁ receptors were also recalculated with inclusion of the muscarinic antagonist scopolamine. The in vitro data used for these correlation analyses were derived exclusively in this laboratory from a direct comparison under identical experimental conditions of all drugs. These findings will be published elsewhere as part of an extensive multiparametric analysis of antipsychotic agents (Millan, M.J.M., M. Brocco, V. Audinot, S.Q., and A. Newman-Tancredi, in preparation). Pearson-Product moment correlation coefficients were calculated in all cases.

Drug Sources. All drug doses are in terms of the base. Drugs were dissolved in sterile water with a few drops of lactic acid. The pH was adjusted to as close to neutrality as possible (>5.0). Drugs were injected s.c. unless otherwise specified (injection volume = 1 ml/kg). Drug salts and sources were as follows: {(±)-[7-(N,N-dipropylamino)-5,6,7,8-tetrahydro-naphtho(2,3b)dihydro,2,3-furane]} ((±)-S-11566) HCl; (+)-S-14297 dibenzoyltartrate; S-16924 HCl; {2-[4-(2,3-dihydrobenzo [1,4]dioxin-6-yl)piperazin-1-yl methyl]indan-2-yl} methanol (S-18126) 2HCl; 3-(4-[4-chlorophenyl]piperazin-1-yl)methyl-1H-pyrrolo[2,3b]pyridine (L-745,870); [R(+)--(2,3-dimethoxyphenyl)-1-[2-(4-fluorophenylethyl)]-4-piperidine-methanol] (MDL-100,907); ondansetron HCl; 1-(1-methylindol-5-yl)-3-(3-pyridyl)urea (SB-200,646) HCl; 5 methyl-1-(3-pyridil-carbamoyl)-1,2,3,5-tetrahydropyrrolo[2,3-f]indole (SB-206,553); quetiapine hemifumarate; risperidone; sertindole; WAY-100,635 3HCl; and ziprasidone HCl were synthesized by Servier chemists (G. Lavielle, C. Mallen, and J.L. Peglion). Clozapine, (±)-8-dihydroxy-2-(di-n-propylamino)tetralin [(±)-8-OH-DPAT] hydrobromide, and raclopride tartrate were purchased from Research Biochemicals International (Natick, MA). Prazosin HCl and scopolamine HCl were purchased from Sigma (Chesnes, France); olanzapine was obtained from Lilly Research Laboratories (Indianapolis, IN); ORG-5222 fumarate was obtained from Organon (Oss, the Netherlands); fananserin was obtained from Rhône Poulenc (Vitry, France); ()-trans-6,7,7a,8,9,13b-hexahydro-3-chloro-2hydroxyl-N-methyl-5H-benzo[d]-naphto-[2,1-benazazepine] (SCH39166) HCl was obtained from Schering Plow Corp. (Bloomfield, NJ); {1(Z)-[2-(dimethylamino)ethoxyimino]]-1-(2-fluorophenyl)-3-(4-hydroxyphenyl)-2(E)-propene] (SR-46349) was obtained from Sanofi (Montpellier, France).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Acquisition of S-16924- and Clozapine-Induced Discriminative Stimuli. As shown in Fig. 1, both S-16924 and clozapine rapidly resulted in robust and stable DS, with the discrimination criterion attained after 53 ± 6 and 68 ± 7 sessions, respectively.

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Percent correct responding during acquisition of the DS effects of S-16924 (2.5 mg/kg i.p.) and clozapine (5 mg/kg i.p.).

Generalization Studies with Dopaminergic and Adrenergic Ligands. As indicated in Table 1, haloperidol showed only partial generalization to clozapine and S-16924, and dose-dependently and markedly decreased RR. The D₂/D₃ antagonist raclopride dose-dependently and fully generalized to S-16924, but showed only partial generalization to clozapine, even at the highest dose tested, which markedly diminished RR (Table 1). The selective D₃ receptor antagonist (±)-S-11566, and its active eutomer (+)-S-14297, failed to generalize to S-16924, and (±)-S-11566 likewise did not generalize to clozapine (Table 2). (+)-S-14297 was not tested against clozapine because we observed problems of cutaneous toxicity in the S-16924-tested animals; as it is inactive by other (i.p. and p.o.) routes of administration, these could not be used. The selective dopamine D₄ receptor antagonist S-18126, which showed only modest generalization to S-16924, likewise provoked cutaneous toxicity which prompted a limitation to only one dose tested in clozapine-trained animals in which, similarly, no generalization was seen (Table 2). A further selective D₄ antagonist, L-745,870, tested by the oral route owing to problems of cutaneous toxicity also showed only modest generalization to S-16924, and no generalization to clozapine (Table 2). In contrast to haloperidol and raclopride, (±)-S-11566, (+)-S-14297, S-18126, and L-745,870 did not significantly modify RR. The D₁/D₅ antagonist SCH-39166 failed to generalize to either S-16924 or clozapine, despite a pronounced reduction in RR (Table 2). Finally, the selective ₁-AR antagonist prazosin, which did not modify RR, did not generalize either to S-16924 or to clozapine (Table 2).

View this table: [in this window] [in a new window]   TABLE 2 Lack of generalization of selective antagonists at D3, D4, D1/D5, and 1-adrenergic receptors to S-16924 and clozapine discriminative stimulus

Generalization Patterns with Serotonergic Ligands. The selective agonist at 5-HT_1A receptors, 8-OH-DPAT, dose-dependently and fully generalized to S-16924 and, at the highest dose tested, significantly decreased RR (Table 3 and Fig. 2). In contrast, even up to a dose markedly decreasing RR, 8-OH-DPAT showed no generalization to clozapine (Table 3 and Fig. 2). Pretreatment with the selective 5-HT_1A antagonist WAY-100,635 did not significantly antagonize the effects of the training dose of either S-16924 or clozapine (Table 4). The selective 5-HT_2A receptor antagonist, MDL-100,907, dose-dependently and fully generalized to S-16924, and two additional 5-HT_2A antagonists, fananserin and SR-46369, also partially generalized to S-16924 (Table 3 and Fig. 3). MDL-100,907 showed partial generalization to clozapine at a dose of 0.04 mg/kg s.c. (Table 3 and Fig. 3). However, this effect was not statistically significant and the MDL-100,907 dose-response curve was markedly biphasic with both lower and higher doses eliciting minimal generalization. Fananserin and SR-46349 did not show generalization to clozapine (Table 3 and Fig. 3). None of these 5-HT_2A antagonists significantly decreased RR (Table 3). The 5-HT_2B/2C antagonist SB-200,646 dose-dependently and fully generalized to S-16924, and an additional 5-HT_2B/2C antagonist, SB-206,553, also partially generalized to S-16924. However, SB-200,646 and SB-206,553 only partially generalized to clozapine (Table 3 and Fig. 3). Neither SB-200,646 nor SB-206,553 modified RR. The selective 5-HT₃ antagonist ondansetron (0.63 mg/kg s.c.) generalized neither to S-16924 (0% n = 4) nor to clozapine (0%, n = 4) and failed to modify RR (0 and 5% decrease for S-16924 and clozapine, respectively).

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Generalization of the 5-HT1A agonist 8-OH-DPAT to a discriminative stimulus elicited by S-16924, but not by clozapine. N = 4 to 5 per value. Top: drug lever selection. Data are percentage of animals selecting drug lever. Asterisks indicate significance of differences (*P < .05; Fisher's exact probability test) as compared to control value (0% drug lever selection). Bottom: response rates. Data are means ± S.E.M. of percentage of control response values obtained during the most recent saline-training session. Asterisks indicate significant decreases in response rates (*P < .05; paired t test) as compared to the control training session.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Generalization of the 5-HT2A antagonists MDL-100,907, fananserin, and SR-46349, and of the 5-HT2B/2C antagonist SB-200,646, to discriminative stimuli elicited by S-16924 as compared to clozapine. Doses are in mg/kg s.c. or p.o. (SB-200,646). N = 4 to 6 per value. Top: drug lever selection. Data are percentage of animals selecting drug lever. Asterisks indicate significance of differences (*P < .05; Fisher's exact probability test) as compared to control value (0% drug lever selection). Bottom: response rates. Data are means ± S.E.M. of percentage of control response values obtained during the most recent saline-training session. Asterisks indicate significant decreases in response rates (*P < .05; paired t test) as compared to the control training session.

Generalization Patterns of Multireceptorial Antipsychotics. S-16924 and clozapine showed dose-dependent and full mutual generalization at doses not significantly modifying RR (Table 5). The clozapine congener olanzapine potently and fully generalized to S-16924 (Table 5 and Fig. 4). Olanzapine also fully generalized to clozapine, although only at higher doses, and it markedly decreased RR at the highest dose tested (Table 5 and Fig. 4). Another clozapine analog, quetiapine, on the other hand, fully generalized to both S-16924 and clozapine without suppressing RR (Table 5 and Fig. 4). However, although the structurally-related ORG-5222 fully generalized to S-16924, it only weakly generalized to clozapine even up to doses markedly diminishing RR (Table 5 and Fig. 4). The phenylindole derivative sertindole (10.0 mg/kg s.c.) generalized neither to S-16924 (20%, n = 5) nor to clozapine (0%, n = 4). It did not markedly suppress RR (14% reduction for S-16924 and clozapine, in each case). In analogy, the benzisoxazole risperidone (0.63 mg/kg s.c.) did not significantly generalize to either S-16924 (40%, n = 5) nor clozapine (25%, n = 4). It did not markedly affect RR in S-16924-trained animals (16% reduction) but these were markedly reduced (by 78%) in clozapine-trained animals. In contrast, however, a further benzisoxazole derivative, ziprasidone, potently and fully generalized to S-16924, whereas it did not show any generalization to clozapine (Table 5 and Fig. 5).

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Generalization of the multireceptorial antipsychotics olanzapine, ORG-5222, and quetiapine to discriminative stimuli elicited by S-16924 and clozapine. N = 4 to 6 per value. Top: drug lever selection. Data are percentage of animals selecting drug lever. Asterisks indicate significance of differences (*P < .05; Fisher's exact probability test) as compared to control value (0% drug lever selection). Bottom: response rates. Data are means ± S.E.M. of percentage of control response values obtained during the most recent saline-training session. Asterisks indicate significant decreases in response rates (*P < .05; paired t test) as compared to the control training session.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Generalization of ziprasidone to a discriminative stimulus elicited by S-16924, but not by clozapine. N = 4 to 7 per value. Top: drug lever selection. Data are percentage of animals selecting drug lever. Asterisks indicate significance of differences (*P < .05; Fisher's exact probability test) as compared to control value (0% drug lever selection). Bottom: response rates. Data are means ± S.E.M. of percentage of control response values obtained during the most recent saline-training session. Asterisks indicate significant decreases in response rates (*P < .05; paired t test) as compared to the control training session.

Generalization with the Muscarinic Antagonist Scopolamine. The selective muscarinic antagonist scopolamine dose-dependently generalized to S-16924 attaining 60% generalization at a dose of 0.08: ED₅₀ (95% CL) = 0.06 (0.04-0.10). Higher doses could not be evaluated owing to the marked suppression of RR: ID₅₀ (95% CL) = 0.02 (0.01-0.06; Fig. 6). Scopolamine also dose-dependently and fully (100%) generalized to clozapine: ED₅₀ (95% CL) = 0.01 (0.004-0.02). It similarly decreased RR: ID₅₀ (95% CL) = 0.05 (0.02-0.12; Fig. 6).

View larger version (14K): [in this window] [in a new window]   Fig. 6.   Generalization of the muscarinic antagonist scopolamine to discriminative stimuli elicited by S-16924 and clozapine. N = 4 to 5 per value. Top: drug lever selection. Data are percentage of animals selecting drug lever. Asterisks indicate significance of differences (*P < .05; Fisher's exact probability test) as compared to control value (0% drug lever selection). Bottom: response rates. Data are means ± S.E.M. of percentage of control response values obtained during the most recent saline training session. Asterisks indicate significant decreases in response rates (*P < .05; paired t test) as compared to the control training session.

Correlation Analyses. The potency of multireceptorial antipsychotics to generalization to S-16924 as compared to clozapine tended to correlate. However, this did not attain statistical significance, owing to the low number (clozapine, S-16924, quetiapine, and olanzapine) of drugs that could be incorporated in this analysis (r = 0.68, n = 4, P > .05). Similarly, there was no significant correlation between potency to generalize to clozapine and drug affinity for any of the receptors specified in Tables 6 and 7, although hD₂ and hD₃ and, less markedly, D₁ and 5-HT_2A receptors all showed a tendency. Again, it should be noted that these analyses were based only upon the limited number of drugs (4) for which full generalization to clozapine was achieved. Upon addition of scopolamine to the analyses, for which binding data was available at hM₁ sites, the r value for these receptors was 0.54, n = 5, P > .05. As concerns S-16924, the number of antipsychotic drugs that could be analyzed was more extensive (clozapine, S-16924, quetiapine, olanzapine, ORG-5222, and ziprasidone). Indeed, significance was achieved for 5-HT_2A, and less markedly, for D₁ and D₂, receptors (Table 7). When scopolamine was added to the analyses, no significant correlation was seen for hM₁ receptors, r = 0.34, n = 7, P > .05. In view of the limited number of drugs that could be incorporated, the results of the correlation analyses should be interpreted with caution.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Generation of a DS with S-16924. Although Porter and Strong (1996) showed that the clozapine congener olanzapine, likewise, sustains a DS in rats, the present study with S-16924 provides a unique and extensive comparative drug discrimination comparison of a potential antipsychotic agent with clozapine. Indeed, inasmuch as neuroleptics fail to induce DS, the ability of S-16924 to elicit a robust and stable DS is itself a notable finding. Furthermore, S-16924 and clozapine DS display full and mutual generalization, suggesting that they share interoceptive properties, an observation underpinning findings showing that S-16924 mimics the "atypical" profile of clozapine in experimental models of potential antipsychotic properties (Millan et al., 1998d).

Generalization Patterns to Other Multireceptorial Antipsychotics. With the exception of S-16924 (this report) and olanzapine (Moore et al., 1992; Porter and Strong, 1996), bidirectional generalization with clozapine has not been reported. Interestingly, olanzapine generalized both to clozapine and to S-16924, suggesting common DS characteristics. This full generalization of olanzapine to clozapine is of note in the light of previous conflicting data. Although Moore et al. (1992) observed generalization of olanzapine to clozapine (5.0 mg/kg i.p.), Goudie and Taylor (1998) did not. However, upon a reduction of the clozapine training dose to 2.0 mg/kg i.p., full generalization of olanzapine was achieved in the latter study. The D₂ antagonist properties and corresponding motor actions of olanzapine, which are more pronounced than those of clozapine, may interfere with its evaluation (Moore et al., 1992; Meltzer, 1995; Millan et al., 1998d). Indeed, olanzapine only partially generalized to clozapine in the monkey, whereas, in the presence of a dopaminergic agonist, full generalization was obtained (Carey and Bergman, 1997). Another dibenzepine derivative related to clozapine, quetiapine (Ellenbroek et al., 1996), likewise generalized to S-16924 and clozapine herein, consistent with common DS properties. This observation corroborates reports that quetiapine generalizes to clozapine in rats (Goudie and Taylor, 1998) and monkeys (Carey and Bergman, 1997). Furthermore, these data are in line with other studies suggesting that clozapine congeners frequently generalize to clozapine (Bruhwyler et al., 1997; Carey and Bergman, 1997; Goudie and Taylor, 1998). Nevertheless, certain drugs structurally related to clozapine, such as loxapine and clothiapine, fail to generalize (Carey and Bergman, 1997). Furthermore, herein, ORG-5222 (Boer et al., 1993) showed only modest generalization to clozapine even at doses higher than those fully generalizing to S-16924. Presumably, the precise pattern of receptorial interactions, rather than the chemical structure, determines whether drugs generalize or not. Indeed, the reciprocal generalization of clozapine to the structurally-distinct yet receptorially-similar benzodioxopyrrolidine derivative, S-16924, supports this argument. Furthermore, the piperidine analog PNU-96415E, which shows a binding profile similar to that of clozapine, fully generalized to a clozapine DS (Tang et al., 1997). Nevertheless, certain other classes of multireceptorial antipsychotic with binding profiles less similar to that of clozapine, including the benzisoxazole risperidone and the phenylindole derivative sertindole, failed to generalize to clozapine or S-16924, even at doses exerting marked antipsychotic activity in functional model in rodents (Results, Table 5, Carey and Bergman, 1997; Millan et al., 1998d; M.J.M. and M. Brocco, unpublished observations). Interestingly, another benzisoxazole, ziprasidone (Seeger et al., 1995), likewise failed to generalize to clozapine even at doses higher than those potently generalized to S-16924. This distinction is of interest inasmuch as ziprasidone possesses, in analogy to S-16924, partial agonist properties at 5-HT_1A receptors (Seeger et al., 1995).

Significance of Training Doses. It should be noted that the present pattern of data was obtained for a single training dose of clozapine (and S-16924). It is possible, in particular for such multireceptional compounds, that differences may have been observed at alternative training doses. Goudie et al. (1998) have explored this issue in showing that quetiapine and olanzapine more potently and fully generalize to a lower (2.0) than a higher (5.0) dose of clozapine. These data illustrate the utility of such a double-dose strategy. Nevertheless, as emphasized by these authors, such differences appear to be quantitative rather than qualitative. This suggests that the generalization patterns seen with the single doses of clozapine and S-16924 used herein are of broad pertinence to their interoceptive properties.

Receptorial Mechanisms Underlying S-16924 and Clozapine DS: Serotonergic Receptors. That the 5-HT_1A agonist actions of ziprasidone are implicated in its generalization to S-16924 is supported by generalization of the selective 5-HT_1A agonist 8-OH-DPAT. In contrast, 8-OH-DPAT did not generalize to clozapine, and previous studies have similarly suggested that 5-HT_1A receptors are not involved in the DS properties of clozapine (Hoenicke et al., 1992). Although clozapine is a partial agonist at human 5-HT_1A receptors, its affinity at these sites is 50-fold lower than that of S-16924 (Newman-Tancredi et al., 1996; Millan et al., 1998b). The inability of the selective 5-HT_1A antagonist WAY-100,635 to inhibit S-16924 or clozapine DS suggests that, even in the absence of 5-HT_1A receptor stimulation, their DS properties can be recognized. Activation of 5-HT_1A receptors is, then, sufficient, but not necessary for reproduction of the interoceptive properties of S-16924.

Receptorial Mechanisms Underlying the S-16924 and Clozapine DS: Dopaminergic Receptors. The inability of haloperidol and substituted benzamides, such as amisulpride, to generalize to clozapine suggests that dopamine D₂ receptor blockade is not an important component of its DS properties (Hoenicke et al., 1992; Wiley and Porter, 1992; Carey and Bergman, 1997; Goudie and Taylor, 1998); this interpretation is reinforced by the present observations with haloperidol and the benzamide raclopride (Ögren et al., 1986; Millan et al., 1998c). Although haloperidol partially generalizes to S-16924, raclopride showed full generalization, suggesting that, in contrast to clozapine, D₂ receptor blockade may be a more marked feature of the S-16924 cue. Raclopride is a potent antagonist at D₂ and D₃, receptors, but a role of the latter could be excluded inasmuch as the selective D₃ antagonist (±)-S-11566 and its active eutomer (+)-S-14297 (Millan et al., 1995) generalized to neither S-16924 nor clozapine. Both clozapine and S-16924 share preferential antagonist activity at D₄ versus D₂ receptors. Nevertheless, the selective D₄ antagonists S-18126 and L-745,870 (Bristow et al., 1997; Millan et al., 1998c) generalized to neither clozapine nor S-16924, suggesting that they are not involved in their DS properties. Finally, although S-16924 and clozapine show modest affinity at D₁/D₅ receptors, the D₁/D₅ antagonist SCH-39166 failed to generalize to clozapine and S-16924, suggesting that these sites do not play a predominant role in their DS properties.

Receptorial Mechanisms Underlying S-16924 and Clozapine DS: Muscarinic Receptors. In the pigeon, Hoenicke et al. (1992) reported that the cholinergic antagonists atropine and scopolamine fail to generalize to clozapine, and interpreted the generalization seen with cyproheptadine and other nonselective drugs in terms of their 5-HT_2C antagonist properties. In contrast, Kelley and Porter (1997) noted that their ability to generalize to clozapine in rats correlated with antagonist potency at muscarinic receptors. Both these authors and Nielsen (1988) obtained full generalization to clozapine with scopolamine in rats, an observation confirmed herein. The generalization of muscarinic antagonists to clozapine has been interpreted as reflecting its marked affinity (5 nM) at muscarinic receptors (Cunningham-Owens, 1996; Millan et al. 1998b). However, notwithstanding the negligible (>1.0 µM) affinity of S-16924 for muscarinic receptors (Millan et al., 1998b), scopolamine also showed generalization. This observation suggests that blockade of muscarinic receptors triggers interoceptive effects resembling those provoked independently of muscarinic sites by S-16924.

Correlation Analyses. The potency of several antipsychotics for generalization to clozapine did not significantly correlate with affinity at any sites examined (Tables 6 and 7). The small number of drugs attaining the criterion of full generalization limits the power of this analysis. Nevertheless, together with results acquired with receptor-selective agents, this analysis suggests that the clozapine DS represents a "compound" DS reflecting a complex pattern of interactions at multiple receptor types (Goudie and Taylor, 1998). Overall, a similar conclusion seems warranted for S-16924, and significant correlations between 5-HT_2A, hD₂, and D₁ receptor affinities and generalization were apparent. S-16924 islike clozapinean antagonist at these sites (Millan et al., 1998b). Although the selective D₁ antagonist SCH-39166 did not generalize to S-16924, the significant correlation coefficients for 5-HT_2A and hD₂ sites correspond to the generalization to S-16924 of selective antagonists at these receptors. Thus, 5-HT_2A and D₁/D₂ antagonist properties may be of particular importance in the generalization of antipsychotics to S-16924. Furthermore, the doses of antipsychotic that blocked the DS elicited by S-16924 are close to those that block 5-HT_2A receptor-mediated behaviors such as phemcyclidine-induced locomotion (Millan et al., 1998d).

Conclusions. In conclusion, S-16924 generates a robust and stable DS in rats. Furthermore, notwithstanding their chemical distinctiveness, S-16924 and clozapine share certain communalities in their DS properties. They display, thus, full and reciprocal generalization, and both show generalization with the clozapine congeners olanzapine and quetiapine. Furthermore, they both show generalization with the muscarinic antagonist scopolamine, notwithstanding the negligible affinity of S-16924 for these sites. Furthermore, 5-HT_2C receptor antagonist properties may be involved in each case. However, there are also certain differences, notably an apparent involvement of 5-HT_1A agonist as well as 5-HT_2A and D₂ antagonist actions in the DS properties of S-16924. These "compound" DS properties of S-16924 and clozapine likely reflect their global patterns of interaction at multiple monoaminergic receptors. A more precise identification of the key receptorial actions underlying their interoceptive properties, as well as those of other antipsychotic agents, remains a challenge for future studies.