The 5-hydroxytryptamine 2C (5-HT2C) receptor subtype has received considerable attention as a target for drug discovery, having been implicated in a wide variety of disorders. Here, we describe the in vitro pharmacological profile of the novel 5-HT2C receptor-selective agonist vabicaserin [(−)-4,5,6,7,9,9a,10,11,12,12a-decahydrocyclopenta[c] [1,4]diazepino[6,7,1-ij]quinoline hydrochloride] (SCA-136), including a comprehensive strategy to assess 5-HT2B receptor selectivity using diverse preparations and assays of receptor activation. Vabicaserin displaced 125I-(2,5-dimethoxy)phenylisopropylamine binding from human 5-HT2C receptor sites in Chinese hamster ovary cell membranes with a Ki value of 3 nM and was >50-fold selective over a number of serotonergic, noradrenergic, and dopaminergic receptors. Binding affinity determined for the human 5-HT2B receptor subtype using [3H]5HT was 14 nM. Vabicaserin was a potent and full agonist (EC50, 8 nM; Emax, 100%) in stimulating 5-HT2C receptor-coupled calcium mobilization and exhibited 5-HT2A receptor antagonism and 5-HT2B antagonist or partial agonist activity in transfected cells, depending on the level of receptor expression. In rat stomach fundus and human colonic longitudinal muscle endogenously expressing 5-HT2B receptors, vabicaserin failed to induce a 5-HT2B receptor-dependent contraction and produced a rightward shift of the 5-HT and α-methyl-5-HT concentration-response curves in these preparations, respectively, consistent with 5-HT2B competitive antagonism. Likewise, vabicaserin failed to induce a 5-HT2B receptor-mediated contraction in arteries from deoxycorticosterone acetate-salt-treated rats, a model of hypersensitized 5-HT2B receptor function, and produced a rightward shift in the 5-HT-induced response that was consistent with 5-HT2B receptor antagonism. In summary, vabicaserin is a novel, potent, and selective 5-HT2C receptor agonist.
At least 14 distinct 5-HT receptor subtypes have been cloned and classified based on sequence similarity and common signal transduction pathways (Barnes and Sharp, 1999). The 5-HT2 receptor subfamily accommodates three subtypes designated 5-HT2A, 5-HT2B, and 5-HT2C, and these receptors belong to the large family of seven transmembrane domain G protein-coupled receptors. They display high sequence homology with each other and signal transduction initially believed to be principally via activation of phospholipase C (Baxter et al., 1995). Coupling through a wide array of signal transduction pathways has also been reported (Berg et al., 1998; Miller, 2005; Werry et al., 2008).
Pharmacological interest within the 5-HT2 receptor family in the context of drug discovery has largely focused on the 5-HT2A and 5-HT2C receptor subtypes. In the case of the 5-HT2A receptor subtype it has been recognized that drugs used to treat schizophrenia and depression, in addition to hallucinogenic agents, display affinity for this target. The 5-HT2C receptor has been implicated in a wide variety of conditions including obesity, anxiety, depression, obsessive compulsive disorder, schizophrenia, migraine, and erectile dysfunction (Wacker and Miller, 2008). As a consequence, it has received significant attention as a target for drug discovery. Several novel 5-HT2C receptor agonists have been described, including [7bR,10aR)-1,2,3,4,8,9,10,10a-octahydro-7bH-cyclopenta-[b][1,4]diazepino[6,7,1hi]indole] (WAY-163909) (Dunlop et al., 2005; Marquis et al., 2007; Rosenzweig-Lipson et al., 2007), 2-[(3-chlorophenyl)methoxy]-6-(1-piperazinyl)pyrazine hydrochloride (CP-809101) (Siuciak et al., 2007), and lorcaserin (Thomsen et al., 2008), and shown to have activity in preclinical animal models used to evaluate potential antipsychotic, antidepressant, and antiobesity activity. Moreover, lorcaserin has demonstrated clinical efficacy in obesity (Smith et al., 2009). As with any small-molecule drug discovery effort, selectivity is important, and a particular consideration in 5-HT2C receptor agonist targeted drug discovery is the selectivity of compounds toward the related 5-HT2B receptor subtype, specifically 5-HT2B receptor agonist activity. Activation of the 5-HT2B receptor has been implicated in primary pulmonary hypertension (Launay et al., 2002) and valvulopathy (Fitzgerald et al., 2000; Rothman et al., 2000).
As part of a strategy to develop novel 5-HT2C receptor agonists as a potential therapeutic for psychiatric disorders, vabicaserin [(−)-4,5,6,7,9,9a,10,11,12,12a-decahydrocyclopenta[c] [1,4]diazepino[6,7,1-ij]quinoline hydrochloride] (SCA-136) has been identified as a potent and selective ligand for the 5-HT2C receptor. This article describes its in vitro pharmacological profile determined using in vitro functional assays of 5-HT2A, 5-HT2B, and 5-HT2C receptor activation and a comprehensive strategy for assessing functional activity toward the related 5-HT2B receptor subtype. Data demonstrate vabicaserin to be a novel, potent, and selective 5-HT2C receptor agonist.
Materials and Methods
Vabicaserin (SCA-136) was synthesized at Wyeth Research (Princeton, NJ). α-Methyl-5-(2-thienylmethoxy)-1H-indole-3-ethanamine hydrochloride (BW723C86) (Tocris Bioscience, Ellisville, MO), deoxycorticosterone acetate (DOCA), and 5-HT (Sigma-Aldrich, St. Louis, MO), 6-methyl-1,2,3,4-tetrahydro-1-[3,4-dimethoxyphenylmethyl-9H-pyrido[3,4b]indole] hydrochloride (LY272015) (a gift from Eli Lilly & Co., Indianapolis, IN), and fluo-3-AM (Invitrogen, Carlsbad, CA) were obtained from external sources.
5-HT2 Receptor Cell Lines.
Human 5-HT2C, 5-HT2A, and 5-HT2B receptors were expressed in stably transfected Chinese hamster ovary (CHO) cell lines. Cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum, nonessential amino acids, penicillin/streptomycin, and the following selection markers (maintenance medium): 5-HT2C (50 fmol/mg expression level), 5 μg/ml mycophenolic acid, 0.25 mg/ml xanthine, 100 μM sodium hypoxanthine, and 16 μM thymidine; 5-HT2A (200 fmol/mg), 800 μg/ml neomycin and 500 μg/ml zeocin; 5-HT2B: 400 μg/ml neomycin and 1000 nM methotrexate for 5-HT2B receptor low expression line (500 fmol/mg), and 100 nM and 200 nM methotrexate for the high (5000 fmol/mg) and intermediate (1500 fmol/mg) expression lines, respectively.
5-HT2C Receptor Radioligand Binding.
Agonist and antagonist binding experiments were performed in a 96-well microtiter plate format using a total volume of 200 μl. Agonist binding studies were conducted using 60 μl of incubation buffer made in 50 mM Tris·HCl buffer, pH 7.4 and containing 4 mM CaCl2; 20 μl of [125I] DOI (specific activity, 2200 Ci/mmol; PerkinElmer Life and Analytical Sciences, Waltham, MA). The dissociation constant, KD, of [125I] DOI at the human 5-HT2C receptor was 0.4 nM as determined by saturation binding analysis. Antagonist binding studies were conducted using 80 μl of incubation buffer (50 mM Tris·HCl buffer, pH 7.4 and containing 0.1% ascorbic acid, 10 mM pargyline, and 4 mM CaCl2, 20 μl of [3H]mesulergine) (specific activity, 92.0 Ci/mmol; GE Healthcare, Chalfont St. Giles, Buckinghamshire, UK), at a final concentration of 0.5 nM. The dissociation constant, KD, of [3H]mesulergine at the human serotonin 5-HT2C receptor was 0.8 nM as determined by saturation binding analysis. The reaction was initiated by the addition of 100 μl of tissue suspension. Nonspecific binding was measured in the presence of 1 μM unlabeled DOI (added in 20 μl volume) or 1 mM mianserin, for agonist and antagonist binding, respectively. The reactions proceeded for 60 or 120 min at room temperature. After incubation, the bound ligand-receptor complex was filtered using a 96-well unifilter with a Packard Filtermate 196 Harvester (PerkinElmer Life and Analytical Sciences). The bound complex caught on the filter disk was air-dried, and the radioactivity was measured in a Packard TopCount (PerkinElmer Life and Analytical Sciences) equipped with six photomultiplier detectors, after the addition of 40 μl of Microscint-20 scintillant to each shallow well. The unifilter plate was heat-sealed and counted in a Packard TopCount with a tritium efficiency of 31.0%.
Receptor binding affinity for a range of monoaminergic receptor subtypes including human 5-HT1A, 5-HT1B, 5-HT1D, 5-HT2A, 5-HT2B, 5-HT6, 5-HT7, dopamine D2, D3, and D4 and α1 adrenergic was assessed using similar protocols to those described for the 5-HT2C receptor studies. In all cases, except the α1 adrenergic receptors, recombinant receptors expressed in CHO cells (or HeLa in the case of 5-HT6) were used as receptor source. Rat cortical membrane homogenates were used as a source of α1 adrenergic receptors. Table 1 includes the radioligands used and agents for determination of nonspecific binding for these receptor subtype selectivity assessments.
Radioligand binding results were analyzed by constructing log concentration response curves to generate IC50 estimates. Ki values were calculated from the equation described by Cheng and Prusoff (1973). Single concentration experiments determined the displacement of specific binding at predetermined drug concentrations. Data are mean values (± S.E.M.) from two to three independent experiments.
Measurement of Intracellular Ca2+ Mobilization.
Mobilization of intracellular calcium upon receptor activation was measured using the fluorometric imaging plate reader (FLIPR). Cells were plated at a density of 50,000 cells/well into black 96-well plates with clear bottoms 16 to 24 h before the assay. Hanks' balanced salt solution containing 20 mM HEPES and 2.5 mM probenecid was prepared fresh on the day of the assay and used as assay buffer. The assay buffer was also used to prepare dye-loading buffer that contained a final concentration of 4 μM fluo-3-AM, 0.2% pluronic acid, and 1% fetal bovine serum. Dye-loading buffer (100 μl) was added to each well after removing the culture media, and the loading lasted 1 h at 37°C in a CO2 incubator. The cells were then washed twice with the assay buffer and kept in the buffer (<1 h) at room temperature until transferred to the FLIPR for functional assay. Dye loading was monitored by basal fluorescence signal test in each experiment, with laser intensity adjusted to a suitable level to achieve 8000 to 15,000 fluorescence units at baseline. The fluorescence was detected with an excitation wavelength of 488 nM and the emission filter at 515 nM. The argon laser power was adjusted between 0.3 and 1.2 W, and the camera F/stop was set at 2, and exposure time was at 0.4 s. Calcium signal acquisitions for each assay plate was approximately 1.5 min, and fluorescence counts were recorded at 1-s intervals for the first 60 s and every 6 s for the remainder of the run. After 10-s basal signal readings, drug additions were made by a FLIPR-equipped 96-well pipettor loaded with black tips to obtain 1:10 dilution, from a ×10 compound plate. Generally, 20 μl of compound was added to each well containing 180 μl of assay buffer, from a height of 150 μl and at a rate of 40 μl/s. Agonist-elicited increase in fluorescence counts correlated with the increase in intracellular calcium. For antagonist studies, an additional 30- to 60-min preincubation period followed the dye-loading procedure. Cells were exposed to the antagonist during preincubation time and throughout agonist activation.
Measurement of Inositol Monophosphate Accumulation.
Confluent cells were harvested and plated in 11-mm-diameter wells (24-well plate) in maintenance medium at an initial density of 1.2 × 105 cells per well and labeled with 2 μCi myo-[3H]inositol/ml for 18 to 24 h. The cells were then preincubated with DMEM containing 25 mM HEPES and 10 mM LiCl for 30 min to inhibit the monophosphatase activity. At the end of the preincubation, the medium was removed, and the cells were incubated with test compounds for an additional 30 min. The reaction was terminated by aspiration of the incubation medium and addition of 0.5 ml of ice-cold 5% perchloric acid to extract the accumulated inositol phosphates. After 15 min at 4°C, 200 ml of 0.5 M Tes/1.5 M K2CO3 was added to each sample to neutralize to pH 7, and samples were centrifuged to separate the liquid and the precipitated salt. The supernatant samples were applied to columns (Dowex AG 1-X8 resin, formate form, 100–200 mesh) to elute the [3H]IP1 fraction. [3H]IP1 in the eluates was quantified with liquid scintillation counting. Compounds undergoing antagonism studies were included during the 30-min preincubation with DMEM/LiCl and throughout the 30-min agonist exposure. The increase in [3H]IP1 counts corresponded to the receptor-mediated agonist response.
Native 5-HT2B Receptor Functional Studies.
Vabicaserin was evaluated in three isolated tissue bath systems that were 5-HT2B receptor-dependent for the 5-HT-stiumlated smooth muscle contraction. Contraction of the rat stomach fundus, the tissue source for the original cloning of the 5-HT2B receptor, was used to assess the effect of vabicaserin on native rat 5-HT2B receptor functional response. Longitudinal strips of stomach fundus from normal male Sprague-Dawley rats (250–300 g) were mounted into isolated tissue baths for the measurement of isometric contraction. Tissues were originally challenged with 67 mM KCl, and all responses were normalized to this contraction. Responses to cumulative addition of agonists were measured in the fundus. Longitudinal muscle preparations from human colon were used to assess the effect of vabicaserin on native human 5-HT2B receptor functional response. Tissue from three donors was mounted in organ baths and evaluated under conditions of submaximal electrical field stimulation (EFS) where neuronally mediated contractile response are potentiated by 5-HT in a concentration- and 5-HT2B-receptor dependent manner (Borman et al., 2002). EFS-evoked contractions (1–2 Hz, 1-ms pulse width, 10-s duration, 15 v, repeated every 60 s for 3 min) were evaluated in the presence of vabicaserin alone (0.1 and 1 μM) to assess its agonist activity or in the presence of agonist stimulation with α-methyl-5HT in combination with vabicaserin to assess its antagonist activity. Contractions to α-methyl-5-HT, in the absence or presence of vabicaserin, were expressed as a percentage of the α-methyl-5-HT maximum response.
Mesenteric artery from the DOCA salt hypertensive rat was used to assess the effect of vabicaserin on pathologically up-regulated 5-HT2B receptors. This preparation exhibits a 5-HT receptor pharmacology that is consistent with activation of 5-HT2B receptors, whereas the 5-HT2A receptor seems to maintain normal functionality in tissue derived from nonhypertensive animals. Superior mesenteric arteries were dissected from sham normotensive rats (systolic blood pressure <130 mm Hg) or rats made hypertensive (> 160 mm Hg) by the administration of DOCA and salt (1% NaCl and 0.2% KCl) in their drinking water. Arterial helical strips, denuded of endothelium, were pair-mounted (sham and DOCA artery strip in the same bath) and originally challenged with phenylephrine (10 μM), and all responses were normalized to this contraction. Responses to cumulative addition of agonists were measured in these tissues.
Receptor Binding Profile of Vabicaserin.
Vabicaserin (Fig. 1) displayed high-affinity binding at the cloned human 5-HT2C receptor (Ki = 3 nM) using the agonist radioligand [125I]DOI. In contrast, vabicaserin exhibited lower affinity at the 5-HT2C antagonist binding site (22 nM) labeled with [3H]mesulergine. Additional binding studies indicated that vabicaserin possessed affinity for the 5-HT2B and 5-HT1A receptors with Ki values of 14 and 112 nM, respectively. Thus vabicaserin was approximately 4-fold selective, in terms of binding affinity, over these related receptors. Vabicaserin was greater than 50-fold selective over other monoamine receptors (5-HT2A, 5-HT1B, 5-HT1D, 5-HT6, 5-HT7, dopamine D2, D3, D4, and the α1 adrenergic receptor binding site) (Table 1) and over a range of 76 receptor, ion channel, and neurotransmitter uptake sites determined in a Novascreen selectivity profile (Supplemental Table 1).
Functional Assessment of Vabicaserin with Recombinant Receptors.
Vabicaserin was a potent and full agonist in stimulating 5-HT2C receptor-coupled calcium mobilization with an estimated EC50 of 8 nM and Emax of 100%, relative to a maximally effective concentration of 5-HT (Fig. 2A). The 5-HT2B receptor potency and efficacy of vabicaserin in transfected cells was found to highly depend on receptor expression levels. Although vabicaserin exhibited no 5-HT2B receptor agonism in cells expressing approximately 500 or 1500 fmol/mg protein, vabicaserin stimulated calcium mobilization in cells expressing the 5-HT2B receptor at levels of 5000 fmol/mg protein (Fig. 2A and Table 2). In contrast, vabicaserin antagonized responses to 5-HT in both 5-HT2A and 5-HT2B (500 fmol/mg) receptor-expressing cells with IC50 values of 1650 and 29 nM, respectively (Fig. 2B). The ability of vabicaserin to stimulate the accumulation of IP1 in 5-HT2B receptor-transfected cells was also found to critically depend on the level of receptor expression. Vabicaserin failed to stimulate IP1 accumulation in 5-HT2B receptor-expressing cells at a level of receptor expression of 500 fmol/mg (data not shown) and exhibited partial agonist activity (Emax, 50%) in cells expressing 1500 or 5000 fmol/mg 5-HT2B receptor (Fig. 3). Consistent with the calcium mobilization studies, vabicaserin was a potent (EC50, 32 nM) and full 5-HT2C receptor agonist for the stimulation of IP1 accumulation and lacked agonist activity at the 5-HT2A receptor subtype in this assay (Fig. 3). Taken together, these functional studies of 5-HT2 receptor subtype activation in transfected cells demonstrate vabicaserin to be a potent 5-HT2C receptor agonist, 5-HT2A antagonist, and 5-HT2B receptor partial agonist, dependent on the level of receptor expression.
Functional Assessment of Vabicaserin with Native 5-HT2B Receptors.
To further understand the activity of vabicaserin at 5-HT2B receptors, a series of studies were conducted in native systems to assess its 5-HT2B receptor functional activity with physiologically and pathophysiologically expressed 5-HT2B receptors. In the isolated rat stomach fundus preparation, 5-HT produced a concentration-dependent contraction with a maximum contraction comparable with that observed with 67 mM KCl (Fig. 4A). In contrast, vabicaserin (0.1 nM–10 μM) did not induce contraction in the rat stomach fundus under identical experimental conditions to those used for 5-HT (Fig. 4A). In experiments evaluating antagonist activity, vabicaserin (0.1 and 1 μM) produced a concentration-dependent rightward shift of a 5-HT concentration effect curve with no suppression of the maximum functional response to 5-HT in rat stomach fundus (Fig. 4B), as did the positive control 5-HT2B receptor antagonist LY272015 (Fig. 4C). The −logEC50 values for 5-HT were as follows: vehicle = 7.80; 0.1 μM vabicaserin = 6.80; 1 μM = 5.97. Taken together, these results indicate that vabicaserin is a competitive antagonist at the rat stomach fundus 5-HT2B receptor.
In the isolated longitudinal muscle preparation from human colon, vabicaserin (0.1 and 1 μM) did not induce contraction under conditions of submaximal EFS (data not shown), conditions known to support a 5-HT2B receptor-mediated contractile response. Vabicaserin antagonized the effects of α-methyl-5-HT on EFS-evoked contractions, causing a parallel rightward shift of the α-methyl-5-HT concentration-effect curve (Fig. 5). The mean pEC50 of α-methyl-5-HT was 8.2 ± 0.1 S.E.M., and the mean dose of antagonist that produced a 2-fold shift of the agonist (pA2) of vabicaserin was 7.5 ± 0.3 S.E.M. The rightward shift in the agonist concentration-response curve was not associated with a suppression of the maximum contractile response to α-methyl-5-HT, suggestive of competitive antagonism, consistent with the profile obtained in the rat stomach fundus preparation.
In the isolated superior mesenteric artery preparation vabicaserin (0.1 nM–10 μM) did not induce contraction in sham-treated rat superior mesenteric artery, but did cause a small, but highly variable, contraction in arteries from DOCA-salt rat superior mesenteric artery (Fig. 6A), with a response observed in only half of the tissues. The contraction observed in the DOCA-salt superior mesenteric artery was less than 20% of that observed with 5-HT or the 5-HT2B receptor agonist BW723C86 and was not blocked by the 5-HT2B antagonist LY272015 (Fig. 6B). In antagonist studies, vabicaserin (1 μM) produced an approximate 6-fold rightward shift in the 5-HT concentration-effect curve with no suppression of the maximum response to 5-HT in sham-treated rat superior mesenteric artery (Fig. 7A), indicative of competitive 5-HT2A receptor antagonism (Watts et al., 1996). Vabicaserin also produced a rightward shift in the 5-HT concentration-effect curve accompanied by a reduced maximum functional response to 5-HT in DOCA-salt-treated rat superior mesenteric artery (Fig. 7B) (Watts et al., 1996; Watts and Fink, 1999). The −logEC50 values for 5-HT were as follows: vehicle = 7.06; 1 μM = 5.97. Taken together, these results suggest that vabicaserin functions as an antagonist at native 5-HT2A and 5-HT2B receptors expressed in the sham normotensive and DOCA-salt hypertensive-derived rat superior mesenteric artery, respectively.
The 5-HT2C receptor subtype has been implicated as a target for therapeutic intervention in a broad range of disorders, leading to intense drug discovery efforts toward the development of 5-HT2C receptor-selective agonists (Wacker and Miller, 2008). WAY-163909 (Dunlop et al., 2005), CP-809101 (Siuciak et al., 2007), and lorcaserin (Thomsen et al., 2008) have been reported as potent and selective 5-HT2C receptor agonists with preclinical profiles consistent with their potential therapeutic utility in schizophrenia, depression, and obesity. Specifically, WAY-163909 (Marquis et al., 2007) and CP-809121 (Siuciak et al., 2007) exhibit activity in a broad range of animal models predictive of antipsychotic like activity, and, in addition, WAY-163909 is efficacious in a number of rodent models of antidepressant-like effects (Rosenzweig-Lipson et al., 2007). One of the greatest challenges for discovery programs aimed at this target is identifying compounds with selectivity over the 5-HT2B receptor, specifically compounds devoid of 5-HT2B receptor agonist activity. This is important because activation of the 5-HT2B receptor has been implicated in both primary pulmonary hypertension (Launay et al., 2002) and valvulopathy (Fitzgerald et al., 2000; Rothman et al., 2000). Based on our expansion of synthetic efforts around WAY-163909, we designed a novel class of tetrahydroquinoline-fused diazepines leading to the discovery of vabicaserin, a potent and selective 5-HT2C receptor agonist without 5-HT2B agonism (Ramamoorthy et al., 2006).
As with any drug discovery effort it is important to appropriately assess the activity and selectivity of the candidate molecule, in particular with respect to closely related receptor targets. In studies using heterologously expressed 5-HT2 receptor subtypes in stably transfected cell lines, vabicaserin bound with high affinity to the 5-HT2C receptor (Ki = 3 nM) and demonstrated >50-fold selectivity versus the 5-HT2A receptor (5-HT2A Ki = 152 nM) and only 4-fold selectivity versus the 5-HT2B receptor (Ki = 14 nM). Initial characterization of the functional activity of vabicaserin at the 5-HT2A and 5-HT2B receptor demonstrated that vabicaserin did not demonstrate agonist activity at either receptor subtype (EC50 >5 μM) and functioned as either a potent antagonist at the 5-HT2B receptor (IC50 = 29 nM) or a weakly potent antagonist at the 5-HT2A receptor (IC50 = 1.65 μM).
Because activation of the 5-HT2B receptor has been implicated in primary pulmonary hypertension and valvulopathy (Fitzgerald et al., 2000; Rothman et al., 2000), a clear understanding of the functional activity of a compound at this receptor is critical, and our subsequent evaluation of vabicaserin focused on this key question. However, it is important to point out that although we did not evaluate 5-HT2A receptor-mediated functional responses at higher heterologous expression levels, we found the 5-HT2A FLIPR assay to be highly predictive of subsequent 5-HT2A receptor-mediated agonist response in the sham superior mesenteric artery, with compounds exhibiting >20% agonist activity in the FLIPR assay inducing sham artery contraction (J. Dunlop and S. Rosenzweig-Lipson, unpublished observations). Despite the initial demonstration that vabicaserin functioned as an antagonist and not an agonist at the 5-HT2B receptor, the absence of a high level of binding selectivity of vabicaserin for the 5-HT2C receptor relative to the 5-HT2B receptor led to the development and execution of a comprehensive strategy for assessing the activity of a compound at the 5-HT2B receptor using both recombinantly and natively expressed 5-HT2B receptors. Our initial characterization used a number of CHO cells lines exhibiting different expression levels of the 5-HT2B receptor. In these studies vabicaserin exhibited 5-HT2B receptor antagonist activity in CHO cells where the receptor expression level was 500 fmol/mg. In cells where the receptor was expressed at higher levels (1500 and 5000 fmol/mg), vabicaserin exerted partial agonist activity. Significantly, in the calcium mobilization functional assay, vabicaserin was a potent 5-HT2B receptor antagonist in cells with the lowest level of 5-HT2B receptor expression, with an IC50 value determined as 29 nM, in excellent correlation with its measured binding affinity in ligand displacement studies for the 5-HT2B receptor. In studies of IP1 accumulation in 5-HT2B receptor-expressing cells, vabicaserin lacked activity in cells with the low level of receptor expression and was a partial 5-HT2B receptor agonist in cells expressing 1500 and 5000 fmol/mg, respectively. The explanation for the mixed 5-HT2B receptor pharmacology of vabicaserin in transfected cells probably is based on the well established property in G protein-coupled receptor pharmacology that overexpression of receptors can frequently lead to increases in apparent agonist potency and/or efficacy, providing highly sensitive assays of agonist functional responses. However, this has led to some caution in the interpretation of agonist-dependent effects in transfected systems, especially when taken in the absence of effects in native systems. Because our studies in the recombinant cell system with the lowest level of receptor expression supported the pharmacological activity of vabicaserin as a 5-HT2B receptor antagonist, whereas partial agonist activity was apparent in cells with higher levels of receptor expression, we adopted a strategy using a number of systems with native expression of the receptor to further investigate these findings.
Native expression systems where receptors are expressed in a physiologically relevant environment allow for an evaluation of agonist intrinsic activity in the absence of receptor overexpression. In the case of the 5-HT2B receptor subtype, the isolated rat stomach fundus preparation, a key tissue source for the 5-HT2B receptor, has been well validated as a model of 5-HT-dependent receptor activation that is mediated by the 5-HT2B receptor (Baxter et al., 1994). Using this model we found that although 5-HT produced a concentration-dependent contraction, vabicaserin failed to exhibit agonist activity, but behaved as a classic competitive 5-HT2B receptor antagonist producing parallel rightward shifts in the 5-HT concentration-response curve. To assess the effect of vabicaserin on the human 5-HT2B receptor in native tissue, we used the isolated colonic longitudinal muscle preparation. Studies using mRNA and protein localization, in combination with functional receptor pharmacology, have provided evidence for a predominant 5-HT2B receptor-dependent functional contraction in this preparation (Borman et al., 2002). Similar to our data obtained using the rat stomach fundus preparation, α-methyl-5-HT, but not vabicaserin, produced a concentration-dependent contraction of the colon isolated longitudinal muscle, demonstrating a lack of agonist activity of vabicaserin in this native model of human 5-HT2B receptor activation. In contrast, vabicaserin produced parallel rightward shifts in the α-methyl-5-HT concentration response curve, which is consistent again with its property as a 5-HT2B receptor competitive antagonist.
In the rat mesenteric artery, 5-HT receptors mediating contractile responses have been shown, based on pharmacological characterization, to represent a 5-HT2A receptor-dependent response under physiological conditions. In contrast, under pathophysiological conditions arising from the administration of DOCA and salt (1% NaCl and 0.2% KCl) in their drinking water, rats become hypertensive, and the contractile response measured in isolated mesenteric artery from DOCA-salt-treated animals represents a 5-HT2B receptor-dependent response (Watts et al., 1996). The role of the 5-HT2B receptor in the pathophysiology of hypertension in these animals is further supported by the demonstration of the antihypertensive activity of the 5-HT2B receptor antagonist LY-272015 in DOCA-salt treated rats (Watts and Fink, 1999). Using this model we were able to evaluate vabicaserin for functional effects on native 5-HT2A receptors under physiological conditions and functional effects on up-regulated 5-HT2B receptors under pathophysiological conditions. Vabicaserin failed to induce contractions in sham normotensive mesenteric artery and produced small variable contractions in the DOCA-salt-hypertensive mesenteric artery in a fraction of the preparations studied. Follow-up studies indicated these responses to be small in magnitude compared with the 5-HT2B receptor agonist BW723C86 and relatively unresponsive to the presence of the 5-HT2B receptor antagonist LY272015. In experiments evaluating antagonist activity, vabicaserin produced a rightward-shift in the 5-HT concentration-response curve in both the sham and DOCA-salt-derived mesenteric artery, consistent with 5-HT2A and 5-HT2B receptor antagonism, respectively. In the latter case, this effect was accompanied by a suppression in the maximum response to 5-HT, in contrast to the profiles observed in rat stomach fundus and human colon, where the vabicaserin profile was more reminiscent of a competitive 5-HT2B receptor antagonist. The reasons for the different profile observed in the DOCA-salt mesenteric artery preparation are unclear, but a speculation is that the enhanced efficacy of 5-HT in causing arterial contraction is purely 5-HT2B receptor mediated, hence the significant reduction in 5-HT-induced contraction in arteries from the DOCA-salt but not sham animals. Finally, it is important to highlight that whereas the studies in rat mesenteric artery provide a robust in vitro assessment of the functional activation of 5-HT2B receptors in this cardiac preparation, they do not definitively address activation of 5-HT2B receptors in situ in the heart, studies that are beyond the scope of the current study.
Taken in combination, the functional studies with native 5-HT2B receptors in a number of different preparations indicate vabicaserin to be a 5-HT2B receptor antagonist, in agreement with studies performed in 5-HT2B receptor-transfected cells expressing 500 fmol/mg receptor. In 5-HT2B receptor-transfected cells with higher levels of receptor expression vabicaserin exhibited 5-HT2B receptor partial agonist activity. Ultimately, the pharmacology of vabicaserin will depend on many characteristics of the tissue and cells with which it interacts, including receptor expression level, the complement and expression level of components of the receptor signaling pathways, and the efficiency of receptor and effector coupling. In summary, we developed a comprehensive strategy for the assessment of agonist selectivity of 5-HT2C receptor agonists with respect to the related 5-HT2B receptor subtype and described vabicaserin as a novel and selective 5-HT2C receptor agonist.
Participated in research design: Dunlop, Watts, Barrett, Coupet, Pangalos, Schechter, and Rosenzweig-Lipson.
Conducted experiments: Mazandarani, Nawoschik, Smith, J. Zhang, and G. Zhang.
Contributed new reagents or analytic tools: Harrison, Ramamoorthy, and Stack.
Wrote or contributed to the writing of the manuscript: Dunlop, Watts, and Rosenzweig-Lipson.
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
- α-methyl-5-(2-thienylmethoxy)-1H-indole-3-ethanamine hydrochloride
- deoxycorticosterone acetate
- fluorometric imaging plate reader
- inositol monophosphate
- 6-methyl-1,2,3,4-tetrahydro-1-[3,4-dimethoxyphenylmethyl-9H-pyrido[3,4b]indole] hydrochloride
- vabicaserin (SCA-136)
- (−)-4,5,6,7,9,9a,10,11,12,12a-decahydrocyclopenta[c] [1,4]diazepino[6,7,1-ij]quinoline hydrochloride
- Chinese hamster ovary
- Dulbecco's modified Eagle's medium
- 2-[(3-chlorophenyl)methoxy]-6-(1-piperazinyl)pyrazine hydrochloride
- electrical field stimulation.
- Received January 18, 2011.
- Accepted February 22, 2011.
- Copyright © 2011 by The American Society for Pharmacology and Experimental Therapeutics