Introduction

Top Abstract Introduction Methods Results Discussion References

 Antidepressants and benzodiazepine-type anxiolytics are among the most commonly prescribed therapeutic agents. Although selective serotonin reuptake inhibitors, such as fluoxetine, induce fewer side effects than the "classic" antidepressants that were introduced in the 1950s (i.e., MAO inhibitors and tricyclic antidepressants), their clinical efficacy is similarly limited: that is, they have a slow onset of action (i.e., 2-4 weeks) and they are effective in only ~70% of depressed patients, less than half of whom show a full response (see reviews by, for example, Broekkamp et al., 1995; Möller and Volz, 1996; Frazer, 1997). For the treatment of anxiety, the benzodiazepines, introduced in the 1960s, have largely replaced barbiturates and other nonbenzodiazepine sedatives, primarily because of their safety; their clinical efficacy, however, is limited (see Rickels et al., 1978; Greenblatt et al., 1983a, 1983b; Rosenbaum, 1982). Thus, currently available treatments for depression and anxiety, although being safer and having fewer side effects, do not appear to be markedly more efficacious than the antidepressants and anxiolytics that were discovered in the 1950-60s. That advances in terms of improved clinical efficacy have as yet to be made is perhaps not surprising inasmuch as the mechanisms of antidepressant and anxiolytic drug action have so far remained essentially unchanged [i.e., indirect stimulation of monoaminergic systems to exert antidepressant effects, and potentiation of -aminobutyric acid (GABA)-ergic systems to exert anxiolytic effects]. Perhaps, agents acting by other mechanisms will prove more efficacious than the currently available treatments.

One novel molecular target has become apparent with the introduction of buspirone for the treatment of anxiety, and also, of depression (see reviews by, for example, Tunnicliff et al., 1991; Fulton and Brogden, 1997). Buspirone, and its analogues gepirone and ipsapirone, activate directly a particular subtype of 5-HT receptor, the 5-HT_1A receptor. Their clinical efficacy, however, is not an improvement over that of other available treatments (e.g., Deakin, 1993). Several hypotheses have guided the search for 5-HT_1A agonists with enhanced activity in animal models of anxiety and depression. One such hypothesis is that the combination of 5-HT_1A agonist with 5-HT₂ antagonist properties confers a more advantageous therapeutic profile (Barrett and Vanover, 1993; Millan and Brocco, 1993), and several compounds thought to possess such a combination have been described [e.g., WY 50,324 (Barrett and Vanover, 1993), and S 14671 (Millan et al., 1992)]. Recent results (Kleven and Koek, 1996), however, failed to confirm empirically the alleged superiority of mixed 5-HT_1A agonists/5-HT₂ antagonists and showed, for example, that the anxiolytic-like effects of 5-HT_1A agonists in animals were not markedly enhanced by the 5-HT₂ antagonist, ritanserin.

Our research has been guided by the hypothesis that the magnitude of the psychotropic activity of 5-HT_1A receptor ligands (i.e., their clinical efficacy) is a direct, positive function of the magnitude of their intrinsic activity at the receptor. This relationship has been established, in the laboratory, for the anxiolytic-like effects of 5-HT_1A ligands (Colpaert et al., 1992), and evidence presented here suggests that this relationship exists also for their antidepressant-like effects (see also De Vry, 1995, 1996). This relationship between clinical efficacy and intrinsic activity may explain why buspirone and its analogues, gepirone and ipsapirone, have limited clinical efficacy; these compounds activate the 5-HT_1A receptor only weakly and are generally considered to be partial 5-HT_1A agonists. In contrast with buspirone, greater activation is achieved with flesinoxan (e.g., Boddeke et al., 1992). The latter compound is in phase III trials; it will be of interest to learn about its clinical efficacy, but flesinoxan's intrinsic activity may still be limited and inferior to that of 5-HT itself (McCall et al., 1994; Hadrava et al., 1995).

The purpose of our drug discovery effort has been to identify novel 5-HT_1A ligands possessing an intrinsic activity higher than that of buspirone and flesinoxan. Here, we report results obtained with F 11440 (4-methyl-2-[4-(4-(pyrimidin-2-yl)-piperazino)-butyl]-2H,4H-1,2,4-triazin-3,5-dione; fig. 1), which is shown to be a potent, selective, high efficacy 5-HT_1A receptor agonist with marked antidepressant- and anxiolytic-like activity in preclinical models and, thus, with the potential to exert antidepressant and anxiolytic activity in humans that may be more substantial than that of currently available treatments.

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Chemical structure of 4-methyl-2-[4-(4-(pyrimidin-2-yl)-piperazino)-butyl]-2H,4H-1,2,4-triazin-3,5-dione (F 11440), in comparison with those of buspirone, flesinoxan and 8-OH-DPAT.

Most cases of depressive disorders occur in people with a history of other disorders, most frequently anxiety (e.g., Kessler et al., 1996). It has been hypothesized that impaired serotonergic mechanisms underlie depression preceded by anxiety, and that serotonergically acting anxiolytics, such as 5-HT_1A agonists, will be the treatment of choice in such "anxiety-driven" depression (Van Praag, 1996). If well tolerated in humans, F 11440 may be one of the most pertinent, currently available tools to test this hypothesis.

	Methods

Top Abstract Introduction Methods Results Discussion References

Animals. Male Sprague-Dawley rats (Ico: OFA SD [IOPS Caw], Iffa Credo, l'Arbresle, France), weighing 160 to 200 g on arrival, were group housed (five animals per cage) with food and water freely available in a quarantine room for 4 to 8 days before being used in the experiments. Thereafter, they were housed individually in hanging cages (length × width × height: 28 × 21 × 18 cm; Iffa Credo, France) with metal grid floors, in the room where the experiments were conducted, with unlimited access to filtered (0.22 µm) water and, except when stated otherwise, standard laboratory food (UAR A03; UAR, Epinay/s/Orge, France). For the forced swimming test, male Wistar rats (Rj: WISTAR (AF), Centre d'Elevage Roger Janvier, Le Genest-Saint-Isle, France) were used. They were housed in groups of five in plastic cages (41 × 25 × 14 cm) containing wood shavings, with free access to food (UAR 113) and water until tested.

Radioligand binding. Binding affinities for the different receptors were determined by means of ligand displacement assays using the conditions summarized in table 1. The reactions were stopped by rapid filtration through Whatman GF/B glass-fiber filters, and the filters were washed with appropriate buffer. The radioactivity retained on the filters was measured by scintillation spectroscopy in 4 ml scintillation fluid (Emulsifier Safe, Packard). All experiments were performed in triplicate, except H₁, h 5-HT_1B and h 5-HT_1D binding, which were performed in duplicate. IC₅₀ values were determined using nonlinear regression, except the IC₅₀ values for h 5-HT_1B and h 5-HT_1D receptors, which were obtained by interpolation. K_i values were calculated using the equation K_i = IC₅₀ (1 + [C]/K_D), and the results were expressed as mean pK_i values ± S.E.M. of three independent determinations (two determinations if pK_i < 5.0).

cAMP in HA7 cells. The HeLa cell line permanently transfected with the human 5-HT_1A receptor gene and permanently expressing the 5-HT_1A receptor protein (HA7) as described by Fargin et al. (1989) was commercially obtained from Duke University (Durham, NC). HA7 cells were grown in DMEM (GIBCO) supplemented with 10% fetal calf serum, gentamicin (100 µg/ml), geneticin (G418) (400 µg/ml) in 5% CO₂ at 37°C in a water-saturated atmosphere. The cells were plated in six-well culture plates and used in the experiments at a confluency of 80% to 90%. Culturing medium (DMEM, 10% fetal calf serum, gentamicin 100 µg/ml, G418 400 µl/ml) was replaced by DMEM supplemented with 10% fetal calf serum without antibiotics 24 hr before experimentation.

5

4

In vivo microdialysis. The methods have been described previously (Assié and Koek, 1996). A probe (CMA/12, 2 mm length, 0.5 mm diameter, CMA, Microdialysis AB) was implanted into the left hippocampus [stereotaxic coordinates: rostral -4.8 mm, lateral +4.6 mm, ventral -7.5 mm, from bregma and dura surface according to Paxinos and Watson (1986)] of anesthetized rats (chloral hydrate, 400-500 mg/kg i.p., and supplementary doses were used to maintain anesthesia). The probe was continuously perfused (1.1 µl/min) with artificial cerebrospinal fluid containing 1 µM of the selective 5-HT uptake inhibitor, citalopram. Starting approximately 2 hr after implantation, perfusates were collected every 20 min and analyzed directly for 5-HT content using high performance liquid chromatography with electrochemical detection (DECADE detector, ANTEC Leyden, The Netherlands).

Stimulation of corticosterone secretion. Twenty-four hours before use in the experiments, rats were housed individually in a restricted area (accessible only to the experimenter) and received 15 g standard laboratory food (water continued to be available freely). Experiments, consisting of drug treatments after which animals were decapitated and trunk blood was collected, were conducted between 8:00 a.m. and 10:30 a.m. Drugs (or vehicle) were administered 60 min before decapitation when given p.o., and 30 min before decapitation when given i.p. Antagonists were administered s.c., 30 min before the i.p. administration of agonists. Immediately before decapitation, animals were observed for the presence or absence of LLR.

LLR and elements of the 5-HT syndrome. Twenty-four hours before use in the experiments, the animals were transferred to individual cages (described above) with water, but not food, freely available. The methods used were essentially the same as those described previously (Kleven et al., 1995). During the experiments, which took place between 9:00 a.m. and 12:00 a.m., rats were observed at two time points, centered at 15 and 60 min after the injection, and each lasting 10 min. During a 10-min period, an animal was observed once every min for 10 sec and the presence (1), or absence (0) of LLR and FPT was recorded. To be considered present, a sign had to occur uninterrupted for 3 sec. Because an animal was observed 10 times during a 10-min period, the incidence of a particular sign could vary from 0 to 10. FBP was scored present (1) if it occurred during the entire 10-min observation period; otherwise, the score was 0.

Drug discrimination. In addition to the food pellets obtained during the conditioning sessions, which were held between 9:00 a.m. and 5:00 p.m. (Monday through Friday), rats received 10 g standard laboratory food per day, except during weekends when food was freely available from 5:00 p.m. on Friday until 2:00 p.m. on Sunday.

Duration of action. The methods used were identical to those described above for the 5-HT syndrome, except that behavioral observations were conducted during six 10-min periods, centered at 10, 60, 120, 240 and 480 min after the injection. Dose-response functions were determined from the percentage of rats showing LLR scores of 1 or more. This criterion was based on the finding that an LLR score of 1 or more occurred in <5% of the control animals receiving saline orally (n = 130) during the observation period centered at 60 min after administration of saline. ED₅₀ values and their associated confidence limits were calculated by means of the Litchfield and Wilcoxon procedure. Duration of action was estimated by calculating the ED₅₀ value and 95% confidence limits at each time interval (10 min and 1, 2, 4 and 8 hr), and by plotting the log-ED₅₀ values as a function of time. The duration of effect, defined as the time during which the ED₅₀ was <4 times the minimum ED₅₀ value, was estimated by linear interpolation of the empirical time-log-ED₅₀ function.

Alpha-1 adrenergic antagonist properties. Experiments were carried out in rats (270-390 g) that had free access to water and standard laboratory food. On the day of the experiment, animals were anesthetized with pentobarbital (60 mg/kg, i.p.) and placed on a heated table to maintain the rectal temperature at 37 ± 0.5°C. Animals underwent tracheotomy to allow spontaneous breathing. Catheters were inserted into the penile vein for infusing drugs and in the right carotid artery for continuous measurement of arterial pressure via a Statham P10EZ pressure transducer (Viggo-Spectromed, Oxnard, CA) connected to a Gould amplifier (Gould Instruments, France). The analog arterial pressure signal was digitized and simultaneously recorded by means of data acquisition software (AcqKnowledge, Biopac Systems Inc., Goleta, CA). Rats received an i.p. injection of either vehicle or a single dose of the test drug, and 30 min later, 0.16 mg/kg methoxamine or its vehicle was injected i.v.

Dopamine antagonist properties. Twenty-four hours before use in the experiments, the rats were transferred to individual cages (described above) with water, but not food, freely available. The methods used were essentially the same as those described previously (Koek and Colpaert, 1993). During the experiments, the animals were observed during a 10-min period starting 30 min after the i.p. injection of 40 mg/kg methylphenidate or vehicle. During the 10-min period, an animal was observed once every min for 10 sec and the presence or absence of locomotion (with all four legs moving), rearing (standing on hindlegs, body fully extended), sniffing, gnawing (the cage or body) and licking (the cage) was recorded. To be considered present, a particular behavior had to occur uninterrupted for 3 sec. Because an animal was observed 10 times during the 10-min period, the incidence of each particular behavior could vary from 0 to 10. In addition, the presence of FBP was noted if it occurred during the entire 10-min observation period. Tremor and convulsions were noted if they occurred at any time during the observation period. At the end of the 10-min period, the animals were tested for catalepsy (the animals maintained for 30 sec, either (1) an abnormal cross-limbed position, imposed by the observer, in which the hindpaws were extended forward and placed on the forepaws, which were extended backward, or (2) both forepaws on a cylindrical metal bar 1.25 cm in diameter, 10 cm elevation), akinesia (absence of movement after handling), and loss of righting (the animal remained in position for 10 sec when placed on its back).

Histamine antagonist properties. The in vivo assay of histamine H₁ antagonist activity was based on that described previously (Rocha E Silva and Antonio, 1978). Briefly, 24 hr before use in the experiments, guinea pigs, weighing 310 to 490 g, were transferred to the room where the experiments were conducted, and were maintained under the same conditions as during quarantine. During the experiment, the animal was placed in a glass cylinder (26 cm wide and 15 cm high) into which the histamine aerosol was introduced by a nebulizer (Jouan, Paris, France) that by means of compressed air (pressure: 1 bar) produced dry particles (diameter: 1 to 3 µm) at a flow rate of 28 ml/hr. The concentration of histamine in the nebulizer was 1% (w/v) in 0.9% NaCl. The histamine aerosol induced bronchoconstriction, loss of righting, and, on continued exposure, convulsions. The animal was removed from the cylinder after it showed loss of righting and before the occurrence of convulsions, or when 3 min had elapsed, whichever occurred first. The dependent variable was the preconvulsion time, that is, the latency (in sec) from the start of exposure to the aerosol until the occurrence of loss of righting.

Pigeon conflict procedure. Male pigeons had free access to food pellets until their body weight was stable during 5 consecutive days, at which time their free-feeding weight was calculated. Thereafter, they were fed 5 g of pellets per day, until the body weight was reduced to 80% of its free-feeding value. From then on, the animals were maintained at approximately 80% of their free-feeding weight by giving, on weekdays, a quantity of pellets equal to the difference between the 80% value and the actual body weight, in g. During weekends, generally between 10 and 20 g of food per day was given.

Forced swimming test. The procedure has been described in detail by Porsolt et al. (1978), and involves placing a rat in a cylinder (height: 40 cm, diameter: 20 cm) containing 13 cm water (25°C) for 15 min on the first day of the experiment, and placing the animal again in the cylinder 24 hr later for 5 min. The duration of immobility during the 5-min period was measured by an observer who was unaware of the treatment conditions.

Drugs. The compounds used were [³H]8-OH-DPAT (160-240 Ci/mmol), [³H]mesulergine (70-85 Ci/mmol) and [³H]prazosin (65-85 Ci/mmol) (Amersham, France); [³H]ketanserin (60-90 Ci/mmol), [³H]-5-CT (15-30 Ci/mmol), [³H]pyrilamine (20-30 Ci/mmol) and [³H]YM-09151-2 (70-87 Ci/mmol) (New England Nuclear, France); IBMX, 5-HT creatinine sulfate, 5-HIAA, astemizole, buspirone hydrochloride, haloperidol, ketotifen fumarate, methoxamine hydrochloride, phentolamine mesylate, prazosin hydrochloride, promethazine hydrochloride, pyrilamine maleate, terfenadine, tripolidine hydrochloride and tryptamine (Sigma Chemical Company, St Louis, MO); 5-methyl-urapidil, (±)-8-OH-DPAT hydrobromide, BMY 7378 dihydrochloride, (+)-butaclamol hydrochloride, m-CPP dihydrochloride, DOI, forskolin, imipramine hydrochloride, methysergide maleate, mianserin hydrochloride, NAN-190 hydrobromide, pargyline, quipazine dimaleate, SKF 38393 hydrochloride, TFMPP and urapidil hydrochloride (Research Biochemicals, Natick, MA); ethanol (99.85%; Prolabo Sud-Ouest, Gradignan, France); chlordiazepoxide hydrochloride and diazepam (Interchim, Montluçon, France); levocabastine hydrochloride and noberastine dimaleate (Janssen Research, Beerse, Belgium); cetirizine dihydrochloride (UCB Pharma, Braine L'Alleud, Belgium); abecarnil (Schering AG, Berlin, Germany); citalopram (Lundbeck, Copenhagen, Denmark); gepirone hydrochloride (Bristol Myers Squibb, Wallingford, CT); ipsapirone hydrochloride (Troponwerke, Cologne, Germany); LY-228729 (Eli Lilly and Co., Indianapolis, IN); methylphenidate hydrochloride (Ciba-Geigy Co., Basel, Switzerland); tandospirone citrate (metanopirone; Sumitomo, Osaka, Japan); flesinoxan hydrochloride, lesopitron dihydrochloride, paroxetine hydrochloride, S 14506, WAY-100635 dihydrochloride and WY-50324 hydrochloride (adatanserin) (synthesized by J.-L. Maurel, Centre de Recherche Pierre Fabre).

	Results

Top Abstract Introduction Methods Results Discussion References

Radioligand Binding

F 11440 had nanomolar affinity for 5-HT_1A binding sites (K_i: 4.8 nM, pK_i, 8.33); its affinity for these sites was ~7-fold higher than that of buspirone (pK_i, 7.50), and ~4-fold lower than that of flesinoxan (pK_i, 8.91). F 11440 was more selective for 5-HT_1A sites than either buspirone or flesinoxan (fig. 2). Whereas the affinity of F 11440 for 5-HT_1A sites was 100-fold higher than its affinity for the other serotonin receptor subtypes examined here [i.e., h 5-HT_1B, h 5-HT_1D, 5-HT_2A and 5-HT_2C: pK_i < 5.0), for dopamine D₂ receptors (pK_i, 5.75), for alpha-1 noradrenergic receptors (pK_i, 6.16), and for histamine H₁ receptors (pK_i < 5.0)], this was not the case for buspirone or for flesinoxan. Buspirone had relatively high affinity for dopamine D₂ sites (pK_i, 7.43), and its affinity for alpha-1 adrenergic and H₁ histamine receptors (pK_i, 6.04 and 6.41, respectively) was <100-fold lower than its 5-HT_1A affinity. Compared with buspirone, flesinoxan was more 5-HT_1A selective with respect to alpha-1 adrenergic and histamine H₁ receptors (pK_i, 6.50 and 6.75, respectively). Compared with F 11440, however, flesinoxan was less selective with respect to h 5-HT_1D, h 5-HT_1B, D₂ (pK_i, 7.86, 7.19 and 7.05, respectively) and H₁ receptors.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Affinity of F 11440, flesinoxan and buspirone for 5-HT1A and other binding sites (i.e., alpha-1-adrenergic, H1 histaminergic, h 5-HT1D and h 5-HT1B serotonergic, D2 dopaminergic). The solid bar indicates the affinity, expressed as pKi value, for 5-HT1A binding sites; the concentration that is 100-fold higher than this pKi value is indicated for each compound by a vertical dashed line.

cAMP in HA7 Cells

In cells transfected with human 5-HT_1A receptors, F 11440 inhibited forskolin-induced stimulation of cAMP with a pEC₅₀ value of 6.80 ± 0.11 (mean ± S.E.M.) (fig. 3, top). The pEC₅₀ values obtained with other compounds that exerted 5-HT_1A agonist activity [i.e, 5-HT (7.48 ± 0.19), 8-OH-DPAT (7.59 ± 0.22), buspirone (6.70 ± 0.17), ipsapirone (7.34 ± 0.24), flesinoxan (7.69 ± 0.18), methysergide (5.61 ± 0.06) and tryptamine (5.36 ± 0.01) (not shown)] correlated positively (r = .96, P < .001) with values reported elsewhere (Pauwels et al., 1993).

View larger version (12K): [in this window] [in a new window]   Fig. 3.   Inhibition of forskolin-stimulated cAMP production in HA7 cells stably expressing human 5-HT1A receptors by F 11440 (top; first series of experiments), by F 11440, the 5-HT1A receptor agonists, ipsapirone, buspirone, 8-OH-DPAT and flesinoxan, and by 5-HT, each tested at a concentration of 105 M (middle; second series of experiments), and by F 11440 (104 M) and flesinoxan (104 M) in comparison with the effects obtained with 5-HT (105 M) (bottom; third series of experiments). Values are means ± SEM of four independent determinations (top) or of nine independent determinations (middle and bottom), each performed in triplicate. Within the middle and bottom, means represented by bars with a different fill pattern differ significantly (P < .05); means represented by bars with the same fill pattern are not significantly different.

The maximal inhibition of forskolin-induced stimulation of cAMP produced by a compound in HA7 cells offers one measure of its intrinsic activity at 5-HT_1A receptors. The partial 5-HT_1A agonists buspirone and ipsapirone inhibited forskolin-induced stimulation, but only to ~60% of control, whereas the full agonist 5-HT inhibited the effects of forskolin more extensively, to ~10% of control (fig. 3, middle). At a concentration of 10⁵ M, F 11440 inhibited forskolin-stimulated cAMP to a significantly greater extent than buspirone and ipsapirone. At a concentration of 10⁴ M, F 11440 inhibited cAMP levels to 9.4 ± 1.4% (mean ± S.E.M.) of control, a value that was significantly lower than that obtained with 10⁴ M flesinoxan (i.e., 17 ± 2.2%) and not significantly different from the value obtained with 5-HT (i.e., 10.7 ± 1.8%) (fig. 3, bottom).

Expressing the mean (n = 3-9) maximal inhibition of forskolin-stimulated cAMP production observed with each compound tested in the pigeon conflict procedure (see below) as a proportion of the maximal effect produced by 5-HT, the following values were obtained: F 11440: 1.0, S 14506: 0.95, flesinoxan: 0.93, LY-228729: 0.88, tandospirone: 0.87, 8-OH-DPAT: 0.81, WY 50,324: 0.79, gepirone: 0.77, lesopitron: 0.70, buspirone: 0.49, ipsapirone: 0.46, BMY 7378: 0.0 and NAN 190: 0.0. The intrinsic activities of flesinoxan, 8-OH-DPAT, gepirone, buspirone, and ipsapirone, correlated positively (r = .76, r_s = .90) with those reported previously in Chinese hamster ovary cells expressing the human 5-HT_1A receptor (McCall et al., 1994).

In Vivo Microdialysis

The mean basal extracellular level of 5-HT in ventral hippocampus was 34 ± 1.2 fmol/20 µl (n = 55). F 11440, administered i.p., dose-dependently decreased the extracellular level of 5-HT (fig. 4). Its ED₅₀ value, calculated from the percent AUC data, was 0.049 mg/kg. F 11440 was ~30-fold more potent than buspirone, and ~4-fold more potent than flesinoxan (ED₅₀ values: 1.6 and 0.23 mg/kg, buspirone and flesinoxan, respectively; its maximal effect (AUC = 34%) was similar to that of buspirone (37%) and flesinoxan (37%) (data not shown). The effects of 0.16 mg/kg F 11440 were blocked completely by the 5-HT_1A antagonist, WAY-100635 [ED₅₀: 0.021 mg/kg; fig. (4].

View larger version (30K): [in this window] [in a new window]   Fig. 4.   Effects of F 11440 alone (top), and together with WAY-100635 (bottom), on extracellular 5-HT levels in microdialysates from the ventral hippocampus of chloralhydrate-anesthetized rats. The mean basal extracellular level of 5-HT was 34 ± 1.2 fmol/20 µl. Data shown are mean ± SEM for 5 animals per group. Top left, 5-HT levels are expressed as a percentage of the mean absolute amount of 5-HT in the four samples collected before the injection of saline (s.c.) at t-40 min; () vehicle, () 0.01 mg/kg, () 0.04 mg/kg, () 0.16 mg/kg, or () 0.63 mg/kg F 11440 were administered i.p, at t = 0 min. Top right, dose response curve for the inhibitory effect of F 11440 expressed as the mean percent AUC (area under the curve) for the 140 min post-injection period. Significant differences from dose 0 are indicated as ** P < .01 (one way analysis of variance followed by Dunnett's test). Bottom, ( or dotted line) vehicle, ( or dashed line) 0.16 mg/kg F 11440, () WAY-100635, () WAY-100635 + 0.16 mg/kg F 11440. Saline or WAY-100635 were administered s.c, at t-40 min, and vehicle or F 11440 were administered i.p, 40 min later, at t = 0 min.

Stimulation of Corticosterone Secretion

In vehicle control rats (n = 175, i.p.; n = 150, p.o.), the mean plasma corticosterone level was 73 ± 10 (mean ± S.E.M.) ng/ml after i.p. administration and 45 ± 9.1 ng/ml after p.o. administration. Levels >410 ng/ml occurred in <5% of the i.p. controls, and levels >250 ng/ml occurred in <5% of the p.o. controls. Drugs were therefore considered to produce an effect in an individual animal if the observed corticosterone level was >410 ng/ml when administered i.p., or >250 ng/ml when administered p.o.

F 11440 dose-dependently increased corticosterone levels (fig. 5), and the ED₅₀ value to produce these effects when administered p.o. (i.e., 0.16 mg/kg) was ~3-fold higher than the ED₅₀ obtained after i.p. injection (i.e., 0.057 mg/kg). When administered i.p., F 11440 was ~60-fold more potent than buspirone (ED₅₀: 3.4 mg/kg) and ~20-fold more potent than flesinoxan (ED₅₀: 1.2 mg/kg); its maximal effect (680 ± 41 ng/ml) was similar to that of buspirone (610 ± 59 ng/ml) and flesinoxan (540 ± 38 ng/ml) (data not shown). The effects of 0.16 mg/kg F 11440 (i.p.) on corticosterone levels were blocked completely by WAY-100635 (ED₅₀: 0.0027 mg/kg; fig. (5); this dose of F 11440 also produced LLR (see below), which was similarly blocked by WAY-100635 (ED₅₀: 0.0025 mg/kg; data not shown).

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Ability of F 11440 to increase plasma corticosterone levels in rats after i.p. () and after p.o. administration (), and its antagonism by the 5-HT1A antagonist, WAY-100635. The results are expressed as the percentage of rats in which corticosterone levels were higher than a criterion value (see text) that was met in <5% of vehicle control animals. F 11440 dose-dependently increased the percentage of animals with significantly elevated corticosterone levels, both after i.p. (t-30 min) and after p.o. (t-60 min) administration (ED50 values: 0.057 and 0.16 mg/kg, respectively; left panel). WAY-100635 (administered s.c, t-60 min) dose-dependently antagonized (ED50: 0.0027 mg/kg) the increase of corticosterone levels produced by the i.p. administration of 0.16 mg/kg F 11440 (right).

LLR and Elements of the 5-HT Syndrome

An LLR or FPT score of 1 or more or the presence of FBP [scored as (1)] each occurred in <5% of the vehicle control animals (n = 225, i.p.) during the observation period centered at 15 min after the administration of vehicle. Drugs were therefore considered to produce an effect in an individual animal when scores higher than zero were obtained.

After i.p. administration in rats, F 11440 produced LLR and, at higher doses, FPT and FBP (table 2). This profile was similar to that of 5-HT_1A agonists, such as 8-OH-DPAT, but differed from that of partial 5-HT_1A agonists (e.g., buspirone) and putative 5-HT_1A antagonists (e.g., NAN-190), which did not produce FPT, and from that of 5-HT agonists lacking substantial 5-HT_1A affinity (i.e., DOI, mCPP, TFMPP and quipazine), which did not produce LLR (data not shown). F 11440 was ~5 times more potent than flesinoxan, and ~18 times more potent than buspirone, in producing LLR.

At various intervals after its administration, F 11440 produced LLR with a p.o. potency that was similar to its i.v. potency (data not shown). A comparison of the ED₅₀ values after p.o. and i.v. administration at 60 min (i.e., 0.26 and 0.30 mg/kg, respectively) and at 120 min (i.e., 0.79 and 1.4 mg/kg, respectively) indicated that these values did not differ by more than a factor of 1.8.

DS Effects of 8-OH-DPAT

The discrimination of 8-OH-DPAT (0.16 mg/kg) from saline was acquired within 100 sessions by 23 of the 27 rats (median sessions to criterion: 32; range: 17-88). Administration of the training drug under test conditions produced dose-related increases in drug-appropriate responding (ED₅₀: 0.065 mg/kg, table 3) that reached a 100% level of effect at the training dose. A significant decrease of the overall rate of responding (79% of control) was produced only by the training dose.

F 11440 produced dose-dependent responding on the 8-OH-DPAT lever (ED₅₀: 0.021 m/kg) that reached a 100% level of effect at a dose of 0.08 mg/kg. Only a higher dose (i.e., 0.16 mg/kg) decreased significantly the overall rate of responding (55% of control). F 11440 was ~30-fold more potent than flesinoxan and buspirone to produce 8-OH-DPAT lever selection. In contrast to F 11440, flesinoxan and buspirone, which produced 8-OH-DPAT lever selection in 89% of the animals, chlordiazepoxide and ethanol produced DL selection in no more than 67% and 50% of the animals, respectively, at doses that decreased the overall rate of responding.

Duration of Action

F 11440 produced LLR at each of the pretreatment intervals (fig. 6). The lowest ED₅₀ value (0.085 mg/kg) was observed at the first observation time, whereupon it increased to 19 mg/kg at 8 hrs after administration. Based on the time during which the ED₅₀ was <4 times the lowest ED₅₀, the duration of action of F 11440 was estimated to be 1.3 hr, similar to that of flesinoxan and buspirone, and shorter than the other 5-HT_1A agonists shown.

View larger version (28K): [in this window] [in a new window]   Fig. 6.   Potency of F 11440 and prototypical 5-HT1A agonists in producing LLR at various times after oral administration. Values shown are ED50 values (mg/kg) calculated for each observation time point. Closed symbols represent ED50 values that are larger than the highest dose tested (40 mg/kg). Shown in parentheses are the estimated durations of action based on the time during which the ED50 was <4 times the lowest ED50 value.

In Vivo Selectivity

Alpha-1 adrenoceptor antagonist activity. Inhibition of the increase in blood pressure induced in anesthetized rats by the i.v. administration of the alpha-1-adrenoceptor agonist, methoxamine, was used as an in vivo measure of peripheral alpha-1 antagonist properties. Methoxamine (0.16 mg/kg, i.v.) increased mean arterial pressure from a base-line value of 115 ± 5 mm Hg to a peak value of 175 ± 3 mm Hg (n = 18). Prazosin, 5-methyl-urapidil and urapidil dose-dependently antagonized the increase in mean arterial pressure induced by methoxamine, with ID₅₀ values of 0.16, 0.36 and 1.0 mg/kg i.p., respectively. According to the measure used here, F 11440 exerted alpha-1 antagonist activity in vivo, but only at doses that were ~70-fold higher than those that produced 5-HT_1A agonist effects (e.g., increases in corticosterone levels; fig. 7). In contrast, flesinoxan, although it had high selectivity for 5-HT_1A receptors in vitro, exerted alpha-1 antagonist activity in vivo at doses <3-fold higher than those needed to produce 5-HT_1A agonist effects.

View larger version (19K): [in this window] [in a new window]   Fig. 7.   Ability of F 11440 (top) and flesinoxan (bottom) to exert 5-HT1A agonist activity (increase of corticosterone levels; ) and alpha-1 adrenergic antagonist activity (inhibition of the pressor response induced by methoxamine; ) after i.p. administration in rats. The figure inserts show in vitro (ratio of pKi values) and in vivo (ratio of ED50 values) measures of the selectivity for 5-HT1A as compared with alpha-1-adrenergic receptors. Corticosterone data obtained with F 11440 are replotted from fig. 5.

Dopamine antagonist activity. A total of 140 animals received a vehicle control treatment via either the s.c. or i.p. route, 30 min before the injection of methylphenidate (40 mg/kg i.p.). Gnawing was prominent after methylphenidate administration (9.7 ± 0.02; mean ± S.E.M.); sniffing, rearing and locomotion, although increased at doses from 0.63 to 10 mg/kg (Koek and Colpaert, 1993), did not occur after administration of the 40 mg/kg dose. Gnawing scores of 9 or 10 were observed in >95% of the animals treated with the 40 mg/kg dose of methylphenidate. Thus, a pretreatment was considered to have decreased the incidence of gnawing in an individual animal when a score <9 was observed. In control animals not receiving methylphenidate (n = 104), gnawing, locomotion, rearing, sniffing and licking each appeared with an average incidence of <1, and the scores obtained in <5% of the animals were as follows: gnawing (1), locomotion (5), sniffing (9), rearing (7), and licking (1). Test compounds were therefore considered to have normalized methylphenidate-induced behaviors in individual animals if the following composite criterion was met: gnawing < 2, locomotion < 6, sniffing < 10, rearing < 8, and licking < 2.

Histamine antagonist activity. The average preconvulsion time in guinea pigs exposed to a histamine aerosol after being pretreated with vehicle control (n = 66) was 47 ± 0.5 sec, and 95% of these controls showed preconvulsion times shorter than 55 sec. Thus, a pretreatment was considered to protect an individual animal if the preconvulsion time was longer than 55 sec. In addition, two other levels of protection from the effects of histamine were defined: (1) preconvulsion time longer than 70 sec (i.e., prolongation of the mean control preconvulsion time by >50%), and (2) preconvulsion time longer than 180 sec (i.e., protection from the effects of histamine during the entire 3 min exposure period).

View larger version (21K): [in this window] [in a new window]   Fig. 8.   Ability of F 11440 (top) and flesinoxan (bottom) to exert 5-HT1A agonist activity (increase of corticosterone levels in rats; ) and histamine H1 antagonist activity (protection from histamine-induced preconvulsions in guinea pigs; other symbols, one for each duration of protection) after p.o. administration. The figure inserts show in vitro (ratio of pKi values) and in vivo (ratio of ED50 values) measures of the selectivity for 5-HT1A as compared with histamine H1 receptors. Corticosterone data obtained with F 11440 are replotted from figure 5.

DS effects of chlordiazepoxide and ethanol. The discrimination between chlordiazepoxide (5 mg/kg) and saline was acquired within 100 sessions by all 29 rats trained (median sessions to criterion: 26; range: 14-60). Chlordiazepoxide produced dose-dependent responding on the DL [ED₅₀: 1.3 mg/kg; table (3)] that reached a 100% level of effect. At doses ~7-fold higher than those that produced 8-OH-DPAT-like DS effects, F 11440 produced DL selection (ED₅₀: 0.14 mg/kg), but with a maximal effect of 60% at a dose (0.16 mg/kg) at which lever selection occurred in only five of the seven animals tested and that was higher than the lowest dose (0.08 mg/kg) that significantly decreased the overall rate of responding. Ethanol and 8-OH-DPAT produced DL selection in a maximum of only 14 and 25% of the animals, respectively. Buspirone and flesinoxan did not produce DL selection.

Pigeon Conflict Procedure

For the 58 pigeons used in this study, the median shock intensity was 4.5 mA (range: 2-5 mA) and body weights were maintained at an average of 79 ± 1% of control (mean ± S.E.M.). Mean rates of responding under the different components of the multiple schedule before testing began were 2.19 ± 0.06 and 0.07 ± 0.01 responses/sec, unpunished and punished responding, respectively. Thus, under control conditions, response rates during the punished components were suppressed by >90% relative to shock-free response rates. Among the groups tested in the present experiment, mean control rates of punished responding ranged from 0.003 ± 0.001 to 0.07 ± 0.022 responses/sec and the number of control punished responses emitted in individual pigeons ranged from 0 to 247.

The prototypical 5-HT_1A agonists as well as the benzodiazepines, chlordiazepoxide and abecarnil, produced dose-related percentage increases in punished responding (fig. 9), and did so, in general, along inverted U-shaped dose-response functions, with the descending limb corresponding to the onset of effects on unpunished responding. F 11440 produced significant and substantial increases of punished responding over a wide range of doses that did not significantly affect the rate of unpunished responding. In comparison with other 5-HT_1A agonists (fig. 9, table 4), F 11440 was (1) among the most potent compounds to increase punished responding, (2) produced its effects on punished responding over a range of doses that was wider than that of all other compounds tested, except 8-OH-DPAT, (3) had maximal effects on punished responding similar to those of 8-OH-DPAT, but higher than all other compounds, (4) increased punished responding in a percentage of pigeons higher than any of the other 5-HT_1A agonists, except 8-OH-DPAT, and (5) exerted its effects on punished responding at doses lower than those affecting unpunished responding. In particular, the antipunishment effects of F 11440 were more substantial than those of buspirone, ipsapirone and flesinoxan, which significantly increased punished responding at best at only one dose, with maximal effects lower than those of F 11440 and at doses similar to those that decreased unpunished responding.

View larger version (36K): [in this window] [in a new window]   Fig. 9.   Effects of F 11440, 5-HT1A agonists, and the benzodiazepines, chlordiazepoxide and abecarnil, on punished and unpunished responding in the pigeon. For all drugs, n = 7 per dose. Values represent the mean ± SEM percentage change from control sessions conducted immediately before the test session. Where absent, error bars are contained within the symbol. Closed symbols: punished responding; Open symbols: unpunished responding. *Values significantly (P < .05) higher than (punished responding) or different from (unpunished responding) performance under control conditions.

The potencies with which the 5-HT_1A agonists increased punished responding in pigeons were correlated positively (r = .7, P < .05; fig. 10) with their ability to produce lower lip retraction in rats. The magnitude of the effects on punished responding observed here with the 5-HT_1A agonists, 8-OH-DPAT, S 14506, flesinoxan, buspirone BMY 7378, and NAN-190, correlated positively (r = .81, P < .05; data not shown) with the magnitude of their antipunishment effects in pigeons reported previously (Colpaert et al., 1992). Further, the maximal effects on punished responding of all 5-HT_1A agonists examined here correlated positively (r = .74, P < .005; r_s = .78, P < .005; fig. 10) with their intrinsic activity at 5-HT_1A receptors, which was measured by expressing the mean maximal inhibition of forskolin-stimulated cAMP production observed with each compound as a proportion of the maximal effect produced by 5-HT.

View larger version (15K): [in this window] [in a new window]   Fig. 10.   Ability of F 11440 and various other 5-HT1A agonists to increase punished responding in pigeons: relationships with other effects. Top, relationship between potency to produce LLR in rats and potency to increase punished responding. Bottom, relationship between intrinsic activity at 5-HT1A receptors (maximal inhibition of forskolin-stimulated cAMP production in HA7 cells, expressed as a proportion of the maximal effect of 5-HT) and anxiolytic-like activity (maximum effect on punished responding). "r" represents Pearson's product-moment correlation coefficient.

Forced Swimming Test

The results obtained with F 11440, flesinoxan, ipsapirone, buspirone, imipramine and paroxetine are summarized in figure 11 and are expressed as percentage inhibition of the median immobility time observed in control animals (control medians ranged from 140 to 162 sec).

View larger version (21K): [in this window] [in a new window]   Fig. 11.   Ability of F 11440 to produce antidepressant-like effects (inhibition of immobility in the forced swimming or "behavioral despair" test in rats) in comparison with effects of other 5-HT1A agonists (top), and the antidepressants, imipramine (middle) and paroxetine (bottom). F 11440 and the other 5-HT1A agonists were administered once, 60 min before testing, whereas imipramine and paroxetine were administered once, at 60 min, or three times, at 24 hr, 4 hr and 60 min before testing. All compounds were administered p.o. *Significantly (alpha = .05) different from vehicle control values.

F 11440, administered once, p.o., 60 min before the test session, potently inhibited immobility over a range of doses, and the apparent maximal effect was higher than that of flesinoxan, ipsapirone, and buspirone. F 11440 significantly inhibited immobility at all five doses tested. In contrast, flesinoxan and ipsapirone significantly inhibited immobility at no more than two doses. Buspirone did not decrease immobility, but actually increased it.

Both the classical tricyclic antidepressant, imipramine, and the more recently described selective serotonin reuptake inhibiting antidepressant, paroxetine, were able to inhibit immobility. Their ability to inhibit immobility was statistically significant, yet moderate, after a single administration, but was more pronounced after three administrations, at 24, 4 and 1 hr before the test session. When administered repeatedly, imipramine inhibited immobility at doses ranging from 5 to 80 mg/kg. The ability of higher doses to inhibit immobility could not be examined, because the repeated administration of 320 mg/kg imipramine had lethal effects. Paroxetine, when administered repeatedly, dose-dependently and significantly inhibited immobility, and attained its maximum effect at a dose of 80 mg/kg. Repeated administration of 320 mg/kg (i.e., 4-fold higher than the dose of paroxetine that produced the maximum effect after one administration) had lethal effects. Only paroxetine was able to produce antidepressant-like effects of a magnitude similar to that produced by a single administration of F 11440 but only when paroxetine was administered repeatedly.

	Discussion

Top Abstract Introduction Methods Results Discussion References

F 11440 is a potent, selective and efficacious 5-HT_1A receptor agonist. Its affinity for 5-HT_1A binding sites was higher than that of the partial agonist, buspirone, which is clinically available as an anxiolytic, and somewhat lower than that of the reportedly higher-efficacy 5-HT_1A agonist, flesinoxan, which is in phase III of clinical development as an anxiolytic/antidepressant. In vivo, F 11440 was ~4 to 10 times more potent than flesinoxan, and nearly 100 times more potent than buspirone, in exerting 5-HT_1A agonist activity after oral administration. F 11440 did not have detectable antidopaminergic or antihistaminergic activity (unlike buspirone and flesinoxan, respectively) and had a 70-fold separation between its 5-HT_1A agonist and alpha-1 antagonist properties in vivo, unlike flesinoxan, which showed a <3-fold separation. F 11440 had intrinsic activity at human 5-HT_1A receptors higher than that of buspirone, ipsapirone, and flesinoxan. Thus, F 11440 appears to be a more potent, a more selective and a more efficacious 5-HT_1A receptor agonist than buspirone and flesinoxan.

F 11440 is the outcome of a research effort guided by the hypothesis that high intrinsic activity is necessary for 5-HT_1A agonists to produce large anxiolytic and antidepressant effects (Colpaert et al., 1992; De Vry, 1996). This hypothesis was supported by the finding that intrinsic activity at 5-HT_1A receptors correlated positively with the magnitude of anxiolytic-like effects and the finding of a positive relation between intrinsic activity at 5-HT_1A receptors and the magnitude of antidepressant-like effects. Indeed, F 11440, which showed higher intrinsic activity at 5-HT_1A receptors than buspirone, ipsapirone and flesinoxan, exerted, in the animal models used here more substantial anxiolytic- and antidepressant-like effects than buspirone, ipsapirone and flesinoxan. Furthermore, the only compound able to produce antidepressant-like effects of a magnitude similar to that produced by a single administration of F 11440 was paroxetine, but p