Both preclinical evidence and clinical evidence suggest that α7 nicotinic acetylcholine receptor activation (α7nAChR) improves cognitive function, the decline of which is associated with conditions such as Alzheimer’s disease and schizophrenia. Moreover, allosteric modulation of α7nAChR is an emerging therapeutic strategy in an attempt to avoid the rapid desensitization properties associated with the α7nAChR after orthosteric activation. We used a calcium assay to screen for positive allosteric modulators (PAMs) of α7nAChR and report on the pharmacologic characterization of the novel compound RO5126946 (5-chloro-N-[(1S,3R)-2,2-dimethyl-3-(4-sulfamoyl-phenyl)-cyclopropyl]-2-methoxy-benzamide), which allosterically modulates α7nAChR activity. RO5126946 increased acetylcholine-evoked peak current and delayed current decay but did not affect the recovery of α7nAChRs from desensitization. In addition, RO5126946’s effects were absent when nicotine-evoked currents were completely blocked by coapplication of the α7nAChR-selective antagonist methyl-lycaconitine. RO5126946 enhanced α7nAChR synaptic transmission and positively modulated GABAergic responses. The absence of RO5126946 effects at human α4β2nAChR and 5-hydroxytryptamine 3 receptors, among others, indicated selectivity for α7nAChRs. In vivo, RO5126946 is orally bioavailable and brain-penetrant and improves associative learning in a scopolamine-induced deficit model of fear conditioning in rats. In addition, procognitive effects of RO5126946 were investigated in the presence of nicotine to address potential pharmacologic interactions on behavior. RO5126946 potentiated nicotine’s effects on fear memory when both compounds were administered at subthreshold doses and did not interfere with procognitive effects observed when both compounds were administered at effective doses. Overall, RO5126946 is a novel α7nAChR PAM with cognitive-enhancing properties.
Nicotinic acetylcholine receptors (nAChRs) belong to the superfamily of neurotransmitter-gated ion channel receptors. Functionally distinct nAChR channels can be formed either as homomeric pentamers, such as those that contain exclusively α7 subunits, or heteromeric pentamers, such as α4β2, α4β4, or α1βγδ subunits (Changeux et al., 1984; Gotti et al., 1997; Dajas-Bailador and Wonnacott, 2004). α7nAChRs are present in areas of the brain that play critical roles in memory formation and cognition, including the hippocampus, cerebral cortex, and mesocortical limbic system. In these regions, they modulate neurotransmission both presynaptically and postsynaptically as these receptors are expressed on axonal terminals as well as on cell somata and dendrites (Frazier et al., 1998a,b; Fabian-Fine et al., 2001; Marchi et al., 2002).
The considerable interest in developing α7nAChR-targeted therapies stems from reports that activation of these receptors can improve the performance of rats in tests designed to quantify learning and cognition (Arendash et al., 1995; Meyer et al., 1998; Levin et al., 1999; van Kampen et al., 2004; Wallace and Porter, 2011) and findings that selective blockade of α7nAChRs impairs performance (Felix and Levin, 1997; Bettany and Levin, 2001). Observations of a reduction in α7nAChR protein expression have also been reported in Alzheimer patients (Burghaus et al., 2000). Furthermore, human studies of healthy volunteers have shown that the α7nAChR-selective partial agonist GTS-21 increased performance in attention and memory tests (Kitagawa et al., 2003). The demonstration of functional effects of α7nAChR agonists in humans makes this receptor an attractive target for drug development.
α7nAChRs are highly permeable Ca2+ channels, pharmacologically identifiable by their high affinity for the antagonists α-bungarotoxin (α-BTX) and methyl-lycaconitine (MLA) and biophysical properties that include both rapid activation and rapid decay of current. Additionally, α7nAChRs readily desensitize (loss of response) as a result of exposure to agonist (Couturier et al., 1990; Seguela et al., 1993). This attenuation of functional response with repeated agonist occupancy is thought to be an important mechanism for controlling the number of functional receptors and thus modulation of synaptic efficacy.
Development of orthosteric-, or primary site–directed, nicotinic acetylcholine receptor agonists has been an important focus of therapeutic discovery efforts; however, such approaches may also have clear limitations. Because of sustained receptor activation, orthosteric site–directed agonists typically cause progressive desensitization and loss of function. Long-term consequences of sustained nicotine exposure are known to include altered receptor properties and expression levels (Lukas, 1991; Hsu et al., 1996).
Positive allosteric modulators (PAMs) offer an alternative strategy with a potential advantage of increasing endogenous neurotransmission. For example, the α7nAChR receptor PAM PNU-120596 [N-(5-chloro-2,4-dimethoxyphenyl)-N′-(5-methyl-3-isoxazolyl)-urea] reduces desensitization and even restores activity of the agonist-desensitized channel (Hurst et al., 2005). Although there is not necessarily an a priori argument indicating that α7nAChR receptor desensitization limits cognitive performance, such compounds offer the possibility of a differentiated therapeutic mechanism. They may prove useful for treatment of cognitive impairing diseases (e.g., Alzheimer’s disease) by altering desensitization processes, which results in loss of functional receptors and potentially development of tolerance (Changeux et al., 1984; Marks et al., 1985; Ksir et al., 1987; Wonnacott, 1990). Other benefits of allosteric modulation may include improved target selectivity and enhancement of receptor function when cholinergic tone is physiologically upregulated. Many molecules, such as ivermectin, 5-hydroxyindole, NS-1738 [N-(5-chloro-2-hydroxyphenyl)-N′-[2-chloro-5-(trifluoromethyl)phenyl]urea], SB-206553 (3,5-dihydro-5-methyl-N-3-pyridinylbenzo[1,2-b:4,5-b′]dipyrrole-1(2H)-carboxamide hydrochloride), and PNU-120596, have been reported to act as PAMs at α7nAChRs (Krause et al., 1998; Zwart et al., 2002; Hurst et al., 2005; Timmermann et al., 2007; Dunlop et al., 2009; reviewed in Gopalakrishnan et al., 2007) and exhibit procognitive effects in vivo (Timmermann et al., 2007; Wallace and Porter, 2011; Callahan et al., 2013). PNU-120596 has also been shown to improve the auditory gating deficit caused by amphetamine, a model proposed to reflect disturbances associated with schizophrenia (Hurst et al., 2005) and more recently to augment the effects of donepezil on learning and memory in aged rodents and nonhuman primates (Callahan et al., 2013).
The aim of the present study was to characterize the in vitro and in vivo pharmacologic properties of a novel and selective α7nAChR-positive allosteric modulator, RO5126946 (5-chloro-N-[(1S,3R)-2,2-dimethyl-3-(4-sulfamoyl-phenyl)-cyclopropyl]-2-methoxy-benzamide), discovered through screening of small-molecule libraries using a cell-based calcium transient assay.
Materials and Methods
Calcium Transient Response Assay.
GH4C1 cells are a suspension culture-adapted continuous strain of rat pituitary cells originally developed from a pituitary tumor and do not contain endogenous nicotinic receptors (Tashjian et al., 1968; Quik et al., 1996). An engineered line of GH4C1 cells with stable recombinant expression of human α7nAChR was generated after transfection of naïve cells with a mixture of a DNA expression vector and cationic lipid transfection reagent. The vector that was used contained the wild-type form of human α7 nicotinic receptor subunit cDNA cloned into a plasmid designed to support constitutive mammalian cell expression in cultures selected for stable integration of the dominant selectable marker hygromycin resistance gene. After transfection, cells were cultured for 24 hours, then collected by centrifugation, and resuspended in fresh media containing 0.25 mg/ml hygromycin. After 2 to 4 weeks of growth in the presence of antibiotic selection, pressure cultures were cloned using limited dilution cloning in 96-well plates. Stock cultures of the GH4C1-α7nAChR cells were cultured at 37°C, in a humidified atmosphere supplemented with 4% CO2, in growth medium consisting of F10 medium (Invitrogen, Carlsbad, CA), 2.5% fetal bovine serum (Summit Biotechnology, Fort Collins, CO), 15% heat-inactivated donor horse serum (Invitrogen), 250 µg/ml Hygromycin B (Invitrogen), and 100 nM MLA (Sigma-Aldrich, St. Louis, MO). Culturing cells in the presence of MLA augmented functional responses to nicotinic agonists measured by increases in intracellular calcium. On the day of the experiment, cells were dispensed into sample wells of a 96-well poly-d-lysine–coated black plastic plate with a clear bottom at 0.5 × 105 cells/well. Sample wells were then supplemented with 1 µM FLUO-3 AM dye (TEFLabs, Austin, TX) in assay buffer: Hanks’ balanced salt solution (Invitrogen) supplemented with 2 mM CaCl2 (Sigma-Aldrich), 10 mM HEPES (Invitrogen), 2.5 mM Probenecid (Sigma-Aldrich), and 0.1% BSA (Sigma-Aldrich). Extracellular dye was removed using an automated plate washer (BioTek Instruments, Winooski, VT) and replaced with assay buffer. By using a 96-well fluorescence intensity plate reader (MDS Analytical Technologies, Sunnyvale, CA), responses to application of PNU-120596 or RO5126946 were monitored for 5 minutes, after which acetylchonline (ACh) (30 μM, final assay concentration) was added to all wells. Similar studies were also conducted using untransfected human embryonic kidney (HEK) cells and PC-12 cells, with changes to cell culture methods and growth media only. PC-12 neurons were differentiated as described previously (Henderson et al., 1994; Takahashi et al., 1999).
[125I]α-BTX Binding Studies.
Affinity at human α7nAChR was determined via competition binding experiments using [125I]α-BTX and membranes derived from GH4C1 cells expressing recombinant, wild-type human α7nAChR. Frozen cell pellets were thawed and homogenized in 10 v/w of homogenization buffer consisting of 50 mM Tris-HCl and 2 mM EDTA (pH 7.4, 4°C) and centrifuged at 500g for 10 minutes at 4°C. Supernatants were collected and membranes pelleted by centrifugation at 48,000g for 20 minutes at 4°C. Supernatants were removed and pellets washed by rehomogenization in 20 ml of fresh buffer (50 mM Tris-HCl, pH 7.4, 4°C). Samples were then centrifuged again (48,000g for 20 minutes at 4°C), and after resuspension, pellets were stored in buffer or used immediately for binding experiments. Binding assays were conducted in assay buffer consisting of 50 mM Tris-HCl, pH 7.4, 25°C. Test compounds and 0.5 nM [125I]α-BTX were added to 96-well assay plates, followed by the addition of radioligand and membrane to all wells. Nonspecific binding was determined in independent replicate wells in which binding reactions were carried out in the presence of 10 µM epibatidine. Reactions were incubated at room temperature for 1 hour and then stopped by rapid filtration over UniFilter 96-well microplates with bonded GF/C filters (PerkinElmer, Inc., Waltham, MA). After the addition of Microscint scintillation cocktail to all filter wells, receptor-bound radioligand was detected using a Top Count liquid scintillation counter (Packard, Waltham, MA). The Kd for [125I]α-BTX at human α7nAChR was determined to be 0.2 ± 0.9 nM, with a Bmax determination of 23 ± 11 pmol/mg membrane protein.
Primary Cultures of Hippocampal Neurons.
Newborn rat hippocampal neurons were cultured as previously described (Biton et al., 2007) with a slight modification. Briefly, hippocampi were isolated from brains of newborn Sprague-Dawley rats (postnatal day 1–3, n = 75; Charles River Laboratories, Hollister, CA), followed by dissociation using papain (1 mg/ml) and gentle triturating. Cells were diluted in primary cell culture medium (Neurobasal-A medium, 2% B27 supplement, 0.5 mM l-glutamine, 100 U/ml penicillin, and 100 mg/ml streptomycin) at a density of 6.5e5 cells/ml; 1 ml was added to a 35-mm dish containing coverslips coated with poly-d-lysine and laminin, which resulted in approximately 700 cell/mm2. Media were changed 48 hours after seeding, and cytosine β-d-arabinoside (10 µM) was added to prevent the proliferation of non-neuronal cells. Subsequently, media were changed every 3–5 days. Neurons were cultured for at least 10 days before experiments were done.
Stable GABAA Receptor Cell Line Generation and Cell Culture.
The stable HEK293 cell line expressing the rat α1β2γ2 subunit combination of the GABAA receptor was derived by transfection of pIRES plasmids containing the subunit cDNAs into HEK cells (CRL 1573, HEK293; American Type Culture Collection, Manassas, VA) as described by the manufacturer. cDNAs encoding the α1 subunit were cloned in the pIRESpuro vector, the β2 subunit cloned in the pIRESneo vector, and the γ2 (short) subunit cloned in the pIREShygro vector (Clontech Laboratories, Inc., Mountain View, CA). The cells were grown in minimal essential medium (Gibco, Grand Island, NY) supplemented with 10% fetal bovine serum, X1 penicillin/streptomycin (Gibco), 200 µg/ml G418 (Gibco), 0.3 µg/ml puromycin (Clontech), 400 µg/ml hygromycin B (Roche Applied Sciences, Indianapolis, IN), and 20 mM HEPES (Gibco). For electrophysiologic experiments, the cells were plated onto glass coverslips coated with 0.1 mg/ml poly-d-lysine.
Whole-Cell Patch-Clamp Electrophysiology.
Experiments performed on cultured rat hippocampal neurons were done at room temperature using postnatal P1-3 neurons, prepared and cultured as described already. Whole-cell currents in response to compounds were recorded from a holding potential of −60 mV. For rat hippocampal neurons, recording electrodes (4–8 MΩ) were filled with a solution consisting of 126 mM CsCH3SO3, 10 mM CsCl, 4 mM NaCl, 1 mM MgCl2, 0.5 mM CaCl2, 5 mM EGTA, 10 mM HEPES, 3 mM Mg2+ATP, 0.3 mM Na+GTP, and 4 mM phosphocreatine. The pH of this solution was adjusted to 7.2. Neurons were perfused with a solution containing 140 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 10 mM HEPES, and 10 mM glucose, adjusted to pH 7.4 (external osmolarity: 310–320, internal osmolarity: 290–300). For GH4C1 cells, Patch pipettes (4–8 MΩ) were filled with a solution consisting of 120 mM CsF, 10 mM CsCl, 10 mM EGTA, 10 mM HEPES, 4 mM Na+-ATP, 0.3 mM Na+-GTP. The pH of this solution was adjusted to 7.3, and external osmolarity was 294. GH4C1 cells were perfused with a solution containing 135 mM NaCl, 5 mM KCl, 2 mM MgCl2, 2 mM CaCl2, 10 mM HEPES, and 25 mM glucose. The pH of this solution was adjusted to 7.4, and external osmolarity was 308. The liquid junction potential for the pair of solutions was 8.6 mV at 20°C calculated using pCLAMP 8.0 (Molecular Devices, Sunnyvale, CA). Compounds were applied in a rapid application technique directly to neurons or cells from the tip of a 250-μm pipette connected to a Y-tube (Kawa, 1996). Recordings were performed using an Axopatch 200A amplifier, filtered at 1 kHz with a sampling frequency of 10 Hz, and analyzed with pCLAMP 8.0 (Molecular Devices). Data are shown as net charge to represent the time integration (60-second interval) of channel activity that occurs in response to drug application (Papke and Porter Papke, 2002).
Whole-Cell Patch-Clamp of GABAA Receptors Expressing Cells.
Experiments were performed with HEK293 cells stably expressing the rat α1β2γ2 subunit combination of the GABAA receptor. Cells were perfused with a solution consisting of 150 adjusted to 7.4 with 1M NaOH. Patch pipettes with (2 or 3 MΩ) were pulled from borosilicate glass (GC150TF-10; Clark; Harvard Apparatus, Holliston, MA) with a DMZ-Universal Puller (DMZ; Dagan Corporation, Minneapolis, MN) and filled with a solution containing 140 mM CsCl, 10 mM HEPES, 11 mM EGTA, 1 mM CaCl2, 1 mM MgCl2, 4 mM Mg-ATP and 25 mM sucrose, pH adjusted to 7.2 with 1 M CsOH. Cells were patch-clamped in whole-cell mode (Vhold = −60 mV) using a MultiClamp 700B patch-clamp amplifier (Axon Instruments/Molecular Devices, Downingtown, PA), and their currents were digitized at 20 KHz and filtered at 1 kHz using pCLAMP 8.0. GABA, in the presence or absence of RO5126946, was applied to the investigated cell for 1 second at 1-minute intervals using a multibarreled microapplicator pipette controlled by a stepping motor (RSC-200; BioLogic Science Instruments, Grenoble, France). For each experiment, at least three GABA control applications were generated, and only cells showing stable GABA current responses were selected for drug testing. Before application of a GABA-RO5126946 mixture, the same concentration of the RO5126946 alone was applied by bath perfusion for 1 minute. After or before the drug testing, β-CCM (methyl β-carboline-3-carboxylate) was applied with and without GABA to the same cell as a control for correct receptor expression. During the HEK293 whole-cell patch-clamp experiments, all solutions contained 0.1% dimethylsulfoxide, which by itself was without detectable effect on the GABA responses.
For data analysis, the pClamp data acquisition program set (ClampFit; Axon Instruments) was used. The maximum current amplitude was measured for each application of GABA in the presence and absence of RO512694. The relative effect of RO5126946 and β-CCM was calculated as Idrug/IGABA, where IGABA is the GABA-induced current in the absence of drug, and Idrug is the GABA-induced current in the presence of RO512694 or β-CCM and expressed as percentage inhibition.
Oocyte Patch-Clamp Electrophysiology.
All experiments were carried out at receptors expressed in Xenopus oocytes using the method of cDNA expression. Xenopus oocytes were prepared and injected using standard procedures. Briefly, ovaries were harvested from female Xenopus laevis that were deeply anesthetized and pitted in accordance with the animal rights rules of the Geneva canton. A small piece of ovary was isolated for immediate preparation, and the remaining part was placed at 4°C in a sterile Barth solution containing in containing 88 mM NaCl, 1 mM KCl, 2.4 mM NaHCO3, 10 mM HEPES, 0.82 mM MgSO4.7H2O, 0.33 mM Ca(NO3)2.4H2O, 0.41 mM CaCl2.6H2O, at pH 7.4, and supplemented with 20 µg/ml of kanamycine, 100 U/ml penicillin, and 100 µg/ml streptomycin. All recordings were performed at 18°C and cells superfused with oocyte ringer medium containing 82.5 mM NaCl, 2.5 mM KCl, 5 mM HEPES, 1.8 mM CaCl2.2H2O, 1 mM MgCl2.6H2O, pH 7.4. Injections of cDNAs encoding human α7nAChR, were performed in at least one hundred oocytes using a proprietary automated injection device (Hogg et al., 2008). To prevent contamination by calcium-activated chloride currents in α7nAChR recordings, oocytes were placed for at least 2 hours in a medium containing 100 μM BAPTA-AM [1,2-bis(2-aminophenoxy)ethane-N,N,N,′N′-tetraacetic acid-acetoxymethyl ester] (Boton et al., 1989). Currents evoked by ACh or other agonists were recorded using an automated process equipped with standard two-electrode voltage-clamp configuration). Unless indicated, cells were held at −80 mV. Data were captured and analyzed using Matlab (Mathworks Inc., Natick, MA). Concentration-activation curves were fit using the empirical Hill equation: where X is ligand concentration, Y is the fraction of evoked current, EC50 is concentration for 50% activation, and nH is the apparent cooperativity.
The pharmacokinetic properties of RO5126946 were evaluated in male Sprague-Dawley rats (Charles River Laboratories). RO5126946 was administered by intravenously or orally at a dose of 3 mg/kg in a Roche formulated vehicle (TG13: hydromellose, USP; polysorbate 80, NF; benzyl alcohol, NF, and sterile water, USP) for the initial analysis. Blood samples were drawn at 0.08, 0.25, 1, 2,4 6, 8, 24 hours for the intravenous dosing and 0.25, 0.5, 1, 2, 4, 6, 8, 24 hours for the p.o. dosing, and the samples were analyzed by standard extraction procedures followed by liquid chromatography–mass spectrometry analysis for the RO5126946 content. Evaluation of cerebrospinal fluid (CSF) concentration and plasma free concentration was measured as described previously (Liu et al., 2006, 2009). The concentration in the CSF was measured in male Sprague-Dawley rats after a dose of 3 mg/kg i.v. Bound versus free RO512646 was determined using heparinized rat plasma obtained from Pel-Freez Biologicals (Rogers, AR) as described previously (Gever et al., 2010). Centrifree Micropartition Devices (Millipore, Bedford, MA) were used to separate the unbound from protein-bound material. RO5126946 was added to ultrafiltered plasma to yield a final concentration between 200 and 5000 ng⋅ml−1 (n = 3); 1 ml of the plasma solutions and 0.3 ml of the ultrafiltrate solution were added to the filtration device and centrifuged (fixed angle) for 20 minutes at 2000g. Protein binding was calculated according to percent of bound = [(mean filtrate concentration − mean plasma concentration)/mean plasma concentration] × 100. In separate experiments, the pharmacokinetic properties at oral doses of 1 and 10 mg/kg were determined similarly in male Sprague-Dawley rats.
Male Sprague-Dawley rats (8–10/group) (Charles River Laboratories) were acclimated in test cages (12 inches wide × 10 inches deep × 12 inches high) (Habitest; Coulbourn Equipment; Harvard Apparatus) equipped with a grid floor (i.e., 7-mm stainless steel bars spaced approximately 1.75 cm apart) for 2 minutes before hearing a 30-second tone, during the last 2 seconds of which the rats received a 1.0-mA scrambled foot shock. The rats received three pairings of the tone and the shock, and 30 seconds after the last pairing were then returned to their home cages. Approximately 24 hours later, the rats were placed into an altered context (i.e., clear Plexiglas mouse cage with a citrus odor in dim lighting) in which they acclimated for 2 minutes before hearing a 3-minute tone (cue test). Freezing behavior (i.e., total lack of motion except for minimal respiratory movements) was measured in the cue test in 5-second increments for a total of 5 minutes (i.e., 2 minutes before the tone and 3 minutes during the tone). RO5126946 (3, 10, and 30.0 mg/kg) was suspended in a Roche-formulated vehicle consisting of hypromellose (2910, 50 cps), USP; polysorbase 80, NF benzyl alcohol, NF; purified or sterile water, USP; and it was administered orally 60 minutes before the training session. Scopolamine hydrobromide (Sigma-Aldrich) (0.18 mg/kg) was dissolved in saline (0.9% NaCl) and administered s.c. 30 minutes before training. Nicotine bitartrate (Sigma-Aldrich) (0.03 and 0.3 mg/kg) was dissolved in saline and administered i.p. immediately before training.
All in vitro data were plotted and analyzed in GraphPad Prism (GraphPad Software Inc., San Diego, CA). Data are presented as means ± S.E.M. EC50 values were estimated from curve fits generated with GraphPad software using nonlinear regression analysis methods. A four-parameter logistic that describes modulation of ACh responses as a function of RO5126946 concentration was used: where X is RO5126946 concentration in log10-transformed M units, Y is measured modulation of ACh response, bottom is extrapolation to the low plateau of measured ACh response, top is extrapolation to the high plateau of measured ACh response, and EC50 is the concentration of modulator that produces half-maximal response. For the fear-conditioning studies, a comparison of the scopolamine control group with the vehicle control group using a two-sample t test with unequal variance assumption was performed to ensure significant memory impairment was produced by scopolamine. RO5126946 dosing groups were then compared in separate analyses with the scopolamine control group and the vehicle control group using a one-way analysis of variance (ANOVA) followed by a Dunnett’s post hoc analysis test. P ≤ 0.05 constituted a significant difference.
RO5126946 Chemical Structure.
The molecular structure of RO5126946 is shown in Fig. 1.
RO5126946 Potently Enhances the Function of α7nAChRs.
RO5126946 increased the magnitude of ACh-induced intracellular Ca2+ transients in GH4C1 cells stably expressing wild-type human α7nAChR. Examination of representative fluorescence intensity plate reader traces (Fig. 2A) reveals that RO5126946 at concentrations of 30 and 100 nM had no agonist activity. Similar to vehicle traces, those concentrations did not result in a change from baseline calcium levels. When ACh was subsequently applied at a final concentration of 30 μM, a large increase of the calcium response was observed in RO5126946-treated cells relative to those subject to vehicle pretreatment. Peak response amplitudes in the presence of RO5126946 (at concentrations ranging from 0.1 to 10,000 nM) were considered relative to 30 μM ACh alone. This analysis reveals that RO5126946 potentiated α7nAChR responses by a maximum of 401% ± 115 with an EC50 for positive modulation of response ∼0.06 μM ± 0.05 (n = 19; Fig. 2B). We found that this potentiation was fully suppressed by pretreatment with 10 nM of the selective α7nAChR antagonist MLA for 10 minutes, as no increase in calcium from baseline levels was observed when ACh was applied to RO5126946 pretreated cells (data not shown). We also examined untransfected GH4C1 cells and found that RO5126946 in the absence or presence of 30 μM ACh elicited no increase in intracellular calcium (data not shown).
RO5126946 Does Not Bind to the α7nAChR Orthosteric Site.
We performed competition binding studies to determine whether RO5126946 interacted with the ACh binding site of α7nAChR. For this, we determined specific binding of the orthosteric site probe radioiodinated α-bungarotoxin ([125I]α-BTX) in the presence of RO5126946 at concentrations that varied from 0.1 to 100,000 nM. RO5126946 maximally displaced ∼20% of the radiolabeled α-BTX associated with membranes at the highest concentration tested (100 μM); in contrast, the α7nAChR orthosteric agonist epibatidine showed concentration-dependent displacement of [125I]α-BTX with maximal (∼95%) inhibition of binding at concentrations ≥0.3 μM and Ki ∼6 × 10−9 M (Fig. 3A). Because RO5126946 does not compete for [125I]α-BTX binding, this suggests that this compound does not act at the orthosteric binding site. To determine whether RO5126946 altered the affinity of ACh for its binding site at α7nAChR, we also tested ACh competition at concentrations that varied from 1 to 10,000 nM for [125I]α-BTX binding in the presence and absence of 10 μM RO5126946 (Fig. 3B). Using a t test for statistical analysis, there was no significant shift (P value = 0.48) in ACh affinity alone (Ki = 6.7 ± 1.8 × 10−6 M, n = 5) compared with its affinity in combination with RO5126946 (Ki = 5.3 ± 2.1 × 10−6 M, n = 3).
Potentiation of Currents at Recombinant α7nAChRs by RO5126946.
To measure the effects of RO5126946 at the α7nAChR, whole-cell currents were recorded using patch-clamp electrophysiologic techniques in GH4C1 cells expressing the α7nAChR receptors. Addition of RO5126946 profoundly increased the peak amplitude of the initial response (Fig. 4A). RO5126946 potentiated α7nAChR whole-cell current responses maximally by 9200 ± 1338%, n = 4. The apparent EC50 of the modulatory effect was 8.4 ± 1.5 × 10−7 M, n = 4 (net charge analysis, normalized to the effect of 1 mM ACh alone; Fig. 4B).
Potentiation of Currents at Native α7nAChR in Cultured Rat Hippocampal Neurons by RO5126946.
We next evaluated the properties of RO5126946 at native α7nAChRs expressed in cultured rat hippocampal neurons (Fig. 5). In these neurons, nicotine-evoked currents were abolished by 10 nM MLA (n = 3, data not shown), which suggests that the currents were mediated mainly by α7 receptors. Currents evoked by 15 µM nicotine (a concentration slightly lower than the EC50 for activation of α7nAChR) were potentiated by the addition of RO5126946. At a concentration of 1 µM, RO5126946 increased the peak amplitude of initial response to 15 µM nicotine by 2800 ± 730% (n = 3). Representative traces of Fig. 5A show the response to 15 µM nicotine, followed by large potentiation of response with coapplication of RO5126946 and full reversibility of the RO5126946 effect after washout and reapplication of nicotine. Consistent with RO5126946 action as a PAM of α7nAChRs the potentiation of whole-cell currents in the hippocampal neuronal cultures is sensitive to blockade by the α7nAChR antagonist methyl-lycaconitine (Fig. 5B). Determination of the concentration-effect relation for the PAM action of RO5126946 (Fig. 5C) revealed an EC50 of 7.7 ± 2.0 × 10−7 M, n = 3 (net charge analysis, normalized to the effect of 15 µM nicotine alone).
Effects of RO5126946 Are Independent of Glutamate and GABAA Receptors.
Since the rat hippocampal neuron preparations used contain both GABAergic and glutamatergic neurons (Reno et al., 2004), it is possible that the α7nAChR potentiation effect observed with RO5126946 was a consequence of indirect activation of glutamate or GABAA receptors. To investigate this possibility, we added selective antagonists to the bath solution for recordings designed to assess the role of these non-nicotinic receptor channels. The addition of the kainate/AMPA (α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid) receptor antagonist DNQX (6,7-dinitroquinoxaline-2,3-dione) (5 μM) and the N-methyl-D-aspartate (NMDA) receptor antagonist APV [D-(−)-2-amino-5-phosphonopentanoic acid] (10 μM) did not significantly decrease the ability of RO5126946 to potentiate nicotine-evoked currents (P = 0.286) (Fig. 5D). To investigate potentiation of GABAA channel function, GABA-activated currents were recorded with the open channel blocker picrotoxin added to the hippocampal neuron superfusion. The addition of 10 µM picrotoxin did not significantly change the 15 µM nicotine-induced current in the presence of a broad range of RO5126946 concentrations (1 μM RO5126946 ± picrotoxin P = 0.95), indicating that the observed effects do not require function of GABAA receptors (Fig. 5E). Additional whole-cell patch-clamp studies using a HEK cell line recombinantly expressing α1β2γ2 subunits of the GABAA receptor (described as a predominant heteromeric form of GABAA channel in the mammalian brain (Whiting, 2003) were performed to assess RO5126946 pharmacologic properties. Whereas the benzodiazepine site GABAA receptor inverse agonist β-CCM (0.1 μM) inhibited GABA currents by 53.6 ± 4.0% (n = 3), RO5126946 (10 µM) had no effect on the same cells (−0.5 ± 1.6%, n = 3, Fig. 5F).
Effects of RO5126946 on the Concentration Activation Curve.
The characteristic signature of a positive allosteric modulator is best recognized when examining its effects on the agonist concentration activation curve. Typically, a positive modulator should cause an increase in the maximal current amplitude, reduction of EC50 and an increase in the slope of the curve, which is a higher Hill coefficient (Edelstein and Changeux, 2010). To examine whether RO5126946 fulfills these characteristics, concentration activation curves to ACh from 1 to 3 μM were determined for human α7nAChR receptors expressed in Xenopus oocytes. To avoid contamination by calcium-activated chloride channels, oocytes were incubated for at least 2 hours in 100 μM BAPTA-AM, a condition that has been shown to provide adequate calcium chelation (see Materials and Methods). Determination of the concentration-activation curves were effectuated on the same cell first in control conditions (ACh alone) and then with coapplication of ACh and 1 μM RO5126946. Typical currents evoked by ACh in the absence or presence of RO5126946 are illustrated in Fig. 6A. A plot of the peak ACh-evoked current as a function of the logarithm of the ACh concentration yielded a typical concentration activation curve as shown in Fig. 6B. These data clearly illustrate that RO5126946 causes a large increase in the current amplitude (from 3.2 to 25 μA), the expected reduction in the EC50 (from 120 to 5 μM), and increase in the Hill coefficient (from 1.7 to 3).
Analysis of the area under the curve (AUC) reveals a net charge transport of 10.2 ± 6.64 micro-Coloumbs (μCb) for the control conditions, and that exposure to 1 μM RO5126946 increases this value to 166 ± 48.2 μCb. The increase in net charge is approximately 16-fold, and the increase in current amplitude is approximately 4-fold. This suggests that exposure to RO5126946 prolongs the α7nACh response and provides indirect evidence of an action to reduce channel desensitization or slow current decay. In addition, computation of the AUC data reveals that the response time course appears modified by exposure of 1 μM RO51261946 to 160 μM ACh but cannot be fitted by a single exponential decay (Supplemental Fig. 2). Together, these data illustrate that RO5126946 acts as a type II modulator at the human α7nACh receptors.
Specificity of RO5126946.
To assess the specificity of RO5126946, experiments similar to those effectuated for the α7nACh receptors were carried out at the human α4β2 receptor. Typical ACh-evoked currents measured in oocytes expressing the human α4β2 receptors in the absence or presence of 1 μM RO5126946 are shown in Fig. 7. These data illustrate that exposure to RO5126946 causes no significant modification of the amplitude of the ACh-evoked current or on the response time course (n = 4). The AUC analysis for the response to 10 μM ACh was 180.65 ± 39.09 μA compared with 168.5 ± 38.49 μA with the addition of 1 μM RO5126946 with a P value = 0.8 (Supplemental Table 1). As substances such as 5-hydroxyindole have been shown to potentiate both the 5-hydroxytryptamine 3 (5-HT3) receptor and the α7nAChR receptor, it is important to assess the effects of RO5126946 at these receptor subtypes. Experiments were conducted in Xenopus oocytes expressing 5-HT3 receptors and are summarized in Fig. 7. Exposure to RO5126946 showed no detectable effects on the 5-HT–evoked current either in amplitude or time course (n = 5) (Fig. 7, left). The AUC analysis for the response to 1 μM 5-HT was 5.5 ± 1.47 μA compared with 6.34 ± 2.13 μA with the addition of 1 μM RO5126946 with a P value of 0.6 (Supplemental Table 2). Additionally, the effect on 5-HT3 response was measured with 10 μM RO5126946, and four of five cells showed no changes with the addition of RO5126946 (Supplemental Fig. 1; Supplemental Table 2). Target selectivity was examined further by measuring binding to a large panel of >70 common receptor, channel, and enzyme targets. No significant effects of RO5126946 were observed as 10 μM RO5126946 failed to reduce binding of probe ligand by >50% at any of the panel targets. (data not shown).
RO5126946 Does Not Interact with Desensitized nAChRs.
Our initial pharmacologic characterization of RO5126946 suggested that this compound behaves similarly to previously described type II PAMs, such as PNU-120596 (Hurst et al., 2005; Timmermann et al., 2007). A key property of such compounds is their ability to interact with desensitized nAChRs and recover their current gating activity. To test whether RO5126946 behaved in this fashion at α7nAChRs, 1 mM nicotine was applied to hippocampal neurons every 5 seconds until all currents disappeared and likely all α7nAChRs became desensitized. In a control experiment, 0.3 µM PNU-120596 and 1 mM nicotine coapplication to desensitized α7nAChRs resulted in complete signal recovery (Fig. 8, top set of traces, n = 2), comparable to potentiation levels recorded before receptor desensitization. When PNU-120596 was substituted by 1 µM RO5126946 in an identical experiment, no recovery of desensitized receptors was observed (Fig. 8, bottom set of traces, n = 4). A variation of this study was also performed, by which continuous application of nicotine to the bath solution induced α7nAChR desensitization. In these studies, the addition of 0.1 μM PNU-120596 to the perfusion buffer after a long interval of continuous exposure to nicotine resulted in activation of a large current response, whereas the addition of 0.3 µM RO5126946 failed to recover any response (data not shown).
Effects of RO5126946 on Spontaneous Inhibitory Postsynaptic Currents in Cultured Rat Hippocampal Neurons.
Activation of α7nAChR located on inhibitory interneurons increases GABAergic neurotransmission (Maggi et al., 2001; Zaninetti et al., 2002). To investigate the ability of RO5126946 to influence synaptic transmission, spontaneously occurring GABAergic synaptic events were recorded from hippocampal neurons. Since neuronal nAChRs can mediate direct postsynaptic effects when located on cell somata and dendrites or modulate neurotransmitter release on axon terminals, the positive modulation of α7nAChR by RO5126946 could occur in either or both locations (Maggi et al., 2001; Zaninetti et al., 2002). To investigate this further, we evaluated the effects of RO5126946 and nicotine on both spontaneous and miniature inhibitory postsynaptic currents (IPSCs) in cultured hippocampal neurons. Concentrations of nicotine (1 µM) and RO5126946 (0.3 µM), which were determined to be subthreshold or minimally effective for increasing the frequency of IPSC current responses when applied alone, were chosen for these studies. Drugs were delivered via a continuous bath application and applied until response stabilized (∼5 minutes). The representative traces and analyzed frequency data show the effects of nicotine and RO5126946 on IPSC activity and indicate 0.3 µM RO5126946 alone does not significantly increase the frequency of spontaneous IPSC events (P = 0.3). However, the coapplication of 0.3 µM RO5126946 and 1 µM nicotine increased the number of IPSC events by approximately 500% (Fig. 9, A and B). When miniature IPSC activity was measured in the presence of 300 nM tetrodotoxin, the application of either 0.3 µM RO5126946 alone (P = 0.7) or 0.3 µM RO5126946 coapplied with 1 µM nicotine (P = 0.9) did not result in a significant increase in event frequency compared with control conditions (Fig. 9, C and D), indicating a greatly diminished effect on neurotransmitter release from the axon terminal. The large increase in IPSC events observed in the absence of tetrodotoxin indicates a direct effect of RO5126946 on synaptic transmission.
Pharmacokinetic Analysis in Rats.
After oral administration of RO5126946 to rats (n = 2 or 3) (see Materials and Methods for specific collection times), plasma exposures of the compound reach peak concentrations between approximately 2.5 and 3.3 hours and exhibited linear pharmacokinetic properties (Table 1). RO5126946 exhibits oral bioavailability of approximately 100%. Protein binding of RO5126946 is high (approximately 99%); however, measurable levels of RO5126946 were quantified (12.1 ± 3.9 ng/ml) in CSF after 3 mg/kg i.v. administration. The t1/2 of RO5126946 is between 2.6 and 5.2 hours, depending on dose, after oral administration. The log P for compound RO5126946 is 3.44 as calculated by ChemDraw (CambridgeSoft, Waltham, MA).
RO5126946 Improves Associative Learning and Memory in Rats.
To determine whether RO5126946 (3, 10, 30 mg/kg p.o.; 60-minute pretreatment before training) had any cognitive enhancing properties in vivo, it was tested in rats using the fear-conditioning model of associative learning and memory after administration of the cognitive impairing agent scopolamine (0.18 mg/kg s.c.; 30-minute pretreatment before training). In this paradigm, several presentations of a conditioned stimulus (CS-tone) were paired to an unconditioned stimulus (unconditioned stimulus–foot shock) on the training day, which elicited a conditioned freezing response indicative of the learned association between the CS and US when assessed 24 hours later on the testing day. During the CS presentation on the test day, animals (n = 8–10) previously treated with scopolamine on the training day exhibited a significant reduction in the percentage of time freezing compared with vehicle-treated animals (P < 0.001). Overall, there was a main effect of treatment in the scopolamine treated groups [F(3, 34) = 15.46; P = 0.0001], with post hoc analyses revealing a significant reversal of scopolamine-induced deficits with all doses of RO5126946 (P ≤ 0.008) (Fig. 10). Comparison of vehicle-treated animals to the RO5126946-treated groups also showed a main effect of treatment [F(3, 34) = 3.16; P = 0.0371]. Post hoc analyses indicated that only the 3-mg/kg dose of RO5126946 was significantly different from vehicle (P = 0.01). During the CS-alone period, no statistically significant differences were observed between treatment groups (P > 0.20). The mean percentage of time spent freezing for each of the groups in the pre-CS period was less than 4% across treatment groups.
Assessment of RO5126946 Interactions with Nicotine on Associative Learning and Memory in Rats.
To assess the effects of RO5126946 in the presence of nicotine, two studies were conducted using the fear-conditioning model. The first study assessed whether coadministration of subthreshold doses of both RO5126946 and nicotine would result in potentiation of the fear memory. Subthreshold doses were identified as the highest noneffective dose of nicotine and/or RO5126946 tested in previously determined dose response curves for each compound (data not shown). This study was designed to look at the synergistic effects of combining a nicotinic agonist and a nicotinic α7 PAM. In this first study, rats (n = 9 or 10) were coadministered RO5126946 (1.0 mg/kg p.o.; 60-minute pretreatment) and nicotine (0.03 mg/kg i.p.; 0-minute pretreatment) before the fear-conditioning training and tested 24 hours later. Scopolamine (0.18 mg/kg s.c.; 30-minute pretreatment before training) produced a significant decrease in the percentage of time spent freezing during the CS compared with vehicle-treated animals (P < 0.05) as observed previously (Fig. 11A). ANOVA results indicated a main effect of treatment comparing the various scopolamine-treated groups [F(3, 34) = 3.35; P = 0.03], with post hoc analyses revealing a significant reversal of the scopolamine-induced deficit in fear memory when RO5126946 and nicotine were coadministered (P = 0.02). Both nicotine and RO5126946 had no effect on reversing the scopolamine-induced deficit when administered alone (P = 0.99 for both treatments). In addition, there was a main effect of treatment comparing the vehicle and scopolamine-treated groups [F(3, 34) = 7.27; P = 0.0007] with post hoc analyses indicating significant difference between vehicle and RO5126946/scopolamine (P = 0.0025) and vehicle- and nicotine/scopolamine- (P = 0.0028) treated groups. In addition, there were no significant treatment effects were identified for the percentage of time freezing during the pre-CS test period for vehicle and scopolamine-treated groups [F(3, 34) = 0.70; P = 0.5587] or for the scopolamine-treated groups [F(3, 34) = 1.13; P = 0.3518] by ANOVA. The mean percentage of of time spent freezing for each of the groups in the pre-CS period was less than 2% across treatment groups.
The second study was designed to investigate whether coadministration of effective doses of RO5126946 and nicotine interfered with the fear memory. In this study, before FC training, RO5126946 (3 mg/kg p.o.; 60-minute pretreatment) and nicotine (0.3 mg/kg i.p.; 0-minute pretreatment) were administered to rats (n = 9 to 10). Twenty-four hours later, animals were tested for fear memory in response to the CS. Scopolamine (0.18 mg/kg s.c.; 30-minute pretreatment) produced a significant decrease in the percentage of time spent freezing compared with vehicle-treated animals (P < 0.05) (Fig. 11B). A main effect of treatment was identified for scopolamine-treated groups [F(3, 34) = 9.5; P = 0.0001]. Post hoc analyses of these data indicated that RO5126946 (P = 0.0002), nicotine (P = 0.0001), and the combination of RO5126946 and nicotine (P = 0.0031) significantly reversed the scopolamine-induced deficit in fear memory. No significant main treatment effect was identified for the vehicle- and scopolamine-treated groups [F(3, 34) = 2.7; P = 0.0612]. In addition, no significant main effects of treatment on the % time freezing during the pre-CS test period were found for any of the scopolamine-treated groups [F(3, 34) = 1.95; P = 0.1402] or vehicle and scopolamine groups [F(3, 34) = 0.75; P = 0.5273]. Mean percentage of time spent freezing for each of the groups in the pre-CS period was less than 3% across the treatment groups.
α7nAChR PAMs represent a promising approach for augmenting channel function while limiting liabilities associated with agonist-induced desensitization or poor selectivity at the α7nAChR channel. Although it is unclear how complex animal behaviors such as learning and memory are affected by agonist-induced desensitization of α7nAChRs, the development of α7nAChR PAMs is of significant interest since such drugs may effectively avoid receptor desensitization processes and the consequent loss of receptor function and development of tolerance as found in studies of agonists in various preclinical models (Ksir et al., 1987; Wonnacott, 1990; Lloyd and Williams, 2000; Papke and Porter Papke, 2002; Uteshev et al., 2002).
The homomeric α7nAChRs displays unique properties with a very fast desensitization, which resembles that of the glutamate AMPA receptors. The high degree of desensitization of α7 receptors is highly conserved throughout evolution ranging from avian up to mammals, suggesting that this inherent property of the receptor presents functional advantages. Although the functionality of desensitization remains to be elucidated, it is important to consider also that α7 displays a fast recovery after desensitization that could be equally relevant. Current knowledge about the distribution of the α7 receptors suggests that these proteins are expressed presynaptically or in different cell areas but rarely in the postsynaptic organization (Fabian-Fine et al., 2001; Jones and Wonnacott, 2004). The extrasynaptic localization of the α7 receptor raises a question regarding the ACh concentration to which these receptors will be exposed in vivo together with the time course of such changes. Nonetheless, the relevance of the time-dependence function of α7 receptors was illustrated using a different experimental paradigm, as recently shown by different studies in both the cortex and cerebellum (Gu and Yakel, 2011; Gu et al., 2012; Prestori et al., 2013).
Modulation of the α7nAChR activity was shown to affect cognitive function and that agonist can restore drug-induced cognitive deficit (Freedman et al., 2008; Wallace et al., 2011; Prickaerts et al., 2012). Whereas α7nAChR agonists have repeatedly exhibited procognitive properties, it is well recognized that increasing concentrations of agonist lead to a loss of effect, and inverted U-shaped dose-effect curves have been reported in mice, rats, nonhuman primates, and humans (Picciotto, 2003; Wallace et al., 2011). In initial clinical studies in schizophrenic patients tested with the α7nAChR agonist GTS-21, improvement in cognitive measures at the lower doses were observed (Olincy et al., 2006), although it should be noted that in a subsequent study in which a different test battery was used, the procognitive effects were not confirmed at any dose (Freedman et al., 2008). Most recently, in clinical studies with EVP-6124, it is the low dose of the molecule that exhibits procognitive properties in schizophrenic patients, whereas the higher dose is less robust or ineffective, depending on the measure, consistent with the preclinical in vivo data (unpublished presentation by Dana Hilt, American College of Neuropsychopharmacology, 2011). The loss of procognitive effects observed at high concentrations of agonist has been attributed to receptor desensitization based on evidence of these properties from in vitro studies. To our knowledge, nAChR-mediated desensitization has not been measured in vivo; however, the narrow therapeutic ranges observed with the nAChR agonists have limited their investigation in the clinic.
Whereas allosteric modulators have so far received less attention, these molecules are progressively developing and have also shown beneficial outcomes in animal models (for review, see Pandya and Yakel, 2013). As both type I allosteric modulator, such as NS-1738, which is known to increase the α7 response without affecting the desensitization, as well as type II modulator, such as PNU-120596, display beneficial effects in animal models, these data do not provide further clues about the relevance of the fast desensitization of the α7nAChRs. Remarkably, however, low doses of PNU-1205956 were shown to augment the effects of donezepil in aged rodent and nonhuman primates. These data highlight the benefit of α7 modulation, regardless of desensitization (Callahan et al., 2013). Positive modulators offer the potential to exhibit a wider effective dose range that could be investigated in clinical studies to explore the therapeutic potential of activating the α7nAChRs.
This report provides pharmacologic characterization of RO5126946, a potent PAM of the α7nAChR. Application of RO5126946 alone to α7nAChR-expressing GH4C1 cells (at concentrations up to ∼100 nM) had negligible activity for triggering a Ca2+ transient and yet potentiated subsequent ACh-mediated responses that were fully sensitive to blockade by MLA. In competition binding studies, RO5126946 failed to reduce [125I]α-BTX binding to membranes prepared from α7nAChR-expressing GH4C1, consistent with positive modulation of agonist-dependent channel function through an allosteric mechanism. Patch-clamp electrophysiology measurements in recombinant and native systems revealed that coapplication of RO5126946 with ACh increased net charge transport (current AUC) far more than peak amplitude of agonist response, an observation that may reflect an effect to delay response deactivation. The potency of RO5126946 for modulating intracellular Ca2+ transient responses in GH4C1 cells was ∼10-fold higher than that observed for increasing agonist-mediated channel gating in patch-clamp recordings. This difference may be attributed to differences in experimental technique (cell population averages for Ca2+ transient responses versus current recordings from single, patch-clamped cells). Additionally, endogenously expressed voltage-gated calcium channels in GH4C1 cells are presumably recruited after activation of α7nAChR to effectively amplify the intracellular calcium signal (Hinkle et al., 1987; Nelson et al., 1994; Feuerbach et al., 2005) and cause the observed potency shift. In GH4C1 cells recombinantly expressing human wild-type α7nAChRs, agonist-like effects of RO5126946 were observed at concentrations above 1 µM (in the absence of ACh). However, effects of RO5126946 alone (≥1 μM) were not present in electrophysiologic recordings in those same cells or in calcium measurements of parental GH4C1 cells. With regard to potency of the modulatory effects of RO5126946, whole-cell patch-clamp measurements of recombinantly expressed human α7nAChR in GH4C1 cells and those performed in cultured rat hippocampal neurons resulted in potency estimates that were indistinguishable. Taken together with our findings that RO5126946 is noncompetitive with [125I]α-BTX, these findings indicate RO5126946 is not an orthosteric agonist at α7nAChR.
In rat hippocampal neurons, RO5126946 potentiation of nicotine is both fully reversible and completely blocked by MLA. These findings suggest that RO5126946 modulation occurs at the α7nAChR and not downstream of Ca2+ entry into the cell. In rat hippocampal neurons, the addition of NMDA and kainate/AMPA antagonists did not alter ability of RO5126946 to potentiate nicotine evoked currents, suggesting that it is unlikely that NMDA/AMPA receptors have contributed to the effects of RO5126946 on nicotine-evoked currents. Further elucidation of RO5126946 effects on NMDA/AMPA channels would require testing effects in the presence of selective AMPA/NMDA agonists. Finally, in patch-clamped HEK cells recombinantly expressing human GABAA receptors, no effects of RO5126946 on GABA-mediated currents were observed.
Data obtained in Xenopus oocytes expressing the human α7nAChR confirm both the allosteric nature of RO5126946 and its specificity for this receptor subtype. Coapplication of 1 μM RO5126946 with ACh caused a marked increase of ∼8-fold in the α7nAChR response, a leftward shift in the EC50 of ∼24-fold, and an increased Hill coefficient. Concomitant change in these three parameters is typical of an allosteric modulator (Gopalakrishnan et al., 2007; Edelstein and Changeux, 2010). RO5126946 also caused an increase in the AUC reminiscent of the effects of PNU-120596 (Hurst et al., 2005). As experiments were carried out in BAPTA-AM treated oocytes, these effects are attributable to modulation of α7nAChRs. Further experiments carried out in oocytes expressing human α4β2nAChR or human 5-HT3 receptors confirmed the selectivity of RO5126946 for the α7nAChR receptor.
Two types of α7nAChR PAM classes have been reported (Gopalakrishnan et al., 2007; Timmermann et al., 2007). Type I modulators, such as 5-hydroxyindole, NS1738, and ivermectin, increase agonist-evoked peak current but have no effect to delay current inactivation. Type II modulators such as PNU-120596 increase peak current and have a clear effect to slow the rapid rate of α7nAChR channel desensitization. In part, RO5126946 fits the profile exemplified by PNU-120596 since it increases ACh-evoked peak currents and vastly increases net current flux (Supplemental Fig. 2). However, when RO5126946 is applied in the presence of agonist after channel desensitization, no recovery of α7nAChR activity is observed. Because PNU-120596 application under comparable conditions results in rapid emergence of high levels of channel activity, this property clearly distinguishes the activities of RO5126946 and PNU-120596 (Hurst et al., 2005).
In the absence of nicotine, RO5126946 had no effect on sIPSCs, suggesting the presence of little or no endogenous agonist and consistent with data reported for other α7nAChR PAMs (Hurst et al., 2005; Mok and Kew, 2006; Arnaiz-Cot et al., 2008). However, in the presence of nicotine, RO5126946 increased sIPSC frequency in hippocampal pyramidal neurons and positively modulated GABAergic synaptic transmission, thus indicating increased synaptic neurotransmission. Our data suggest that the nicotine-dependent effect of RO5126946 to increase sIPSCs is mediated by α7nAChRs on cell somata and dendrites and requires membrane depolarization.
Given the promising in vitro pharmacologic properties of RO5126946, it was important to determine the in vivo activity of RO5126946. Pharmacokinetic properties of RO5126946 are favorable for oral administration (oral bioavailability ∼100%) and evidence of brain penetration as assessed by detecting measurable levels of the compound in the CSF of rats. Because of the recognized involvement of the α7nAChR in mediating cognitive processes (Rezvani and Levin, 2001; Bitner et al., 2007), we focused our efforts on identifying potential cognitive-enhancing effects of RO5126946. Using the fear-conditioning model of associative learning (for review, see LeDoux, 1992), we showed that vehicle-treated animals exhibited strong fear memory (i.e., freezing ∼80% of the measured time) when represented with conditioning stimulus (tone) 24 hours after undergoing the initial training period. This effect was significantly impaired in animals pretreated with the muscarinic antagonist scopolamine. RO5126946 administration reversed scopolamine-induced deficits in a dose-related manner. The lowest dose tested (3.0 mg/kg) appeared to be a threshold dose as it was significantly different from both vehicle- and scopolamine-treated animals. Doses at or above 10 mg/kg completely reversed the scopolamine-induced impairments as there was no statistical difference between these groups and vehicle-treated animals. After a 10 mg/kg dose, the free plasma concentration of RO5126946 was 34.7 ng/ml, indicating that the effective concentration of drug needed for behavioral effects was low. The procognitive effects of RO5126946 are in agreement with published data by Ng et al. (2007) and Timmermann et al. (2007), which reveal improvements in cognition-based assays after acute administration of the α7nAChR PAMs compound 6 and NS 1738, respectively.
The procognitive effects observed with RO5126946 are also similar to some reports for direct α7nAChR agonists. In particular, nonselective agonists such as nicotine and α7nAChR selective compounds have been shown to improve performance in learning and memory tests in rodents and humans (Levin and Simon, 1998; Rezvani and Levin, 2001; Kitagawa et al., 2003; Wallace et al., 2011). These findings have raised interest in the development of selective α7nAChR agonists for the treatment of neurologic disorders characterized by cognitive decline, such as Alzheimer’s disease and schizophrenia. However, it is possible that nicotinic agonists may not be efficacious in patients who smoke, as up to 70% of schizophrenics normally do (Glassman, 1993; de Leon, 1996; Khantzian, 1997; Dalack et al., 1998; Leonard et al., 2001; Leonard and Freedman, 2003). This finding could potentially be due to direct competition of potential therapeutics with nicotine at the α7nAChR or the desensitization caused by its constant presence. Conversely, allosteric modulators could possibly interact with the high nicotine levels found in smokers in the same way modulators could theoretically synergize with endogenous cholinergic tone.
We investigated RO5126946 interactions with nicotine using the fear-conditioning model. In our initial study, we were interested to determine whether RO5126946 would exhibit positive allosteric modulatory effects in the presence of nicotine in vivo. Although it may be possible to see freezing behavior more than the vehicle response, in our experience, we have not observed the percentage of time freezing to exceed ∼70–80%; therefore, to determine whether additive effects could be demonstrated, we selected ineffective doses of both compounds (as established previously). In this study, neither nicotine (0.03 mg/kg) nor RO5126946 (1.0 mg/kg) was sufficient to reverse the memory impairment induced by scopolamine as predicted; however, both compounds combined produced a significant improvement in the fear memory. These data suggest that RO5126946 potentiated the effects of nicotine.
Since nicotine is a nonselective nAChR agonist, competition for the α7nAChR site, or subsequent desensitization of the receptor after activation may limit the therapeutic effects of the direct agonist approach. Nicotine concentrations are reported to range from 20 to 40 ng/ml in the plasma (Benowitz et al., 1982) and as high as 50 to 100 ng/ml in the arterial blood (Henningfield et al., 1990; Gourlay and Benowitz, 1997) of a typical smoker. To this end, we selected a behaviorally active dose of nicotine (0.3 mg/kg) in rat that yielded plasma exposures of approximately 150 ng/ml at Cmax. In this study, RO5126946 (3 mg/kg) and nicotine (0.3 mg/kg) were both able to reverse the scopolamine-induced deficit in freezing behavior in the fear-conditioning model alone. In addition, when both compounds were coadministered at these same doses, the procognitive effects were maintained compared with the scopolamine-treated group, suggesting that administration of an α7nAChR PAM may have utility as a procognitive agent in a smoking population.
In summary, this study presents in vitro and in vivo pharmacologic profiles of RO5126946, a novel PAM of α7nAChRs. RO5126946 potentiates ACh and nicotine responses in a cell line recombinantly expressing human wild-type α7nAChRs as well as in oocytes and rat hippocampal neurons. RO5126946 increases the peak amplitude of agonist-evoked current and delays channel desensitization, as is typical of type II modulators. However, unlike PNU-120596, RO5126946 appears unable to modulate desensitized α7nAChRs. Importantly, RO5126946 exhibits procognitive properties in vivo both in the presence and absence of nicotine. The pharmacologic properties of RO5126946 described here may enable novel therapeutic approaches for addressing neuronal disorders.
The authors thank Sonia Bertrand, Fabrice Marger, and Sonia Rocha for excellent technical support on this project.
Participated in research design: Sahdeo, Wallace, Hirakawa, Knoflach, Bertrand, Maag, Misner, Tombaugh, Santarelli, Milla, Button.
Conducted experiments: Sahdeo, Wallace, Hirakawa, Bertrand, Misner, Tombaugh.
Contributed new reagents or analytic tools: Brameld.
Performed data analysis: Sahdeo, Wallace, Hirakawa, Knoflach, Bertrand, Maag, Misner.
Wrote or contributed to the writing of the manuscript: Sahdeo, Wallace, Hirakawa, Knoflach, Tombaugh, Bertrand, Milla, Button.
- Received November 20, 2013.
- Accepted May 16, 2014.
↵1 Current affiliation: Molecular Biosciences, School of Veterinary Medicine, University of California, Davis, California.
↵2 Current affiliation: SRI International, Menlo Park, California.
↵3 Current affiliation: Gilead Sciences Inc., Palo Alto, California.
↵4 Current affiliation: Chemistry & Drug Discovery Consulting LLC, Portland, Oregon.
↵5 Current affiliation: Principia Biopharma, South San Francisco, California.
↵6 Current affiliation: Janssen Research and Development, La Jolla, California.
↵7 Current affiliation: Acorda Therapeutics, Inc., Ardsley, New York.
All studies were funded by F. Hoffmann La Roche.
- analysis of variance
- α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid
- D-(−)-2-amino-5-phosphonopentanoic acid
- area under the curve
- 1,2-bis(2-aminophenoxy)ethane-N, N,N,′N′-tetraacetic acid-acetoxymethyl ester
- methyl β-carboline-3-carboxylate
- conditioned stimulus
- cerebrospinal fluid
- human embryonic kidney (cells)
- inhibitory postsynaptic current
- α7 nicotinic acetylcholine receptor
- positive allosteric modulator
- 3,5-dihydro-5-methyl-N-3-pyridinylbenzo[1,2-b:4,5-b′]dipyrrole-1(2H)-carboxamide hydrochloride
- Copyright © 2014 by The American Society for Pharmacology and Experimental Therapeutics