Abstract
Histamine H3 receptor inverse agonists are known to enhance the activity of histaminergic neurons in brain and thereby promote vigilance and cognition. 1-{3-[3-(4-Chlorophenyl)propoxy]propyl}piperidine, hydrochloride (BF2.649) is a novel, potent, and selective nonimidazole inverse agonist at the recombinant human H3 receptor. On the stimulation of guanosine 5′-O-(3-[35S]thio)triphosphate binding to this receptor, BF2.649 behaved as a competitive antagonist with a Ki value of 0.16 nM and as an inverse agonist with an EC50 value of 1.5 nM and an intrinsic activity ∼50% higher than that of ciproxifan. Its in vitro potency was ∼6 times lower at the rodent receptor. In mice, the oral bioavailability coefficient, i.e., the ratio of plasma areas under the curve after oral and i.v. administrations, respectively, was 84%. BF2.649 dose dependently enhanced tele-methylhistamine levels in mouse brain, an index of histaminergic neuron activity, with an ED50 value of 1.6 mg/kg p.o., a response that persisted after repeated administrations for 17 days. In rats, the drug enhanced dopamine and acetylcholine levels in microdialysates of the prefrontal cortex. In cats, it markedly enhanced wakefulness at the expense of sleep states and also enhanced fast cortical rhythms of the electroencephalogram, known to be associated with improved vigilance. On the two-trial object recognition test in mice, a promnesiant effect was shown regarding either scopolamine-induced or natural forgetting. These preclinical data suggest that BF2.649 is a valuable drug candidate to be developed in wakefulness or memory deficits and other cognitive disorders.
The cerebral histaminergic neurons seem to play a critical role in the maintenance of wakefulness and higher cerebral functions, e.g., attention or learning (for review, see Schwartz et al., 1991; Haas and Panula, 2003). Hence, drug-induced activation of histaminergic neurotransmission in the central nervous system represents a promising therapeutic target in a large variety of neuropsychiatric disorders in which these functions are compromised and for which available therapeutic opportunities are limited in this respect (Schwartz and Arrang, 2002).
Stimulation of postsynaptic H1 and/or H2 receptors by agonists is, however, not acceptable due to unavoidable and detrimental actions of these drugs at peripheral, i.e., mainly cardiovascular and gastric targets. In contrast, presynaptic H3 receptors are almost exclusively expressed in the central nervous system, and their blockade by drugs such as thioperamide markedly enhances the activity of histaminergic neurons, as shown namely by the increases in histamine (HA) release and turnover in rodent brain (Arrang et al., 1987; Garbarg et al., 1989; Itoh et al., 1991; Mochizuki et al., 1991). The use of this prototypical autoreceptor antagonist in numerous studies to enhance brain HA release has been instrumental in establishing the roles of tuberomammillary neurons in behavior (for review, see Onodera et al., 1994; Hill et al., 1997; Brown et al., 2001). So far, however, neither this drug nor any other H3 receptor antagonist has been approved for clinical use, in spite of numerous potent compounds being synthesized and tested preclinically (for review, see Celanire et al., 2005; Leurs et al., 2005).
Whereas thioperamide and the first generation of its congeners were imidazole derivatives, a structure to which are attributed toxicity and bioavailability problems, a second generation of compounds was designed in which the imidazole nucleus could be replaced to avoid these drawbacks (Ganellin et al., 1998). In addition, during the recent years, the pharmacochemistry in this field has shown important evolutions.
In agreement, with the identification of a cDNA encoding the human receptor (Lovenberg et al., 1999), the testing of large numbers of compounds became easier than when it depended upon the availability of fresh tissue (Arrang et al., 1988); large differences in pharmacology between rodent and human H3 receptors, attributable to only two-amino acid difference in sequences, were uncovered (Ligneau et al., 2000); the association of this heptahelical receptor with a Gi/Go protein (Clark and Hill, 1996; Takeshita et al., 1998) was confirmed; modeling of the H3 receptor protein, derived from the crystal structure of bovine rhodopsin, and docking of compounds became feasible (Stark et al., 2001; Lorenzi et al., 2005); constitutive activity of the native receptor in brain and the requirement for inverse agonism in drugs to enhance endogenous HA release were evidenced (Morisset et al., 2000).
BF2.649 (Fig. 1), a potent H3 receptor antagonist/inverse agonist, was designed and selected, taking into account all of this recent information to meet the challenge of bringing one such compound to clinical evaluation.
Materials and Methods
[3H]Histamine Release from Synaptosomes. [3H]HA release experiments were performed as described previously (Ligneau et al., 1998). In brief, a crude synaptosomal preparation from rat cerebral cortex was preincubated for 30 min with 0.4 μM [3H]l-histidine at 37°C. After extensive washing, synaptosomes were resuspended in fresh 2 mM K+ Krebs-Ringer medium in the presence of the appropriate drugs. After a 5-min incubation, synaptosomes were depolarized bringing the K+ concentration to 30 mM for 2 min. Incubations were ended by a rapid centrifugation, and [3H]HA levels in the supernatant were determined after an ion exchange chromatography purification. Release was expressed as percentage of total [3H]HA initially present in the synaptosomal preparation. Typically total [3H]HA represented 3500 dpm/mg protein, and total radioactivity approximately 100,000 dpm/mg protein in the test tube.
Cloning and Stable Expression of the Human H3 Receptor cDNAs. The human H3 receptor was cloned and expressed as described previously (Ligneau et al., 2000; Gbahou et al., 2006). In brief, a human striatal cDNA library was screened using primers based on the sequence of the human H3 receptor (Lovenberg et al., 1999). Clones exhibiting a full-length cDNA sequence corresponding to the human H3 receptor were obtained. The cDNA was ligated into the mammalian expression vector pCIneo (Promega, Charbonnières, France).
CHO-K1 or rat glioma C6 cells were transfected using Superfect (Qiagen, Courtaboeuf, France), and HEK-293 cells were transfected using Lipofectamine (Invitrogen, Cergy Pontoise, France). Stable transfectants were selected with 1 mg/ml Geneticin, tested for their expression levels of [125I]iodoproxyfan binding sites, and maintained in the presence of 1 mg/ml Geneticin.
[35S]GTPγS Binding to the Human Recombinant H3 Receptor. [35S]GTPγS binding assays were performed according to Rouleau et al. (2002). CHO-K1 cells stably expressing the human H3 receptor (∼400 fmol/mg protein) were homogenized in ice-cold buffer (50 mM Tris/HCl, pH 7.4). Homogenates were centrifuged twice (20,000g for 10 min at 4°C), and the final pellet was resuspended in 50 volumes of buffer. Membranes (550 μg of protein) were pretreated with adenosine deaminase (1 U/ml) and incubated for 60 min at 25°C with 0.1 nM [35S]GTPγS and the drugs to be tested in a final volume of 1 ml of assay buffer (50 mM Tris/HCl, 50 mM NaCl, 5 mM MgCl2, 10 μM GDP, and 0.02% bovine serum albumin, pH 7.4). The nonspecific binding was determined using 10 μM nonradioactive GTPγS. Incubations were stopped by rapid filtration under vacuum through GF/B glass fiber filters (Whatman, Maidstone, UK). After washing with ice-cold water, the radioactivity trapped on filters was counted by liquid scintillation spectrometry.
A similar assay was used to assess competitive antagonism. In brief, membranes (10 μg of protein) of HEK-293 cells stably expressing the human H3 receptor (∼600 fmol/mg protein) were preincubated in presence of the drugs (imetit and BF2.649) in the buffer (50 mM Tris/HCl, pH 7.4, 10 mM MgCl2, 100 mM NaCl, and 10 μM GDP) in a 96-well microplate under gentle agitation at room temperature (19–20°C) for 30 min before the addition of 0.1 nM [35S]GTPγS (final volume 200 μl). The nonspecific binding was determined using a 10 μM concentration of nonradioactive GTPγS. After 30 min, incubations performed in triplicate were stopped by rapid filtration under vacuum on a Multiscreen MAFCOB50 microplate (Millipore Corporation, Billerica, MA). Radioactivity trapped on filters was counted by liquid scintillation spectrometry.
Histamine H3 Receptor Assay on Guinea Pig Ileum. The procedure used was as described previously (Ligneau et al., 1998). In brief, longitudinal muscle strips from small intestine from guinea pig were dissected out and incubated in a gassed O2/CO2 (95%/5%) modified Krebs-Ringer bicarbonate medium at 37°C in presence of 1 μM mepyramine to block H1 receptors. After equilibration, contractile activity under stimulation (rectangular pulses of 15 V, 0.5 ms, and 0.1 Hz) was recorded. Concentration-response curves of the effect of (R)-α-MeHA alone or together with BF2.649 were performed.
[125I]Iodoproxyfan Binding Assay. Native human brain cortex membranes were prepared as follows. A human cortical sample collected during surgery of the firing area of a drug-resistant epileptic patient was obtained from the Neurosurgery Service of Sainte-Anne Hospital, Paris (courtesy of Dr. J.-P. Chodkiewicz). Immediately after surgery, the tissue was homogenized in the ice-cold phosphate buffer (50 mM Na2HPO4/KH2PO4, pH 7.5), centrifuged (140g for 10 min at 4°C), and the supernatant obtained was centrifuged (23,000g for 30 min at 4°C). The final pellet was resuspended in 3 ml of the phosphate buffer by a rapid sonication to get the membrane preparation. Cells expressing the H3 receptor were washed and homogenized with a Polytron (Kinematica, Lucerne, Switzerland) in ice-cold phosphate buffer and centrifuged (23 000g, 30 min, +4°C).
Mouse brain cortex was homogenized with a Polytron in the ice-cold phosphate buffer, centrifuged (140g for 10 min at 4°C), and the supernatant was centrifuged (23,000g for 30 min at 4°C). Final pellets were resuspended in the phosphate buffer to get membrane preparations.
Binding experiments were performed according to Ligneau et al. (1998) with slight modification. Membranes (5–15 μg of protein/incubation) were incubated for 1 h at 37°C alone or together with BF2.649 in increasing concentrations in a final volume of 200 μl in the presence of 25 pM [125I]iodoproxyfan. Nonspecific binding was determined using 1 μM imetit, a specific H3 receptor agonist. Incubations, performed at least in triplicate, were stopped by rapid filtration on Whatman GF/B glass fiber membranes. Radioactivity trapped on filters was directly counted using a gamma counter.
[3H]Mepyramine Binding Assay. The human H1 receptor was amplified from human genomic DNA and cloned into the expression vector pCDNA3.1-HisC (Invitrogen). HEK-293 cells were transfected, and stable transfectants were selected. Cells were washed using a phosphate-buffered saline buffer and disrupted using a Polytron in ice-cold phosphate buffer (50 mM Na2HPO4/KH2PO4, pH 7.5). The homogenate was centrifuged (23,000g for 30 min at 4°C), and the pellet was resuspended in the phosphate buffer by a short sonication. Membranes were stored at –80°C until use. After thawing and rapid sonication, membranes (10 μg of protein/incubation) were incubated in a 96-well microplate under gentle agitation for 1 h at room temperature (19–20°C), alone or together with BF2.649 in increasing concentrations in a final volume of 200 μl in the presence of 1 nM [3H]mepyramine. Nonspecific binding was determined using 3 μM triprolidine, a selective H1 receptor antagonist. Incubations performed at least in triplicate were stopped by rapid filtration on 1450-521 glass fiber membranes (PerkinElmer Wallac, Turku, Finland). The trapped radioactivity was counted by scintillation spectrometry using a solid scintillant. The expression level of the human H1 receptor was 1652 ± 74 fmol/mg protein, and [3H]mepyramine as radioligand presented a Kd value of 513 ± 29 pM.
[125I]Iodoaminopotentidine Binding to the H2 Receptor. A membrane preparation of guinea pig striatum obtained in ice-cold phosphate buffer (50 mM Na2HPO4/KH2PO4, pH 7.4) was incubated together with 10 pM [125I]iodoaminopotentidine and increasing BF2.649 concentrations in the same buffer. Nonspecific binding was determined using 100 μM tiotidine, a selective H2 receptor antagonist (MDS Panlabs assay 239700, Spectrumscreen report 1009464; MDS Panlabs, Taipei, Taiwan).
[3H]HA Binding to the H4 Receptor. The human histamine H4 receptor was cloned and transiently expressed in COS-1 cells. Cells were harvested and homogenized in ice-cold buffer (50 mM Tris/HCl and 5 mM MgCl2, pH 7.5) using a Polytron. Membranes (100 μg of protein/incubation) were incubated for 1 h at 37°C in the buffer together with 10 nM [3H]HA and increasing BF2.649 concentrations (final volume 1 ml). Nonspecific binding was determined using 1 μM clobenpropit. Incubations were stopped by rapid filtration on Whatman GF/C glass fiber membrane presoaked in 0.3% of polyethylenimine. Radioactivity trapped on filters was counted by liquid scintillation spectrometry.
Radioreceptor Assay of H3 Receptor Ligands in Serum. Male OF1 mice (18–20 g; Charles River, L'Arbresle, France) were fasted for 16 h before BF2.649 administration. At various times after treatment, animals were sacrificed, and the blood was collected at 4°C. Serum obtained after centrifugation (100g for 10 min at 4°C) was stored at –20°C before BF2.649 measurement by a radioreceptor assay (Ligneau et al., 1998) using [3H](R)-α-MeHA as radioligand and the recombinant human H3 receptor. In brief, 1-ml aliquots of a membrane preparation of CHO-K1 cells stably expressing the human H3 receptor were incubated for 60 min at 25°C with 1 nM [3H](R)-α-MeHA alone or together with different concentrations of BF2.649 in diluted serum of BF2.649-free mice (standardization curve) or with diluted serum samples of BF2.649-treated mice. Specific binding was defined as that inhibited by 10 μM ciproxifan. Incubations performed at least in triplicate were ended by rapid filtration on Whatman GF/C filters, and radioactivity trapped on filters after extensive washings was counted by liquid scintillation spectrometry.
Assay ofTele-Methylhistamine in Brain. Male OF1 mice (18–20 g; Charles River) were fasted for 16 h before p.o. administration. After treatments, animals were sacrificed. The brain was dissected out and homogenized in 10 volumes (w/v) of ice-cold 0.4 N perchloric acid. The clear supernatant obtained after centrifugation (2000g for 30 min at 4°C) was stored at –20°C before measuring the t-MeHA level by enzymoimmunoassay as described previously (Ligneau et al., 1998).
Microdialysis. Male Wistar rats (275–325 g; R. Janvier, Le Genest Saint Isle, France) were anesthetized with chloral hydrate (Carlo Erba; 400 mg/kg i.p.) and further mounted in a Kopf stereotaxic frame. A CMA/12 guide cannula (CMA Microdialysis; Phymep, Paris, France) was implanted into the medial prefrontal cortex (AP, +3.2 from bregma; ML, +0.6; DV, –2.0 mm from dura) according to the atlas of Paxinos and Watson (1998) and further secured with dental cement and anchor screws into the skull. Animals were housed singly, and 5 days was allowed for postoperative recovery.
Between 9:00 and 10:30 AM on the day of the experiment, each animal was transferred into a freely moving animal system, and a CMA/12 microdialysis probe (3-mm membrane length; CMA Microdialysis) was slowly lowered into the guide cannula and allowed to recover for 2 h. Probes were perfused with artificial cerebrospinal fluid perfusion fluid (147 mM NaCl, 2.7 mM KCl, 1.2 mM CaCl2, and 0.85 mM MgCl2, pH 7.4; CMA Microdialysis) at a flow rate of 1 μl/min (model PHD 2000 syringe pump; Harvard Apparatus Inc., Holliston, MA). In experiments aimed at determination of microdialysate acetylcholine (ACh), a 50 nM concentration of the acetylcholinesterase inhibitor neostigmine was added to the cerebrospinal fluid perfusion fluid. The first three 30-min dialysate samples were discarded, and the three subsequent samples were collected to provide a baseline. Ten minutes before the end of the basal fraction collection, vehicle (1 ml/kg saline) or BF2.649 (10 mg/kg) was administered intraperitoneally, and five further 30-min dialysates were collected.
Electrochemical Quantification of Extracellular Levels of Dopamine and ACh. The 30-μl dialysate samples were analyzed either for dopamine or ACh levels with high-performance liquid chromatography-electrochemical detection (Waters, St. Quentin en Yvelines, France). For dopamine detection, the mobile phase (3% methanol, 97% buffer containing 50 mM citric acid, 70 μM EDTA, 90 μM octane sulfonic acid, and 2.3 mM NaCl, pH 2.9) was pumped through an Atlantis dC18 analytical column (2.1 × 150 mm, 3 μm; Waters) at a flow rate of 0.25 ml/min. Eluates were quantified at 40°C using an amperometric detector set at 600 mV [2465 electrochemical detector fitted with a 2-mm glassy carbon working cell with an in situ Ag/AgCl (ISAAC; Waters) reference electrode and a 50-μm spacer; Waters]. For ACh detection, the mobile phase (50 mM NaH2PO4, 500 μM EDTA, and 0.5% ProClin in liquid chromatography-grade water, pH 8.5) was pumped through an UniJet microbore ACh/choline analytical column (530 × 1 mm) coupled to an enzymatic postcolumn (50 × 1 mm) (Bioanalytical Systems, Kenilworth, UK) at a flow rate of 0.13 ml/min. Eluates were quantified at 40°C with the amperometric detector set at 500 mV (with a platinum cell, an in situ Ag/AgCl reference electrode, and a 50-μm spacer). A standard sample was included every six to eight samples to enable quantification and check for reproducibility. The limit of detection for dopamine and ACh was 0.65 and 3.4 fmol/μl, respectively.
Analysis of Sleep-Wake Cycle and Neocortical Power Spectral Density in Cats. Animals were chronically implanted with cortical and muscle electrodes to record neocortical and hippocampal electroencephalogram (EEG), ponto-geniculo-occipital activity, electromyogram, and electrooculogram, respectively, to allow monitoring their sleep-wake cycle. After recovery from surgery, animals were housed individually in a sound-proof and dimly illuminated recording cage maintained at a 25 ± 2°C on a 12-h light/dark cycle, and they were fed daily at 6:00 PM. Polygraphic recordings started 7 days after surgery, and they were performed during 4 days to collect reference qualitative and quantitative data of sleep-wake control in each individual. Then, evaluation of responses to BF2.649 was performed over 24-h sessions and analyzed by successive 30-s epochs.
Sleep-wake recording criteria included wakefulness (W), light slow-wave sleep (SWS1), deep SWS (SWS2) and paradoxical sleep (PS) according to Lin (2000). Spectral analysis of EEG was performed by Fast-Fourier Transform in the frequency bands 0.8 to 2.5 Hz (δ slow waves), 8 to 15 Hz (spindles), and 20 to 45 Hz (β and γ rapid waves). BF2.649 was administered orally (at 11:00 AM) at doses of 5 or 10 mg/kg.
Two-Trial Object Recognition Test in Mice. This test was performed using male C57BL/6J mice (12 weeks old; R. Janvier) according to Bertaina-Anglade et al. (2006) under two different experimental conditions: first, in a situation of natural forgetting; and second, in a situation where animals were treated with scopolamine to induce memory deficit. In the two experimental conditions, mice were submitted to a habituation session the day before the object recognition test, and they were allowed to explore a wooden arena for 30 min in presence of two objects (objects not used later for evaluation of memory performance). On experimental day, mice were submitted to two trials spaced by an intertrial interval.
In the natural forgetting situation, during the first trial (acquisition trial, T1), mice were placed in the arena containing two identical objects for an amount of time necessary to explore these two objects for a total of 8 s. Exploration was defined as directing the nose at distance lower than 2 cm from the object and/or touching the object. Three hours later, the second trial (testing trial, T2) was performed by introducing the mice in the same arena with one of the two objects presented in the first trial being replaced by an unknown novel object. Mice were left in the arena for 5 min, and the time they spent in exploration of each object together with locomotor activity was determined. The following parameters were measured: time required to achieve 8 s of object exploration during T1, time spent in active exploration of the familiar object during T2, time spent in active exploration of the novel object on T2, and locomotor activity on T2 (assessed by the number of lines crossed per minute). The ability of BF2.649 (15 mg/kg i.p., 30 min before T1) to delay the natural forgetting was investigated. Tacrine (0.25 mg/kg i.p.), an acetylcholinesterase inhibitor, was administered just after T1 completion.
In the situation of scopolamine-induced cognitive deficit, mice were submitted to two trials spaced by an interval of 60 min. During the first trial (T1), mice were placed in the arena containing the two identical objects for an amount of time necessary to explore the two objects for a total of 20 s. For the second trial (T2), one of the objects presented in the first trial was replaced by an unknown novel object. Mice were placed in the arena for 5 min, and the time they spent in exploration of each object together with locomotor activity was determined. The following parameters were measured: time required to achieve 20 s of object exploration during T1, time spent in active exploration of the familiar object during T2, time spent in active exploration of the novel object on T2, and locomotor activity on T2. The ability of BF2.649 (5 and 15 mg/kg i.p.; 40 min before T1) to reverse experimental memory deficit induced by scopolamine (0.3 mg/kg i.p.; 30 min before T1) was investigated.
Analysis of Data. Maximal effects and ED50, EC50, and IC50 values were determined with an iterative computer least-squares method derived from that of Parker and Waud (1971) using the following nonlinear regression:
Ki values were calculated from IC50 values, assuming a competitive antagonism and using the relationship (Cheng and Prussoff, 1973): where S and Kd represent either the concentration and the dissociation constant of the radioligand, respectively (in binding experiments), or the concentration of HA and its EC50, respectively (in [3H]HA release experiments). When a fixed concentration of BF2.649, tested as an antagonist, was added to imetit in increasing concentrations, the Ki value of BF2.649 was calculated using the following equation (Cheng and Prussoff, 1973): where EC50 and EC50′ are the imetit concentrations required to obtain half-maximal inhibition of release in the absence and presence of BF2.649, respectively, and I is the BF2.649 concentration.
Animals. Animals were housed in group under a 12-h light/dark cycle (lights on 7:00 AM) in a temperature-(21 ± 2°C) and humidity (45 ± 15%)-controlled environment with free access to food and water. Drug administrations were performed in either 1% methylcellulose for the oral route or in 0.9% NaCl for intravenous and intraperitoneal routes. All drug doses are expressed as free base of compound except in the study performed in cat in which BF2.649 doses are as hydrochloride salt. All experiments carried out in the present study were conducted in accordance with international European ethical standards (86/609-EEC) and the French National Committee (décret 87/848) for the care and use of laboratory animals.
Radiochemicals and Drugs. [125I]Iodoproxyfan (2000 Ci/mmol) was prepared as described previously (Ligneau et al., 1998). [125I]NaI and [3H](R)-α-MeHA (specific activities at reference date of 2000 and 38 Ci/mmol, respectively) were from GE Healthcare (Orsay, France). [3H]Mepyramine (20 Ci/mmol), [3H]histamine (42 Ci/mmol), and [35S]GTPγS (1250 Ci/mmol) were provided by PerkinElmer (Courtaboeuf, France). Drugs and their sources were as follows. BF2.649 synthesis (Free University, Berlin, Germany or Interquim, Barcelona, Spain) was prepared as described previously (Meier et al., 2001). Ciproxifan, thioperamide, and (R)-α-MeHA were from Laboratoire Bioprojet (Paris, France). Imetit was synthesized at University College (London, UK). Triprolidine was a kind gift from GlaxoSmithKline (Uxbridge, Middlesex, UK). Histamine, dopamine, dihydroxyphenyl acetic acid, and neostigmine were from Sigma (Isle d'Abeau, France). ACh and choline were provided by Bioanalytical Systems. All other chemicals were obtained from commercial sources, and they were of the highest purity available.
Results
Effect of BF2.649 on [125I]Iodoproxyfan Binding. BF2.649 displaced [125I]iodoproxyfan binding from mouse brain cortical membranes with an IC50 value of 26.4 ± 4.5 nM. Taking into account the Kd value of the radioligand (161 ± 9 pM), the deduced Ki value for BF2.649 was 14 ± 1 nM (Table 1).
Using membranes from human cerebral cortex, an IC50 value of 5.3 ± 2.2 nM was found in a single experiment with four drug concentrations studied in triplicates. Likewise, BF2.649 displaced [125I]iodoproxyfan binding from membranes of rat glioma C6 cells stably expressing the human H3 receptor with an IC50 value of 4.2 ± 0.2 nM. Taking into account the Kd value of the radioligand (50 ± 4 pM), the deduced Ki value for BF2.649 was 2.7 ± 0.5 nM (Table 1).
Effects of BF2.649 on H3 Receptor-Mediated Responses in Vitro. Histamine (1 μM) inhibited the depolarization-induced [3H]HA release from rat cerebral cortex synaptosomes with a maximal effect of 47 ± 3%. BF2.649 progressively reversed this response with a Hill coefficient close to unity and an IC50 value of 330 ± 68 nM (Fig. 2), leading to a Ki value of 17 ± 4 nM (Table 1). At high drug concentrations, the release was over (by 15–20%) the basal release measured in the absence of added HA.
The basal-specific [35S]GTPγS binding to membranes of HEK-293 cells stably expressing the human H3 receptor represented 25.9 ± 0.7 fmol/mg protein. Imetit, a selective H3 receptor agonist, induced a concentration-dependent increase in [35S]GTPγS binding with a maximal effect (152 ± 5%) elicited at a 10 nM concentration, an EC50 of 0.79 ± 0.55 nM, and a pseudo-Hill coefficient not significantly different from unity (0.83 ± 0.44) (Fig. 3). BF2.649 (1, 3, and 10 nM) induced a parallel rightward shift of the concentration-response curve of imetit without modifying its maximal response (Fig. 3). The IC50 values obtained for imetit with the different BF2.649 concentrations (1, 3, and 10 nM) allowed the calculation of a Ki value of 0.16 ± 0.03 nM. This value for BF2.649 was in line with the pKB value of 9.5 (KB = 0.31 nM) provided by the Schild plot analysis (Fig. 3, inset).
The effect of BF2.649 in this functional assay was also investigated using CHO-K1 cells stably expressing the human H3 receptor at a level of ∼400 fmol/mg protein. BF2.649 elicited a dose-dependent decrease of the basal-specific [35S]GTPγS binding to membranes with a maximal effect corresponding to 75 ± 1% of the basal-specific binding and an EC50 value of 1.5 ± 0.1 nM (Fig. 4; Table 1). In comparison, the prototypic H3 receptor antagonist/inverse agonist ciproxifan maximally reduced the specific [35S]GTPγS binding to 83 ± 2% of the basal value. Compared with the maximal inverse agonism of ciproxifan, the intrinsic activity of BF2.649 was then approximately 50% higher.
Receptor Selectivity. The compound presented low affinities at the non-H3 histamine receptors with Ki values in the micromolar range at the H1 receptor and even higher at the H2 and H4 receptors, ensuring a selectivity ratio of at least 230 in favor of the H3 receptor (Table 2). In addition, BF2.649 displayed Ki values higher than 1 μM in a variety of radioligand binding assays and functional tests (Pharmascreen; MDS Panlabs) (data not shown).
Pharmacokinetics of BF2.649 in Mice. After i.v. administration of 10 mg/kg BF2.649 to mice (Fig. 5), the H3 receptor ligand concentration decreased progressively, fitting a typical biexponential decay model with half-times (t1/2) values of 13 and 126 min for the distribution and elimination phases, respectively. At 6 h, the serum BF2.649 concentration was still detectable with a value of 390 ± 72 nM. When BF2.649 was given orally at the same dose (Fig. 5), serum ligand concentrations rose rapidly being maximal at 60 min with a maximal concentration (Cmax) of 1994 ± 222 nM. Then, BF2.649 concentration decreased with a value measurable at 6 h (566 ± 86 nM). The AUCs were 7916 and 9449 nM · h after i.v. and p.o. administrations, respectively, leading to an oral bioavailability coefficient (AUCp.o./AUCi.v.) of 84%.
Changes in Braint-MeHA Level after Acute or Chronic BF2.649 Administration to Mice. Administration of BF2.649 to mice elicited a dose-dependent and marked increase of brain t-MeHA levels, an index of histaminergic neuron activity (Schwartz et al., 1991). Ninety minutes after oral administration, the maximal increase was by 91 ± 5%, similar to that elicited by ciproxifan (86 ± 4%), and the ED50 was 1.5 ± 0.2 mg/kg p.o. (Fig. 6). Similar values were found with female OF1 and male C57BL/6J mice (data not shown). The study of the time course indicated that the increase elicited by 3 mg/kg was maximal after 90 min (119 ± 9%, in this experiment), with the level decreasing slowly and still being significantly enhanced after 4 h 30 min but no more so after 6 h (Fig. 7). Similar results were obtained in male Wistar rats with BF2.649 showing an estimated ED50 value of 3 mg/kg p.o. on the t-MeHA index in cerebral cortex (data not shown). In addition, the effect on t-MeHA elicited by a 10-mg/kg oral dose of BF2.649 was not significantly modified after a 17-day chronic treatment with BF2.649 (Table 3) in mice.
Changes in Dopamine and Acetylcholine Release in Rat Prefrontal Cortex. Basal dialysate concentrations of dopamine, taken as the mean of the three values preceding drug administration, were 3.38 ± 0.80 and 3.05 ± 0.31 fmol/μl for vehicle- and BF2.649-treated group, respectively. Basal dialysate concentrations of ACh, taken as noted above, were 32.6 ± 5.0 and 31.8 ± 6.7 fmol/μl for vehicle- and BF2.649-treated group, respectively. Basal levels of dopamine or ACh were not significantly different between the two groups. BF2.649 elicited a significant elevation in dialysate ACh levels with a maximal increase (193% of basal levels) within 60 min after drug administration and a return to baseline ∼120 min after drug administration (Fig. 8). Extracellular level of dopamine was progressively increased after BF2.649 administration (133% of basal levels 150 min after administration).
Effects of BF2.649 on Neocortical EEG Power Spectral Density and Sleep-Wake Cycle in Cats. Administration of BF2.649 caused a dose-dependent reduction of neocortical slow activity (at δ range, 0.8–5 Hz) and spindles (8–15 Hz), resulting in a total cortical activation, i.e., low-voltage electrical activity with dominant waves in the β and γ bands (mainly 25–45 Hz). Furthermore, BF2.649 increased the power density of these neocortical fast rhythms (Fig. 9)
Analysis of hypnograms (4 h after BF2.649 oral administration) evidenced a dose-dependent increase in time spent in the waking state with a doubling of wake-phase duration for cats receiving 10 mg/kg BF2.649 compared with that following placebo treatment in the same cats (Fig. 9). This increase in waking occurred at the expense of paradoxical and deep slow-wave sleeps, which both almost fully disappeared, and of the light slow-wave sleep, which was reduced by ∼50% (Fig. 10).
Promnesiant Effect of BF2.649 in Natural Forgetting and Scopolamine-Induced Deficit in Mice. In the natural forgetting test, duration of acquisition trial T1 was similar in vehicle (285 ± 40 s) and tacrine groups (284 ± 20 s), but BF2.649 (15 mg/kg i.p.) significantly increased this parameter (452 ± 45 s) (Fig. 11). Memory performance was evaluated by the comparison between the time spent exploring the novel object and the time spent exploring the familiar object during the test trial T2. Control mice explored similarly the novel and the familiar object (2.6 ± 0.4 and 2.2 ± 0.3 s) with a difference (D) of 0.4 ± 0.3 s. Tacrine induced a more pronounced exploration of the novel object compared with the familiar object (3.7 ± 0.4 versus 1.6 ± 0.3 s, respectively; P = 0.0002) with a D of 2.1 ± 0.4 s significantly higher (P = 0.0133) than that of control mice. BF2.649 also induced a significantly more pronounced exploration of the novel object compared with the familiar object (3.6 ± 0.3 versus 2.4 ± 0.3 s, respectively; P = 0.0036) with a D of 1.2 ± 0.3 s. Locomotor activity in control and in tacrine-treated mice was similar (15.0 ± 1.9 versus 19.0 ± 2.5 lines/min, respectively) whereas BF2.649-treated mice presented a significantly enhanced locomotor activity (24.1 ± 2.4 lines/min) (P = 0.0120).
In the scopolamine-induced deficit test, scopolamine (0.3 mg/kg i.p.) had no effect on duration of the acquisition trial T1 (314 ± 21 s) compared with vehicle (323 ± 30 s) (Fig. 12). Whereas BF2.649 at a 5-mg/kg dose had no effect on T1 duration, at a higher dose (15 mg/kg i.p.), it significantly increased this parameter (488 ± 29 s; P = 0.0007 versus control). Memory performance was evaluated by the comparison between the time spent exploring the novel object and the time spent exploring the familiar object during the test trial T2. During T2, control mice explored the novel object significantly more than the familiar object (6.7 ± 1.2 and 2.7 ± 0.6 s; P = 0.0013) with a D of 4.0 ± 0.9 s. When scopolamine was administered, mice spent the same amount of time with the two objects (3.4 ± 1.1 s for the novel object and 3.4 ± 0.6 s for the familiar object) with a D of 0.0 ± 1.1 s. BF2.649 administered at the lowest dose tested (5 mg/kg i.p.) to scopolamine-treated mice induced a nonsignificant (P = 0.1087) tendency to explore the novel object more than the familiar object with a D of 1.4 ± 0.8 s. When BF2.649 was administered at higher dose (15 mg/kg), scopolamine-treated mice spent significantly more time with the novel object than with the familiar object (4.3 ± 0.7 versus 1.6 ± 0.5 s) (P = 0.0056) with a D of 2.7 ± 0.7 s. Scopolamine significantly increased locomotor activity (18.2 ± 1.7 lines/min in scopolamine-treated versus 12.2 ± 1.8 lines/min in control) (P = 0.0285). In scopolamine-treated mice, BF2.649 administration (5 and 15 mg/kg) did not significantly modify locomotor activity (24.6 ± 3.3 and 22.9 ± 3.1, respectively, versus 18.2 ± 1.7 lines/min in scopolamine-treated mice) in scopolamine-treated mice.
Discussion
From the present study, BF2.649 seems to be a high-affinity, competitive antagonist and potent inverse agonist at the human H3 receptor. It elicits, after oral administration, an activation of histaminergic neurons and a variety of neurochemical and behavioral responses that allow to predict some therapeutic applications for this compound.
The structure of BF2.649, like that of many other H3 receptor antagonists/inverse agonists described recently (for review, see Cowart et al., 2004), is derived from rational pharmacochemical investigations (Ganellin et al., 1998; Meier et al., 2001) that initially led to the replacement of the imidazole ring by cyclic or aliphatic N-substituted amines; indeed, for a long time, the presence of an imidazole ring was considered as an obligatory feature of molecules interacting potently with the H3 receptor. Animal toxicity studies performed on BF2.649 (data not shown) have confirmed that hepatic or ocular toxicities displayed by imidazole-containing compounds, e.g., thioperamide or ciproxifan (J.-M. Lecomte and J.-C. Schwarz, unpublished data), could be avoided by replacement of the imidazole nucleus in this drug.
The apparent affinity of BF2.649, derived from radioligand binding studies slightly differed from values derived from functional tests, namely, because of different ionic composition of buffers. In addition, the affinity in radioligand binding studies was clearly superior at the human H3 receptor, at which Ki values were in the low nanomolar range, compared with the rodent receptors, at which the Ki values were nearly five to six times higher (Table 1). Interestingly, the apparent affinity of the drug did not markedly differ at the recombinant human receptor from that at the native receptor from a sample of human cerebral cortex, indicating a lack of any major influence of associated proteins modulating G protein-coupled receptors (Christopoulos and Kenakin, 2002), at least on the binding parameter. The rather large difference in affinity between rodent and H3 receptors is not unique to the present drug, with some compounds, e.g., ciproxifan, contrarily to BF2.649, showing preference for the rodent receptor; such differences were fully attributed by targeted mutation studies to two amino acids in the third transmembrane segment that differ in rat and human (Ligneau et al., 2000). The significantly lower affinity of BF2.649 at the rodent receptor was taken into account when extrapolating the active doses in rat or mice to therapeutic doses in currently ongoing clinical trials in humans.
BF2.649 displayed a high selectivity toward the H3 receptor. Among the four histamine receptor subtypes, it was more than ∼220-fold selective for the human H3 receptor, in marked contrast with imidazole-containing compounds, e.g., thioperamide, ciproxifan, or clobenpropit, which rather potently interact with the H4 receptor, with the latter compound even displaying potent H4 receptor agonist activity (Oda et al., 2000). A similarly high selectivity ratio was found in a panel of ∼110 receptors, channels, or enzymes explored in the MDS Panlabs Pharmascreen.
BF2.649 seemed to be a competitive antagonist and inverse agonist at a variety of H3 receptor-mediated signaling pathways. It counteracted the histamine-induced inhibition of [3H]HA release from depolarized rat cortical synaptosomes, a classic autoreceptor model (Schwartz et al., 1991), and it induced a progressive rightward shift of the dose-response curve to imetit, a highly selective H3 receptor agonist, on the enhancement of [35S]GTPγS binding to the human H3 receptor (Rouleau et al., 2002).
On these two functional models, the compound showed evidence for inverse agonist activity. In agreement, it diminished in a dose-dependent manner the basal [35S]GTPγS binding to the human receptor with a maximal effect ∼50% higher than that of ciproxifan, illustrating the concept of variable intrinsic activity for inverse agonism. Hence, on our “artificial” cell system expressing a rather high receptor density (∼500 fmol/mg protein), a substantial fraction of the H3 receptor seems already coupled to G proteins in the absence of agonist. The physiological relevance of the process is indicated by the enhanced [3H]HA release over basal value (by ∼20%) when the response to the agonist was completely antagonized (Fig. 2), an effect already noticed on the same “natural” model with other inverse agonists (Morisset et al., 2000; Gbahou et al., 2003). Also in support of the physiological relevance of a uniquely high constitutive activity in the H3 receptor, presumably linked to the presence of a short sequence at the end of its third intracytoplasmic loop, is that it can be observed on the native receptor of cerebral membranes using the [35S]GTPγS binding test, whereas this is not true for most other receptors as assessed in the same preparation (Rouleau et al., 2002). The importance of the process from a pharmacological and also, presumably, therapeutic point of view is indicated by the fact that only inverse agonists and not neutral antagonists are able to activate histaminergic neurons in vivo (Morisset et al., 2000). Nevertheless, classification of H3 receptor ligands depends on the test system as shown, namely, with the existence of “protean ligands”, i.e., compounds behaving as partial or full agonists, neutral antagonists, or inverse agonists according to the test system (Gbahou et al., 2003).
Considering the rather modest potency of BF2.649 on the rodent H3 receptor in vitro, the compound was quite potent in vivo, in increasing cerebral t-MeHA levels, an effect occurring at a low oral dose and maintained during several hours. This is presumably the consequence of its good oral absorption and brain penetration with an oral bioavailability ratio, assessed using a radioreceptor assay (Ligneau et al., 1998), of ∼0.84 in mice. The enhanced activity of histaminergic neurons elicited by BF2.649, as assessed by the augmentation of histamine main extracellular metabolite level, persisted upon repeated administration, suggesting that tolerance failed to develop on this response; this is obviously a prerequisite for the application of this class of drugs in chronic neurological or psychiatric disorders.
Other aminergic neurons were also activated in brain of rats receiving BF2.649. In agreement, neurotransmitter release from dopaminergic and cholinergic neurons projecting to the prefrontal cortex was enhanced approximately 2-fold, as shown in microdialysates from this region. Similar changes were recently reported for other H3 receptor inverse agonists, indicating that this corresponds to a class effect (Fox et al., 2005). Presynaptic H3 receptors inhibiting the release of a variety of neurotransmitters from synaptosomes, or even from intact nerve endings in vivo, were evidenced following their stimulation by exogenous agonists; however, this does not mean that they are, in all cases, in a state of tonic activation or, thus, that they display constitutive activity or are submitted, under basal conditions, to activation by endogenous histamine, as discussed previously (Schwartz and Arrang, 2002).
Interestingly, BF2.649, while enhancing dopamine release and dihydroxyphenyl acetic acid /dopamine ratio in the rodent prefrontal cortex, did not significantly affect the latter index of dopaminergic neuron activity in striatum (data not shown), as also shown with another H3 antagonist (Fox et al., 2005). Such a selective enhancement of the activity of prefrontal cortical dopaminergic and cholinergic afferents may be predictive of beneficial effects in psychotic disorders or dementias in which these systems are currently thought to be blunted.
The marked waking effect of BF2.649, presumably consequent to the HA release it induces, is consistent with a large body of experimental evidence showing that histaminergic neurons play a prominent role in cortical activation and arousal (for review, see Schwartz et al., 1991; Lin, 2000; Schwartz and Arrang, 2002), with observations of similar effects with several H3 receptor inverse agonists (Vanni-Mercier et al., 2003). The waking effect being suppressed in animals receiving an H1 receptor antagonist or in mice lacking this receptor (Lin 2000; Lin et al., 2002), it is likely to result from histamine release onto, e.g., thalamic relay nuclei, basal forebrain neurons, or mesopontine tegmentum neurons involved in the control of wakefulness and sleep (Lin, 2000).
The enhancement by BF2.649 of fast cortical EEG rhythms, known to be associated with high level of vigilance, in cats displaying a calm behavior, is consistent with an ability to improve performance in cognitive tests. Here, we show that mice pretreated with this drug perform better in a learning and memory model, the two-trial object recognition test in mice, in which retrieval was compromised either by administering scopolamine or by augmenting the delay between learning and testing. Interestingly, mice facing a novel object displayed an enhanced exploration activity that was not attributable to general locomotor activation (the latter being not observed in mice introduced in a “poor” environment) and that may reflect an enhanced attention or “curiosity”, a function that is largely mediated by brain histamine. In agreement, mice lacking histamine synthesis exhibit deficit of waking and exploration in a new environment (Parmentier et al., 2002).
H3 receptor inverse agonists, starting with the prototypical drug thioperamide, have shown various procognitive properties, including promnesiant and proattentional effects in a variety of rodent models (for review, see Hancock and Fox, 2004), leaving little doubt about their therapeutic potential in cognitive deficits, e.g., in Alzheimer's disease, other dementias, or attention-deficit and hyperactivity disorder. Such procognitive effects could also find therapeutic applications in psychotic states in which cognitive deficits are modestly responsive to currently used antipsychotic agents. The various therapeutic opportunities suggested by animal studies with BF2.649 and other H3 receptor inverse agonists are currently explored in ongoing clinical trials.
Acknowledgments
We thank Prof. Sigurd Elz for investigations on BF2.649 effects in the H3 receptor assay on guinea pig ileum and S. Rouanet for technical assistance. We also acknowledge the experimental and technical contributions of C. Buda, J.-P. Sastre, and G. Guidon.
Footnotes
-
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
-
doi:10.1124/jpet.106.111039.
-
ABBREVIATIONS: HA, histamine; BF2.649, 1-{3-[3-(4-chlorophenyl)propoxy]propyl)piperidine, hydrochloride; CHO, Chinese hamster ovary; HEK, human embryonic kidney; GTPγS, guanosine 5′-3-O-(thio)triphosphate; (R)-α-MeHA, (R)-α-methylhistamine; t-MeHA, tele-methylhistamine; ACh, acetylcholine; EEG, electroencephalogram; W, wakefulness; SWS1, light slow-wave sleep; SWS2, deep slow-wave sleep; PS, paradoxical sleep; T1, acquisition trial; T2, testing trial; AUC, area under the curve; D, difference.
- Received July 20, 2006.
- Accepted September 26, 2006.
- The American Society for Pharmacology and Experimental Therapeutics