Abstract
The relative contribution of α4β2, α7 and other nicotinic acetylcholine receptor (nAChR) subtypes to the memory enhancing versus the addictive effects of nicotine is the subject of ongoing debate. In the present study, we characterized the pharmacological and behavioral properties of the α7 nAChR agonist N-[(3R)-1-azabicyclo[2.2.2]oct-3-yl]-7-[2-(methoxy)phenyl]-1-benzofuran-2-carboxamide (ABBF). ABBF bound to α7 nAChR in rat brain membranes (Ki = 62 nM) and to recombinant human 5-hydroxytryptamine (5-HT)3 receptors (Ki = 60 nM). ABBF was a potent agonist at the recombinant rat and human α7 nAChR expressed in Xenopus oocytes, but it did not show agonist activity at other nAChR subtypes. ABBF acted as an antagonist of the 5-HT3 receptor and α3β4, α4β2, and muscle nAChRs (at higher concentrations). ABBF improved social recognition memory in rats (0.3–1 mg/kg p.o.). This improvement was blocked by intracerebroventricular administration of the α7 nAChR antagonist methyllycaconitine at 10 μg, indicating that it is mediated by α7 nAChR agonism. In addition, ABBF improved working memory of aged rats in a water maze repeated acquisition paradigm (1 mg/kg p.o.) and object recognition memory in mice (0.3–1 mg/kg p.o.). Rats trained to discriminate nicotine (0.4 mg/kg s.c.) from vehicle did not generalize to ABBF (0.3–30 mg/kg p.o.), suggesting that the nicotine cue is not mediated by the α7 nAChR and that selective α7 nAChR agonists may not share the abuse liability of nicotine. Our results support the hypothesis that α7 nAChR agonists may provide a novel therapeutic strategy for the treatment of cognitive deficits with low abuse potential.
Nicotine enhances cognitive functions, such as attention, learning, consolidation, and retention, in both animals and humans, through activation of brain nicotinic acetylcholine receptors (nAChRs) (Levin et al., 1999, 2006). These ligandgated ion channels are homopentamers formed by five identical subunits (α7 nAChR) or heteropentamers of multiple α and β subunits. Various isoforms of these subunits have been identified (α2–α10; β2–β4; for reviews, see Paterson and Nordberg, 2000; Gotti et al., 2006). The most common nAChRs found in the brain are the α7 subtype with a low affinity for nicotine and the α4β2 subtype with a high affinity for nicotine. Evidence from neuroanatomical, electrophysiological, and behavioral studies support a role for both of these receptor subtypes in processes of learning and memory.
Studies using [125I]α-bungarotoxin and [3H]cytisine to label α7 and α4β2 nAChRs, respectively, have identified high densities of these receptors in the hippocampus, a brain area that plays an important role in learning and memory (Paterson and Nordberg, 2000). The α4β2 nAChR agonist A-85380, the α7 nAChR agonist AR-R 17779, and the α7 nAChR agonist (and weak α4β2 antagonist) DMXB (GTS-21) modulated the induction of hippocampal long-term potentiation in rats, indicating a role in neuronal plasticity (Hunter et al., 1994; Gordon et al., 1998; Chen et al., 2000; Fujii et al., 2000). Coapplication of the α7 nAChR antagonist methyllycaconitine (MLA) inhibited the effects of AR-R 17779, whereas MLA alone had no effects on long-term potentiation (Chen et al., 2000).
Both α7 and α4β2 nAChR agonists have been shown to improve performance in learning and memory tasks (Levin et al., 2006). GTS-21 improved performance in long-delay trials of a delayed matching-to-sample test in monkeys (Briggs et al., 1997), in the Lashley III maze and one-way active avoidance, and increased general learning and reference memory in the 17-arm radial maze in rats (Arendash et al., 1995). In clinical trials with healthy volunteers, GTS-21 improved attention and working and episodic memory (Kitagawa et al., 2003). GTS-21 is a weak partial agonist of human α7 nAChRs, and it inhibits α4β2 nAChRs and 5-HT3 receptors (Briggs et al., 1997; Kem et al., 2004). The more selective α7 nAChR agonist, AR-R 17779, improved long-term win-shift acquisition in the eight-arm radial maze (Levin et al., 1999). Infusion of MLA into the hippocampus impaired working memory in a similar radial arm maze task (Felix and Levin, 1997). However, AR-R 17779 failed to improve performance in a five-choice serial reaction time task (Grottick and Higgins, 2000). Recently, we described improvements in social recognition memory after treatment with AR-R 17779 (van Kampen et al., 2004). Together, these results indicate a role for α7 nAChRs in learning and memory, rather than in attention processes. The observation that the β2-subunit is critical for the addictive effects of nicotine (Picciotto et al., 1998), and that α7 nAChRs are apparently not involved in the rewarding effects of nicotine (Brioni et al., 1996; Grottick et al., 2000), suggests that selective α7 nAChR agonists may have no, or only low, abuse potential. This property would add to the value of selective α7 nAChR agonists for the treatment of memory disorders.
We examined the effects of the novel α7 nAChR agonist N-[(3R)-1-azabicyclo[2.2.2]oct-3-yl]-7-[2-(methoxy)phenyl]-1-benzofuran-2-carboxamide (ABBF; Fig. 1A) on the performance of normal adult rats in a social memory task and on object recognition in mice. In addition, we examined the effects of treatment with ABBF on the performance of aged (33- to 34-month-old) rats in a working memory-specific version of the Morris water escape task. Previous studies have suggested that the discriminative stimulus effect of nicotine is closely related to its positive reinforcing stimulus effect and that a failure to generalize to the nicotine cue could be indicative for a lack of nicotine-like abuse potential (for discussion, see Merlo Pich et al., 1999; Stolerman et al., 1999). Therefore, we tested whether ABBF would generalize to the nicotine cue in rats.
Materials and Methods
Materials
Ethanol absolute, 99.8%, was obtained from Riedel-de Haen (Seelze, Germany). Solutol HS 15 (12-hydroxystearic-acid ethoxylate) was obtained from BASF (Ludwigshafen, Germany). Tylose MH300P (methylhydroxyethyl-cellulose) was obtained from Hoechst AG (Frankfurt, Germany). MLA was obtained from Research Biochemicals International/Sigma-Aldrich (Deisenhofen, Germany). Acetylcholine and (–)-nicotine hydrogen tartrate were purchased from Sigma-Aldrich (St. Louis, MO). ABBF (Fig. 1A) was synthesized by the Medical Chemistry (Bayer HealthCare AG, Wuppertal, Germany).
Membrane Preparation and Binding Assays
Rats were decapitated, and the brains were rapidly removed and placed in ice-cold homogenization buffer [10% (w/v) 0.32 M sucrose, 1 mM EDTA, 0.1 mM phenylmethylsulfonyl fluoride, and 0.01% (w/v) NaN3, pH 7.4 at 4°C]. Brains were homogenized at 600 rpm in a Potter glass Teflon homogenizer. The resulting suspension was centrifuged (1000g at 4°C for 10 min), and the supernatant was collected. The pellet was resuspended [20% (w/v)], and the suspension was recentrifuged (1000g at 4°C for 10 min). Both supernatants were combined and centrifuged (15,000g at 4°C for 30 min). The resulting pellet (P2 fraction) was resuspended in binding buffer (50 mM Tris-HCl, 1 mM MgCl2, 120 mM NaCl, 5 mM KCl, and 2 mM CaCl2, pH 7.4) and centrifuged (15,000g at 4°C for 30 min). The resuspension and centrifugation were repeated once.
The final pellet was resuspended in binding buffer and incubated in a final volume of 250 μl (0.2 mg of membrane protein/assay) in the presence of 2 nM [3H]methyllycaconitine, 0.1% (w/v) bovine serum albumin, and different concentrations of the test substance for 2.5 h (21°C). Nonspecific binding was determined in the presence of 10 μM MLA.
The incubation was terminated by the addition of 4 ml of phosphate-buffered saline (20 mM Na2HPO4, 5 mM KH2PO4, and 150 mM NaCl, pH 7.4, at 4°C) and rapid filtration using a cell harvester (Brandel Inc., Gaithersburg, MD) and type A/E glass fiber filters (Gelman Instrument Co., Ann Arbor, MI), pretreated for 3 h with 0.3% (v/v) polyethylenimine. Filters were washed twice with 4 ml of phosphate-buffered saline (4°C), and bound [3H]methyllycaconitine was determined by scintillation counting. All tests were performed in triplicate.
For 5-HT3 receptor binding assays, membranes from HEK293 cells expressing human recombinant 5-HT3 receptor (RB-HS3; Receptor Biology, Inc., Beltsville, MD) were used. Membranes were diluted according to manufacturer's instructions in incubation buffer (50 mM Tris base, pH 7.4, 5 mM MgCl2, 0.5 mM EDTA, 0.1% ascorbic acid, and 10 μM pargyline) and incubated in a volume of 200 μl (membrane protein concentration, 3 μg/assay) for 60 min at 21°C in the presence of 0.5 nM selective 5-HT3 receptor radioligand [3H]GR65630 (PerkinElmer Life and Analytical Sciences, Boston, MA) and different concentrations of test compound. Nonspecific binding was determined in the presence of 100 μM 5-HT. The incubation was terminated by filtration through type A/E glass fiber filters (Gelman Instrument Co.) or GF/B filters (Whatman, Maidstone, UK), that were pretreated for at least 1 h with 0.3% (v/v) polyethylenimine. Filters were washed three times with 3 ml of buffer (50 mM Tris-HCl, pH 7.4 at 4°C), and bound radioactivity was determined by scintillation counting. All tests were performed in triplicate.
The IC50 values were determined from plots of binding activity versus log compound concentration using a sigmoidal curve fit (Prism software, version 2.0; GraphPad Software Inc., San Diego, CA). The dissociation constant Ki of test compounds was determined from their IC50 values, the dissociation constant KD, and the concentration L of [3H]methyllycaconitine or [3H]GR65630 as appropriate, using the equation Ki = IC50/(1 + L/KD).
Electrophysiological Assays
Preparation of Xenopus oocytes, injection of receptor cDNA, and electrophysiological measurements of receptor activity were performed as described previously (Methfessel et al., 1986; Schnizler et al., 2003). Pieces of ovary were excised from anesthetized adult female Xenopus laevis. The tissue was treated with 2 mg/ml collagenase (Research Biochemicals International/Sigma-Aldrich) to release the oocytes from the follicle. Intact stage V oocytes were selected manually and placed into individual wells of 96-well plates (Greiner, Frickenhausen, Germany) filled with modified Barth's solution [88 mM NaCl, 1 mM KCl, 0.82 mM MgSO4, 0.41 mM Ca(NO3)2, 2.4 mM NaHCO3, and 5 mM Tris-HCl, pH 7.4, with 50 μg/ml gentamicin]. Approximately 30 nl of cDNA solution, containing expression plasmids with inserts coding for the target receptors, were injected into the germinal vesicle of each oocyte using an automated system. Injected oocytes were incubated at 19°C for 3 to 8 days in modified Barth's solution before the measurements were taken.
For electrophysiological recording, oocytes were impaled with two glass microelectrodes filled with pipette solution (1.5 M potassium acetate and 0.1 M KCl). Voltage clamp was performed with a standard voltage-clamp amplifier (Gene-Clamp 500 amplifier; Molecular Devices, Sunnyvale, CA). For automated recording, the position of the cells and the recording head stage, the amplifier, the solution exchange, and the data acquisition were under full computer control using a software package developed by D. Bertrand (Department of Neuroscience, Centre Medical Universitaire, Geneva, Switzerland) as described by Schnizler et al. (2003). All the recordings reported here were obtained at a membrane potential of –80 mV. Oocytes were superfused with normal frog Ringer's solution (115 mM NaCl, 2.5 mM KCl, 1.8 mM CaCl2, and 10 mM HEPES, pH 7.2). Solutions of test compounds in normal frog Ringer's solution were superfused through the recording chamber for 20 s while the voltage-clamp current was recorded. The washout time between applications of test solutions was 5 min.
Each measurement of a test solution was preceded by the application of a standard concentration of acetylcholine (for nAChRs) or serotonin (for 5-HT3 receptors). Standard concentrations producing approximately 10 to 20% of the maximum response for each receptor were used as internal controls for differences in expression levels between different oocytes and changes of the response during an experiment. Current amplitudes evoked by the test solutions were normalized to that of the preceding acetylcholine or serotonin application. The standard concentration of acetylcholine was different for each receptor subtype, depending on the sensitivity of the subtype to acetylcholine. For rat and human α7 nAChRs, 50 μM acetylcholine was used as reference standard. The concentrations for rat α3β4, α4β2, and muscle nAChRs were 3.2, 0.32, and 0.1 μM acetylcholine, respectively, and 0.5 μM serotonin was used as reference standard for the human 5-HT3A receptor.
As initial assay for agonistic activity, 100 μM ABBF was applied to an oocyte expressing a high level of the corresponding receptor subtype. If an agonistic effect was noted, a full concentration-response curve was obtained. To test for antagonistic effects and potentiator activity of ABBF, various concentrations of the compound were applied together with the standard concentration of acetylcholine or 5-HT, and the resulting inward current was compared with the current level elicited by the standard concentration of agonist applied alone. EC50 values were calculated using GraphPad Prism.
Functional Assays on Muscarinic Acetylcholine Receptors
Recombinant Chinese hamster ovary cells stably expressing the human muscarinic acetylcholine receptor subtypes M1 to M5 were used as cell-based functional in vitro test. These cells stably expressed the calcium-sensitive photoprotein aequorin and the Gα16 protein, thus allowing the coupling of all muscarinic receptors to phospholipase C and luminometric detection of the calcium release induced by agonist stimulation.
Cells were seeded 2 days before testing. The cell culture media (Dulbecco's modified Eagle's medium, 10% fetal calf serum, 2 mM glutamine, and 10 mM HEPES) were changed on test day to Tyrode's solution (140 NaCl mM, 5 KCl mM, 1 mM MgCl2, 2 mM CaCl2, 20 mM glucose, and 20 mM HEPES) containing 50 μM coelenterazin for 4 h. Full dose-response curves of test compounds were prepipetted in microtiter plates and transferred by a CyBio-pipetting robot to the cell plates. The light signals were detected immediately by a luminometer (Lumibox; Bayer AG, Leverkusen, Germany), and the resulting dose-response curves and EC50 values were calculated using GraphPad Prism.
Behavioral Experiments
All animal experiments followed the principles of laboratory animal care and were in accordance with the guidelines given by European regulations and the German government and approved by the local authorities (Regierungspräsidium, Düsseldorf, Germany).
Social Recognition. Adult (4- to 5-month-old) male Wistar rats (HsdCpb:WU) and juveniles (4–5 weeks old) were supplied by Harlan-Winkelmann (Borchen, Germany). The animals were housed in groups of three in type IV (adult rats) and type III (juvenile rats) Makrolon cages, under a 12-h light/dark schedule (lights on at 7:00 AM). Food (ssniff, Soest, Germany) and water were available ad libitum, except during testing. Ambient room temperature (22°C) and relative humidity (55 ± 5%) were kept constant. The animals were randomly assigned to their respective treatment groups. Animals were adapted to laboratory housing conditions for 1 week before behavioral testing. One habituation session was performed under essentially similar conditions as the test session (see below). The social recognition test was performed as described by van Kampen et al. (2004).
The task consisted of two trials, separated by a 24-h retention interval. Adult animals were individually housed 30 min before testing. An enclosure (63 × 41 × 40 cm; aluminum side walls, Plexiglas front) was put over the cage 4 min before testing with the lid of the cage removed. During the first trial (T1), a juvenile was placed into the cage, and the social investigation by the adult was measured cumulatively for 2 min by a trained observer. Sniffing and grooming of body parts, anogenital sniffing, and close following were scored. ABBF (1 mg/kg p.o.) dissolved in 10% ethanol/20% Solutol/70% distilled water in an application volume of 1 ml/kg was given 30 min, 1.5 h, or 4 h before T1 to assess the duration of the effect after a single administration. After a retention interval of 24 h, social investigation time was measured in a second trial (T2) for 2 min, with the same juvenile being placed into the observation cage as in T1. A decreased duration of exploring the juvenile indicates social recognition. The difference was expressed as percentage reduction of the social investigation time during T2 in comparison with T1. This measure was subjected to analysis of variance (ANOVA) with the factor treatment, supplemented with Fisher's least significant difference (LSD) post hoc comparisons between treatment groups.
To show that the effect is mediated by stimulation of α7 nAChRs, adult rats were treated with ABBF (1 mg/kg p.o., 30 min before T1) and injected i.c.v. with the α7 nAChR antagonist MLA (10 μg injected 4 min before T1). T2 was measured after a retention interval of 24 h, using the same juvenile as in T1.
A further experiment was carried out to reveal potential nonspecific effects of ABBF, which may influence investigation of a conspecific, by placing a different (novel) juvenile into the observation cage at T2. Adult rats were injected p.o. with 0.3 mg/kg ABBF in saline (0.9% NaCl), 30 min before T1. T2 was conducted 24 h later, with either the same or a different juvenile as interaction partner. Nonspecific effects may result in a shorter investigation time of the novel juvenile. Compounds with specific effects on cognition are expected to have no effect on the duration of investigating the novel juvenile.
Object Recognition Task. Thirty (experiment 1) or 60 (experiment 2) male OF1 [(Ico:OF1(IOPS Caw)] mice, weighing 22 to 26 g, were supplied by Iffa Credo (l'Arbresle, France). The mice were housed in groups of 10 in standard Makrolon type III cages, and they were allowed to adapt to the laboratory for 1 week. During the course of behavioral testing, they were housed individually in standard Makrolon type II cages. The observation arena of the object recognition test consisted of a circular open field, 480 mm in diameter. The wall (400 mm in height) and the floor consisted of transparent Makrolon. Three different sets of objects, made of aluminum, were used.
During two consecutive days, the mice were allowed to explore the empty apparatus twice for 5 min each day. The mice were pretrained in pairs of two trials that were separated by a retention interval of 1 h. During the first trial (T1), the apparatus contained two identical objects. A mouse was taken from its home cage and placed into the apparatus, equidistant from the two objects, facing the wall in front of the experimenter. After T1, the mouse was transferred to its home cage. One hour later, the mouse was again placed into the apparatus for the second trial (T2). Now, the exploration arena contained two different objects, a familiar object from T1 and a novel object. The time spent exploring the two objects during T1 and T2 was recorded.
Exploration was scored whenever the mouse directed its nose to the object at a distance ≤20 mm and/or whenever it touched the object with its nose. Sitting on the object was not considered as exploratory behavior. To remove olfactory cues the objects were thoroughly cleaned after each trial. All combinations and locations of objects were used in a balanced manner. This reduces the effects of the individual preference of mice for particular locations or particular objects. The object recognition test provides measures for exploration and discrimination, i.e., noncognitive effects of a drug can be distinguished from effects on memory performance (for details, see Prickaerts et al., 2005). The times spent exploring the familiar and new object during T2 were represented as a and b, respectively. The discrimination index d2 was calculated as d2 = (b – a)/(a + b).
As soon as the mice had reached a good discrimination performance (i.e., a discrimination index d2 ≥ 0.15; see below), pretraining continued with a retention interval of 24 h. Now, the retention performance stabilized around a d2 of 0 (=no discrimination). Each trial lasted 180 s. To habituate the mice to the p.o. administration of test compounds, they routinely received saline (10 ml/kg body weight) after the first trial (T1) of each pair. This phase of pretraining consisted of two trial pairs.
In the first experiment, the six mice with the lowest object exploration times were omitted from further testing. The remaining 24 animals were assigned to one of two groups, using a matched random assignment procedure. Matching was based on the inspection times in the last pretraining trial. Drug testing consisted of two pairs of trials (T1, T2), separated by a 24-h retention interval. ABBF (1 mg/kg) or vehicle (0.5% Tylose) was administered p.o., 30 min before T1, in an application volume of 5 ml/kg. During the first pair of trials, half of the mice were treated with vehicle, and the other half received ABBF. During a second pair of trials after a 2-day washout period, the animals treated previously with vehicle received ABBF, whereas the animals treated previously with ABBF received vehicle only. Thus, each animal was treated once with vehicle and with ABBF. Consequently, results are based on 24 observations per treatment condition. Treatment effects were analyzed using a within-subjects (i.e., repeated measures) ANOVA.
In the second experiment, the 12 mice with the lowest object exploration times were omitted from further testing. The remaining 48 animals were assigned to one of four groups (n = 12 per group), a vehicle control group (0.5% Tylose), or a group treated with either 0.1, 0.3, or 1 mg/kg ABBF in 0.5% Tylose, using a matched random assignment procedure. Matching was based on the means of the inspection times during T2 of the trial pairs with a 24-h retention interval. ABBF or vehicle was administered p.o. 30 min before T1. After the first pair of trials, all mice remained undisturbed for 17 days. Then, all mice were tested again in a second pair of trials. Thus, each animal was treated twice with vehicle or ABBF. The measures obtained during the two trial pairs were averaged per animal. These means were evaluated statistically by ANOVA with the factor treatment (vehicle versus 0.1, 0.3, or 1 mg/kg ABBF).
Repeated Acquisition (Working Memory) Performance ofAged Fischer 344 × Brown Norway Rats in the Water Maze. Thirty-two aged male Fischer × Brown Norway (F344/NHsd × BN/RijHsd) F1 hybrid rats (FBNF1) were supplied by Harlan (Indianapolis, IN). They were housed in groups of four in standard Makrolon type IV cages with sawdust bedding in an air-conditioned room (22°C). The lights were on from 6:00 AM to 6:00 PM. Food (ssniff) and water were available ad libitum, except during testing. The rats were habituated to the local animal facilities for about 1.5 months before behavioral testing started. The Morris water escape training and the assessment of the effects of ABBF on spatial working memory were performed when the rats were 33 to 34 months old.
The rats were pretrained in the standard Morris water escape task in five daily sessions, with four successive trials per session. For a detailed description of the standard task, see van der Staay (2006). The 20 rats that were able to locate the escape platform were selected and allocated by matched random assignment on the escape latencies in the last acquisition session of the standard Morris water escape task to one of three treatment groups for the repeated acquisition task: a vehicle (0.5% Tylose) control group (n = 6), a group receiving 0.3 mg/kg ABBF (n = 7), or a group receiving 1 mg/kg ABBF (n = 7). Each day, ABBF was freshly suspended in 0.5% Tylose. Vehicle or compound was administered p.o. in an application volume of 2 ml/kg 30 min before each repeated acquisition session on four successive days. The animals received three pairs of trials in a session. Within a pair of trials, both trials were run in close succession. When a rat had completed a trial pair, it was gently dried with crêpe paper and returned to its home cage. The animals were kept warm under a 150-W infrared bulb (Original Hanau Solilux; Heraeus, Hanau, Hessen, Germany) fixed approximately 60 cm above the floor of the cage. After an interval of approximately 15 min, the next pair of trials was given. The start positions were at the virtual borders between quadrants at the rim of the pool. Out of the four alternative start positions, three positions were randomly selected per animal. For each trial pair within a session, a different start position was used. Within a daily session, the escape platform remained in the same position. Over the series of the four sessions, the platform was moved to a different position in each session.
For each rat, the traveled distance was averaged per session separately for the first and second trials of the pairs [average of first swims: (trial1,1 + trial2,1+ trial3,1)/3; average of second swims: (trial1,2 + trial2,2 + trial3,2)/3; the first subscript represents the number of the trial pair within a session, and the second subscript represents the trial within trial pairs]. The acquisition curve of the repeated acquisition task was analyzed further with a two-way ANOVA with the factor treatment (vehicle versus 0.3 or 1 mg/kg ABBF) and the repeated measures factors sessions (sessions 1–4) and trial pairs (average of the first versus average of the second trials of the three trial pairs within a session).
Drug Discrimination. Male Wistar rats (HsdCpb:WU) were purchased from Harlan-Winkelmann. Body weight upon arrival at the laboratory was around 220 g, which gradually increased up to approximately 450 g during the course of the study. Rats were individually housed in Makrolon type III cages under a normal 12-h light period (lights on at 7:00 AM). The animals had restricted access to food (approximately 13 g per day, standard pellets; ssniff), and they were offered water ad libitum. Room temperature was maintained at 20 to 22°C. Experiments were performed in sound- and light-attenuated standard operant chambers (Coulborn Instruments, Lehigh Valley, PA). The chambers were equipped with two levers equidistant from a food tray between the levers. Food reinforcement (45 mg of precision pellets; Bio-Serv, Frenchtown, NJ) was delivered by an automated food dispenser located outside of the chamber. Data collection and experimental contingencies were programmed using OPN software (Emmett-Oglesby et al., 1982) on a PC interfaced with the operant chamber. Ventilation and masking noise were provided by a fan mounted on the wall of the chamber. A white houselight was switched on during the sessions, which were conducted between 9:00 AM and 12:00 AM.
In general, the procedure described by De Vry and Jentzsch (1998) was followed. After initial shaping to lever press for food reinforcement, the rats (n = 16) were trained to discriminate 0.4 mg/kg nicotine hydrogen tartrate (s.c., t –15 min) from vehicle in a standard two-lever fixed ratio:10 operant procedure. Daily sessions were conducted that were terminated either after the rat hat gained 50 reinforcers or when 10 min had elapsed, whichever event came first. For half of the animals, responses on the left lever were reinforced after nicotine; for the other half of the animals, responses on this lever were reinforced after vehicle. The rats were injected with drug or vehicle according to the following sequence: D-D-V-D-V // V-D-V-V-D // D-V-D-V-V // D-D-V-D-V (D, drug; V, vehicle, and //, no sessions during the weekends) with repetition. Discrimination criterion consisted of 10 consecutive sessions in which no more than nine responses occurred on the nonreinforced lever before the first reinforcer was obtained. After reaching discrimination criterion, generalization tests were interspersed between the training sessions. During test sessions, responding on the selected lever, i.e., the lever on which 10 responses accumulated first, was reinforced for the remainder of the session. Generalization tests were separated by at least three training sessions in which vehicle and drug were correctly discriminated, i.e., less than five incorrect responses before the first reinforcer, and when at least 20 reinforcers were obtained per session. The animals were tested with different doses of nicotine hydrogen tartrate (0, 0.05, 0.1, 0.2, 0.4, and 0.8 mg/kg s.c., 15 min before test) and ABBF (0, 0.3, 1, 10, and 30 mg/kg p.o. 30 min before test).
(–)-Nicotine hydrogen tartrate was dissolved in 0.9% NaCl (saline) for s.c. administration. ABBF was suspended in 0.5% Tylose and distilled water for p.o. administration. Compounds were administered in an application volume of 2 ml/kg body weight for p.o. administration and 1 ml/kg for s.c. administration.
Test results were expressed as the percentage of rats that selected the drug lever (% drug lever selections). Generalization was considered to be complete if at least 80% drug lever selections was obtained. In addition, the percentage of animals that selected a lever (either drug or vehicle lever) was determined as an index of behavioral disruption (i.e., % lever selections). Least-square linear regression analysis was used to estimate ED50 values (and their 95% confidence limits) after log-probit conversion of the data.
Results
In Vitro Pharmacology
ABBF had high affinity (Ki = 62 ± 20 nM, mean ± S.E.M.; n = 4; Fig. 1B) for α7 nAChR in rat brain membranes labeled with the α7 nAChR radioligand [3H]methyllycaconitine. The compound was approximately 50-fold more potent than the natural agonist acetylcholine (Ki = 3 μM) and 10-fold more potent than nicotine (Ki = 770 nM) measured using the same conditions.
ABBF inhibited binding of the 5-HT3 receptor-selective radioligand [3H]GR65630 to the human recombinant 5-HT3 receptor expressed in HEK293 cells with similar affinity (Ki = 60 ± 10 nM; Fig. 1F).
The application of ABBF to Xenopus oocytes expressing recombinant rat or human α7 nAChRs produced a strong and reversible, rapidly desensitizing inward current, which is typical for a nicotinic agonist. The concentration-response curves for ABBF and acetylcholine (Fig. 1C) suggested that ABBF was a full agonist of the rat and the human α7 nAChRs with an EC50 value of 3 μM (pEC50 ± S.E.M. = 5.47 ± 0.20 and 5.51 ± 0.06, respectively), >50-fold more potent than the activity of acetylcholine in this assay format (EC50 = 170 μM) (data for the human α7 nAChRs are shown in Fig. 1C). Coapplication of ABBF with a reference concentration of 50 μM acetylcholine increased the acetylcholine response, shifting the concentration-response curve for ABBF to the left (EC50 = 0.5 μM; pEC50 ± S.E.M. = 6.29 ± 0.09; Fig. 1D). The most sensitive assay for the functional activity of ABBF was a brief (1-min) preincubation with ABBF that desensitizes the receptor and thus reduces subsequent responses to acetylcholine. Using this assay format, ABBF showed an IC50 value of 0.1 μM (pIC50 ± S.E.M. = 7.01 ± 0.37; Fig. 1E).
ABBF (tested at concentrations of up to 100 μM) had no agonist activity at recombinant α4β2, α3β4, and muscle nAChRs or 5-HT3 receptors. Coapplication of different concentrations of ABBF with constant reference concentrations of acetylcholine producing approximately 10 to 20% of the maximal response for each receptor showed that high concentrations of ABBF had an inhibitory effect on α3β4 (IC50 = 1.5 μM; pIC50 ± S.E.M. = 5.83 ± 0.08; acetylcholine reference concentration, 3.2 μM), α4β2 (IC50 = 7.6 μM; pIC50 ± S.E.M. = 5.12 ± 0.08; acetylcholine reference concentration, 0.32 μM) and muscle nAChRs (IC50 = 6.4 μM; pIC50 ± S.E.M. = 5.19 ± 0.42; acetylcholine reference concentration, 0.1 μM). ABBF was selective versus the muscarinic AChRs (EC50 and IC50 > 10 μM at M1–M5 muscarinic acetylcholine receptors).
Behavioral Experiments
Social Recognition. ABBF did not significantly change social investigation time during the first encounter (T1). After a 24-h delay, vehicle-treated adult rats no longer showed a reduced social investigation time at the second encounter (T2) with the same juvenile rat that they had inspected at T1, indicating that social recognition memory was lost after 24 h. ABBF significantly improved the social recognition performance of adult rats. To assess the duration of the effect, ABBF was administered at different times before T1. Treatment of adult rats with 1 mg/kg p.o. ABBF between 30 min and 4 h before T1 improved the recognition performance (F3,28 = 4.99; p < 0.01). Post hoc LSD comparisons confirmed that at all intervals between drug administration and T1 (0.5, 1.5, and 4 h), ABBF reduced the percentage of social investigation time at T2 (Fig. 2A).
To exclude possible noncognitive effects of ABBF on the adult rat that might change its investigation of the familiar juvenile at T2, the social investigation time was measured in parallel groups with a familiar or a novel juvenile. A two-way ANOVA followed by post hoc LSD comparisons showed that only the group that was treated with 0.3 mg/kg p.o. ABBF and was confronted with the familiar juvenile at T2 had a significant reduction of the percentage of social investigation time at T2 (p < 0.05; Fig. 2B).
To confirm that the improvement of recognition memory by ABBF was due to stimulation of α7 nAChR receptors, the effect of i.c.v. administration of the α7 nAChR antagonist MLA at 10 μg was tested. A two-way ANOVA revealed a significant effect of ABBF (F1,23 = 4.67; p < 0.05), MLA (F1,23 = 5.93; p < 0.05), and an interaction between the treatment with ABBF and MLA (F1,23 = 7.92; p < 0.01). Post hoc t tests confirmed that treatment with 1 mg/kg ABBF improved social recognition performance; this effect was antagonized by i.c.v. administration of 10 μg of MLA (Fig. 2C). Object Recognition. In both experiments, mice displayed a similar level of exploration of the objects during T1 after vehicle or drug treatment (F1,46 = 0.73; N.S. and F3,44 = 1.23, N.S.; data not shown). During T2, vehicle-treated mice spent as much time exploring a novel object as exploring a familiar object that they had been exposed to 24 h previously, resulting in a discrimination index of 0 (=no object recognition memory). Treatment with 1 mg/kg ABBF improved the object discrimination performance of the OF1 mice at T2 (d2: F1,46 = 15.89; p < 0.01; Fig. 3A). The exploration times at the novel object and the familiar one at T2 are depicted in Fig. 3B.
In a second experiment, treatment with 0.3 and 1 mg/kg ABBF improved the object discrimination performance of the OF1 mice at T2 (d2: F3,44 = 3.02; p < 0.05; Fig. 3C), but not treatment with 0.1 mg/kg ABBF or vehicle. The exploration times at the novel object and the familiar one at T2 are depicted in Fig. 3D.
Repeated Acquisition (Working Memory) Performance of Aged Fischer 344 × Brown Norway Rats in the Water Maze. Swimming speed was not affected in any of the sessions, i.e., the treatment did not affect sensorimotor performance or motivation of the rats. Treatment with ABBF did not affect the distance swum to reach the platform, averaged over the four repeated acquisition sessions (general mean: F2,17 = 0.73; N.S.; Fig. 4). All groups learned to decrease the distance traveled across training sessions (sessions: F3,51 = 10.65; p < 0.01). This improvement was similar for the three groups of rats (sessions × treatment interaction: F6,51 = 0.54; N.S.).
There was an overall effect of trial pairs, i.e., the average performance in the first trials differed from that in the second trials (F1,17 = 7.30; p < 0.05.). Treatment with ABBF affected the change in performance from the first to the second trials of pairs (trial pairs × treatment interaction: F2,17 = 8.02; p < 0.01). There were no effects of sessions on this effect (sessions × trial pairs interaction: F3,51 = 0.55; N.S.; sessions × trial pairs × treatment interaction: F6,51 = 0.56; N.S.).
The treatment effect on trial pairs was further evaluated by an analysis that compared the difference scores between the mean of all first trials and the mean of all second trials over the four repeated acquisition sessions. The overall difference score between first and second trials was affected by the drug treatment (F2,17 = 8.02; p < 0.01). Post hoc comparisons confirmed that the group treated with 1 mg/kg ABBF reduced the distance swum to reach the platform during the second trials compared with the first trials. This was not the case for the vehicle-treated control group and the group treated with 0.3 mg/kg ABBF. Additional t tests confirmed that the difference score of the group treated with 1 mg/kg (but not vehicle or 0.3 mg/kg) exceeded zero, i.e., the second trials had shorter escape latencies than the first trials (vehicle control: t5 = 1.49, N.S.; 0.3 mg/kg ABBF: t6 =–1.80, N.S.; and 1 mg/kg ABBF: t6 = 3.34; p < 0.05).
Drug Discrimination. Fourteen of 16 rats learned to discriminate nicotine (0.4 mg/kg s.c.) from vehicle, the median number of sessions to reach criterion being 41 (range, 26–89 sessions). As assessed at the 15-min injection-test interval used during training, the generalization obtained with nicotine was dose-dependent [ED50 value (95% confidence limits): 0.11 (0.06–0.20) mg/kg s.c.], and complete (≥80%) at 0.2, 0.4, and 0.8 mg/kg (Fig. 5). Generalization occurred in the absence of behavioral disruption (100% lever selections at each dose tested). ABBF did not cause any behavioral disruption and only reached a maximal level of 33% drug lever selection, i.e., did not generalize to the nicotine cue in the dose range tested (0.3–30 mg/kg; Fig. 5).
Discussion
The present study characterizes ABBF as a potent α7 nAChR agonist that improves recognition and working memory in rodents and that does not generalize to a nicotine cue. ABBF competed with [3H]methyllycaconitine for binding to α7 nAChRs in rat brain membranes with a Ki value of 60 nM, which is approximately 50-fold more potent than the natural agonist acetylcholine (Ki = 3 μM) and 10-fold more potent than nicotine (Ki = 770 nM).
The functional studies showed that ABBF is a selective agonist specific for the α7 nAChR subtype relative to the other nAChRs tested (α3β4, α4β2, and muscle nAChR). The nAChR subtypes used for the selectivity tests were chosen as representative of the major classes of heteropentameric nicotinic receptors: αβγδ for the muscle nAChR, α3β4 for the ganglionic nAChR family, and α4β2 as the major variety of nAChR in the central nervous system. ABBF also showed high affinity for the 5-HT3 receptor in binding assays and antagonist activity in electrophysiological experiments. 5-HT3 receptor antagonists are in clinical use as antiemetics, and they are generally well tolerated (Sorbe, 1996). Therefore, 5-HT3 receptor antagonist activity would not preclude potential clinical application of a combined α7 nAChR agonist/5-HT3 receptor antagonist such as ABBF. GTS-21, a weak partial α7 nAChR agonist that has cognitive-enhancing effects in healthy volunteers, is also an antagonist at 5-HT3 receptor and it inhibits α4β2 nAChRs (Kitagawa et al., 2003; Kem et al., 2004). The observation that several ligand classes show high affinity to both α7 nAChRs and 5-HT3 receptors may be due to the relatively high degree of sequence similarity.
The activity of ABBF at the rat and human α7 nAChR is very similar, in contrast to several other ligands that show species differences in their potency and efficacy due to four amino acids changes in the ligand binding site of the rat compared with the human α7 nAChR (Stokes et al., 2004). On coapplication with acetylcholine, the concentration-response curve of ABBF was shifted to the left, suggesting that low concentrations of ABBF can potentiate the responses to acetylcholine. The most sensitive assay for the functional activity of ABBF was preincubation that desensitized the α7 nAChR in the oocyte assay and reduced subsequent responses to acetylcholine. This assay format, which is probably the most relevant for the effects observed after application in vivo (where low and constant levels of ABBF are expected to occur at the synapses) resulted in an IC50 value of 100 nM, in good agreement with the affinity measured in binding assays with the α7 nAChR radioligand [3H]MLA at 60 nM.
To further elucidate the role of α7 nAChR in learning and memory, we examined the effects of ABBF in the social recognition test. This test measures the difference of the investigation time an adult animal displays during the first and the second encounter with a juvenile animal. Rodents have an innate interest in their conspecifics, and their olfactory discrimination capabilities result in recognition of a juvenile they have previously examined. This recognition leads to a shorter investigation time during the second encounter. In a previous study, we demonstrated that the selective α7 nAChR agonist AR-R 17779 improves the performance of rats in the social recognition test (van Kampen et al., 2004). Normal rats do not remember a previously inspected juvenile if the second encounter takes place after 24 h: the investigation time is similar in both encounters. However, if rats are treated with the α7 nAChR agonist AR-R 17779 or the acetylcholinesterase inhibitor metrifonate before the first encounter, the social investigation time is significantly reduced during the second encounter.
In the present study, we have demonstrated that another α7 nAChR agonist, ABBF, can also improve performance of rats in the social recognition test, confirming and extending the initial observations. Improved memory performance was observed after administration of ABBF for up to 4 h before T1, suggesting that no rapid desensitization of the effect after agonist application is taking place. In contrast to the in vitro experiments with recombinant receptor, desensitization may occur to a lower extent in vivo, or the desensitized receptor conformation may still activate signal transduction pathways. The behavioral changes observed are memory-specific, because treatment with ABBF did not reduce the investigation time after 24 h, if a novel juvenile was presented instead of the same as in the first inspection period. The α7 nAChR antagonist MLA blocked the improved memory performance induced by ABBF, confirming that it is mediated by stimulation of α7 nAChRs, rather than functional antagonism caused by desensitization of the α7 nAChR. Moreover, it suggests that 5-HT3 receptors are not involved in the cognitive effect of ABBF.
As a second test to assess the effects of ABBF on learning and memory, we used the object recognition task, which was first described for rats (Ennaceur and Delacour, 1988) and subsequently adapted to mice (Dodart et al., 1997). It is based on the spontaneous behavior of rodents to explore a novel object more than a familiar object that they have already explored several minutes or hours before. After a 24-h retention interval, the animals will no longer discriminate between a known and an unfamiliar object. Performance in this task is improved by phosphodiesterase 5 inhibitors (Prickaerts et al., 2005) and phosphodiesterase 2 inhibitors (Boess et al., 2004), and it is impaired by scopolamine (Dodart et al., 1997). The α7 nAChR agonist ABBF (0.3 and 1 mg/kg p.o.) significantly increased exploration of the novel object, consistent with an improved memory of the familiar object.
To examine the effects of ABBF on working memory in the Morris water maze, we used a modification of the repeated acquisition paradigm described by Whishaw (1985). In previous studies, we had found that aged (26-month-old) FBNF1 rats showed profound performance deficits in the Morris water escape task, compared with adult counterparts, and that the age-related decline was even more profound in rats aged 33 to 34 months (van der Staay, 2006). Moreover, 12 of the 32 aged rats in the present study were unable to acquire the standard task. These rats were excluded from training in the repeated acquisition task. We found that 1 mg/kg ABBF improved working memory in the very old FBNF1 rats, as evidenced by the decrease in distance traveled from the first to the second trials in trial pairs. This test can be considered a working memory variant of the Morris water escape task because within a session, the animal must use the information gathered in the first trial from a particular start position to the submerged platform to improve performance in the second. We have previously investigated in detail whether the improvement within a repeated acquisition session is a consequence of an overall improvement from trial to trial and found that there is true within trial-pairs improvement. The latency from the second trial in the pair to the first trial in the next pair increased in adult rats, indicating that a new start position represents a new problem (van der Staay and de Jonge, 1993). Although α7 nAChR knockout mice did not show impairments in the basic Morris water maze (Paylor et al., 1998), a deficit was reported in a delayed matching-to-place modification, pointing to a role of α7 nAChRs in working memory, in agreement with our observations (Fernandes et al., 2006).
Although we have shown that the α7 nAChR antagonist MLA prevents the effect of ABBF in the social recognition test, we cannot exclude a contribution of 5-HT3 receptor inhibition to the efficacy of ABBF in the other behavioral tests. However, the observations with other α7 nAChR agonists that cause varying degrees of 5-HT3 receptor inhibition suggest that α7 nAChR activation is also responsible for the improvements in the object recognition and water maze tests. During the preparation of this article, two other groups reported on the activity of α7 nAChR agonists in models of learning and memory. SSR180711 enhanced performance in the object recognition task and restored MK-801-induced deficits in the water maze (Pichat et al., 2007). PHA-543,613 improved object recognition (Wishka et al., 2006). These studies support our present results and confirm the previous observations with GTS-21 and AR-R 17779 (Arendash et al., 1995; Briggs et al., 1997; van Kampen et al., 2004).
The finding that ABBF failed to induce a significant level of generalization to the nicotine cue (i.e., maximal level of 33% generalization) is in accordance with previous findings obtained with another α7 nAChR agonist (AR-R 17779) in a similar drug discrimination assay (Kaiser et al., 1998). Although we cannot rule out a contribution of the different routes of administration (s.c. for nicotine versus p.o. for ABBF), the fact that ABBF does not generalize to the nicotine cue at doses that are active in the memory tasks supports the suggestion that the discriminative effect of nicotine is not (primarily) mediated by the α7 nAChR. It was previously reported that the nicotine cue is blocked by antagonists of the β2-subunit of the nAChR (Shoaib et al., 2000) but not by selective α7 nAChR antagonists (Brioni et al., 1996). In addition, transgenic mice that lack the β2-subunit of the nAChR fail, or hardly learn to detect, the nicotine cue (Shoaib et al., 2002), whereas α7 nAChR knockout mice acquire discrimination of nicotine at a rate similar to wild-type mice (Stolerman et al., 2004). It has been argued that the discriminative stimulus effect of nicotine is closely related to its positive reinforcing stimulus effect (for discussion, see Merlo Pich et al., 1999; Stolerman et al., 1999) and that testing the extent of generalization to the nicotine cue offers an opportunity to assess whether such a compound has nicotine-like abuse potential. Interestingly, self-administration studies with nicotine have already indicated that the α7 nAChR most probably does not play a (major) role in the positive reinforcing stimulus of nicotine (Epping-Jordan et al., 1999; Merlo-Pich et al., 1999; Grottick et al., 2000). Therefore, it is suggested that selective α7 nAChR agonists do not share the abuse potential characteristics of nicotine.
Taken together, the present study showed that the α7 nAChR agonist ABBF can improve performance in several learning and memory tests in both rats and mice without producing nicotine-like discriminative stimulus effects. Therefore, α7 nAChR agonists may provide a novel therapeutic strategy for the treatment of cognitive deficits in patients suffering from memory disorders, such as Alzheimer's disease and other dementias.
Acknowledgments
We acknowledge the assistance of P. Ammelung, D. Chodor, H. Dresen, K. Jentzsch, M. Keil, S. Kellermann, D. Klankers, and M. Kuster. The work described was conducted in the Pharma Research CNS Department of Bayer HealthCare AG (Wuppertal, Germany) and Bayer Technology Services (Leverkusen, Germany) from August 2002 to June 2003.
Footnotes
-
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
-
doi:10.1124/jpet.106.118976.
-
ABBREVIATIONS: nAChR, nicotinic acetylcholine receptor; MLA, methyllycaconitine; 5-HT, 5-hydroxytryptamine; ABBF, N-[(3R)-1-azabicyclo[2.2.2]oct-3-yl]-7-[2-(methoxy)phenyl]-1-benzofuran-2-carboxamide; HEK, human embryonic kidney; D, drug; V, vehicle; ANOVA, analysis of variance; LSD, least significant difference; MK-801, 5H-dibenzo[a,d]cyclohepten-5,10-imine (dizocilpine maleate); A-85380, 3-(2(S)-azetidinylmethoxy) pyridine; AR-R 17779, (–)-spiro[1-azabicyclo[2.2.2]octane-3,5′-oxazolidin-2′-one; GTS-21 (DMXB), 3-[2,4-dimethoxybenzylidene]anabaseine; GR65630, 3-(5-methyl-1H-imidazol-4-yl)-1-(1-methyl-1H-indol-3-yl)-1-propanone; SSR180711, 1,4-diazabicyclo[3.2.2]nonane-4-carboxylic acid, 4-bromophenyl ester; PHA-543,613, N-[(3R)-1-azabicyclo[2.2.2]oct-3-yl]furo[2,3-c]pyridine-5-carboxamide.
-
↵1 Current affiliation: Lilly Research Centre, Windlesham, United Kingdom.
-
↵2 Current affiliation: Grünenthal GmbH, Aachen, Germany.
-
↵3 Current affiliation: Merz GmbH, Frankfurt, Germany.
-
↵4 Current affiliation: Department of Pharmacology, University of Iowa City, Iowa City, Iowa.
-
↵5 Current affiliation: Animal Sciences Group, Wageningen University and Research Center, Lelystad, The Netherlands.
-
↵6 Current affiliation: Schwarz BioSciences GmbH, Monheim, Germany.
-
↵7 Current affiliation: EnVivo Pharmaceuticals, Watertown, Massachusetts.
- Received December 21, 2006.
- Accepted February 12, 2007.
- The American Society for Pharmacology and Experimental Therapeutics