Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

During the last 15 yr, considerable evidence has accumulated which suggests that 5-HT_1A receptor ligands may have a broad spectrum of therapeutic benefits (for reviews, see De Vry, 1995; Ennis, 1993). Initial clinical experience was collected with the pyrimidinylpiperazine derivatives buspirone, ipsapirone, gepirone and tandospirone, and indicated that such compounds have mixed anxiolytic/antidepressive properties (De Vry et al., 1991, 1992; Pecknold, 1994; Stahl et al., 1992; Traber and Glaser, 1987). However, in addition to therapeutic applications in the field of psychiatry, more recent preclinical studies have suggested that 5-HT_1A receptor agonists have also pronounced neuroprotective properties, and thus may have additional potential in the field of neurology (e.g., Bielenberg and Burkhardt, 1990; Bode-Greuel et al., 1990; Prehn et al., 1991; for additional references, see De Vry et al., 1997a).

Studies aimed at elucidating the mechanism of action underlying the therapeutic effects of 5-HT_1A receptor ligands have suggested that the level of intrinsic activity, as well as the relative efficacy of these compounds at different 5-HT_1A receptor populations determines their particular therapeutic profile (De Vry, 1995). Typically, 5-HT_1A receptor populations are subdivided in presynaptic or somatodendritic receptors and postsynaptic receptors, and the pyrimidinylpiperazine derivatives have been characterized as full agonists at the former receptor population, and partial agonists or antagonists at the latter receptor populations (for reviews, see Glaser et al., 1991; Glaser and De Vry, 1992). Therefore, the pyrimidinylpiperazine derivatives have been referred to as being mixed agonist/antagonists, or partial agonists at 5-HT_1A receptors. The level of intrinsic activity of the pyrimidinyl piperazines appears to be situated between that of the aminotetralin derivative 8-OH-DPAT (Arvidsson et al., 1981), which is generally considered as a 5-HT_1A receptor full agonist, and that of the "silent" antagonist WAY-100635 (Forster et al., 1995). Based on extensive preclinical studies, it was hypothesized that the therapeutic spectrum of a 5-HT_1A receptor ligand could be optimized by increasing the apparent intrinsic activity at pre- and postsynaptic 5-HT_1A receptor populations (De Vry, 1995, 1996).

Besides the prototypical 8-OH-DPAT, a number of compounds with a level of intrinsic activity at 5-HT_1A receptors higher than that of the pyrimidinylpiperazine derivatives have been described in recent years. Among these are the aminotetralin derivative LY274601 (Foreman et al., 1995), the phenylpiperazine derivative flesinoxan (Schipper et al., 1991), the naphthylpiperazine derivative S 14671 (Millan et al., 1992), the benzodioxan derivatives MDL 72832 (Mir et al., 1988) and MKC-242 (Abe et al., 1996), the chroman derivative S 20499 (Kidd et al., 1993), and the tetrahydrobenzindole derivatives BAY r 1531 (Glaser et al., 1987), LY228729 (Foreman et al., 1993), LY293284 (Foreman et al., 1994) and U-92016A (McCall et al., 1994). However, several of these compounds have not been developed clinically due to inappropriate pharmacokinetics. Moreover, although the level of intrinsic activity at 5-HT_1A receptors appears to be higher than that of the pyrimidinylpiperazine derivatives, it is likely that for some of these compounds the level is still considerably less than that of 5-HT itself and therefore, such compounds should still be considered as partial agonists.

Recently, the novel aminomethylchroman derivative BAY × 3702 (fig. 1) was described as a high affinity 5-HT_1A receptor agonist with pronounced neuroprotective properties (De Vry et al., 1997a, b; Horváth and Augstein, 1997; Horváth et al., 1997). We report on the receptor binding profile of BAY × 3702 and describe the effects of the compound in a number of biochemical, electrophysiological, behavioral and endocrinological models of 5-HT_1A receptor function. Functional models included models of somatodendritic receptor activation, such as 5-HT neuronal firing activity in brain slices containing the DRN (Greuel and Glaser, 1992; Schechter et al., 1990; VanderMaelen and Aghajanian, 1983), or recorded in vivo (Jolas et al., 1995; Sprouse and Aghajanian, 1987; Sommermeyer et al., 1993b) and 8-OH-DPAT drug discrimination (Glennon, 1986; Schreiber and De Vry, 1993; Schreiber et al., 1995), and models of postsynaptic receptor activity, such as the adenylate cyclase assay in hippocampal membranes (De Vivo and Maayani, 1986; De Vry et al., 1991; Hamon et al., 1988) and stimulation of plasma ACTH levels (Critchley et al., 1994; Di Sciullo et al., 1990; Gilbert et al., 1988; Pan and Gilbert, 1992); as well as models that may reflect activation of multiple receptor populations, such as hypothermia (Glaser et al., 1991; but, see Higgins et al., 1988; Hjorth, 1985; Millan et al., 1993; for extensive discussion on models of pre- and postsynaptic function, see De Vry et al., 1991; Glaser et al., 1991). In general, effects of BAY × 3702 were compared with those induced by the reference compounds 8-OH-DPAT (Arvidsson et al., 1981) and ipsapirone (Traber et al., 1984), and involvement of 5-HT_1A receptors in the effect was ascertained by antagonism experiments with WAY-100635 (Forster et al., 1995).

View larger version (7K): [in this window] [in a new window]   Fig. 1.   Structure of BAY × 3702.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Animals. Male Wistar rats weighing 150 to 250 g were obtained from Winkelmann (Borchen, FRG). Animals were group-housed (except in the case of drug discrimination, where they were individually housed) in standard Macrolon cages for at least 1 wk before the start of the experiments. Room temperature was maintained at 21.5 ± 1.5°C and lights were on from 7:00 A.M. to 7:00 P.M.. A standard laboratory food diet and tap water were available ad libitum; except in the case of drug discrimination, where the rats were maintained at about 80% of their free-feeding weight by restricting their daily diet to 13 to 15 g. Experimental protocols and conditions were conform with the local regulations on animal welfare.

Receptor binding. The methodologies used for measuring binding to various neurotransmitter receptors are summarized in table 1. For membrane preparations fresh calf brains from a local slaughterhouse trimmed free of blood vessels and meninges and brains of decapitated Wistar rats were used; whereas human cortex membranes were obtained from ABS (Analytical Biological Services Inc., Wilmington, DE). Regions of interest were dissected out, homogenized in 10 volumes of 0.32 M sucrose and centrifuged for 10 min at 700 × g. The supernatant was centrifuged at 43,500 × g for 10 min and the pellet resuspended in 50 mM Tris-HCl (pH: 7.7, 25°C) by a 10-sec polytron treatment. Aliquots of the membrane preparations were stored at 140°C over liquid nitrogen. To remove endogenous serotonin, brain membranes were incubated at 37°C for 10 min before the experiment. Assays were terminated by rapid filtration over Whatman GF/C filters using a Brandel cell harvester. K_i values were calculated according to the Cheng-Prusoff equation from the respective IC₅₀ values obtained from dose-response curves with at least six concentrations. Those receptors for which no K_i values are given were assayed at Panlabs Inc., Bothell, WA.

5-HT_1A- and 5-HT_1D receptor coupled adenylate cyclase assay in membranes. 5-HT_1A- and 5-HT_1D receptor coupled adenylate cyclase activity was measured in rat hippocampal or calf substantia nigra membranes, respectively, according to the methods described by De Vivo and Maayani (1986) and Schoeffter et al. (1988). Hippocampi or substantia nigra tissue were homogenized in 25 volumes of 0.3 M sucrose containing (in mM) EGTA (1), EDTA (5), dithiothreitol (5) and Tris-HCl (20), (pH: 7.4, 25°C) in a motor-driven Potter-Elvehjem homogenizer. The homogenate was centrifuged for 10 min at 1,000 × g and the supernatant was centrifuged at 39,000 × g for 10 min. The pellet was resuspended in homogenization buffer at a protein concentration of about 1 mg/ml and aliquots were stored at 140°C. Before use the membranes were rehomogenized in a Potter-Elvehjem homogenizer. A total of 50 µl of the membrane suspension was added to the incubation mixture containing (in mM) NaCl (100), magnesium acetate (2), ATP (0.2), cAMP (1), GTP (0.01), forskolin (0.01), Tris-HCl (80), creatine phosphate (5), 0.8 U/µl creatine phosphokinase, IBMX (0.1), 1-2 µCi -[³²P]ATP, test compounds in various concentrations and approximately 50 µg membrane protein. Incubations (10 min at 30°C) were started by adding the membrane solutions to the prewarmed (5 min at 30°C) incubation mixtures. The formed [³²P]cAMP was determined according to Salomon (1979). Protein was measured as described by Bradford (1976).

Dopamine D₂- and alpha₂-adrenergic receptor-linked cAMP formation in cultured cell lines. The mouse fibroblast cell line LZR-1 expressing rat dopamine D₂ receptors was obtained from Dr. O. Civelli; whereas alpha₂-adrenergic receptor containing NG108-15 neuroblastoma x glioma hybrid cells were obtained from Dr. B. Hamprecht. Cell culture and cAMP measurements were performed as described (Sommermeyer et al., 1993a; Traber et al., 1975). For measuring cAMP levels the cells were washed twice with incubation buffer, containing (in mM) HEPES (20), NaCl (145), KCl (5.4), CaCl (1.8), MgCl (1), glucose (20), pH: 7.4 and preincubated for 10 min at 37°C. After addition of effectors and test compounds cells were incubated for 30 min. Incubations were terminated by removing the incubation medium followed by addition of 0.5 ml ethanol (96% v/v). Supernatants were transferred to test tubes, the ethanol was evaporated, and the residues were dissolved in H₂O. Determination of cAMP was performed according to Gilman (1970) using a protein binding assay.

Alpha₁-adrenergic receptor-linked phosphoinositide breakdown. Accumulation of InsP_n was measured in the presence of LiCl essentially as described by Sommermeyer et al. (1990) in cultured brain astrocytes. Cells were prelabeled with 2 µCi/ml [³H]myo-inositol for 24 hr in serum-free DMEM medium. One hour before incubation with noradrenaline and test compounds LiCl in a final concentration of 10 mM was added. The 30-min incubations were terminated by pouring 1.25 ml of ice-cold methanol directly onto the cells. Cells were scraped off, added to 1.25 ml chloroform containing tubes and kept over night at 4°C. The upper phase was applied to AG-1X8 (Bio Rad, Munich, FRG) columns. The lower chloroform-phase was washed once with 500 µl of the upper phase from a mixture containing chloroform/methanol/100 mM sodium cyclohexan-1,2-diamintetraacetate 16/8/5 (v/v/v). The collected upper phases were applied to the corresponding columns. After washing with 10 ml 5 mM disodium tetraborate, 60 mM sodium formate total InsP_n were eluted with 6 ml of 100 mM formic acid, 1 mM ammonium formate and radioactivity was determined by liquid scintillation counting.

Alpha₂-adrenergic receptor-regulated 5-HT release in rat brain slices. Superfusion experiments with rat occipitoparietal cortex slices were performed according to Goethert et al. (1981). Slices were incubated in Krebs-Ringer solution containing 0.1 µM [³H]5-HT for 60 min, transferred to superfusion chambers and superfused for 155 min at a rate of 1.1 ml/min. The superfusates were collected in 5-min samples from the 60th min on. At the end of the superfusion the slices were solubilized with tissue solubilizer (Bilute S, Zinsser, FRG), and the radioactivity in the slices and in the collected fraction was determined. Each slice was electrically stimulated (20 mA, 2 msec, 3 Hz) for 2 min 70 (S₁) and 115 min (S₂) after onset of the superfusion. Drugs were present throughout the superfusion or were added 85 min after onset. Stimulation evoked [³H]-overflow was expressed as the ratio of the overflow evoked by S₂ to that evoked by S₁.

In vitro electrophysiological recordings in DRN. In general the methodology described by Greuel and Glaser (1992) was followed. After decapitation under light ether anaesthesia, the brainstems of male Wistar rats (100-120 g) were dissected and mounted on a vibratome chuck for cutting 350-µm slices. During cutting, the brainstem was submerged in cold (1-3°C) carbogenated buffer. The slices were then transferred into a submersion-type slice chamber and allowed to recover for at least 60 min. The ACSF contained (in mM): NaCl (124), KCl (5), CaCl₂ (2), NaH₂PO₄ (1.25), MgSO₄ (2), NaHCO₃ (26), D-glucose (10); pH: 7.4. (L)-Noradrenaline (5 µM) was added to the ACSF to activate the otherwise silent neurons (VanderMaelen and Aghajanian, 1983). Conventional extracellular recording techniques were used. Briefly, tungsten electrodes (impedance: 12 M) were advanced by means of a hand-driven Leitz micromanipulator into the tissue until the activity of a neuron was detected on the oscilloscope. Electrode signals were passed through a high impedance probe, amplified by a factor of 1000 and bandpass filtered at 1 to 3 kHz (3 dB). The signals were monitored on an oscilloscope and audiosystem, and finally fed through a window discriminator into an A/D converter (Coulbourn Instruments, Allentown, PA) of a PC-based digital system. Data were collected with a bin width of 5 sec. Putative 5-HTergic neurons were identified by their regular, low frequency, spontaneous activity (1-5 Hz) and by their response to application of 8-OH-DPAT (VanderMaelen and Aghajanian, 1983). When the neuronal activity was reduced by the 5-HT_1A receptor agonist, only those experiments that showed at least a partial recovery of activity after wash out of the drugs were considered for analysis. The drugs tested were dissolved to the desired concentrations and introduced into the slice chamber by means of a stopcock arrangement. The temperature of the ACSF was held constant at 35°C.

In vivo electrophysiological recordings in DRN. In general, the methodology described in Sommermeyer et al. (1993b) was followed. For extracellular single unit recordings, rats were placed in a stereotactic head holder after being anaesthetized with a mixture of xylazine (2 mg/kg), ketamine (100 mg/kg) and chlorpromazine (10 mg/kg) administered i.m. This mixture allowed the use of low doses of each anaesthetic compound and, in addition, allowed recordings for more than 2 hr without reinjection of anaesthetics. A tungsten electrode (impedance: 12 M, 10 Hz) was advanced in the DRN (interaural line coordinates: F 1.2, H 3.5, L 0.0; Paxinos and Watson, 1982, at an angle of 15° from the vertical plane). Inside the DRN, the electrode was advanced slowly in steps of 2 µm until the spontaneous activity of a neuron was detected. Conventional electronic equipment was used to process the signals (see higher). Putative 5-HTergic neurons were identified according to the above mentioned criteria. Compounds were administered i.v.

8-OH-DPAT drug discrimination. Rats were trained to discriminate 8-OH-DPAT (0.1 mg/kg, i.p.; T, 15 min, n = 20) from vehicle in a standard two-lever fixed ratio 10 food-reinforced operant procedure according to the method described by Schreiber and De Vry (1993). Daily sessions were terminated after 50 reinforcers or after 10 min, whichever came first. Discrimination criterion consisted of 10 consecutive sessions in which not more than nine responses occurred on the nonreinforced lever before the first reinforcer was obtained. Generalization tests were performed when the number of incorrect responses before the first reinforcer was not more than four on three consecutive sessions and when at least 20 reinforcers were obtained per session. During testing, responding on the selected lever, i.e., the lever on which 10 responses accumulated first, was reinforced for the remainder of the session. All animals were tested with different doses of the training compound (0.01-0.4 mg/kg, i.p.). Generalization tests were performed 15 min after application of the test compound (except for a time-dependency study, where 0.1 mg/kg BAY × 3702 was tested 15-120 min after application, and the oral dose-dependency study, where BAY × 3702 was tested 30 min after application). In general, each dose of a test compound was tested in six to eight rats, randomly allocated to each test condition. Test results were expressed as the percentage of rats that selected the drug lever (% drug lever selections). Behavioral disruption was expressed as the percentage of rats which failed to select a lever.

Body temperature. Different groups of rats (n = 5-7 per group) were treated with vehicle or various doses of a test compound and their body temperature was oesophagally measured repeatedly at fixed time points. Time points measured included: 5 min before, and either 7.5, 15, 30 and 60 min (i.v. dose-response determination), or 15, 30, 60, 120 and 240 min (i.p. and p.o. dose-response determination) after drug administration. For graphical presentation (fig. 10, lower panel) and calculation of ED₅₀ values, results were expressed as temperature change in °C relative to baseline value, and corrected for the temperature change observed in the vehicle control group. In an antagonism experiment, the 5-HT_1A receptor antagonist WAY-100635 (1 mg/kg, i.p.) or vehicle was administered 60 min before administration of BAY × 3702 (0.03 mg/kg, i.v.) or vehicle. Body temperature was measured 5 min before, and 7.5, 15, 30 and 60 min after the i.v. bolus injection.

Determination of plasma ACTH concentrations. Endocrinological experiments were performed between 10:00 A.M. and 2:30 P.M.. At selected time points after p.o. administration of drug or vehicle (n = 5-6 per group), rats were killed by decapitation. Trunk blood was collected in prechilled tubes containing 0.5 ml of 0.3 M EDTA, pH: 7.4. Blood samples were centrifuged at 1000 × g for 20 min (4°C) and plasma was stored at 20°C until analysis. Plasma levels of ACTH were determined using commercially available radioimmunoassay kits. The sensitivity limit of the assays was 7 pg/ml. The inter- and intraassay variation coefficients were less than 10% in all cases. For graphical presentation and calculation of ED₅₀ values, results were expressed as % change as compared to baseline values.

Data analysis. In general, all in vitro biochemical experiments were performed in triplicates and repeated at least twice; representative experiments were selected for data presentation. Also in the case of the electrophysiological experiments, representative experiments were selected for data presentation. Hypothermia and endocrinological data were analyzed by ANOVA; followed, where appropriate, by Tukey's post hoc comparisons. In general, the MED was defined as the lowest dose which induced a statistically significant effect (P < .05). Least-square linear regression analysis was used to estimate ED₅₀, ID₅₀ and T_1/2 values and the corresponding 95% CL after log-probit conversion of the data. ED₅₀ values with nonoverlapping 95% CL were considered to be significantly different.

Drugs and chemicals. Radioactive ligands and reagents were from Amersham Buchler, Braunschweig, FRG, Du Pont NEN, Bad Homburg, FRG or New England Nuclear, Dreieich, FRG. Ketanserin was a gift from Janssen Pharmaceutica, Beerse, Belgium, clonidine from Boehringer, Ingelheim, FRG. Commercially obtained compounds were: 5-HT (Merck, Darmstadt, FRG); (L)-noradrenaline (Fluka, Deisenhofen, FRG); HEAT (Beiersdorf, Hamburg, FRG); (L)-quinpirole (RBI, Cologne, FRG); haloperidol, idazoxan and (D,L)-propranolol (Sigma, Deisenhofen, FRG); sulpiride and methiothepin (Biotrend, Cologne, FRG); ketamine (Parke Davis, Berlin, FRG). BAY × 3702, BAY × 3703, ipsapirone, 8-OH-DPAT and WAY-100635 were synthesized at the Chemistry Department of Bayer AG, Wuppertal, FRG; xylazine and chlorpromazine were obtained from Bayer AG, Leverkusen, FRG. All other chemicals were purchased from Merck, Darmstadt, FRG, or Sigma, Deisenhofen, FRG. Radioimmunoassay kits for the determination of plasma ACTH levels were obtained from DPC Biermann Ltd., Bad Nauheim, FRG. Compounds or their appropriate vehicles were given i.v., i.p. or p.o. in a volume of 1, 2 and 5 ml/kg body weight, respectively. BAY × 3702, BAY × 3703 and 8-OH-DPAT were dissolved in HCl or lactic acid and distilled water or a 0.9% NaCl solution; ipsapirone and WAY-100635 were dissolved in distilled water. Doses were expressed as the salts.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Receptor binding. BAY × 3702 bound with high affinity to 5-HT_1A receptors of different species, including human, rat and calf, and different brain areas, including, cortex and hippocampus; with K_i values ranging from 0.19 (calf hippocampus) to 0.58 nM (rat hippocampus; table 2). The binding is considered to be stereoselective as it was found that the affinity of the (+)-enantiomer, BAY × 3703, was about 10-fold lower than that of BAY × 3702 (fig. 2). 5-HT_1A receptor affinity of the reference compounds 8-OH-DPAT and ipsapirone was generally lower; the potency difference vs. BAY × 3702 being about factor 1.5 to 10 (depending on particular brain tissue) and factor 5 to 15, for 8-OH-DPAT and ipsapirone, respectively (table 2; fig. 2). As measured in calf hippocampus, affinity of BAY × 3702 to 5-HT_1A receptors was about 20 times higher as compared with 5-HT (K_i: ~4 nM). BAY × 3702 exhibited also a relatively high to moderate affinity for alpha₁- and alpha₂-receptors (K_i values: 6 and 7 nM, respectively), 5-HT₇ and 5-HT_1D receptors (7 and 36 nM, respectively) and dopamine D₂ and D₄ receptors (48 and 91 nM, respectively; table 2). In more than 50 other receptor binding assays, weak affinity was only detected for 5-HT_2(c) receptors (310-380 nM) and sigma binding sites (176 nM, table 2). The binding profile of BAY × 3702 is considered to be relatively selective because at least one order of magnitude differentiates its binding to the 5-HT_1A receptor from its binding to other receptors.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Inhibition by BAY × 3702 and its (+)-enantiomer BAY × 3703 of [3H]ipsapirone binding to calf hippocampal membranes in comparison to 8-OH-DPAT and ipsapirone. Data are from one representative experiment performed in triplicates, which was repeated twice. The concentration of [3H]ipsapirone was 2 nM.

5-HT_1A receptor coupled adenylate cyclase in rat hippocampal membranes. BAY × 3702 inhibited forskolin-stimulated adenylate cyclase activity in rat hippocampal membranes with an IC₅₀ value of 1.87 ± 0.74 nM (mean ± 1 S.E.M., n = 5, fig. 3). The maximal level of inhibition was similar to that caused by 8-OH-DPAT and 5-HT (data not shown), but more than that of ipsapirone (fig. 3). Therefore, BAY × 3702 and 8-OH-DPAT can be characterized as full agonists; whereas ipsapirone behaves as partial agonist in this assay. In accordance with the 5-HT_1A receptor binding data, BAY × 3702 was more potent than 8-OH-DPAT (IC₅₀: 25.25 ± 0.06 nM, n = 4) and ipsapirone (IC₅₀: > 1 µM, fig. 3). The (+)-enantiomer BAY × 3703 was about 10-fold less potent than BAY × 3702 (IC₅₀: 21.5 ± 1.5 nM, n = 2, fig. 3). The inhibitory effect of BAY × 3702 on cyclase activity could be completely reversed by the selective 5-HT_1A receptor antagonist WAY-100635 (fig. 4).

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Inhibition by BAY × 3702 and its (+)-enantiomer BAY × 3703 of forskolin-stimulated adenylate cyclase activity in rat hippocampal membranes in comparison to 8-OH-DPAT and ipsapirone. Data are from one representative experiment performed in triplicates, which was repeated two to four times. The concentration of forskolin was 10 µM.

View larger version (11K): [in this window] [in a new window]   Fig. 4.   Antagonism by the 5-HT1A receptor antagonist WAY-100635 of the inhibition by BAY × 3702 of the forskolin-stimulated adenylate cyclase activity in rat hippocampal membranes. Data are from one representative experiment performed in triplicates, which was repeated twice.

5-HT_1D receptor coupled adenylate cyclase in calf substantia nigra membranes. BAY × 3702 inhibited the forskolin-stimulated adenylate cyclase in calf substantia nigra membranes by approximately 13% with an IC₅₀ value of 100 nM, whereas 5-HT induced an inhibition of about 16% (IC₅₀: ~20 nM). Therefore BAY × 3702 can be characterized as a relatively weak 5-HT_1D receptor agonist (data not shown).

Dopamine D₂ receptor regulated cAMP levels. The dopamine D₂ receptor agonist quinpirole reduced the cAMP levels in forskolin-stimulated LZR-1 cells with an IC₅₀ value of 20 nM (fig. 5, left panel). BAY × 3702 had no inhibitory effect on the cAMP concentration in prestimulated cells. However, similar to the D₂ receptor antagonist sulpiride, BAY × 3702 antagonized the quinpirole (10 µM) induced inhibition of cAMP formation in forskolin-treated cells (fig. 5, right panel), and thus behaved like a dopamine D₂ receptor antagonist.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Left panel, Effects of BAY × 3702 and the dopamine D2 receptor agonist quinpirole on cAMP levels in forskolin-(10 µM) stimulated LZR-1 cells. Right panel, Effects of BAY × 3702 and the dopamine D2 receptor antagonist sulpiride on the quinpirole (10 µM) induced inhibition of cAMP formation in forskolin-treated cells.

Alpha₁-adrenergic receptor linked phosphoinositide breakdown. Activation by noradrenaline of alpha-1-adrenergic receptors in primary cultures of rat brain astrocytes led to a stimulation of PI turnover and thus to an increased formation of inositol phosphates. BAY × 3702 by itself did not affect PI-turnover. It antagonized, however, the noradrenaline-induced stimulation of inositol phosphate formation with an IC₅₀ value of 0.29 µM (data not shown). Therefore BAY × 3702 can be considered as a relatively weak antagonist at alpha-1-adrenergic receptors.

Alpha₂-adrenergic receptor regulated cAMP levels and 5-HT release. Intrinsic activity at alpha-2-adrenergic receptors was investigated in two ways. First, in NG 108-15 neuroblastoma x glioma hybrid cells the PGE₁-induced elevation of the intracellular cAMP level was reduced by the alpha-2-adrenergic receptor agonist clonidine via interaction with alpha-2-adrenergic receptors (fig. 6). BAY × 3702 alone had no effect on the PGE₁-stimulated cAMP formation. The attenuation of cAMP levels by 1 µM clonidine, however, was concentration-dependently reversed by BAY × 3702 as it was by the alpha-2-adrenergic receptor antagonist idazoxan (10 µM, fig. 6). Second, clonidine inhibited by activation of alpha₂-adrenergic receptors the electrically evoked release of 5-HT from rat brain cortical slices, while BAY × 3702 was without effect (fig. 7). The inhibition caused by 10 and 100 nM, but not of 1 µM clonidine was completely antagonized by 1 µM BAY × 3702. The results from these two assay systems characterize the compound as an alpha-2-adrenergic receptor antagonist.

View larger version (13K): [in this window] [in a new window]   Fig. 6.   Antagonism by BAY × 3702 of the clonidine-induced attenuation of the PGE1-stimulated rise in the cAMP formation in NG108-15 neuroblastoma × glioma hybrid cells. Concentrations of clonidine and PGE1 were 1 and 3 µM, respectively. Data are from one representative experiment performed in triplicates, which was repeated twice.

View larger version (14K): [in this window] [in a new window]   Fig. 7.   Antagonism by 1 µM BAY × 3702 of the clonidine-induced inhibition of the electrically evoked [3H]5-HT overflow from rat cortical slices. Data represent the S2/S1 ratio as percentage of the ratio obtained in the absence of clonidine from one representative experiment, which was repeated twice. Each data point results from two superfusion chambers.

In vitro electrophysiological recordings in DRN. The neurons were identified as presumably 5-HTergic by standard criteria (Vandermaelen and Aghajanian, 1983), such as low frequency firing and responsivity to application of 8-OH-DPAT (50 nM, fig. 8, upper panel). Bath application of BAY × 3702 at concentrations as low as 1 to 3 nM produced a complete, relatively long-lasting and reversible inhibition of cell firing in the DRN (fig. 8, upper panel). The recovery of activity was always slower than the recovery after 8-OH-DPAT, possibly indicative of a slower dissociation constant for BAY × 3702. The amount of inhibition, as well as the duration was concentration dependent. It was, however, difficult to obtain the usual graded dose-response relationship due to the narrow margin between the no-effect and the full-effect concentration. BAY × 3702 appeared to act as a full agonist in all experiments and, although no extensive dose-response relationships were established, the compound was clearly more potent than 8-OH-DPAT and ipsapirone; the difference in potency vs. BAY × 3702 being approximately 1 and 2 orders of magnitude, respectively (for ipsapirone, see fig. 8, lower panel). The nonselective 5-HT_1A receptor antagonist propranolol, when tested at a concentration that did by itself not produce any changes in neuronal activity (100 µM), was able to completely antagonize the inhibitory effect of BAY × 3702 (2 nM) when tested in conjunction with the latter compound (data not shown). When propranolol was applied after BAY × 3702 recovery of activity was accelerated; suggesting competitive displacement of BAY × 3702 from its binding sites (data not shown).

View larger version (20K): [in this window] [in a new window]   Fig. 8.   Inhibition of 5-HT cell firing by BAY × 3702 (upper panel) and ipsapirone (lower panel), as compared with 8-OH-DPAT in a rat slice containing the DRN. The compounds were bath applied at the time periods indicated by the horizontal bars above the histogram. The x-axis denotes time in s; the y-axis refers to the extracellularly recorded rate of action potentials.

In vivo electrophysiological recordings in DRN. BAY × 3702 (0.5 µg/kg, i.v.) induced an almost complete and relatively long-lasting inhibition of DRN neuronal firing (fig. 9, upper panel). The onset of inhibition occurred with a delay of 2 to 4 min and the inhibition was completely reversible. Again, BAY × 3702 appeared to be about 10 times more potent than 8-OH-DPAT and about 50 times more potent than ipsapirone (fig. 9, middle and lower panel, respectively).

View larger version (19K): [in this window] [in a new window]   Fig. 9.   Inhibition of 5-HT cell firing in the DRN by BAY × 3702 (upper panel), 8-OH-DPAT (middle panel) and ipsapirone (lower panel) in the anaesthethized rat. The arrow indicates the time of i.v. bolus administration of the compound. The x-axis denotes time in s; the y-axis refers to the extracellularly recorded rate of action potentials.

8-OH-DPAT drug discrimination. Sixteen of 20 rats learned to discriminate 8-OH-DPAT (0.1 mg/kg, i.p.) from vehicle; the median number of sessions to reach criterion being 38 (range: 23-70 sessions). Four rats were discarded after 70 sessions because of insufficient response rates (both under drug and vehicle conditions) or lack of discriminative stimulus control. Dose-dependent and complete generalization was obtained with the selected 5-HT_1A receptor agonists [fig. 10, upper panel: 8-OH-DPAT (ED₅₀ and 95% CL in mg/kg, i.p.: 0.028; 0.020-0.041), BAY × 3702 (0.022; 0.011-0.042), BAY × 3703 (1.38; CL not computable, data not shown) and ipsapirone (0.44; 0.23-0.85). With all compounds, complete generalization (100% drug lever selections) was obtained at doses devoid of behavioral disruptive effects (% lever selections not affected). However, although 8-OH-DPAT still induced complete generalization when tested at doses higher than the training dose, this coincided with severe behavioral disruption (ID₅₀ for % lever selection with 95% CL in mg/kg, i.p.: 0.20; 0.15-0.27, data not shown). Dose-dependent and complete generalization was also obtained after oral administration of BAY × 3702 (0.38; 0.13-1.11; 100% generalization at 3 mg/kg, data not shown). The time-dependency study indicated that generalization of BAY × 3702 (0.1 mg/kg, i.p.) was maximal at 15 min (100% drug lever selections) and reduced to the vehicle level again (14.3% drug lever selections) at 2 h after administration (T_1/2 and 95% CL in min: 58; 36-93; data not shown).

View larger version (19K): [in this window] [in a new window]   Fig. 10.   Upper panel: Generalization of BAY × 3702, 8-OH-DPAT and ipsapirone in rats trained to discriminate 8-OH-DPAT (0.1 mg/kg, i.p.) from vehicle in a standard 2-lever drug discrimination procedure (T, 15 min; n = 6-8 per group, except for 8-OH-DPAT: n = 16). Lower panel, Dose-dependent hypothermia induced by BAY × 3702, 8-OH-DPAT and ipsapirone in rats shown at the time of maximal effect (T, 30 min, except for 8-OH-DPAT: T, 60 min; n = 5-7 per group).

Body temperature. BAY × 3702 induced a dose- and time-dependent reduction in body temperature; the MED values being 0.01, 0.3 and 10 mg/kg after i.v., i.p. and p.o. administration, respectively (temperature reduction of at least 1°C, P < .05; for ED₅₀ values see table 3). In general, the hypothermic effects of BAY × 3702 were pronounced, as a reduction of 2.5°C, compared to vehicle treatment, was obtained at the highest doses tested (fig. 10, lower panel). The time-dependency data indicated that effects were maximal between 0.5 and 1 hr after administration and no longer observed after 1, 2 and 4 hr, after i.v., i.p. and p.o. administration, respectively (fig. 11 and data not shown). After i.p. administration, BAY × 3702 was found to be about 3 times more potent than 8-OH-DPAT, and 30 times more potent than ipsapirone, and BAY × 3702 appeared to be slightly more efficient than the latter compounds (maximal body temperature reductions of 2.5, 1.9 and 1.7°C, for the highest tested doses of BAY × 3702, 8-OH-DPAT and ipsapirone, respectively; fig. 10, lower panel; table 3). After oral administration, BAY × 3702 was about 10 times more potent than ipsapirone (MED values: 3 and 30 mg/kg, respectively; for ED₅₀ values, see table 3). Again, BAY × 3702 appeared to be more effective than ipsapirone (maximal extent of hypothermia as compared to vehicle control: 2.5 and 2.0°C, respectively). The hypothermia induced by BAY × 3702 (0.03 mg/kg, i.v.) was abolished by pretreatment with the 5-HT_1A receptor antagonist WAY-100635 (1 mg/kg, i.p.; T, 60 min; fig. 11).

View larger version (20K): [in this window] [in a new window]   Fig. 11.   Reversal of the hypothermic effects induced by BAY × 3702 by the 5-HT1A receptor antagonist WAY-100635 in rats. WAY-100635 (1 mg/kg) was administered i.p., 60 min before i.v. administration of BAY × 3702 (0.03 mg/kg). Asterisks indicate a statistically significant difference between the vehicle/BAY × 3702 treated group and each of the three other groups (P < .05, n = 7 per group).

Plasma ACTH concentrations. In preliminary experiments, it was found that BAY × 3702 and ipsapirone (30 mg/kg, p.o.) induced a time-dependent stimulation of plasma ACTH levels; the effect being maximal 30 min after application for both compounds (data not shown). Therefore, dose-dependency experiments were performed at the latter postapplication time interval (fig. 12). After oral administration, both compounds induced dose-dependent effects; BAY × 3702 being about five times more potent than ipsapirone (MED values: 10 and 75 mg/kg, respectively, P < .05; for ED₅₀ values see table 3). In addition, BAY × 3702 appeared to be slightly more effective than ipsapirone (maximal extent of ACTH stimulation as compared to vehicle control: 380 and 321%, respectively).

View larger version (14K): [in this window] [in a new window]   Fig. 12.   Stimulation of plasma ACTH in rats by BAY × 3702 and ipsapirone at the time of maximal effect (T, 30 min). Mean (± 1 S.E.M.) control value for vehicle treatment was 48.3 ± 4.8 pg/ml (n = 5-6 per group).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

We characterized the novel aminomethylchroman derivative BAY × 3702 (De Vry et al., 1997a, b) in a variety of in vitro and in vivo models of 5-HT_1A receptor function. Selection of these models was such that it included models of presynaptic or somatodendritic receptor activation, as well as models of postsynaptic receptor activation. This is considered to be important as the apparent intrinsic activity of a 5-HT_1A receptor ligand may depend on the particular assay system (for review, see De Vry 1995). Thus, it is generally found that the intrinsic activity of a 5-HT_1A receptor ligand appears to be relatively higher (or at least of equal magnitude) in models of presynaptic function, as compared to models of postsynaptic function. (e.g., Di Sciullo et al., 1990; Foreman et al., 1995; Glaser et al., 1991; Hamon et al., 1988; McCall et al., 1994; Millan et al., 1992, 1993). Moreover, even in the case of models presumably reflecting activation of postsynaptic receptors, the apparent intrinsic activity of a ligand may vary from one model to another and it has been suggested that such variations are related to different levels of spare receptors or receptor reserve (e.g., Meller and Bohmaker, 1994; Yocca et al., 1992; for discussion, see De Vry, 1995). With these restrictions in mind, BAY × 3702 was characterized in the present study as a highly potent, relatively selective and orally active 5-HT_1A receptor ligand with high intrinsic activity.

As compared with reference 5-HT_1A receptor ligands, BAY × 3702 appears to be about 2 to 20 times more potent than 8-OH-DPAT (Arvidsson et al., 1981), and about 5 to 50 times more potent than ipsapirone (Traber et al., 1984) across the diverse in vitro and in vivo models. In general, the potency difference between 8-OH-DPAT and ipsapirone as obtained in our study is very similar to the potency difference reported in the literature using similar models (e.g., De Vry, 1995; De Vry et al., 1991; Di Sciullo et al., 1990; Gilbert et al., 1988; Hamon et al., 1988; Jolas et al., 1995, McCall et al., 1994; Schechter et al., 1990; Schreiber et al., 1995). It should be realized, however, that the potency ratio obtained with these compounds may depend to some extent on the route of administration, as well as on the model. Although not explicitly compared for each route of administration, potency differences appeared to be higher after i.v. administration (e.g., in vivo electrophysiology model), as compared with i.p. administration and, in particular, p.o. administration (e.g., drug discrimination, hypothermia and ACTH model), and this application-dependent finding reflects pharmacokinetic factors [i.e., absolute bioavailability after oral administration in the rat is about three times higher for ipsapirone than for BAY × 3702 (Langer M, Opitz W and Schöllnhammer G, personal communication) for further pharmacokinetic information on BAY × 3702, see De Vry et al., 1997a; Schwarz et al., 1997]. However, for a given route of administration, variation in the potency difference may also reflect the relative involvement of pre- and postsynaptic 5-HT_1A receptor populations with different levels of spare receptors. Thus, it is conceivable that the potency difference between compounds with different levels of intrinsic activity may appear to be higher if no or relatively few spare receptors are present in the system, and smaller if a relatively high level of spare receptors is present in the system.

In the in vivo models, the same potency order of the 5-HT_1A receptor agonists was maintained, independently of the particular route of administration. Interestingly, however, for each compound large potency differences were obtained dependent on the model. Thus, the most sensitive model appeared to be the drug discrimination model, followed by the hypothermia model and the ACTH model (fig. 10; table 3). These sensitivity differences correlate with the presumed involvement of pre- vs. postsynaptic receptors in the model (for discussion, see De Vry, 1995). Thus, drug discrimination (at least with the relatively low training dose of 8-OH-DPAT used in our study) is supposed to be a model of presynaptic function (Schreiber and De Vry, 1993); whereas hypothermia is thought to be a model of pre- and postsynaptic receptor activation (Glaser et al., 1991), and ACTH stimulation is generally considered to be a model of postsynaptic 5-HT_1A receptor activation (Pan and Gilbert, 1992; but see Bluet-Pajot et al., 1995). In general, potency differences obtained in the in vivo models reflect the differences in affinity for the 5-HT_1A receptor as assessed in the binding studies, as well as the differences in potency in the adenylate cyclase model of 5-HT_1A receptor activation. Moreover, the stereoselectivity observed in the 5-HT_1A receptor binding assay was retained in the adenylate cyclase model and in an in vivo model (i.e., drug discrimination); in each case the ()-enantiomer BAY × 3702 was found to be 10 to 30 times more potent than the (+)-enantiomer BAY × 3703 (figs. 2 and 3).

With respect to intrinsic activity, BAY × 3702 appears to be a 5-HT_1A receptor full agonist as it was found that it was equally, or slightly more efficient, than the full agonist 8-OH-DPAT; and more efficient than the partial agonist ipsapirone. This differentiation tended to be maximal in those models that are supposed to be mediated postsynaptically (or have at least a postsynaptic component). Thus, in the adenylate cyclase model, BAY × 3702, as well as 8-OH-DPAT, was more effective than ipsapirone. The finding that 8-OH-DPAT is more effective than ipsapirone in this model is in accordance with previous studies (e.g., De Vry et al., 1991; but see Hamon et al., 1988 where it is reported that ipsapirone shows the same level of efficacy than 8-OH-DPAT), and is most likely related to the presumed absence of spare receptors (Yocca et al., 1992). Also in a forskolin-stimulated adenylate cyclase assay using chinese hamster ovary (CHO) cells expressing human 5-HT_1A receptors, 8-OH-DPAT has been reported to have a higher intrinsic activity than ipsapirone (e.g., McCall et al., 1994). In the in vivo models with a postsynaptic 5-HT_1A receptor involvement, such as hypothermia and ACTH stimulation, BAY × 3702 and 8-OH-DPAT were slightly more efficient than ipsapirone (see also, Glaser et al., 1991; Di Sciullo et al., 1990; but see Gilbert et al., 1988). In the models of presynaptic function, such as inhibition of DRN neuronal firing, or drug discrimination, estimation of possible differences in intrinsic activity between BAY × 3702 and 8-OH-DPAT, on the one hand, and ipsapirone on the other hand was virtually impossible, as ipsapirone induced already maximal effects (see also: Schechter et al., 1990; Schreiber and De Vry, 1993; Schreiber et al., 1995).

Further evidence for the suggestion that BAY × 3702 is an agonist at 5-HT_1A receptors was obtained by means of antagonism tests. Thus, both in the adenylate cyclase model, the electrophysiological model and the hypothermia model, pharmacological effects induced by BAY × 3702 were successfully blocked by a selective 5-HT_1A receptor antagonist (i.e., WAY-100635; Forster et al., 1995), or a nonselective 5-HT_1A receptor antagonist (i.e., propranolol, Sprouse and Aghajanian, 1986). Although it remains to be determined whether the effects obtained with BAY × 3702 in the drug discrimination model can be reversed by a 5-HT_1A receptor antagonist, it was found that WAY-100635 was able to antagonize the BAY × 3702 cue in rats trained to discriminate this compound (0.1 mg/kg, i.p.) from vehicle in a similar drug discrimination procedure (De Vry et al., 1997a).

The binding profile of BAY × 3702 was characterized by testing the compound in more than 50 binding assays covering a wide range of classical receptors, enzymes and ion channels. The binding profile of the compound is considered to be relatively selective as at least one order of magnitude differentiates its binding to 5-HT_1A receptors from its binding to other receptors. Besides 5-HT_1A receptors (K_i: 0.2-0.6 nM), BAY × 3702 bound also with relatively high to moderate affinity to the following receptors: alpha-1- and alpha-2-adrenergic (K_i: 6 and 7 nM, respectively); 5-HT₇- and 5-HT_1D (7 and 36 nM); dopamine D₂- and D₄ (48 and 91 nM); sigma sites (176 nM) and 5-HT_2C (310 nM); others: >10 µM. At those receptors where BAY × 3702 bound with lower affinity, the compound appeared to be either an agonist (5-HT_1D receptors) or an antagonist (alpha-1, alpha-2 and D₂ receptors). It can be hypothesized, however, that receptor interactions other than those with 5-HT_1A receptors are not essentially involved in the in vivo effects described in the present study, as it was found that the effects induced by BAY × 3702 were virtually completely blocked by pretreatment with 5-HT_1A receptor antagonists. It can even be hypothesized that at least at the doses that are therapeutically relevant (i.e., 1-30 µg/kg, i.v.; 0.01-0.3 mg/kg, i.p.; 0.1-3 mg/kg, p.o.; for ED₅₀ values obtained in animal models of neuroprotection, depression and anxiety, see De Vry et al., 1997a; Horváth et al., 1997), the behavioral effects of the compound are mediated exclusively by interactions with 5-HT_1A receptors. This is suggested by the finding that the BAY × 3702 cue (0.1 mg/kg, i.p.) did not generalize to the alpha-1- and alpha-2-adrenergic antagonists prazosin and idazoxan, nor to the D₂ receptor antagonist raclopride (De Vry J and Jentzsch J, unpublished data), and the finding that the anxiolytic effects observed with the compound in the shock-induced ultrasonic vocalization rat model of anxiety, as well as the antidepressive effects observed in the forced swimming rat model of depression were completely blocked by the 5-HT_1A receptor antagonist WAY-100635 (De Vry et al., 1997a). In line with these findings, preliminary experiments indicate that the neuroprotective effects obtained with BAY × 3702 in a rat model of permanent focal cerebral ischemia (i.e., middle cerebral artery occlusion) can be attenuated by cotreatment with WAY-100635 (De Vry et al., 1997a; Horváth et al., 1997).

Although it was not the aim of the present study to test explicitly the selectivity ratio in vivo, some behavioral data in addition to the data obtained in the drug discrimination procedures mentioned above suggest indirectly that the compound is a fairly selective 5-HT_1A receptor ligand. Thus, it was found that pretreatment with BAY × 3702 (0.01-1 mg/kg, i.p.), as well as with 8-OH-DPAT and ipsapirone, was unable to block the discriminative effects of d-amphetamine in rats trained to discriminate the latter compound (1 mg/kg, i.p.) from vehicle (De Vry J, unpublished data). Because this model is sensitive to detect dopamine D₂ antagonism in vivo (both haloperidol and raclopride attenuate the discriminative effects of d-amphetamine with MED values of 0.1 and 1 mg/kg, i.p., respectively; De Vry J, unpublished data), it can be concluded that the D₂ vs. 5-HT_1A selectivity ratio of BAY × 3702 in vivo is at least factor 30 (comparison of the potency of BAY × 3702 to generalize to the 8-OH-DPAT cue and its potency to block the amphetamine cue). This conclusion is underscored by the observation that BAY × 3702 (1-10 mg/kg, i.p.), as well as 8-OH-DPAT and ipsapirone, are not effective in a conditioned avoidance reaction paradigm in rats and fail to block apomorphine induced climbing in mice; two additional models sensitive to dopamine D₂ antagonists (MED value of haloperidol and raclopride in both models: 0.1-0.3 mg/kg, i.p.; De Vry J, unpublished data). It remains to be determined to what extent possible interactions of BAY × 3702 with adrenergic, D₄ and 5-HT_1D and 5-HT₇ receptors occur in vivo.

Preclinical evidence has suggested that selective, high affinity 5-HT_1A receptor ligands with a high level of intrinsic activity offer an improved therapeutic profile as compared with compounds with a partial agonist profile (see introduction and De Vry, 1995, 1996). BAY × 3702 was derived from a project aimed at identifying selective, high affinity, orally active 5-HT_1A receptor full agonists. Recently, the compound has entered clinical trials to verify the promising neuroprotective efficacy observed in a wide spectrum of preclinical models (De Vry et al., 1997a, b; Horváth and Augstein, 1997; Horváth et al., 1997).