Abstract
Metabotropic glutamate receptor 7 (mGlu7) has been suggested to be a promising novel target for treatment of a range of disorders, including anxiety, post-traumatic stress disorder, depression, drug abuse, and schizophrenia. Here we characterized a potent and selective mGlu7 negative allosteric modulator (NAM) (+)-6-(2,4-dimethylphenyl)-2-ethyl-6,7-dihydrobenzo[d]oxazol-4(5H)-one (ADX71743). In vitro, Schild plot analysis and reversibility tests at the target confirmed the NAM properties of the compound and attenuation of l-(+)-2-amino-4-phosphonobutyric acid–induced synaptic depression confirmed activity at the native receptor. The pharmacokinetic analysis of ADX71743 in mice and rats revealed that it is bioavailable after s.c. administration and is brain penetrant (cerebrospinal fluid concentration/total plasma concentration ratio at Cmax = 5.3%). In vivo, ADX71743 (50, 100, 150 mg/kg, s.c.) caused no impairment of locomotor activity in rats and mice or activity on rotarod in mice. ADX71743 had an anxiolytic-like profile in the marble burying and elevated plus maze tests, dose-dependently reducing the number of buried marbles and increasing open arm exploration, respectively. Whereas ADX71743 caused a small reduction in amphetamine-induced hyperactivity in mice, it was inactive in the mouse 2,5-dimethoxy-4-iodoamphetamine–induced head twitch and the rat conditioned avoidance response tests. In addition, the compound was inactive in the mouse forced swim test. These data suggest that ADX71743 is a suitable compound to help unravel the physiologic role of mGlu7 and to better understand its implication in central nervous system diseases. Our in vivo tests using ADX71743, reported here, suggest that pharmacological inhibition of mGlu7 is a valid approach for developing novel pharmacotherapies to treat anxiety disorders, but may not be suitable for treatment of depression or psychosis.
Introduction
Metabotropic glutamate receptor 7 (mGlu7) has been suggested to be one of the most important receptors for multiple central nervous system (CNS) functions among eight mGluRs, based on it having the widest distribution in the brain and the highest degree of evolutionary conservation across species (Flor et al., 1997). In the brain, mGlu7 is most abundant in anatomic regions involved in emotional reactivity and cognitive functioning, such as hippocampus, amygdala, and the locus coeruleus (Kinoshita et al., 1998; Swanson et al., 2005).
Development in understanding the function of mGlu7 and its use as a potential target for drug discovery has been hampered by the lack of bioavailable and brain-penetrant pharmacological tools. High conservation of the orthosteric binding site has been a serious challenge for developing molecules with a high selectivity at this receptor. Initial evidence on the possible role of mGlu7 in the CNS largely came from a series of studies involving mGlu7 knock-out (KO) mice and those implementing gene silencing techniques (Masugi et al., 1999; Sansig et al., 2001; Cryan et al., 2003). Compared with wild-type controls, mGlu7 KO mice showed consistent reductions in anxiety- and depression-like responses in a variety of behavioral tests (Cryan et al., 2003; Callaerts-Vegh et al., 2006). In addition, mGlu7 KO mice had signs of reduced reactivity of the hypothalamic-pituitary-adrenal (HPA) axis to stress (Mitsukawa et al., 2006), impaired fear extinction response, and a deficit in the conditioned taste aversion paradigm (Masugi et al., 1999; Callaerts-Vegh et al., 2006).
The discovery of N,N′-dibenzhydrylethane-1,2-diamine dihydrochloride (AMN082), presumably the first orally active and brain-penetrant mGlu7 allosteric agonist, promised to provide a much needed pharmacological tool for assessing effects of direct activation of the receptor (Mitsukawa et al., 2005; Conn and Niswender, 2006). However, in vivo evaluation of AMN082 provided conflicting results. On the one hand, AMN082 reduced open arm exploration in the elevated plus maze (EPM) test in rats, indicative of an anxiogenic-like response (Palazzo et al., 2008) and elevated the plasma levels of corticosterone and adrenocorticotropic hormone in mice, indicative of a stress response (Mitsukawa et al., 2005). On the other hand, AMN082 exhibited an anxiolytic-like profile in the rat stress-induced hyperthermia and four-plate tests (Stachowicz et al., 2008) and an antidepressant-like profile in the mouse forced swim test (FST) and tail suspension test (Palucha et al., 2007), an outcome that was not expected from the mGlu7 KO studies. Suggested explanations for these discrepancies are the possibilities that the receptor may be internalized upon activation by AMN082 (Pelkey et al., 2007) resulting in a functional antagonism, as well as activity of its metabolite, Met-1, at serotonin, dopamine, and norepinephrine transporters (Sukoff Rizzo et al., 2011).
Recent discovery of the systemically active negative allosteric modulator (NAM) of mGlu7 6-(4-methoxyphenyl)-5-methyl-3-pyridin-4-ylisoxazolo[4,5-c]pyridin-4(5H)-one (MMPIP) with inverse agonist activity was another step toward uncovering the role of this receptor (Suzuki et al., 2007; Nakamura et al., 2010). However, when tested in vivo, MMPIP impaired nonspatial and spatial memory in the object recognition and in the radial arm maze tests, respectively, and reduced social interaction in rats, while having no effects in a battery of tests relevant for motor function, anxiety, depression, sensorimotor gating, nociception, and seizure threshold (Hikichi et al., 2010). An extensive pharmacological study of MMPIP and close analogs revealed that these compounds show a context-dependent activity when expressed in recombinant cell lines, but were found inactive in a physiologic setup (Niswender et al., 2010). Further studies are needed to understand the in vivo effects, or lack of effects, of MMPIP.
Here we provide a comprehensive characterization of (+)-6-(2,4-dimethylphenyl)-2-ethyl-6,7-dihydrobenzo[d]oxazol-4(5H)-one (ADX71743), a potent, selective, and brain-penetrant mGlu7 NAM (Fig. 1A). It was developed through chemical lead optimization of a hit compound (Tang et al., manuscript submitted), which was identified from a high-throughput screening campaign of our corporate chemical library using a Ca2+ mobilization assay. In vitro, we performed Schild plot analysis and reversibility tests at the target confirming the NAM properties of the compound and demonstrated activity at the native receptor using electrophysiological measures in the mouse hippocampus. After confirming selectivity of ADX71743, we performed pharmacokinetic evaluation of the compound in mice and rats after s.c. administration confirming its suitable profile for in vivo testing. In vivo, after confirming normal motor activity in mice and rats using locomotor activity and rotarod tests, ADX71743 was evaluated in tests relevant for anxiety, such as the mouse marble burying (MB) and EPM tests; depression, such as the mouse FST; and psychosis, such as the amphetamine-induced hyperactivity 2,5-dimethoxy-4-iodoamphetamine (DOI)–induced head twitch tests in mice and the conditioned avoidance response (CAR) test in rats.
The chemical structure of ADX71743 (A), concentration-response curves of ADX71743 and LY341495 on l-AP4–induced intracellular Ca2+ mobilization (B), and concentration-response curves of ADX71743 and LY341495 on intracellular cAMP decrease induced by l-AP4 (C). Data are presented as a percentage of the response to 3 mM l-AP4.
Materials and Methods
Stable Cell Lines.
The cDNAs encoding the human or the rat mGlu7 (hmGlu7 and rmGlu7, respectively) were subcloned into an expression vector also containing the hygromycin resistance gene. For intracellular calcium flux measurement, the cDNA encoding a chimeric Gα protein allowing redirection of the activation signal to intracellular calcium flux was subcloned into a different expression vector also containing the puromycin resistance gene, and both of these vectors were cotransfected into human embryonic kidney (HEK)293 cells with PolyFect reagent (Qiagen, Basel, Switzerland). Subsequently, hygromycin and puromycin treatment allowed selection of antibiotic resistant clones that had stably integrated one or more copies of both plasmids. Alternatively, the hmGlu7 containing expression vector was transfected into HEK293 cells expressing Phoenyx, a cAMP biosensor allowing a dynamic real-time cAMP measurement in live cells described previously (Lütjens et al., 2010). Positive functional cellular clones expressing mGlu7 were identified based on using the reference group III mGluR agonist l-(+)-2-amino-4-phosphonobutyric acid (l-AP4) and the nonselective mGlu orthosteric antagonist (2S)-2-amino-2-[(1S,2S)-2-carboxycycloprop-1-yl]-3-(xanth-9-yl)propanoic acid (LY341495). HEK293 cells expressing rat or human mGlu7 were maintained in media containing Dulbecco’s modified Eagle's medium (DMEM), fetal calf serum (10%), penicillin (100 U/ml), streptomycin (100 μg/ml), gentamycin (Geneticin) (100 μg/ml), hygromycin B (40 μg/ml), and puromycin (1 μg/ml) at 37°C with 5% CO2 in a humidified atmosphere.
Fluorescent Cell-Based Ca2+ Mobilization Assay.
This assay was performed in a pH 7.4 buffered solution containing 20 mM HEPES, 143 mM NaCl, 6 mM KCl, 1 mM MgSO4, 1 mM CaCl2, 0.125 mM sulfinpyrazone, and 0.1% glucose. Twenty-four hours before the pharmacological experiment, hmGlu7 or rmGlu7-transfected HEK293 cells were plated out at a density of 2.5 × 104 cells/well in black-well/clear-bottomed and poly(l-ornithine)–coated 384-well plates in glutamine/glutamate-free DMEM containing 10% fetal bovine serum, 100 U/ml penicillin, and 100 μg/ml streptomycin, supplemented with 1 μg/ml doxycycline. Cells were incubated overnight at 37°C with 5% CO2 in a humidified atmosphere. On the day of the assay, the cells were loaded with a 3-μM dye solution of Fluo4-AM (Invitrogen, Lucerne, Switzerland) in assay buffer containing 0.03% pluronic acid. After 1 hour at 37°C with 5% CO2 in a humidified atmosphere, the extracellular dye was removed by washing the cell plate three times with 1× phosphate-buffered saline (Invitrogen, Lucerne, Switzerland). Assay buffer was added to cells and calcium flux was measured using a fluorometric imaging plate reader (FLIPR) (Molecular Devices, Sunnyvale, CA). After 10 seconds of basal fluorescence recording, compounds to be tested were added to cells in a concentration-dependent manner, and left for incubation on cells for 170 seconds. During that time, changes in fluorescence levels were monitored to detect any agonist activity of the compounds. The cells were then stimulated by l-AP4 EC80 (concentration giving 80% of the maximal l-AP4 response) for an additional 170 seconds to measure inhibiting activities of the compounds. In the reversibility experiments, cells were either washed three times with phosphate-buffered saline (1×) after compound addition, or not. Cells were then stimulated by l-AP4 EC80.
Phoenyx cAMP Assay.
This assay was performed in pH 7.4 (1×) buffered Hanks' balanced salt solution (Invitrogen). Twenty-four hours before the experiment, human mGlu7-Phoenyx-transfected HEK293 cells were plated out at a density of 2 × 104 cells/well in black-well/clear-bottomed and poly(l-ornithine)–coated 384-well plates in media containing DMEM, 10% fetal bovine serum, 100 U/ml penicillin, 100 μg/ml streptomycin, 100 μg/ml gentamycin, 40 μg/ml hygromycin B, and 1 μg/ml puromycin, supplemented with 1 μg/ml doxycycline. Cells were incubated overnight at 37°C with 5% CO2 in a humidified atmosphere. On the day of the assay, the cells were first starved for 90 minutes in the assay buffer at 37°C with 5% CO2 in a humidified atmosphere. The cells were then loaded with a 4-μM solution of coelenterazine H (Dalton Pharma, Toronto, Canada) in assay buffer, and the intracellular cAMP level was measured using a FLIPR (Molecular Devices). After 63 seconds of basal signal recording, compounds to be tested were added to the cells in a concentration-dependent manner, and left for incubation on the cells for 95 seconds. The cells were then costimulated by 10 μM forskolin (Sigma-Aldrich, Buchs, Switzerland), and glutamate or l-AP4 EC80 (concentration giving 80% of the maximal agonist response) for an additional 10 minutes to measure inhibiting activities of the tested compounds. Schild plot experiments were performed using the same protocol, by testing l-AP4 in a concentration-dependent manner in the absence or presence of increasing concentrations of tested compounds.
Selectivity.
ADX71743 was functionally tested up to 30 μM as an agonist, positive allosteric modulator, or NAM of other rat or human members of the mGlu family (mGlu1, mGlu2, mGlu3, mGlu4, mGlu5, mGlu6, mGlu8) using the above-described fluorescent cell-based Ca2+ mobilization assay. In addition, ADX71743 was tested in agonist and antagonist mode in the Cerep P27 cellular functional G-protein coupled receptor (GPCR) profile containing 29 targets (Cerep, Poitiers, France).
Serum Protein Binding.
Serum protein binding was measured by equilibrium dialysis using 96-well plates specifically designed for this purpose (HT Dialysis, Gales Ferry, CT). This reusable 96-well plate is assembled so that each well is divided vertically in two parts by a dialysis membrane. Molecular weight cut-off regenerated dialysis cellulose-membranes (Dialysis Membrane Strips; HT Dialysis) with a molecular weight cut-off 12–14,000 Da were conditioned according to the manufacturer and used for all experiments. One microliter of a 1 mg/ml dimethylsulfoxide (DMSO) solution of ADX71743 was added to mouse serum to reach the final concentration of 1 μg/ml. Portions (150 µl) of the serum solution were added to one side of the membrane and the pH 7.4 phosphate buffer solutions was added to the other side of the well. The addition in the two compartments was made simultaneously. Experiments were carried out in duplicate for each time point. Individual wells were used for each time point. The plates were sealed and set in an incubator at 37°C under gentle shaking. Samples were taken from the serum and buffer compartments at the start of the experiments and they were immediately stored at 4°C after 6 and 7 hours. The serum and buffer samples were subsequently analyzed by a specific liquid chromatography–mass spectrometry (LC-MS) method to measure ADX71743 concentrations. The measured concentration of ADX71743 in serum and buffer was determined. The portion of bound ADX71743 in serum was calculated according to the following equation (Wright et al., 1996):where Cserum is the total (bound + free) concentration of ADX71743 in serum after equilibrium is reached and Cbuffer is the concentration of ADX71743 measured in the buffer solution at the same time. This equation is valid only when the equilibrium between the serum and buffer solutions is completed. Previous experiments demonstrated that, for the reference substances, the equilibrium between the serum and buffer solution was reached after 5–6 hours.
Animals.
Unless otherwise specified below, the studies used adult male C57Bl6/J mice (24–30 g) and Sprague-Dawley rats (250–350 g; Charles River, L’Arbresle, France). Upon arrival to the animal facility, mice were group-housed (n = 5 per cage) in type II cages (16 × 22 × 24 cm), whereas rats were group-housed (n = 2 per cage) in type III cages (22 × 37 × 18 cm). Animals were maintained on a 12-hour light/dark cycle (lights on from 07:00 to 19:00 h) under constant temperature (22°C ± 2°C) and humidity (>45%) conditions. Standard chow and water were available ad libitum. Animals were acclimated for at least 10 days before experimentation. All experimental procedures and conditions were approved by the Ethical Committees of Addex Therapeutics and performed in full compliance with international European ethical standards (86/609-EEC), the French National Committee (décret 87/848) for the care and use of laboratory animals, the UK Animals (Scientific Procedures) Act 1986, and the Declaration of Helsinki.
Hippocampal Electrophysiology.
Adult (>8 weeks) female C57Bl6/J mice (Harlan, Bicester, UK) were killed by decapitation and the brain was removed and placed into ice-cold oxygenated sucrose Krebs medium containing 202 mM sucrose, 2 mM KCl, 1.25 mM KH2PO4, 10 mM MgSO4, 0.5 mM CaCl2, 26 mM NaHCO3, and10 mM glucose. The brain was hemisected along the midline and 300-µm parasagittal slices were prepared with an oscillating microtome (Integraslice; Campden Instruments Ltd., Loughborough, UK). Slices were then transferred to a recovery chamber at room temperature containing oxygenated Krebs solution with 124 mM NaCl, 2 mM KCl, 1.25 mM KH2PO4, 1 mM MgSO4, 2 mM CaCl2, 26 mM NaHCO3, and 10 mM glucose. After at least 1 hour of recovery, individual slices were transferred to an interface recording chamber where they were perfused with Krebs solution (21.5°C–21.9°C). Extracellular field potential recordings were made with an Axoprobe 1A amplifier (Axon Instruments Ltd., Inverurie, UK) via a Krebs-filled glass micropipette (resistance 8–10 MΩ) positioned in the stratum radiatum of the CA1, digitized (5 kHz) via a CED1401 interface, and stored on a computer with Spike2 software (Cambridge Electronic Design Ltd., Cambridge, UK). Field excitatory postsynaptic potential (fEPSP) responses were evoked (0.1-ms pulses applied every 10 seconds; 3.2–4.5 V adjusted to approximately 80% of the maximal spike-free response) by a bipolar stimulating electrode positioned in the stratum radiatum near the CA3-CA1 border.
l-AP4 was prepared as a 100 mM stock solution in water and ADX71743 (10 mM) was prepared in 100% DMSO. Stock solutions were stored in aliquots at −20°C and individual aliquots were thawed on the day of experiment and diluted to the desired concentration in Krebs solution. DMSO was present throughout the recordings at a concentration of 0.1%. For the electrophysiological studies the average peak amplitude of the fEPSPs was measured over six consecutive trials with CED Spike2 software. Data are presented as mean effect ± S.E.M. Statistical significance was determined by paired t test using GraphPad Prism software (GraphPad Inc., San Diego, CA).
In Vivo Pharmacokinetic Studies.
Pharmacokinetic studies after s.c. administration of ADX71743 to mice and rats were performed with a suspension of the drug in a 50% water solution of hydroxyl-propyl-β-cyclodextrin (CD). The volume of administration was 3 ml/kg. ADX71743 was administered at 12.5 and 100 mg/kg to cover the range of doses to be used in in vivo behavioral studies. Blood was collected at 0.25, 0.5, 1, and 2 hours after administration as terminal samples (n = 3 per time point). Samples at 6, 8, and 24 hours were additionally collected in animals receiving the 100 mg/kg dose. In addition, brain and cerebrospinal fluid (CSF) samples were collected in animals receiving the 12.5 mg/kg dose. Blood samples were collected in 1.5-ml polyethylene Eppendorf tubes containing 4-μl 15% ethylenediaminetetraacetic acid (EDTA) solution and immediately placed on ice. Samples were centrifuged at 4°C for 12 minutes at 5900g for 3 minutes (equivalent to 8000g if using an Eppendorf type 5415 centrifuge). Plasma was transferred to 1.5-ml Eppendorf tubes and stored at −20°C until analysis. Brain samples were collected into plastic 24-well plates. CSF was collected into 20-µl heparin-coated capillary tubes. For sample preparation, CSF samples were checked carefully for blood contamination and transferred into Eppendorf tubes containing 20-µl control rat plasma. Subsequently, samples were placed on ice and, if necessary, kept frozen at −20°C until analysis. CSF sample extraction was identical to the procedure described for plasma samples (see below).
Plasma Analysis.
The plasma sample analysis was performed using a tandem LC-MS method (ultra performance liquid chromatography, Waters coupled with API 3200 Qtrap; Applied Biosystems, Zug, Switzerland). The electrospray positive ionization was used in MRM mode (transition 270.09/138.15). The high performance liquid chromatography conditions used were a 1.5-minute gradient with ammonium formate 10 mM pH 3.5/acetonitrile formic acid (15 mM) at 0.8 ml/min starting with 5% of buffer to 100% acetonitrile mobile phase. To prepare plasma samples, 150-µl acetonitrile (protein precipitation) was added to 50 µl of plasma spiked respectively with 10-µl DMSO for unknown samples or 10 µl of ADX71743 for calibration and quality control samples. After vortexing and centrifugation (15 minutes, 4°C, 13,200g), a portion (100 µl) was transferred into the 384-well analytical plate. Five microliters of the supernatant were injected into the system. Quantification process was performed using the Analyst software (AB SCIEX, Framingham, MA). The concentrations of the test item in the samples were calculated from the corresponding peak areas produced and using the calibration equation curve (10 calibration points in duplicate). Three levels of QC samples (in duplicate) were added to validate the run. The assay was qualified valid if at least 75% of the calibration points and at least 66% of the QC samples were determined with accuracy equal or better than 20% of the nominal concentrations. The limits of quantification for the compound in plasma were 1.57 and 0.6 ng/ml at 12.5 and 100 mg/kg doses, respectively.
Rotarod Test in Mice.
A mouse rotarod apparatus (MED Associates, St. Albans, VT) with constant speed (16 rotations per minute) was used in this experiment as described previously (Campo et al., 2011). On the day of the experiment mice (n = 10/group) were treated s.c. with ADX71743 (50, 100, 150 mg/kg) or its vehicle (CD) and were tested on the rotarod in two 3-minutes sessions performed 30 and 90 minutes after treatment. Two additional groups of mice (n = 10/group) were treated via oral gavage (p.o.) with either (R)-4-amino-3-(4-chloro-phenyl) butanoic acid ((R)-baclofen) (10 mg/kg) or its vehicle (saline) and were tested 60 and 120 minutes after treatment. The time spent on rotarod (seconds) was analyzed by the Kruskal–Wallis test followed by Dunn’s multiple comparisons.
Spontaneous Locomotor Activity Test in Mice.
Spontaneous locomotor activity, assessed as horizontal distance traveled (centimeters), was monitored using Plexiglas arenas (35 × 35 × 50 cm) in conjunction with video tracking and computerized analysis systems (Viewpoint, Lyon, France) as described previously (Campo et al., 2011). Mice (n = 10/group) were treated s.c. with ADX71743 (50, 100, 150 mg/kg) or its vehicle (CD). Two additional groups of mice (n = 10/group) were treated p.o. with either (R)-baclofen (10 mg/kg) or its vehicle (saline). Thirty and 60 minutes after administration of ADX71743 and (R)-baclofen, respectively, mice were individually placed into arenas and their locomotor activity was monitored for 60 minutes. The total distance traveled during the test (centimeters) was analyzed by one-way analysis of variance (ANOVA) followed by planned comparisons.
Spontaneous Locomotor Activity Test in Rats.
Spontaneous locomotor activity, assessed as horizontal distance traveled (centimeters), was monitored using Plexiglas arenas (50 × 50 × 50 cm) in conjunction with video tracking and computerized analysis systems (Viewpoint). Rats (n = 9–10/group) were treated s.c. with ADX71743 (50, 100, 150 mg/kg) or its vehicle (CD). Two additional groups of mice (n = 8–9/group) were treated p.o. with either (R)-baclofen (10 mg/kg) or its vehicle (saline). Thirty and 60 minutes after administration of ADX71743 and (R)-baclofen, respectively, mice were individually placed into arenas and their locomotor activity was monitored for 60 minutes. The total distance traveled during the test (centimeters) was analyzed by one-way ANOVA followed by planned comparisons.
MB Test in Mice.
A set of type II cages (with clear Plexiglas covers) used in the experiment contained extra amounts (5 cm high) of sawdust bedding and had 10 marbles evenly spaced against the walls of the cage. Mice (n = 10/group) were treated s.c. with ADX71743 (50, 100, 150 mg/kg), or its vehicle (CD). Two additional groups of mice (n = 10/group) were administered p.o. either chlordiazepoxide (30 mg/kg) or its vehicle (saline). Thirty and 60 minutes after administration of ADX71743 or chlordiazepoxide, respectively, animals were individually placed in experimental cages and were left undisturbed for 30 minutes. At the end of this period, animals were removed from the cage and numbers of buried marbles were counted. The marble was considered to be buried if it had at least two-thirds of its surface covered in sawdust. Terminal blood samples were collected from all ADX71743-treated animals at the end of the experiment and plasma was analyzed as described for the pharmacokinetic studies. The number of buried marbles was analyzed by the Kruskal–Wallis test followed by Dunn’s multiple comparisons.
EPM Test in Mice.
The EPM was made of opaque plastic and consisted of four arms of equal lengths and widths (15 × 4 cm) arranged in the form of a plus sign. Two opposite arms, referred to as closed arms, were enclosed by walls 12 cm high, whereas two remaining arms, referred to as open arms, had no walls. The maze was raised 50 cm above the floor. A mouse was placed in the center of the maze facing one of the closed arms and was left to explore the maze for 5 minutes. The arms were cleaned with 35% ethanol between each test session. The experiment was performed under dim light conditions (approximately 70 Lux). The numbers of open and closed arms entries and time animals spent on open arms were recorded. Mice (n = 10/group) were treated s.c. with ADX71743 (50, 100, 150 mg/kg) or its vehicle (CD). Two additional groups of mice (n = 10/group) were administered p.o. either chlordiazepoxide (30 mg/kg) or its vehicle (saline). Thirty and 60 minutes after administration of ADX71743 and chlordiazepoxide, respectively, animals were individually placed on the maze and left to explore it for 5 minutes. Percept entries into open arms, as well as time (seconds) spent on open and closed arms of the maze, was analyzed by one-way ANOVA followed by planned comparisons.
FST in Mice.
The procedure was performed as described previously by Campo et al. (2011). Briefly, 1 day before the experiment, mice were pre-exposed to the swim session using individual glass cylinders (height: 25 cm; diameter: 10 cm) containing water 10 cm deep at 24C° ± 1C° for 15 minutes. On the test day, 24 hours later, animals were treated s.c. with ADX71743 (50, 100, 150 mg/kg) or its vehicle (CD; n = 10/groups). Two additional groups (n = 10/group) were treated i.p. with imipramine (30 mg/kg) or its vehicle (saline). Thirty or 60 minutes after administration of ADX71743 or imipramine, respectively, animals were exposed to the test swim session for 6 minutes under identical conditions. All test sessions were recorded by a video camera positioned on the side of the cylinder. A trained observer blind to the treatment scored the tapes. The behavioral measures scored included the duration of immobility exhibited during the last 4 minutes of the 6-minute test. An animal was considered to be immobile as it remained floating motionless in the water making only the movements necessary to keep its head above the water. Terminal blood samples were collected from all ADX71743-treated animals at the end of the experiment and plasma was analyzed as described for the pharmacokinetic studies. The time spent immobile (seconds) in the FST was analyzed with a one-way ANOVA, followed by planned comparisons.
Amphetamine-Induced Hyperactivity in Mice.
The experiment was performed in the locomotor arenas used for the assessment of spontaneous locomotor activity in mice. Mice (n = 10/group) were treated s.c. with ADX71743 (50, 100, 150 mg/kg), or its vehicle (50% cyclodextrin). Two additional groups (n = 10/group) received p.o. risperidone (0.3 mg/kg) or its vehicle (saline). After treatment, animals were individually placed into activity arenas for 30 minutes of habituation. At the end of this period they were challenged i.p. with either amphetamine (3 mg/kg) or its vehicle (saline), returned to the arenas, and monitored for activity for the additional 60 minutes. The total distance traveled during the test (centimeters) was analyzed by one-way ANOVA followed by planned comparisons.
DOI-Induced Head Twitches in Mice.
Adult male C57Bl6/J mice were purchased from Charles River (Margate, UK) and group-housed under standard laboratory conditions at RenaSci Ltd. (Nottingham, UK). Mice (n = 8/group) were treated with ADX71743 (50, 100, 150 mg/kg s.c.), its vehicle (CD; s.c.), or olanzapine (0.1 mg/kg i.p.). Sixty minutes (ADX71743 or CD) or 30 minutes (olanzapine) later, animals were challenged (i.p.) with either vehicle (saline) or DOI (3 mg/kg). Each animal was then placed into a novel cage and the number of head twitches was counted for 6 minutes immediately after administration of DOI or saline by an observer blind to drug treatment. Terminal blood samples were collected from all ADX71743-treated animals at the end of the experiment and plasma was analyzed as described for the pharmacokinetic studies. The number of DOI-induced head twitches was analyzed by one-way ANOVA followed by planned comparisons.
CAR Test in Rats.
Adult male Wistar rats were purchased from Charles River and group-housed under standard laboratory conditions at RenaSci Ltd. Conditioned avoidance behavior was assessed using automated shuttle boxes (42 × 16 × 20 cm; MED Associates), partitioned into two compartments and equipped with infrared-sensitive photocells. Each box was placed into a sound-attenuated chamber. Animals were trained to move to the adjacent compartment within 10 seconds upon administration of the conditioned stimulus (tone and light), to avoid exposure to the unconditioned stimulus (footshock, 0.5 mA for a maximum duration of 10 seconds) via the grid floor. In the first phase of training, each animal underwent 30 trials in a 30-minute test session with a variable intertrial interval of 20–30 seconds. If an animal crossed to the other compartment to avoid the shock, this was recorded by the apparatus as an avoidance response. If the animal crossed to the other compartment during presentation of the shock, this was recorded by the apparatus as an escape response. The training continued until 80% avoidance responses (i.e., 24 correct trials) were obtained for the group. In the second phase of training, each animal underwent 10 trials in a 10-minute session with a variable intertrial interval of 20–30 seconds. This was continued until 80% (i.e., eight correct trials) avoidance responses were obtained for the group. A baseline (pretest) session was run the day before the experiment. In this session, all animals were dosed with vehicle 60 minutes before the test. Animals then underwent the test protocol (10 trials in a 10-minute session with a variable intertrial interval of 20–30 seconds). Animals that exhibited stable performance (>80% avoidance responses for the last three drug-free CAR sessions) underwent drug (or vehicle) testing the next day. Any animals that did not meet the success criteria underwent further training on the day of the experiment. As a part of the pharmacological validation of the model, animals trained in the CAR were tested with antipsychotic drugs (haloperidol, aripiprazole, risperidone, and olanzapine) administered in weekly intervals. All drugs dose-dependently inhibited avoidance response and increased escape responses (unpublished data). One week after the last validation experiment with an antipsychotic drug, animals were tested with a vehicle (CD) and the day after received ADX71743 (10, 30, 100 mg/kg s.c.), its vehicle (CD s.c.), or olanzapine (0.1 mg/kg p.o.) 60 minutes before being evaluated in the above-described test protocol. At the end of the experiment blood from all ADX71743-treated animals was collected via tail vein and plasma was analyzed as described for the pharmacokinetic studies.
Drugs.
l-AP4 (suspended in 0.1 N NaOH) was purchased from AbCam Biochemicals (Cambridge, UK). Glutamate, or l-glutamic acid hydrochloride ((S)-2-aminoglutaric acid, (S)-2-aminopentanedioic acid ammonium salt), (R)-baclofen, d-amphetamine, chlordiazepoxide, imipramine, olanzapine and risperidone were purchased from Sigma-Aldrich. LY341495 (mGlu2/3 orthosteric antagonist; Kingston et al., 1998) was purchased from Tocris Bioscience (Bristol, UK). DOI was purchased from Steroplast Ltd. (Manchester, UK).
ADX71743 was synthesized at Addex Therapeutics. The compound was suspended in water containing CD. The suspensions were homogenized with stainless steel balls for 30 minutes at 30 Hz in a 2-ml Eppendorf tube, and then vortexed and sonicated for 10 minutes. d-amphetamine was dissolved in saline and administered i.p. at 3 ml/kg volume. (R)-Baclofen was dissolved in saline and administered p.o. Chlordiazepoxide and risperidone were suspended in saline and administered p.o. Olanzapine was suspended in saline and administered i.p. Imipramine was suspended in distilled water and administered i.p. All drugs dosed s.c. or i.p. were administered at 3 ml/kg volume. All drugs dosed p.o. were administered either at 10 ml/kg or at 5 ml/kg volumes when given to mice and rats, respectively. All solutions and suspensions were prepared fresh daily. All doses of pharmacological agents are expressed as free base.
Results
Identification and In Vitro Pharmacological Characterization of ADX71743 on Rat and Human Recombinant mGluRs and Native Mouse Receptors.
After a high-throughput screening campaign of the Addex corporate library using a HEK293 cell line stably coexpressing mGlu7 with a chimeric Giα protein, several compounds inhibiting the calcium flux induced by l-AP4 were identified. After a hit confirmation and lead optimization processes, ADX71743 was identified (Tang et al., manuscript submitted; Fig. 1A). This compound tested in cell lines expressing hmGlu7 or rmGlu7 together with a chimeric Gα protein, allowing redirection of receptor activation onto calcium signaling, was found to fully inhibit an EC80 of l-AP4 (2 mM) with full efficacy and an IC50 of 63 ± 2 nM and 88 ± 9 nM, respectively (Fig. 1B; Table 1). In comparison, the IC50 of reference compound LY341495, a nonselective orthosteric mGlu2/3 antagonist was 345 ± 9.5 nM and 449 ± 42 nM for the hmGlu7 and rmGlu7 clones, respectively (Fig. 1B; Table 1), in agreement with values previously reported (Kingston et al., 1998).
Summary of functional potency of reference agonists l-AP4 and glutamate (EC50 values), and of ADX71743 and LY341495 (IC50 values) on recombinant mGlu7
EC50 values (μM) and IC50 values (nM) are expressed as mean ± S.E.M. from at least five independent experiments performed in duplicate.
mGlu7 is naturally coupled to the adenylate cyclase through Gi/o. Therefore, to demonstrate activity of ADX71743 on the physiologic signaling of mGlu7, the compound was tested on cells coexpressing hmGlu7 and a cAMP biosensor Phoenyx (Lütjens et al., 2010). In this assay, a full concentration-response curve with glutamate, the natural ligand of the mGlu7 receptor, saturates, allowing calculation of an EC50 of 264 ± 37 μM (Table 1). ADX71743 could therefore be tested against an EC80 of glutamate (IC50 of 22 ± 4 nM) (Table 1) as well as against an EC80 of l-AP4 (IC50 of 125 ± 17 nM) (Table 1; Fig. 1C). The IC50 of LY341495 when tested against l-AP4 was 2262 ± 268 nM (Fig 1C; Table 1).
To further characterize the pharmacological profile of ADX71743, its mode of action was analyzed by Schild plot experiments in the Phoenyx cAMP assay and compared with the effect of the orthosteric antagonist LY341495. ADX71743 induced a concentration-dependent rightward shift of the l-AP4 concentration-response curve (a 3-fold shift of the EC50 of l-AP4 with 0.1 μM of compound on average, and a maximum 13.7-fold shift observed with 10 μM of compound) together with a decrease of l-AP4 efficacy (Fig. 2B). In a similar protocol, LY341495 induced a concentration-dependent rightward shift of the l-AP4 concentration-response curve without any impact on maximal efficacy of the agonist as expected for a competitive antagonist–competitive agonist pair (Fig. 2C; calculated pA2 = 5.92, slope = 0.83). To test the reversibility of ADX71743 activity, experiments were performed in the calcium assay using the hmGlu7 clone. Cells were either washed three times or were kept without washing after addition of ADX71743, followed by measurement of calcium levels. ADX71743 was found active on an EC80 of l-AP4 as expected in nonwashed cells, whereas the inhibitory effect was not observed in washed cells, demonstrating the reversible effect of the compound on mGlu7 (Fig. 2A).
Reversibility of the antagonistic activity of ADX71743 toward hmGlu7 in intracellular Ca2+ mobilization assay (A). Effect of ADX71743 and LY341495 on concentration-response curves of l-AP4 in the Phoenyx cAMP assay (B and C). Agonist concentration-response curves for l-AP4–induced intracellular cAMP decrease were generated in the absence or presence of various concentrations of ADX71743 (B) or LY341495 (C).
We also investigated the selectivity of ADX71743 versus other mGlu-expressing cells in series of FLIPR experiments and observed that it had no detectable activity (agonist or allosteric effects) in cell lines expressing hmGlu3, hmGlu4, rmGlu5, hmGlu6, and hmGlu8. A negligible inhibition of rmGlu1 (32% at 30 µM) and a weak positive allosteric modulator effect on hmGlu2 (EC50 of 11 µM) was measured. When further tested in a functional GPCR screen against 27 targets (Cerep profile P27, excluding muscarinic M2 and M4 receptors, unavailable at the time of the test) in agonist and antagonist modes, no stimulation or inhibition above 27% was observed (unpublished data).
Hippocampal Electrophysiology.
In the presence of l-AP4 (300 µM), the hippocampal CA1 fEPSP amplitude was reversibly depressed to 62% ± 1% (n = 6) of control values. Coapplication of ADX71743 and l-AP4 resulted in a concentration-dependent reversal of the l-AP4–induced depression, with 0.1 µM ADX71743 reversing the effects of l-AP4 by 11% ± 1% (n = 3; P < 0.01) and 10 µM resulting in a 20% ± 3% reversal (n = 6; P < 0.001) (Fig. 3, A and B).
Reversal of l-AP4–induced depression of synaptic transmission in the presence of ADX71743. Representative traces (averaged fEPSPs in response to Schaffer collateral stimulation) from a single hippocampal slice experiment taken at different time points (i–iv) as indicated in the time course plot below (A). For clarity, the stimulus artifacts have been truncated. Mean percentage reversal of the depression of fEPSPs amplitude induced by l-AP4 (300 µM) by ADX71743 (B). n = 3–6. Error bars represent S.E.M. **P < 0.01; ***P < 0.001.
Serum Protein Binding: Plasma and Brain Pharmacokinetic Profile of ADX71743.
In mice and rats, s.c. administration of 100 mg/kg ADX71743 resulted in a similar pharmacokinetic profile (Table 2). After s.c. administration of 12.5 mg/kg ADX71743 in mice, plasma reached high Cmax between 0.25 and 0.5 hours, and then declined rapidly, after a half-life of approximately 0.5 hours (Fig. 4A; Table 2). After s.c. administration of 12.5 mg/kg, CSF and brain concentrations of ADX71743 also reached Cmax rapidly. Thereafter, brain concentrations declined rapidly, whereas the CSF concentrations were sustained for a slightly longer period (Fig. 4A; Table 2). The ratio between the CSF and plasma concentration at Cmax was 5.3, which is consistent with the free plasma concentration available for diffusion in the brain, as determined in the serum protein binding studies (fu = 0.044). The area under the curve and Cmax values of plasma exposure show linear and similar increases in relation to the dose after s.c. administration of 12.5 and 100 mg/kg ADX71743 in mice (Fig. 4B).
Summary of the pharmacokinetic properties of ADX71743 after s.c. administration to mice and rats
Plasma (ng/ml), CSF (ng/ml), and brain (ng/g) concentrations of ADX71743 in mice after s.c. administration at 12.5 mg/kg and sample collection 0.25, 0.5, 1, and 2 h postadministration (A). The Cmax (ng/ml) and area under the curve with extrapolation to infinity (AUCinf; ng⋅h/ml) values after s.c. administration of ADX71743 (0, 12.5, and 100 mg/kg) in mice (B).
Spontaneous Locomotor Activity Test in Mice and Rats.
ADX71743 (50, 100, 150 mg/kg) had no effect on spontaneous locomotor activity in mice, whereas baclofen markedly suppressed (80%; P < 0.001) the total distance traveled by animals compared with vehicle treatment (Fig. 5A). ADX71743 (50, 100, 150 mg/kg) also had no effect on spontaneous locomotor activity in rats, whereas baclofen resulted in a virtually full (P < 0.001) suppression of activity (Fig. 5B).
Locomotor activity (distance traveled; cm) of (A) male C57Bl6/J mice (n = 10/group) and (B) Sprague-Dawley rats (n = 8-10/group) during 60 minutes expressed as total distance. Animals were pretreated with ADX71743 (50, 100, 150 mg/kg, s.c.), CD (s.c.), (R)-Baclofen (10 mg/kg p.o.), or saline (p.o.). ADX71743 and its vehicle were administered 30 min, whereas (R)-baclofen and its vehicle were administered 60 min before testing. Each point represents the observed mean (± S.E.M.). ***P < 0.001 compared with saline (Sal).
Rotarod Test in Mice.
ADX71743 (50, 100, 150 mg/kg) had no effect on latencies to fall from rotarod (seconds) when animals were tested 30 and 90 minutes after administration (Table 3). In contrast, baclofen resulted in robust reduction of latencies to fall (75%; P < 0.001) when animals were tested 60 and 120 minutes after administration (Table 3).
Activity of ADX71743 in the mouse rotarod test
MB Test in Mice.
ADX71743 resulted in an approximately 60% reduction in the number of buried marbles at 50 and 100 mg/kg (P < 0.01 and P < 0.05, respectively), with further reduction (approximately 75%; P < 0.001) at 150 mg/kg compared with its vehicle (Table 4). The corresponding concentrations of ADX71743 in plasma in animals treated at 50, 100, and 150 mg/kg were 3451, 6990, and 10,430 ng/ml, respectively. These plasma concentrations resulted in CSF/IC50 values of 4, 9, and 14, respectively (Table 4) based on fu determined in independent experiments. Chlordiazepoxide-treated mice buried approximately 60% (P < 0.01) fewer marbles compared with corresponding vehicle-treated controls (Table 4).
Activity of ADX71743 in the mouse MB test and measured plasma, CSF concentrations, and CSF concentration/IC50 (in vitro) of ADX71743 in mice at the end of the MB experiment
CSF exposure was calculated from the plasma exposure using mice plasma protein binding (fu = 4.4%).
EPM Test in Mice.
ADX71743 resulted in dose-dependent increases in the percentage of entries into open arms [F(5, 54) = 6.02; P < 0.001]. Specifically, at 100 and 150 mg/kg ADX71743 resulted in 2.3-fold (P < 0.05) and 2.5-fold (P < 0.01) increases in open arm entries compared with corresponding vehicle-treated controls (Fig. 6A). ADX71743 also increased the time spent on open arms [F(5, 54) = 6.65; P < 0.001]. Specifically, at 100 and 150 mg/kg ADX71743 resulted in 2.3-fold and 2.5-fold (both P < 0.05) increases in time spent on open arms compared with corresponding vehicle-treated controls (Fig. 6B). Diazepam-treated animals exhibited similar 2.5-fold (P < 0.001) increases in the percentage of entries into open arms and in time spent on those arms compared with vehicle-treated control (Fig. 6, A and B). There was no effect of treatment on the number of closed arm entries (unpublished data).
Percent entries into open arms (A) and time spent (seconds) on open arms (B) by male C57Bl6/J mice (n = 10/group) during 5 minutes of the EPM test. Animals were pretreated with ADX71743 (50, 100, 150 mg/kg s.c.), CD (s.c.), diazepam (1.5 mg/kg p.o.), or saline (p.o.). ADX71743 and its vehicle were administered 30 minutes, whereas diazepam and its vehicle were administered 60 minutes before testing. Each point represents the observed mean (+S.E.M.). *P < 0.01; **P < 0.001 compared with CD; ###P < 0.001 compared with saline (Sal).
FST in Mice.
ADX71743 (50, 100, 150 mg/kg) had no effect on the time animals spent in immobility, whereas imipramine-treated animals exhibited over 50% reduction (P < 0.001) in the immobility time (Table 5). The corresponding plasma concentrations of ADX71743 in animals treated at 50, 100, and 150 mg/kg were 7265, 12,607, and 13,107 ng/ml, respectively. These plasma concentrations resulted in CSF/IC50 values of 9, 16, and 17, respectively (Table 5).
Activity of ADX71743 in the FST and measured plasma, CSF concentrations, and CSF concentration/IC50 (in vitro) of ADX71743 in mice at the end of the experiment
CSF exposure was calculated from the plasma exposure using mice plasma protein binding (fu = 4.4%).
Amphetamine-Induced Hyperactivity Test in Mice.
ADX71743 (50, 100, 150 mg/kg) dose-dependently reduced amphetamine-induced hyperactivity [F(5, 95)=46.2; P < 0.001; Fig. 7]. Specifically, at 100 and 150 mg/kg, there were approximately 20% (P < 0.01) and 30% (P < 0.001) reductions in hyperactivity compared with corresponding vehicle-pretreated controls (Fig. 7). Risperidone-treated animals exhibited robust reduction in amphetamine-induced hyperactivity (approximately 80%; P < 0.001; Fig. 7).
Total distance (centimeters) traveled by male C57Bl6/J mice (n = 20/group) during 60 minutes of the amphetamine-induced hyperactivity test. Animals were pretreated with ADX71743 (50, 100, 150 mg/kg s.c.), CD (s.c.), or risperidone (0.3 mg/kg p.o.) and habituated to the locomotor arenas for 30 minutes (ADX71743/CD) or 60 minutes (risperidone). At the end of the habituation period, animals were challenged i.p. with either amphetamine (3 mg/kg) or saline and tested for further 60 minutes. Each point represents the observed mean (+S.E.M.). **P < 0.01; ***P < 0.001 compared with the reference group (Ref).
DOI-Induced Head Twitches in Mice.
ADX71743 (50, 100, 150 mg/kg) had no effect on the number of DOI-induced head twitches, whereas olanzapine resulted in nearly 80% reduction (P < 0.001) in this number (Table 6). The corresponding plasma concentrations of ADX71743 in animals treated at 50, 100, and 150 mg/kg were 3233, 4273, and 6311 ng/ml, respectively. These plasma concentrations resulted in CSF/IC50 values of 4, 6, and 8, respectively (Table 6).
Activity of ADX71743 and olanzapine in the DOI-induced head twitches test in mice together with the measured plasma, CSF concentrations and CSF concentration/IC50 (in vitro) of ADX71743 in mice at the end of the experiment
CSF exposure was calculated from the plasma exposure using mice plasma protein binding (fu = 4.4%).
Conditioned Avoidance Response Test in Rats.
ADX71743 (10, 30, 100 mg/kg) had no effect on numbers of avoidances or escapes (Table 7). In contrast, olanzapine resulted in an approximately 60% (P < 0.01) reduction in the number of avoidances and more than 3-fold (P < 0.01) increases in the number of escapes (Table 7). None of the animals, except one, showed escape failures during the test (data not included). The corresponding plasma concentrations of ADX71743 in rats treated at 10, 30, and 100 mg/kg were 1188, 3351, and 9800 ng/ml, respectively. These plasma concentrations resulted in CSF/IC50 values of 1, 3, and 9, respectively (Table 7).
Activity of ADX71743 in the CAR test together with the measured plasma, CSF concentrations, and CSF concentration/IC50 (in vitro) of ADX71743 in rats at the end of the experiment
CSF exposure was calculated from the plasma exposure using mice plasma protein binding (fu = 4.4%).
Plasma Concentration of ADX71743 in In Vivo Studies.
Concentrations of ADX71743 in plasma after s.c. administration at 50, 100, and 150 mg/kg in the DOI-induced head-twitch, MB and the FST in mice are shown in Fig. 8. Plasma concentrations of ADX71743, measured in the FST study, were higher than those measured in the DOI and MB studies, probably due to different postdosing sampling time (30 versus 60 minutes) and the rapid clearance of the compound (Fig. 8).
Plasma concentration (ng/ml) of ADX71743 after s.c. administration at 50, 100, and 150 mg/kg in mice across three experiments including DOI-induced head twitches, MB, and FST. Samples were taken approximately 30 minutes (FST) or 60 minutes (DOI and MB) after administration of the compound.
Discussion
Over the last decade, data supporting the hypothesis that the mGlu7 receptor plays a pivotal role in the CNS have been accumulating. Although initial studies using inactivation of the mGlu7 receptor provided important first evidence of its role, those involving direct and selective engagement of the receptor have not been possible due to the absence of subtype selective and brain-penetrant molecules, like ADX71743.
Here we describe the in vitro and in vivo pharmacological properties of ADX71743, a potent and selective mGlu7 NAM. We identified ADX71743 through chemical lead optimization after a high-throughput screening campaign of the corporate chemical library using a FLIPR assay. ADX71743 was found to completely block the mGlu7 agonist-induced signal, tested in recombinant systems measuring either changes in intracellular calcium or changes in intracellular levels of cAMP. However, in curve shift analysis using the cAMP assay, the degree of inhibition of the l-AP4 response appeared to saturate at around 40% for l-AP4 concentrations above 300 μM, suggesting a weak degree of cooperativity between ADX71743 and l-AP4. It also suggests that at high concentrations of glutamate, some mGlu7-mediated signaling may persist even in the presence of high concentrations of the NAM. The Schild plot experiments clearly indicate the noncompetitive nature of ADX71743 inhibition, whereas simple wash experiments seem to indicate that its binding to the receptor is readily reversible. Activity of ADX71743 at the native mGlu7 was demonstrated in the in vitro preparation of mouse hippocampus, where ADX71743 attenuated the depression of synaptic transmission induced by the group III mGlu agonist l-AP4. ADX71743 also exhibited good exposure after s.c. administration in mice and rats, making it suitable for in vivo testing. The pharmacokinetic profile of ADX71743 was found to be similar in mice and rats when the compound was given s.c. at 100 mg/kg. Considering an IC50 of 125 nM, and an average CSF/total plasma concentration ratio of 0.05 or more, plasma concentration above 1 μg/ml would be necessary to cover 50% of the receptor activity. A pharmacokinetic analysis of the compound showed that a Cmax just above this value was reached for a very short period of time when ADX71743 was administered at 12.5 mg/kg s.c. Therefore, we chose higher doses (50, 100, and 150 mg/kg s.c.) for in vivo studies in mice and rats and confirmed that pharmacological activity can be observed when plasma concentrations are above 1 μg/ml.
In vivo efficacy of ADX71743 was evaluated in rodent models of anxiety, depression, and psychosis based on earlier evidence, but also reflecting regional localization of mGlu7 within the mammalian CNS. Although it is widely distributed throughout the brain, mGlu7 receptor shows particularly high abundance in the neocortex, piriform and entorhinal cortices, hippocampus, amygdala, globus pallidus, ventral pallidum, and the locus coeruleus (Kinoshita et al., 1998). Abnormalities in these regions have been linked to anxiety disorders (Walker and Davis, 2002), depression (Sanacora et al., 2012), and psychosis (Moghaddam and Adams, 1998; Schoepp and Marek, 2002) among other CNS disorders.
ADX71743 showed an anxiolytic-like profile in the MB and the EPM tests in mice. Both tests are known to be sensitive to typical anxiolytic drugs (Pellow et al., 1985; Lister, 1990; Nicolas et al., 2006), the latter also being relevant to obsessive compulsive disorder (Thomas et al., 2009). ADX71743 resulted in robust reductions in numbers of buried marbles to near maximal levels at lower doses (50 and 100 mg/kg) and similar to those produced by an anxiolytic drug, chlordiazepoxide. The plasma analysis confirmed that an approximately 60% reduction in the number of buried marbles corresponded to CSF/IC50 values of 4–9 (at 50 and 100 mg/kg), followed by further reductions (73%) at CSF/IC50 = 14 (at 150 mg/kg). In accordance with our findings, mGlu7 KO mice exhibited an approximately 60% reduction in the number of buried marbles compared with their wild-type controls (Callaerts-Vegh et al., 2006). In the EPM test, ADX71743 dose-dependently increased open arm entries and the time spent on these arms, without producing nonspecific changes in activity on closed arms. In accord with our findings, mGlu7 KO mice showed increases in open arm entries and time spent on these arms compared with their wild-type controls (Callaerts-Vegh et al., 2006).
At this point, we can only speculate on how the reduction in activity of the mGlu7 receptor can lead to reduced anxiety-like reactivity. The mGlu7 receptor is located presynaptically and, depending on the type of neuron it is located on, can regulate the release of glutamate, GABA, or other (e.g., norepinephrine) neurotransmitters (O’Connor et al., 2010). Glutamate has low affinity to the mGlu7 receptor (Okamoto et al., 1994), which remains inactive under normal conditions, only becoming active under the conditions of excessive glutamate release (Ferraguti and Shigemoto, 2006). Although an mGlu7 NAM is expected to reduce receptor-mediated inhibitory control, the net outcome of this disinhibition will depend on a specific brain region. Future studies involving site-specific injections of ADX71743 can aid in further understanding the question of modulation of anxiety by the mGlu7 receptor. In addition, there is a possibility that reduced activity of mGlu7 will impact anxiety-like reactivity via modulating several downstream targets, especially those involved in reactivity to stress. According to Mitsukawa et al. (2006), mGlu7 KO mice exhibit signs of HPA axis dysregulation, including upregulation of glucocorticoid (GR) and 5-HT1A receptors in the hippocampus, increased sensitivity to GR-mediated negative feedback, and increases in brain-derived neurotrophic factor protein in the hippocampus (Mitsukawa et al., 2006). These changes correlate well with reduced anxiety- and depression-like reactivity of mGlu7 KO mice (Cryan et al., 2003). Whether inhibition of mGlu7 with a NAM can lead to alterations in HPA axis seen in mGlu7 KO animals remains to be investigated.
ADX71743 failed to show an antidepressant-like profile in the mouse FST despite reaching CSF/IC50 values of 9, 16, and 17 at 50, 100, and 150 mg/kg doses, respectively. In addition, ADX71743 failed to show a clear antipsychotic-like profile in models relevant to psychosis. The amphetamine-induced hyperactivity, DOI-induced head twitch, and CAR tests in rodents have been shown to be sensitive to typical and atypical antipsychotic drugs (Ellenbroek, 1993; Wettstein et al., 1999; Wadenberg, 2010). The rationale for testing ADX71743 in models of psychosis, in part, came from several independent studies showing a link between polymorphisms in mGlu7 and schizophrenia (Ohtsuki et al., 2008; Ganda et al., 2009; Shibata et al., 2009). ADX71743 resulted in a dose-dependent, albeit modest, reduction of amphetamine-induced hyperactivity in mice. However, the follow-up mouse DOI-induced hyperactivity and the rat CAR tests revealed no activity of the compound despite it reaching adequate concentrations in plasma and CSF for in vivo activity (see Tables 6 and 7).
We can only speculate on reasons why mGlu7 NAM ADX71743 was inactive in tests relevant to depression and psychosis. The outcome of in vivo studies is unlikely to be impacted by regional distribution of mGlu7 in mice and rats, which is virtually identical in these species (Kinoshita et al., 1998). In fact, only in the medial habenula, mGlu7 is present in high abundance in rats, while being absent in mice, whereas the opposite trend is seen in the cerebellar nuclei (Kinoshita et al., 1998). One explanation of in vivo results is involvement of mGlu7 in anxiety-like reactivity, but not in depression or psychosis. Future studies involving more disease-relevant animal models, such as genetic line of “helpless” mice (El Yacoubi et al., 2003) and Flinders Sensitive line rats (Overstreet et al., 2005) for depression as well as rats exposed in utero to mitotoxin methylazoxymethanol acetate for schizophrenia (Lodge and Grace, 2009), can shed light on this question. There is also a possibility that the effect of ADX71743 was linked to mGlu7 concentration in a specific neural circuit recruited in the in vivo test. Such possibility is unlikely, since both amphetamine-induced hyperactivity (where a weak effect was seen) and the CAR test (where there was no effect) rely on the nucleus accumbens and the ventral tegmental area (Geyer and Ellenbroek, 2003; Wadenberg, 2010), anatomic regions that have similar low concentration of mGlu7 (Kinoshita et al., 1998).
In conclusion, we present in vitro and in vivo characterization of ADX71743 as a centrally active compound suitable for investigation of the role of mGlu7 receptor. This compound shows potent NAM activity at the mGlu7 receptor with a clean selectivity profile at other subtypes of the mGlu family and other GPCRs, and a pharmacokinetic profile making it suitable for in vivo profiling. In vivo, ADX71743 shows anxiolytic-like efficacy in the mouse MB and EPM tests. Interestingly, the compound did not demonstrate antidepressant-like activity in the FST, suggesting that reduction of mGlu7 activity via the NAM mechanism is more relevant for anxiety than depression. The compound also was largely inactive in models predictive of antipsychotic-like activity. Further optimization of this and other series of mGlu7 NAMs is underway to provide even more potent and better exposed compounds for future studies. In short, these data suggest that mGlu7 inhibition merits further study as a novel approach for the treatment of obsessive compulsive disorder and other anxiety disorders.
Acknowledgments
The authors thank Sonia Manganiello for determining chirality of ADX71743, as well as Jean-Philippe Rocher and Christelle Bolea for reviewing the manuscript. The authors also thank Steven Vickers and Sharon Cheetham at RenaSci Ltd. for support in DOI-induced head twitch and conditioned avoidance response tests.
Authorship Contributions
Participated in research design: Kalinichev, Girard, Charvin, Campo, Le Poul, Mutel, Poli, Neale, Salt, Lütjens.
Conducted experiments: Rouillier, Girard, Royer-Urios, Bournique, Finn, Neale.
Contributed new reagents or analytic tools: Royer-Urios, Bournique, Finn, Neale.
Performed data analysis: Kalinichev, Rouillier, Girard, Royer-Urios, Bournique, Finn, Poli, Neale, Salt.
Wrote or contributed to the writing of the manuscript: Kalinichev, Rouillier, Poli, Neale, Salt, Lütjens.
Footnotes
Abbreviations
- ADX71743
- (+)-6-(2,4-dimethylphenyl)-2-ethyl-6,7-dihydrobenzo[d]oxazol-4(5H)-one
- AMN082
- N,N′-dibenzhydrylethane-1,2-diamine dihydrochloride
- ANOVA
- analysis of variance
- CAR
- conditioned avoidance response
- CD
- 50% hydroxyl-propyl-β-cyclodextrin
- CNS
- central nervous system
- CSF
- cerebrospinal fluid
- DMEM
- Dulbecco’s modified Eagle’s medium
- DOI
- 2,5-dimethoxy-4-iodoamphetamine
- EPM
- elevated plus maze test
- fEPSP
- field excitatory postsynaptic potential
- FLIPR
- fluorometric imaging plate reader
- FST
- forced swim test
- fu
- function unbound
- GPCR
- G-protein coupled receptor
- HEK
- human embryonic kidney
- HPA
- hypothalamic-pituitary-adrenal
- KO
- knock-out
- l-AP4
- l-(+)-2-amino-4-phosphonobutyric acid
- LY341495
- (2S)-2-amino-2-[(1S,2S)-2-carboxycycloprop-1-yl]-3-(xanth-9-yl)propanoic acid
- MB
- marble burying
- mGlu7
- metabotropic glutamate receptor 7
- MMPIP
- 6-(4-methoxyphenyl)-5-methyl-3-pyridin-4-ylisoxazolo[4,5-c]pyridin-4(5H)-one
- NAM
- negative allosteric modulator
- Received October 3, 2012.
- Accepted December 18, 2012.
- Copyright © 2013 by The American Society for Pharmacology and Experimental Therapeutics