Introduction

Top Abstract Introduction Methods Results Discussion References

The introduction of buspirone as a clinically efficacious anxiolytic (Riblet et al., 1982) has focused attention on the 5-HT_1A receptor as a possible molecular site of anxiolytic drug action. Among a number of 5-HT_1A agonists that have recently been characterized as having anxiolytic-like activity, several are reported to have additional affinity for 5-HT_2A/2C receptors (Abou-Gharbia and Moyer, 1990; Sills et al., 1990; Millan et al., 1992b; Albinsson et al., 1994; Krisch and Bole, 1994). Not only are compounds that possess 5-HT_2A/2C antagonist properties shown to have anxiolytic-like effects in pigeons (e.g., ritanserin and ICI 169,369) (Gleeson et al., 1989; Brocco et al., 1990; Koek et al., 1992; Kleven and Koek, 1996), but several compounds with a putative mixed 5-HT_1A agonist/5-HT_2A/2C antagonist profile (e.g., S 14671 and WY-50,324) are reported to produce effects of a magnitude greater than that typically observed with 8-OH-DPAT or the azapirones, buspirone, ipsapirone and tandospirone (Barrett and Zhang, 1991b). Consequently, it has been suggested that compounds that interact with both 5-HT_1A and 5-HT_2A/2C receptors may have a more advantageous therapeutic profile (Millan et al., 1992a; Barrett and Vanover, 1993).

Several new compounds possessing affinity for both 5-HT_1A and 5-HT_2A/2C receptors have recently been introduced (e.g., BIMT 17, LEK-8804 and FG5893) (Albinsson et al., 1994; Krisch and Bole, 1994; Borsini et al., 1995). However, several of the compounds described as having a mixed profile possess only moderate in vitro affinity for 5-HT_2A/2C receptors (e.g., CGS 18102A, FG5893, WY-50,324 and S14671) (Abou-Gharbia and Moyer, 1990; Sills et al., 1990; Millan et al., 1992b; Albinsson et al., 1994), whereas they generally exhibit more than 100-fold higher affinities for the 5-HT_1A receptor. Furthermore, where available, evidence for in vivo 5-HT_2A antagonist properties consists primarily of their ability to reverse DOI-, 5-HTP, or quipazine-mediated head twitches (Abou-Gharbia and Moyer, 1990; Sills et al., 1990; Albinsson et al., 1994; Krisch and Bole, 1994), a response that is inhibited not only by compounds exhibiting affinity for 5-HT_2A receptors (Kennett and Curzon, 1991; Schreiber et al., 1995), but also by other compounds lacking 5-HT_2A affinity, such as alpha₁-adrenergic antagonists, dopamine antagonists and 5-HT_1A agonists (Arnt et al., 1984; Yocca et al., 1990; Koek et al., 1992; Dursun and Handley, 1993; Schreiber et al., 1995). Accordingly, the ability of mixed 5-HT_1A agonist/5-HT_2A/2C antagonists to suppress 5-HT_2A-mediated head twitches cannot be interpreted as solely due to 5-HT_2A antagonism because 5-HT_1A agonist properties may be responsible. This necessitates additional experiments using 5-HT_1A antagonists to rule out 5-HT_1A actions of mixed compounds.

Although DOI has similar affinity for both 5-HT_2A and 5-HT_2C receptors (Middlemiss and Tricklebank, 1992), several recent studies (Kennett et al., 1994; Schreiber et al., 1995; Dursun and Handley, 1996) indicate that DOI-induced head twitches are mediated by 5-HT_2A receptors. The extensive study conducted by Schreiber and colleagues (1995) is in agreement with previous pharmacological characterizations of 5-HTP-induced head twitches (Kennett and Curzon, 1991; Koek et al., 1992) in that a high correlation was found between the potencies of a variety of 5-HT₂ antagonists in blocking head twitches and affinity for 5-HT_2A, but not 5-HT_2C receptors. Additionally, the report that the putative 5-HT_2C/2B selective antagonist SB 200646A fails to block DOI-induced head twitches (Kennett et al., 1994) is consistent with the involvement of 5-HT_2A receptors in this behavior. Thus, antagonism of DOI-induced head twitches should constitute evidence for in vivo 5-HT_2A antagonist properties. However, because 5-HT_1A agonists are also expected to decrease DOI-induced head twitches, additional studies with 5-HT_1A antagonists are needed to rule out the involvement of 5-HT_1A receptors in the effects of mixed 5-HT_1A agonist/5-HT_2A/2C antagonists.

Previous studies (Glennon, 1986; Koek et al., 1992; Schreiber et al., 1994; Arnt, 1996) indicate that the DS effects of DOI are also mediated by 5-HT_2A receptors, although more recent data (M. S. Kleven, M.-B. Assié and W. Koek, in preparation) suggest that the involvement of 5-HT_2C receptors cannot be ruled out. The conclusion that the DS effects of DOI are mediated by 5-HT_2A receptors (Schreiber et al., 1994), was primarily based on findings that the 5-HT_2A antagonist MDL 100,907 was effective, whereas the 5-HT_2C/2B antagonist SB 200646A was ineffective, in blocking the drug-appropriate responding in DOI-trained rats. A more recent report that the 5-HT_2A antagonist MDL 100151 also antagonizes the DS effects of DOI (Arnt, 1996) further supports the idea that blockade of 5-HT_2A receptors is sufficient to inhibit the effects of DOI. Nonetheless, in an extensive examination of more than twenty compounds possessing 5-HT₂ antagonist properties (M. S. Kleven, M.-B. Assié and W. Koek, in preparation), including the 5-HT_2C/2B antagonists SB 206553 (Kennett et al., 1996) and SDZ SER 082 (Nozulak et al., 1995), behavioral potencies for blocking the DS effects of DOI were significantly correlated with both 5-HT_2C and 5-HT_2A affinity. Although there is ample evidence that 5-HT_2A receptors play a role, the involvement of 5-HT_2C receptors in the DS effects of DOI cannot be dismissed. Thus, antagonism of the DS effects of DOI may be interpreted as evidence for in vivo 5-HT_2A, and perhaps 5-HT_2C antagonist properties. With respect to the pharmacological interactions of the type seen with DOI-induced head twitches, alpha₁-adrenergic- and dopamine D₂- antagonist properties do not appear to explain the ability of ritanserin and risperidone to block the DS effects of DOI (Koek et al., 1992); the effects of pretreatment with 5-HT_1A agonists have not, to our knowledge been reported previously. But, 5-HT_1A agonists have been reported to engender partial substitution in rats trained to discriminate DOI from saline (Colpaert et al., 1992; Koek et al., 1995). The extent to which this phenomenon interferes with the ability to detect in vivo 5-HT_2A antagonist properties of mixed 5-HT_1A agonist/5-HT_2A/2C antagonists in this behavioral assay remains to be determined.

The purpose of the present study was 1) to examine the ability of the selective 5-HT_1A antagonist WAY-100635 to reverse the effects of putative mixed compounds on head twitches induced by the 5-HT_2A/2C agonist DOI; 2) to examine the ability of mixed compounds to inhibit the DS effects of DOI, because it has been suggested that this discrimination may be a more selective means to determine in vivo 5-HT₂ antagonist properties of drugs with mixed actions (Koek et al., 1992); 3) to further characterize in vivo 5-HT_1A agonist properties by examining the effects of mixed compounds on observable behavior and in rats trained to discriminate the 5-HT_1A agonist 8-OH-DPAT from saline; 4) to examine the differential ability of DOI to modify the observable behavioral effects induced by drugs with mixed actions and 5) to conduct more extensive binding studies under the same conditions for all of the compounds examined in this study, because binding affinities of many of the mixed antagonists for other receptors have not been reported.

To characterize further the in vivo 5-HT_2A antagonist properties of putative mixed 5-HT_1A agonist/5-HT_2A/2C antagonists, their ability to antagonize both DOI-induced head twitches and DS effects in rats was compared with effects of the 5-HT₂ antagonists, ketanserin and ritanserin, and the 5-HT_1A agonists 8-OH-DPAT and buspirone. Of the compounds examined in this study, WY-50,324 was of interest because it has been suggested that mixed 5-HT_1A agonist/5-HT_2A/2C antagonist properties may underlie its reportedly exceptional anti-conflict effects in the pigeon (Barrett and Vanover, 1993). The mixed compounds FG5974 (Albinsson et al., 1994), LEK-8804 (Krisch and Bole, 1994) and CGS 18102A (Sills et al., 1990) were recently examined for their anticonflict activity in the pigeon (Kleven and Koek, 1996), but were not found to have a profile of activity superior to that of 8-OH-DPAT. The results of our study indicate that, although concomitant 5-HT_1A agonist properties complicate the in vivo characterization of mixed compounds, 5-HT_2A antagonist properties could be identified for CGS 18102A. That is, its ability to antagonize both DOI-induced head twitches and the DS effects of DOI in rats was comparable to that found with the relatively 5-HT₂-selective antagonists, ketanserin and ritanserin, but not for the remaining compounds, WY-50,324, FG5974 or LEK-8804.

	Methods

Top Abstract Introduction Methods Results Discussion References

Receptor Binding Studies

Binding affinities for 5-HT_1A, 5-HT_2C, 5-HT_2A, dopamine D₁ and D₂ and alpha₁-adrenergic receptors were determined by means of ligand displacement assays using the conditions summarized in table 1. Frozen brains of male Sprague Dawley rats [Ico: OFA SD (I.O.P.S. Caw) Iffa Credo, St. Germain sur L'Arbresle, France] were used in 5-HT_1A, 5-HT_2A, D₁, D₂ and alpha₁-adrenergic receptor binding studies and were stored at -70°C before use. The brains were thawed in ice-cold Tris-HCl (50 mM, pH 7.4 at 25°C) and tissues were dissected and homogenized in 20 volumes (40 volumes for alpha₁-adrenergic) of fresh Tris-HCl buffer. For 5-HT_1A and 5-HT_2A binding, the homogenate was centrifuged at 39,000 × g for 10 min, the pellet resuspended in the same volume of fresh buffer and recentrifuged. After a further resuspension, the tissue was incubated for 10 min at 37°C and centrifuged as before. The final pellet was then resuspended in a 5-HT assay buffer of Tris-HCl incorporating pargyline (10 µM), CaCl₂ (4 mM) and ascorbic acid (0.1%). For alpha₁ adrenoceptor binding, the homogenate was centrifuged at 1,000 × g for 10 min, the supernatant was centrifuged at 30,000 × g for 10 min, and the pellet resuspended in the same volume of fresh buffer and centrifuged as before. The final pellet was then resuspended in Tris-HCl buffer. For D₁ and D₂ binding the homogenate was centrifuged at 20,000 × g for 10 min, the pellet resuspended in the same volume of fresh buffer and recentrifuged. The final pellet was then resuspended in a dopamine receptor assay buffer of Tris-HCl incorporating NaCl (120 mM) and KCl (5 mM). 5-HT_2C binding was carried out in pig cortex purchased from the local slaughterhouse. The tissue was prepared as for 5-HT_1A receptor binding except that it was centrifuged/resuspended two additional times before incubation.

All experiments were carried out at room temperature, with the exception of D₁ binding that was carried out at 37°C. Compounds were tested at five to seven different concentrations. The incubation was stopped by rapid filtration, under vacuum through GF/B filters with two 5-ml washes of Tris-HCl buffer. The radioactivity retained on the filters was measured by scintillation spectroscopy in 4 ml of scintillation fluid (Emulsifier safe, Packard, Rungis, France). All experiments were performed in triplicate. K_i values were determined by use of the nonlinear, least square curve-fitting EBDA/LIGAND program (Biosoft, Cambridge, U.K.). The results are presented as the mean K_i values from three to five determinations.

Materials. ³H-8-OH-DPAT (160-240 Ci/mmol), ³H-prazosin (65-85 Ci/mmol), ³H-SCH-23390 (60-90 Ci/mmol) and ³H-mesulergine (70-85 Ci/mmol) were purchased from Amersham, Les Ulis, France, ³H-YM-09151-2 (70-87 Ci/mmol) and ³H-ketanserin (60-90 Ci/mmol) from Du Pont-New England Nuclear, Les Ulis, France. 5-HT creatinin sulfate and phentolamine mesylate were purchased from Sigma Chemical Co. (St. Louis, MO), (+)-butaclamol HCl, methysergide maleate, mianserin HCl and SKF 38393 HCl, were all purchased from Research Biochemicals Intl. (RBI, Natick, MA).

Behavioral Observation Studies

Animals. Male Sprague Dawley rats weighing between 160 to 200 g at the beginning of the studies were used. Animals used in observational experiments were housed individually in plastic hanging cages (28 × 21 × 18 cm) with metal grid floors (RC Iffa Credo), in air-conditioned (temperature: 21 ± 1°C; hygrometric degree: 55 ± 5%) rooms under a 12-hr light-dark cycle (lights on from 7:00 A.M. to 7:00 P.M.). During an adaptation period ranging from 7 to 14 days, filtered (0.22 µ) water and food (A04, 4R, Epinay sur Orge, France) were freely available. Between 17 to 21 hr before experiments took place, the animals were transferred to the individual cages (described above); at this time, water, but not food, was available. Experiments were conducted between 9:00 A.M. and 1:00 P.M.

Procedure. Observations were made at a single time point, centered at 15 min after the administration of DOI or vehicle and lasting for a total of 10 min. Four animals were observed individually, in turn, every 15 sec during an observation period of 10 sec per animal and the behaviors shown in table 2 were noted. During each of the 10 observation periods, FPT and LLR were considered present if the animal showed uninterrupted signs for at least 3 sec, whereas FBP was scored as present if it occurred during the entire observation period and head twitches were scored as present if any occurred during the observation period. Because there were 10 observation periods, the incidence of FPT, LLR and head twitches could vary from 0 to 10. On each day, no more than two animals in each group received the same treatment.

Drug treatments. DOI was administered i.p.10 min before the start of the observation period. For tests of potential antagonist activity, all drugs, with the exception of FG5974, were dissolved in distilled water and injected s.c. 45 min before the administration of DOI (0.63 mg/kg i.p.); FG5974 was prepared as a suspension in aqueous Tween 80 (2 drops/10 ml distilled water) and injected i.p. 45 min before DOI. For interaction studies, the 5-HT_1A antagonist WAY-100635 (2.5 mg/kg, dissolved in distilled water) was administered s.c. 60 min before administration of DOI, 0.63 mg/kg, i.e., 15 min before the administration of the 5-HT₂ antagonists ritanserin and ketanserin, the putative mixed 5-HT_1A agonist/5-HT₂ antagonists CGS 18102A, WY-50,324, FG5974 and LEK-8804 or the 5-HT_1A agonists, 8-OH-DPAT and buspirone.

Data analysis. Of a total of 343 animals pretreated with saline or Tween 80 (i.p.10 min before the observation period), head twitch, FBP, LLR or FPT scores of 1 or more were never observed. Test compounds were therefore considered to produce an effect in individual animals if head twitch, FBP, FPT or LLR scores of 1 or more were observed. A total of 68 animals received an injection of DOI, 0.63 mg/kg i.p. 10 min before the start of the observation period and 60 of these animals showed one or more head twitches. Test compounds were considered to produce antagonism of head twitches in individual animals if scores of 0 occurred; dose-response functions were determined by calculating the percentage of animals showing antagonism of head twitches, or the presence of FBP, FPT or LLR. ED₅₀ values and their associated confidence limits were calculated with the method of Litchfield and Wilcoxon (1949) (Tallarida and Murray, 1987) with Abbott's formula (Hubert, 1984) used to correct for the expected basal frequency of head twitch scores of zero (i.e., 12%). Similarly, because 17 of 68 animals treated with DOI (0.63 mg/kg) showed FPT, Abbott's formula was used to correct for the expected basal frequency of FPT in interaction studies.

Drug Discrimination Procedure

Animals. Male Sprague Dawley rats, weighing between 240 to 260 g at the beginning of the studies were used. Animals were housed in individual cages (Iffa Credo, 28 × 21 × 18 cm) with metal grid floors in air-conditioned rooms identical to those described above. Filtered (0.22 µ) water was freely available, but access to standard laboratory food (A04, 4AR, Usine d'Alimentation Rationnelle, Epinay sur Orge, France) was limited to 10 g/day, except during weekends when food was freely available between 5:00 P.M. Friday and 2:00 P.M. Sunday. Experiments were conducted between 9:00 A.M. and 5:00 P.M., Monday through Friday. Animals were cared for in accordance with guidelines set by the US Department of Health and Human Services for humane treatment of animals (Guide for the Care and Use of Laboratory Animals, US DHHS, PHS, National Institutes of Health publication no. 85-23, revised 1985) and the experimental protocol (which was assigned no. 009 by our local regulatory committee) was carried out in accordance with French law and the local ethical committee guidelines for animal research.

Apparatus. Experiments were conducted in standard operant conditioning chambers (model E10-10, Coulbourn Instruments, Lehigh Valley, PA) housed in light- and sound-attenuating enclosures that were ventilated by a fan that also produced a masking noise. Each chamber contained a house-light that was mounted above a food pellet receptacle located between two levers that were situated 2.5 cm above the grid floor. Food pellets (45 mg dustless pellets, Bioserv, Frenchtown, NJ) were delivered by a pellet dispenser (model E14-12, Coulbourn Instruments, Lehigh Valley, PA). Scheduling of reinforcement contingencies, reinforcement delivery and data recording were controlled by a SKED-11 system (State Systems, Kalamazoo, MI) implemented on a PDP-11 computer (Digital Equipment Corporation, Maynard, MA). SKED data files were transferred electronically to a DEC micro VAX computer and appended to designated RS/1 tables (Bolt Beranek and Newman Inc., Cambridge, MA) by means of a dedicated procedure that was written using the Research Programming Language of RS/1.

Discrimination training procedure. For pretraining, rats were assigned randomly to operant chambers in which they were trained to lever press for food pellets. The behavior of half of the rats was shaped to press the left lever and of the other half to press the right lever. During daily 15-min sessions, of which only the first began with the noncontingent delivery of five food pellets, responding on one of the two levers produced food according to a fixed ratio of one press per food pellet (FR1); in addition, food pellets were delivered in a non-contingent manner using a random interval 60-sec schedule. These conditions remained in effect until 20 reinforced responses had accumulated during a session, after which the noncontingent food delivery was stopped and an adjusting ratio schedule became operative. When five consecutive ratios of the same value were completed, the next ratio was increased by 2, without exceeding 10 as the upper limit of the ratio. However, when an interresponse time of more than 15 sec occurred, the next ratio was decreased by 2, with 1 as the lower limit of the ratio. To prevent reinforcement of long interresponse times and chaining between responses on the two levers, one additional response was needed to produce food when the last response of a ratio was preceded either by an interresponse time longer than 15 sec or by a response on the non-reinforced lever. Each adjusting-ratio session started with the value of the last ratio of the immediately preceding session. Pre-training criterion performance was defined as the completion of 50 or more FR10 ratios during a session that started with FR10. Once an animal reached criterion for responding on the first lever it was trained until it attained criterion for responding on the second lever. Thereafter, left-lever sessions and right-lever sessions were alternated daily. Drug discrimination training was started only in those animals that met the pretraining criterion during 2 consecutive sessions with different levers being correct in less than 25 sessions since the start of the pretraining phase.

Testing procedure. Test sessions were conducted twice per week on Wednesday and Friday, although training continued on intervening days. During test sessions, the lever on which 10 responses accumulated first was defined as the selected lever. After lever selection, the animal received a food pellet and subsequent reinforcement was made contingent on pressing the selected lever. A test session ended after 15 min. Testing was postponed to the next scheduled test day if 1) on either of the two most recent training days, the FRF value exceeded 15, 2) on either of the 2 most recent training days, the response rate was less than 80% of the rate observed during the preceding training session of the same type (i.e., drug or saline) or 3) during the most recently preceding saline training session, the total number of responses was less than 500. Also, test data were discarded and the test condition later retested if the test session was followed by a training session of which the FRF value exceeded 15.

Data analysis. Test sessions generated data on two variables: 1) the selected manipulandum, i.e., saline lever or drug lever, representing the measure of discriminative responding and 2) the response rate, i.e., the total number of responses made on either lever during the 15-min session expressed as a percentage of the response rate during the most recently preceding saline training session. Lever-selection data were used to calculate the percentage of animals at each treatment condition selecting the drug lever. Drug effects on this quantal variable were analyzed by means of the Litchfield and Wilcoxon procedure (Tallarida and Murray, 1987), implemented by use of the research programming language RS/1 (Bolt Beranek and Newman Inc., Cambridge, MA), to estimate ED₅₀ values and 95% confidence limits. When less than two intermediate effects were observed, 0 and/or 100% effects were transformed by means of Berkson's adjustment (Hubert, 1984) to permit the use of the Litchfield and Wilcoxon procedure. The order of treatment with individual drugs and doses was unsystematic.

Drug treatments. The training drugs DOI and 8-OH-DPAT were administered i.p.15 min before the start of the behavioral sessions. For tests of potential agonist activity, all drugs were dissolved in distilled water and injected i.p.15 min before the session. For tests of antagonist activity, all compounds with the exception of FG5974 were dissolved in distilled water and administered s.c. 60 min before the behavioral session, i.e., 15 min before the administration of either the training drugs or WY-50,324; FG5974 was prepared as a suspension in aqueous Tween 80 (2 drops/10 ml distilled water) and injected i.p. 60 min before the session.

Drugs. The drugs used in this study were 8-OH-DPAT HBr, buspirone HCl, DOI HCl, NAN-190 HBr, ketanserin tartrate and ritanserin HCl (all obtained from Research Biochemicals Intl., Natick, MA), CGS 18102A (Ciba-Geigy, Rueil-Malmaison, France), FG5974 HCl (Kabi Pharmacia AB, Malmö, Sweden), (S)-WAY-100135 HCl, WAY-100635 HCl and WY-50,324 (synthesized by J.-L. Maurel and J. M. Autin, Centre de Recherche Pierre Fabre, Castres, France) and LEK-8804 tartrate (LEK Pharmaceutical and Chemical Co., Ljubljana, Slovenia). Drugs were dissolved in distilled water, with the exception of FG5974, which was prepared as a suspension in aqueous Tween 80 (2 drops/10 ml distilled water); all drugs were injected in a volume of 10 ml/kg and doses are expressed as the free base. Control treatments consisted of saline and/or Tween 80 injections.

	Results

Top Abstract Introduction Methods Results Discussion References

Receptor binding studies. All of the compounds examined in this study, with the exception of ritanserin and ketanserin, exhibited high affinity for 5-HT_1A receptors (table 3); among the 5-HT_1A agonists, buspirone exhibited the lowest 5-HT_1A affinity (K_i = 24.9 nM). Among the compounds examined in this study, ritanserin exhibited the highest affinity for 5-HT_2C receptors, followed by LEK-8804, ketanserin and CGS 181092A, which exhibited intermediate affinity; the remaining mixed 5-HT_1A agonists, FG5974 and WY 50,324, the 5-HT_1A antagonists NAN-190, (S)-WAY-100135 and WAY-100635 exhibited low affinity (501-2030 nM), and the 5-HT_1A agonists 8-OH-DPAT and buspirone had no affinity for this receptor at the highest concentration tested here (i.e., 10,000 nM). With respect to 5-HT_2A receptors, both ritanserin and ketanserin exhibited subnanomolar affinity, whereas the mixed 5-HT_1A agonists exhibited moderate affinity (i.e., affinity values ranging from 42.9 to 278 nM) and the 5-HT_1A antagonists NAN-190, (S)-WAY-100135 and WAY-100635 and the 5-HT_1A agonists 8-OH-DPAT and buspirone had relatively low or nonexistent 5-HT_2A affinity.

DOI-induced head twitches. As shown in figure 1, DOI (0.04-2.5 mg/kg i.p.) produced dose-related increases in head twitch scores and the percentage of animals having scores of one or more during the 10-min observation period (ED₅₀ = 0.53 mg/kg, 95% confidence limits = 0.16-1.8 mg/kg). Because in preliminary studies the lowest dose that had maximal effects on the percentage incidence of head twitch scores higher than zero was found to be 0.63 mg/kg, this dose was chosen for use in further studies. A total of 68 animals was treated with this dose of DOI in the course of the present experiments, and scores of 1 or more were observed in 60 rats. Thus, ED₅₀ values were estimated with a correction for the basal frequency of zero scores (i.e., 8/68) observed in these control animals.

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Effects of DOI on head twitches. Values represent the mean ± SEM head twitch scores (top panel) and the percentage of animals (bottom panel) showing head twitch scores of more than 0 during a 10-min observation period centered 15 min after administration of DOI (0.04-2.5 mg/kg i.p.; n = 7/group).

View larger version (26K): [in this window] [in a new window]   Fig. 2.   Comparison of the effects of 5-HT2A/2C antagonists, mixed 5-HT1A agonist/5-HT2A/2C antagonists or 5-HT1A agonists on head twitches induced by DOI. Values represent the percentage of animals showing head twitch scores of more than 0 during a 10-min observation period centered 15 min after administration of DOI (0.63 mg/kg i.p.), 60 min after administration of the test compounds. Closed triangles represents the value obtained in 68 control rats pretreated with saline (C; 45-min presession) and 0.63 mg/kg DOI (15-min presession).

Other observable behaviors. As shown in figure 3 (open symbols), the prototypical 5-HT_1A agonist 8-OH-DPAT produced dose-related increases in LLR (ED₅₀ = 0.093 mg/kg; table 6), FBP (ED₅₀ = 3.5 mg/kg) and FPT (ED₅₀ = 0.63 mg/kg).

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Comparison of the ability of prototypical 5-HT1A agonists, 8-OH-DPAT and buspirone and mixed 5-HT2A/2C/1A or 5-HT1A compounds to produce lower lip retraction (LLR), flat body posture (FBP) and forepaw treading (FPT) when administered in combination with saline (open symbols) or DOI (0.63 mg/kg; closed symbols). Symbols represent the percentages of animals (n = 7 to 9/dose) showing individual behavioral scores higher than those observed in control animals treated with saline. See methods for details of the scoring procedure.

Antagonism of discriminative stimulus effects of DOI. As shown in figure 4, the 5-HT₂ antagonists ketanserin and ritanserin dose-dependently decreased DOI-lever selection and reversed the rate-decreasing effects (i.e., 68.5 ± 7.9% of control response rate) of the training dose of DOI (0.63 mg/kg), with estimated ED₅₀ values for the former effects, shown in table 7, similar to those obtained in the head twitch experiments (table 4). Of the mixed 5-HT_1A agonist/5-HT_2A/2C antagonists examined in this study, only CGS 18102A (ED₅₀ = 6.2 mg/kg) antagonized the DS effects of the training dose of DOI in more than 50% of the animals, although all of the compounds were tested at doses that also significantly decreased rates of responding. Although 8-OH-DPAT, WY-50,324 and buspirone produced decreases in drug-lever selection at one or more consecutive doses, the maximal inhibition was not more than 33%. The mixed compounds FG5974 and LEK-8804 were ineffective in decreasing drug-lever selection. None of the mixed antagonists, including CGS 18102A, significantly reversed the rate-decreasing effects of DOI.

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Comparison of the ability 5-HT2A/2C, 5-HT2A/2C/1A or 5-HT1A compounds to antagonize the discriminative stimulus effects of DOI in rats trained to discriminate DOI (0.63 mg/kg i.p.) from saline. Symbols represent the percentage of animals (n = 7 to 9/dose) selecting the drug-appropriate lever during test sessions (upper panel) and the mean ± S.E. response rate (lower panel). Values represent results from sessions in which the test compounds were injected s.c. or i.p. 60 min before the session (45 min before administration of the training dose of the training drug).

Substitution in DOI-trained rats. Each of the 5-HT_1A compounds examined in this study engendered intermediate levels of drug-lever selection in rats trained to discriminate DOI from saline (fig. 5). The mixed compound WY-50,324 and the relatively selective 5-HT_1A agonist 8-OH-DPAT produced the highest levels of DOI-appropriate responding (67 and 60%, WY-50,324 and 8-OH-DPAT, respectively). Lower levels were produced by CGS 18102A (40%), buspirone (33%), FG5974 (25%) and LEK-8804 (22%). All of the compounds produced dose-related decreases in response rates, with maximal effects on this measure occurring at doses that produced the highest levels of drug-lever selection. Lever selection occurred in fewer than 50% of the animals when tested with the highest doses of FG5974, LEK-8804 and buspirone.

View larger version (30K): [in this window] [in a new window]   Fig. 5.   Comparison of the ability of 5-HT2A/2C, 5-HT2A/2C/1A or 5-HT1A compounds to substitute for the discriminative stimulus effects of DOI (0.63 mg/kg i.p.) or 8-OH-DPAT (0.16 mg/kg i.p.). Symbols represent the percentage of animals (n = 7 to 9/dose) selecting the drug-appropriate lever (upper panels) or mean ± S.E. rate of responding (lower panels). Values in DOI-trained rats (left panels) represent results from sessions in which the saline was injected s.c. 60 min before the session (45 min before administration of the test drug).

Substitution in 8-OH-DPAT-trained rats. each of the compounds with 5-HT_1A agonist properties engendered dose-related increases in drug-lever selection in rats trained to discriminate 8-OH-DPAT (0.16 mg/kg i.p.) from saline (fig. 5, right panels). Complete substitution for the DS effects of 8-OH-DPAT was observed after administration of 8-OH-DPAT and the mixed compound WY-50,324, whereas the remaining compounds, buspirone, CGS 18102A, FG5974 and LEK-8804 produced from 78 to 89% drug-lever selection. All of the compounds produced dose-related decreases in rates of responding, with maximal effects on this measure at doses corresponding to those producing the highest levels of drug-lever selection.

Antagonism of partial substitution in DOI-trained rats. Consistent with the results obtained with the training dose of DOI (fig. 4), fixed doses of the 5-HT₂ antagonists ketanserin (0.16 mg/kg) and ritanserin (0.16 mg/kg) produced rightward shifts in the DOI dose-response function, whereas the 5-HT_1A antagonist (S)-WAY-100135 was ineffective (fig. 6). Conversely, pretreatment with (S)-WAY-100135 produced rightward shifts in the dose-response functions of both 8-OH-DPAT and WY-50,324 in DOI-trained rats, whereas ketanserin and ritanserin did not consistently block drug-lever selection. However, a decrease in the magnitude of drug-lever selection engendered by 8-OH-DPAT was found after pretreatment with both ritanserin and ketanserin. Further, ketanserin, but not ritanserin, produced an apparent rightward shift of the WY-50,324 dose-response function; however, these effects were observed at doses that produced substantial decreases in rate of responding (data not shown) that could not be reversed by concomitant administration of the 5-HT₂ antagonists.

View larger version (24K): [in this window] [in a new window]   Fig. 6.   Comparison of the ability of the 5-HT1A antagonist (S)-WAY-100135 and the 5-HT2 antagonists ritanserin and ketanserin to block the discriminative stimulus effects of the DOI, 8-OH-DPAT and WY-50,324 in rats trained to discriminate DOI (0.63 mg/kg i.p.) from saline. Values shown are the percentage of animals (n = 7 to 9/dose) selecting the drug lever. Top panels: pretreatment with saline (open symbols) or (S)-WAY-100135 (10 mg/kg s.c.); middle panels: pretreatment with saline (open symbols) or ritanserin (0.16 mg/kg s.c.); bottom panels: pretreatment with saline (open symbols) or ketanserin (0.16 mg/kg s.c.). Dashed lines connecting symbols indicate that the data were redrawn from other panels; e.g., see left panels of figure 5.

	Discussion

Top Abstract Introduction Methods Results Discussion References

The results of this study show that the putative mixed 5-HT_1A agonist/5-HT_2A/2C antagonist CGS 18102A has prominent 5-HT_2A antagonist properties in vivo, demonstrated both by its ability to attenuate the DS effects of DOI and to block DOI-induced head twitches in a manner not reversible by WAY-100635. In contrast, the putative mixed compounds WY-50,324, FG5974 and LEK-8804 exhibited prominent 5-HT_1A agonist properties in vivo, in that their inhibition of DOI-induced head twitches was reversible by WAY-100635 and they failed to inhibit the DS effects of DOI. Further, all of the mixed compounds exhibited in vivo 5-HT_1A agonist properties, in that they produced elements of the 5-HT syndrome and engendered drug-lever selection in 8-OH-DPAT-trained rats. Although putative mixed agonists/antagonists, as well as the relatively selective 5-HT_1A agonist 8-OH-DPAT, were able to attenuate DOI-induced head twitches, our results indicate that compounds with a mixed profile could be identified on the basis of their behavioral effects in the procedures used here.

In this study, DOI-induced head twitches were reversible by pretreatment with the 5-HT₂ antagonists ritanserin and ketanserin, findings that are consistent with the involvement of 5-HT_2A receptors in this behavior. Because relatively selective 5-HT_1A agonists such as 8-OH-DPAT are also able to inhibit head twitches (Yocca et al., 1990; Koek et al., 1992; Dursun and Handley, 1993; Schreiber et al., 1995), a more detailed pharmacological analysis using selective 5-HT_1A antagonists was performed in this study as was done previously for S14671 (Schreiber et al., 1995), LEK-8804 (Krisch and Bole, 1994) and FG5974 (Albinsson et al., 1994). However, putative 5-HT_1A antagonists that have been shown to have weak partial 5-HT_1A agonist properties, such as (S)-WAY-100135 (Assié and Koek, 1996), (±)-pindolol (Yocca et al., 1990; Dursun and Handley, 1993) and NAN-190 (Greuel and Glaser, 1992) are not the most appropriate tools to rule out the involvement of 5-HT_1A receptors. That is, because the absence of blockade by a presumed 5-HT_1A antagonist is interpreted as evidence against 5-HT_1A receptor involvement, compounds that inhibit DOI-induced head twitches when given alone may be limited in their ability to reverse 5-HT_1A-mediated inhibition of head twitches. However, in this study a behaviorally inactive dose of WAY-100635 significantly reversed the effects of 8-OH-DPAT and buspirone, as well as all of the mixed agonists with the exception of CGS 18102A. Taken together with the relatively high selectivity of WAY-100635 (table 3), these results strongly implicate 5-HT_1A receptors in the ability of the former compounds to inhibit DOI-induced head twitches. Conversely, the results indicate that CGS 18102A inhibits head twitches via mechanism(s) not involving 5-HT_1A receptors.

The results obtained with several of the mixed compounds do not entirely agree with those of previous studies in which head twitches could not be reversed by coadministration of compounds having 5-HT_1A antagonist properties. That is, (±)-pindolol did not reverse the ability of FG5893 (the related methyl ester salt of FG5974) to inhibit DOI-induced head twitches (Albinsson et al., 1994) and NAN-190 did not inhibit the effects of LEK-8804 on 5-HTP-mediated head twitches (Krisch and Bole, 1994). These previous findings may be problematical because (±)-pindolol and NAN-190 have partial agonist properties (Yocca et al., 1990; Greuel and Glaser, 1992; Dursun and Handley, 1993; Moore et al., 1993). Nonetheless, the reversal of the effects of LEK-8804 by WAY-100635 in our study was less than that observed with the prototypical 5-HT_1A agonist 8-OH-DPAT (i.e., 29 vs. 86%, LEK-8004 vs. 8-OH-DPAT). This finding suggests that LEK-8804 may to some extent also decrease DOI-induced head twitches via additional mechanisms not involving 5-HT_1A receptors. Apart from its relatively high affinity for 5-HT_2C receptors, the binding profile of LEK-8804 as examined in this study does not differ markedly from the other mixed compounds (table 3). Thus, it remains to be determined what other mechanisms might be responsible for these differences between LEK-8804 and prototypical 5-HT_1A agonists or putative mixed 5-HT_1A agonists/5-HT_2A/2C antagonists; however, because LEK-8804 did not resemble the 5-HT₂ antagonists ketanserin or ritanserin in drug discrimination studies (see below), it is unlikely that they involve 5-HT_2A antagonist properties.

In this study, 8-OH-DPAT and several of the mixed compounds produced FBP and FPTelements of the "5-HT syndrome" produced by direct and indirect 5-HT agonists in rats (Tricklebank et al., 1984)and produced LLR, another effect characteristic of 5-HT_1A agonists (Berendsen et al., 1989). These findings are consistent with previously reported results and therefore demonstrate that 5-HT_1A agonist-like effects of mixed compounds can be readily observed in the rat. Although all of the compounds examined in this study produced a measurable incidence of FBP, there were several apparent differences from the prototypical 5-HT_1A agonist 8-OH-DPAT with respect to the other observable effects. For example, as with buspirone, the mixed compounds CGS 18102A, WY-50,324 and FG5974, did not produce FPT even when administered at relatively high doses, and, interestingly, the mixed agonists CGS 18102A and LEK-8804 failed to produce LLR in more than 50% of animals tested with doses as high as 40 mg/kg. Although these results indicate relatively weak 5-HT_1A agonist properties, LEK-8804 exhibited a more unusual profile because FBP and FPT could be observed at doses that did not induce LLR. Overall, the observational studies suggest that mixed compounds may have effects that differ from those produced by 8-OH-DPAT. Further studies are needed to determine whether such effects involve unrelated mechanisms or result from putative mixed 5-HT_2A/2C antagonist/5-HT_1A agonist properties.

As further evidence of in vivo 5-HT_1A agonist properties, relatively selective or mixed 5-HT_1A agonists substituted for the DS effects of 8-OH-DPAT. The finding that WY-50,324 has 8-OH-DPAT-like DS effects agrees with reported findings in pigeons trained to discriminate 8-OH-DPAT from saline (Barrett and Zhang, 1991a). Whereas DS effects of the mixed compounds CGS 18102A, FG5974 and LEK-8804 have not, to our knowledge, been reported, their ability to substitute for 8-OH-DPAT is consistent with results from observation studies. But, CGS 18102A, FG5974, LEK-8804 and buspirone did not engender drug-lever selection in all of animals tested, even at doses that produced substantial decreases in rates of responding. Whereas for buspirone these results could result from partial agonist properties (Yocca, 1990; Rabin and Winter, 1993), the intrinsic activity of the mixed compounds at 5-HT_1A receptors remains to be determined. Nonetheless, the finding that WY-50,324 engendered complete substitution indicates that differences in magnitude of drug-lever selection among selective and mixed compounds are unlikely to be related to the combination of 5-HT₂ antagonist and 5-HT_1A agonist properties. Therefore, the ability of these compounds to substitute for the DS effects of 8-OH-DPAT is entirely consistent with their 5-HT_1A agonist properties identified from observable behaviors.

In this study, the DS effects of DOI were reversible by pretreatment with the 5-HT_2A/2C antagonists ritanserin and ketanserin, consistent with previous studies demonstrating the involvement of 5-HT_2A receptors (cf., Koek et al., 1992; Schreiber et al., 1994; Arnt, 1996). Nonetheless, this behavioral assay, as with head twitches, is also complicated by interactions between 5-HT_1A agonists and mechanism(s) underlying the DS effects of DOI, evident from the fact that the 5-HT_1A agonist 8-OH-DPAT engendered partial substitution, consistent with previous findings (Colpaert et al., 1992; Koek et al., 1995). It is likely that this effect is mediated by 5-HT_1A receptors, because all of the mixed compounds also engendered low to intermediate levels of DOI-lever selection (i.e., 22 to 67% drug-appropriate responding), with a rank-order potency (LEK 8804 < 8-OH-DPAT < WY-50,324 < FG5974 < CGS 18102A) virtually identical to that observed in substitution tests in 8-OH-DPAT-trained rats. However, the ED₅₀ values were approximately two to six times higher in the DOI-trained rats, a finding that may explain the inability of LEK-8804 to produce greater effects because there was little separation between rate-decreasing effects and substitution in 8-OH-DPAT-trained rats. Both the rank-order potency of the compounds examined here and the finding that (S)-WAY-100135 produced surmountable shifts in DOI-lever selection engendered by 8-OH-DPAT and WY-50,324, indicate that 5-HT_1A receptors are involved in this effect. Furthermore, because (S)-WAY-100135 did not block the DS effects of DOI, it is likely that 5-HT_1A receptors do not play a direct role in the DS effects of DOI. Conversely, the 5-HT₂ antagonists, ketanserin and ritanserin, did not consistently alter the ability of 8-OH-DPAT or WY-50,324 to partially substitute for DOI. That this partial substitution is not related to performance deficits was addressed in a recent study (Koek et al., 1995) wherein intermediate levels of DOI-appropriate responding engendered by WY-50,324 could not be attributed to behavioral mechanisms other than stimulus generalization, unlike results obtained after administration of the NMDA antagonist dizocilpine to DOI-trained rats. Therefore, the ability of 5-HT_1A agonists to engender partial substitution may be attributed to 5-HT_1A receptor-mediated effects that partially mimic the DS effects of DOI.

One of the main goals of the present study was to determine whether in vivo 5-HT_2A antagonist properties could be identified in compounds that may also be 5-HT_1A agonists. The compounds examined here are of current therapeutic interest because of the possibility that they may be exceptional anxiolytics owing to the combination of 5-HT_1A agonist and 5-HT_2A/2C antagonist properties (Millan et al., 1992a; Barrett and Vanover, 1993). However, as summarized in table 8, the results of the different tests examined in this study point to the existence, for all of the compounds except CGS 18102A, of predominant in vivo 5-HT_1A agonist properties and the absence of evidence for 5-HT_2A antagonist properties. All of the compounds decreased DOI-induced head twitches, an assay that does not readily distinguish among different pharmacological classes. Nonetheless, the selective 5-HT_1A antagonist WAY-100635 blocked the ability of all of the compounds, with the notable exception of CGS 18102A, to inhibit DOI-induced head twitches, demonstrating that 5-HT_1A agonist properties are involved in the effects of all of the compounds except CGS 18102A. Further confirmation of in vivo 5-HT_1A agonist properties was obtained by examining the ability of the different compounds to produce elements of the "5-HT-syndrome" and to substitute for the DS effects of 8-OH-DPAT. Although only 8-OH-DPAT and WY-50,324 produced high levels of all three of the behaviors examined in this study, all of the compounds produced one or more of these different behaviors. Finally, among the mixed compounds examined here, only CGS 18102A blocked the DS effects of DOI, further indicating that it has 5-HT_2A antagonist properties. The concordance between the effects of CGS 18102A in two different behavioral models indicative of 5-HT_2A antagonist properties suggests that putative mixed compounds may be differentiated on the basis of effects in these behavioral models, whereas the remaining mixed compounds exhibited mostly 5-HT_1A agonist properties.

The existence of prominent in vivo 5-HT_1A agonist properties is in agreement with the high affinity and relatively higher selectivity for 5-HT_1A receptors exhibited by the mixed compounds examined in this study (table 3) (Abou-Gharbia and Moyer, 1990; Sills et al., 1990; Albinsson et al., 1994; Krisch and Bole, 1994), whereas in vivo 5-HT_2A antagonist properties of CGS 18102A are not readily predictable on the basis of its in vitro binding profile in comparison to that of the other mixed compounds. That is, all of the mixed compounds with the exception of FG5974 show more than a 100-fold lower affinity for 5-HT_2A receptors, relative to 5-HT_1A affinity. In contrast, FG5974 has high affinity for both 5-HT_1A and 5-HT_2A receptors, yet did not block DOI-induced head twitches in a manner different from 8-OH-DPAT. Nor do absolute 5-HT_2A affinities distinguish CGS 18102A from the remaining mixed compounds, because there is very little separation among this group (i.e., K_i values range from 42.9 to 278 nM) (table 3). Thus, it is unexpected that only CGS 18102A is able to exert 5-HT_2A antagonist effects in vivo. However, because WAY-100635 significantly reversed the effects of compounds such as buspirone, LEK-8804 and FG5974 on DOI-induced head twitches, relatively high D₂ and alpha₁-adrenergic affinities are not sufficient to cause decreases in this behavior. The most parsimonious explanation of the in vivo effects of mixed compounds is that 5-HT_1A agonist properties play a determining role in their ability to reverse DOI-induced head twitches; the finding that CGS 18102A is a relatively weak 5-HT_1A agonist in vivo may explain why only this compound can be differentiated on the basis of the behavioral assays used in this study.

Taken in conjunction with reports that both the DS and the head-twitch inducing effects of DOI are mediated predominantly by 5-HT_2A receptors (Millan et al., 1991a; Kennett et al., 1994; Schreiber et al., 1995), our results suggest that 5-HT_2A receptors do not play a role in the reportedly superior effects of mixed compounds in the pigeon conflict-procedure (Barrett and Zhang, 1991a; Millan et al., 1992a). Moreover, the reported superior anti-conflict effects of mixed compounds may be of limited generality, because in our previous study conducted in pigeons (Kleven and Koek, 1996), only WY-50,324 exhibited effects that could be considered more substantial than those produced by buspirone, whereas FG5974, CGS 18102A and LEK-8804 failed to produce comparable increases in punished responding. Consequently, a mixed profile, insofar as it refers to compounds such as CGS 18102A, cannot account for the reported superior anticonflict effects in pigeons (Barrett and Zhang, 1991a; Millan et al., 1992a). Nonetheless, it remains possible that 5-HT_2C antagonist properties of mixed 5-HT_1A agonists/5-HT_2A/2C antagonists play a role in their anxiolytic-like effects. Because compounds with 5-HT_2C antagonist properties exhibit anxiolytic-like effects in both pigeon (Gleeson et al., 1989; Brocco et al., 1990; Koek et al., 1992; Kleven and Koek, 1996) and rodent conflict-procedures (Kennett et al., 1994), the involvement of in vivo 5-HT_2C antagonist properties in the preclinical effects of the mixed compounds examined in this study merits further study.

In summary, of the putative mixed 5-HT_1A agonist/5-HT_2A/2C antagonists examined here, only CGS 18102A could be identified as having 5-HT_2A antagonist properties in vivo on the basis of results from the different tests used in this study (table 8). Foremost, its ability to antagonize both DOI-induced head twitches and the DS effects of DOI in rats was comparable to that found with the relatively 5-HT_2A-selective antagonist ketanserin and the 5-HT_2A/2C antagonist ritanserin. Furthermore, it substituted for the DS effects of 8-OH-DPAT and produced observable behaviors compatible with 5-HT_1A agonist properties. These findings are expected of a compound having mixed 5-HT_1A agonist/5-HT_2A antagonist properties. Other compounds examined in this study could have limited 5-HT_2A antagonist properties in vivo, nonetheless, the finding that WAY-100635 reversed their effects on head twitches mitigates against this possibility. Although there is a need for a behavioral assay of 5-HT_2A antagonist activity that is independent of 5-HT_1A agonist effects, our results indicate that a 5-HT_1A agonist with prominent 5-HT_2A antagonist properties can be distinguished from relatively selective 5-HT_1A agonists.