Abstract
The normalization of excessive glutamatergic neurotransmission through the activation of metabotropic glutamate 2 (mGlu2) receptors may have therapeutic potential in a variety of psychiatric disorders, including anxiety/depression and schizophrenia. Here, we characterize the pharmacological properties of N-(4-((2-(trifluoromethyl)-3-hydroxy-4-(isobutyryl)phenoxy)methyl)benzyl)-1-methyl-1H-imidazole-4-carboxamide (THIIC), a structurally novel, potent, and selective allosteric potentiator of human and rat mGlu2 receptors (EC50 = 23 and 13 nM, respectively). THIIC produced anxiolytic-like efficacy in the rat stress-induced hyperthermia assay and the mouse stress-induced elevation of cerebellar cGMP and marble-burying assays. THIIC also produced robust activity in three assays that detect antidepressant-like activity, including the mouse forced-swim test, the rat differential reinforcement of low rate 72-s assay, and the rat dominant-submissive test, with a maximal response similar to that of imipramine. Effects of THIIC in the forced-swim test and marble burying were deleted in mGlu2 receptor null mice. Analysis of sleep electroencephalogram (EEG) showed that THIIC had a sleep-promoting profile with increased non-rapid eye movement (REM) and decreased REM sleep. THIIC also decreased the dark phase increase in extracellular histamine in the medial prefrontal cortex and decreased levels of the histamine metabolite tele-methylhistamine (t-MeHA) in rat cerebrospinal fluid. Collectively, these results indicate that the novel mGlu2-positive allosteric modulator THIIC has robust activity in models used to predict anxiolytic/antidepressant efficacy, substantiating, at least with this molecule, differentiation in the biological impact of mGlu2 potentiation versus mGlu2/3 orthosteric agonism. In addition, we provide evidence that sleep EEG and CSF t-MeHA might function as viable biomarker approaches to facilitate the translational development of THIIC and other mGlu2 potentiators.
Introduction
Glutamate is the principal excitatory neurotransmitter in the mammalian central nervous system and mediates a diversity of physiological and behavioral actions through ionotropic and metabotropic glutamate (mGlu) receptors. Of the eight mGlu receptor subtypes, the group II mGlu receptors, which include mGlu2 and mGlu3, are particularly attractive as novel drug targets. mGlu2 and mGlu3 receptors function to inhibit neurotransmitter release as autoreceptors on glutamatergic terminals and presynaptic heteroreceptors (Schoepp, 2001) or via modulatory actions on glia (Conn and Pin, 1997). The mGlu2 and mGlu3 receptors are present throughout key forebrain and limbic areas such as the prefrontal cortex, thalamus, striatum, hippocampus, and amygdala (Wright et al., 2001). Consequently, drugs that activate group II mGlu receptors may represent a novel approach for treating a variety of psychiatric and neurological disorders.
To date, only one structural class of selective agonists for group II mGlu receptors has been discovered and is undergoing development. In preclinical studies, mGlu2/3 receptor agonists [(−)-2-oxa-4-aminobicyclo[3.1.0]hexane-4,6-dicarboxylic acid (LY379268), 1S,2S,5R,6S-2-aminobicyclo[3.1.0]hexane-2,6-bicaroxylate monohydrate (LY354740), and (−)-(1R,4S,5S,6S)-4-amino-2-sulfonylbicyclo[3.1.0]hexane-4,6-dicarboxylic acid (LY404039)] have behavioral and neurochemical effects in models predictive of antistress/anxiolytic and antipsychotic efficacy (Schoepp and Marek, 2002; Swanson et al., 2005). Moreover, the anxiolytic and antipsychotic potential of selective mGlu2/3 receptor agonists has been confirmed in early clinical trials (Grillon et al., 2003; Schoepp et al., 2003; Patil et al., 2007; Dunayevich et al., 2008). In addition, evidence suggests that the activation of mGlu2/3 receptors may be beneficial in depression. For example, mGlu2/3 receptors exert tonic control over the hypothalamic-pituitary-adrenal-axis, which is known to be disrupted in depressed patients. mGlu2/3 receptors agonists also induce neurochemical and behavioral changes suggestive of potential antidepressant efficacy, including increased monoamine efflux in limbic regions (Cartmell et al., 2001; Rorick-Kehn et al., 2007b), stimulation of neurotrophic factor production (Di Giorgi-Gerevini et al., 2005), neuroprotection (Bruno et al., 2001), and an antidepressant-like suppression of REM sleep (Ahnaou et al., 2009). Finally, in vitro and in vivo studies suggest the existence of a synergism between clinically effective antidepressants and mGlu2/3 receptor agonists (Matrisciano et al., 2008).
Another approach toward enhancing group II mGlu receptor function is through the development of allosteric modulators of mGlu2. Unlike orthosteric acting agonists, allosteric modulators do not activate the receptor directly but act at an allosteric site to selectively potentiate the response of mGlu2 receptors to glutamate. Using electrophysiological techniques, it has been shown that allosteric modulators have little or no effects on glutamate release under “normal conditions” but act to potentiate negative feedback control under conditions of excessive glutamate release (Johnson et al., 2005). This ability to recognize “pathological states” or networks could result in therapeutic advantages in terms of safety and efficacy. Given the selectively of mGlu2 potentiators for mGlu2 over mGlu3 receptors, such potentiators are also valuable tools for helping to differentiate the biological substrates driven exclusively by physiologically relevant augmentation of mGlu2 receptors.
Multiple highly selective positive allosteric modulators of mGlu2 have been identified, of which N-(4-(2-methoxyphenoxy)-phenyl-N-(2,2,2-trifluoroethylsulfonyl)-pyrid-3-ylmethylamine (LY487379), N-4′-cyano-biphenyl-3-yl)-N-(3-pyridinylmethyl)-ethanesulfonamide hydrochloride (CBiPES), and biphenyl-indanone A (BINA) are the most widely characterized compounds (Galici et al., 2005; Johnson et al., 2005). Like mGlu2/3 receptor agonists, these mGlu2-positive allosteric modulators have shown efficacy in preclinical models used to predict anxiolytic and antipsychotic activity. However, progress in the development of these positive allosteric modulators of mGlu2 has been slow primarily because of their lack of good drug-like features, including modest potency, poor metabolic stability, and brain penetration.
In the present study, we report the discovery of N-(4-((2-(trifluoromethyl)-3-hydroxy-4-(isobutyryl)phenoxy)methyl)benzyl)-1-methyl-1H-imidazole-4-carboxamide (THIIC) (Khilevich et al., 2008), a structurally novel and highly selective positive allosteric modulator of mGlu2 receptors (Fig. 1). In addition to evaluating the in vitro pharmacology of THIIC, we have characterized the effects of THIIC in models used to predict anxiolytic and/or antidepressant activity, including stress-induced hyperthermia, reversal of stress-induced increases in cerebellar cGMP, marble burying, dominant-submissive test, differential reinforcement of low rate (DRL) 72-s behavior, and mouse forced-swim test. A role for mGlu2 receptor actions driving antidepressant- or anxiolytic-like effects was assessed by using mGlu2(−/−) mice. Finally, in an effort to develop a potential translational biomarker for this target, we studied the effects of THIIC for changes in rat sleep electroencephalogram (EEG) patterns and CNS neurochemistry with a focus on histamine release in the brain and histamine metabolites in the CSF.
Chemical structure of THIIC.
Materials and Methods
In Vitro Assays
Human and Rat mGlu2 Potentiator FLIPR Assays.
AV12 cell lines stably expressing the human or rat mGlu2 receptor and cotransfected with the rat glutamate transporter excitatory amino acid transporter 1 and the Gα15 subunit were used for these studies. The expression of Gα15 allows Gi-coupled receptors to signal through the phospholipase C pathway, resulting in the ability to measure receptor activation by a fluorometric calcium response assay. The cell lines were maintained by culturing in Dulbecco's modified Eagle's medium with high glucose and pyridoxine hydrochloride supplemented with 5% heat inactivated, dialyzed fetal bovine serum, 1 mM sodium pyruvate, 10 mM HEPES, 1 mM l-glutamine, and 5 μg/ml blasticidin (all media purchased from Invitrogen, Carlsbad, CA). Confluent cultures were passaged biweekly by using an enzyme-free dissociation solution (Millipore Bioscience Research Reagents, Temecula, CA). Cells were harvested 24 h before assay and dispensed by using a Matrix Well Mate cell seeder (Thermo Fisher Scientific, Waltham, MA) at 85,000 (human) or 65,000 (rat) cells per well into 96-well, black-walled, poly-d-lysine-coated plates (BD Biosciences, San Jose, CA) in medium containing only 250 μM l-glutamine (freshly added).
Intracellular calcium levels were monitored before and after the addition of compounds by using a fluorometric imaging plate reader (Molecular Devices, Sunnyvale, CA). The assay buffer was comprised of Hank's buffered salt solution (Sigma-Aldrich, St. Louis, MO) supplemented with 20 mM HEPES. The medium was removed, and the cells were incubated with 8 μM Fluo-3AM (Invitrogen; 50 μl per well) in assay buffer for 90 min at 25°C. The dye solution was removed and replaced with fresh assay buffer (50 μl per well). A single-addition FLIPR assay generating an 11-point concentration response curve for the agonist glutamate (Thermo Fisher Scientific) was conducted before each experiment. The results were analyzed by using Prism version 4.03 (GraphPad Software Inc., San Diego, CA) to calculate the concentrations of glutamate needed to induce the EC10 responses and monitor the general sensitivity of the cells for the agonist.
The compound was tested in a two-addition FLIPR assay using a 10-point concentration response profile starting at a final concentration of 25 μM (agonist mode) or 12.5 μM (potentiator mode). A 3-fold dilution series in dimethyl sulfoxide (DMSO) was followed by a single dilution into assay buffer; the final concentration of DMSO was 0.625%. After taking an initial 5-s fluorescent read on the FLIPR instrument, compound was added to the cell plate (50 μl per well). Data were collected every second for the first 30 s and then every 3 s for a total of 90 s to detect agonist activity. Immediately thereafter, the second addition consisting of 100 μl of glutamate in assay buffer (typically approximately 1 μM) was added to the cell plate, generating an EC10 response. After the second addition, data were collected every second for 29 images and then every 3 s for 15 images. The maximal response was defined as that induced by ECmax (100 μM glutamate). The compound effect was measured as maximal minus minimal peak heights in relative fluorescent units corrected for basal fluorescence measured in the absence of glutamate. Determinations were carried out with single plates. Agonist effects were quantified as percentage of stimulation induced by compound alone relative to the maximal glutamate response. Potentiation effects were quantified as percentage of increase in the presence of an EC10 response in glutamate relative to the ECmax response. Using calcium mobilization assays analogous to those described above, THIIC was tested for activity in agonist, potentiator, or antagonist modes at the human mGluR1, mGluR3, mGluR4, mGluR5, and mGluR8 receptor subtypes, each stably expressed in AV12 cells. Maximum concentrations were 12.5 μM (antagonist and potentiator modes) and 25 μM (agonist mode). All data were calculated as relative EC50 values by using a four-parameter logistic curve-fitting program (ActivityBase version 5.3.1.22; IDBS, Surrey, U.K.).
The observed ability to potentiate the glutamate activity was confirmed by using glutamate concentration-response curves generated in the presence of increasing concentrations of THIIC. Glutamate was diluted in assay buffer through 12 concentrations ranging from 100 μM to 49 nM, and the compound was three times serially diluted in DMSO for an eight-point concentration response profile starting at a final concentration of 12.5 μM. Data collection and analysis were as described previously.
GTPγ35S Binding Assay.
GTPγ35S binding was determined by using an antibody capture scintillation proximity technique in a 96-well plate format. In brief, 100 μl (20–40 fmol/well) of membrane preparations from AV12 cells that ectopically express either human or rat mGluR2 was incubated for 30 min with 50 μl of test compound. GDP (1 μM final concentration) was added to receptor membranes before incubation. Dose-response curves were achieved with increasing amounts of THIIC in the presence of increasing amounts of the orthosteric agonist glutamate (Tocris Bioscience, Ellisville, MO). After the incubation period, 50 μl of diluted GTP-γ-[35S] (500 pM final concentration; PerkinElmer Life and Analytical Sciences, Waltham, MA) was added to each well and incubated for 30 min. The labeled membranes were then solubilized with 0.27% Nonidet P40 followed by a 3-h incubation with 20 μl (final dilution of 1:200) of anti-Gαi-3 rabbit polyclonal antibody (Covance Research Products, Princeton, NJ) and 50 μl (1.25 mg/well) of anti-rabbit scintillation proximity assay beads (GE Healthcare, Chalfont St. Giles, Buckinghamshire, UK). The plates were centrifuged for 10 min at 700g and counted for 1 min per well by using a Wallac MicroBeta TriLux scintillation counter (PerkinElmer Life and Analytical Sciences). All incubations took place at room temperature in GTP-binding assay buffer (20 mM HEPES, 100 mM NaCl, 5 mM MgC12, pH 7.4). Data were analyzed by using nonlinear regression for a sigmoidal concentration response curve (Prism version 4.03). Background was subtracted from all wells, data were normalized to a full agonist response (100 μM glutamate), and efficacy was expressed as a percentage of maximal response. Mean EC50 were calculated as a mean of three independent determinations ± S.E.M. Kb and cooperativity factor were calculated from EC50-generated dose ratios imported into an allosteric Schild regression algorithm: Y = (Kb + α × B)/(Kb + B), where α = 10L̂ogα, Kb = 10L̂ogKb, and B = 10X̂.
Selectivity Panel.
THIIC was tested for binding to dopamine and serotonin receptors by using receptor binding assays performed according to methods reported previously (Rorick-Kehn et al., 2007a). An additional panel of more than 30 different biogenic amine receptors, neuropeptide receptors, ion channel binding sites, and neurotransmitter transporter binding assays were run by CEREP Inc. (Celle L'Evescault, France). The targets run were: human adenosine A3 receptors; acetylcholinesterase; angiotensin II; adrenergic receptors α1 (nonselective), α2 (nonselective), β1 and β2; norepinephrine transporter; rat brain benzodiazepine receptor; dopaminergic receptors D1 and D2s; histaminergic receptors H1 and H2; muscarinic receptors M2 and M3; neurokinin receptor 1; opioid receptor μ; serotonin receptors 5HT2B agonist site, rat Ca2+ channel verapamil site; rat brain voltage-gated potassium channel (K+V channel); rat brain small-conductance Ca2+-activated K+ channel (SK+Ca channel); rat Na+ channel (site 2); and rat Cl− channel. All assays were run using recombinant human receptors, except where noted. The full methods and references can be found on the Cerep website (www.cerep.fr). The assays were run at 1 μM and 10 μM THIIC, and the percentage of inhibition is given as the average of three determinations. When significant displacement of radioligand was observed (>50% inhibition at 1 μM), complete concentration-dependent displacement curves (in triplicate) were constructed to generate IC50 values.
Animals
All experiments were performed according to the policies of the Animal Care and Use Committee of Eli Lilly and Company, in conjunction with the American Association for the Accreditation of Laboratory Animal Care-approved guidelines and the Guide for Care and Use of Laboratory Animals (Institute of Laboratory Animal Resources, 1996). Animals were individually or group-housed depending on the assay in an environmentally controlled facility where food and water were available ad libitum unless specified otherwise. mGlu2R(−/−) mice and their wild-type littermates were derived and bred as described previously (Fell et al., 2008). For EEG studies, each rat was housed individually within a specially modified microisolator cage having a custom polycarbonate riser to raise the filter top that held an ultra-low-torque slip-ring commutator (Hypnion, Inc., Lexington, MA). A custom flexible tether (Hypnion, Inc.) was connected at one end to the commutator and at the other end to the animal's cranial implant. Each cage was located within a separate compartment of a stainless-steel recording chamber and had an infrared light source and digital video camera to allow a minimum of twice daily remote visual monitoring. The ambient temperature was 23 ± 1°C, and relative humidity averaged 50%. A 24-h light/dark cycle (12:12) was maintained throughout the study using fluorescent light. Light intensity averaged 35 to 40 lux at midlevel inside the cage.
Materials
THIIC (Khilevich et al., 2008) and 3-[(2-methyl-1,3-thiazol-4-yl)ethynyl]pyridine (MTEP) were synthesized at Lilly Research Laboratories. For in vivo experiments, THIIC was suspended in 1% carboxymethylcellulose, 0.25% Tween 80, and 0.05% Dow antifoam. MTEP was dissolved in sterile H2O. Imipramine and alprazolam were purchased from Sigma-Aldrich. Chlordiazepoxide HCl (Sigma-Aldrich) was dissolved in 0.9% NaCl. Alprazolam was suspended in 5% acacia, and imipramine was dissolved in sterile 0.9% NaCl. For all experiments, drugs were mixed fresh before use. Doses of THIIC were administered to rats in a volume of 4 ml/kg; doses were administered to mice in a volume of 10 ml/kg. Imipramine and alprazolam were dosed in a volume of 1 mg/kg to rats and 10 mg/kg to mice. Routes of administration are indicated separately for each experiment.
Assays for Antidepressant-Like Activity
Mouse Forced-Swim Test.
Male, NIH Swiss mice (Harlan, Indianapolis, IN) weighing 20 to 25 g were housed in plastic cages (40.6 × 20.3 × 15.2 cm) with 10 to 12 mice/cage in a vivarium at least 7 days before the experiments. Mice were maintained on a 12-h light/dark cycle (6:00 AM/6:00 PM), and all procedures were performed between noon and 4:00 PM. Animals were removed from the vivarium to the testing area in their home cages and allowed to adapt to the new environment for at least 1 h before testing. The forced-swim test was performed using the original method described by Porsolt et al. (1977). In brief, mice were placed individually in clear plastic cylinders (10 cm in diameter by 25 cm in height) filled to 6 cm with 22 to 25°C water for 6 min. The duration of immobility was recorded during the last 4 min of a 6-min trial. A mouse was regarded as immobile when floating motionless or making only those movements necessary to keep its head above the water. Dose-effect functions for THIIC were performed after both intraperitoneal (30 min before) and oral dosing (60 min before). Imipramine was used as a comparator standard (15 mg/kg i.p., 30 min before). Statistical analyses were performed by analysis of variance (ANOVA) followed by Dunnett's test.
DRL 72-s Behavior in Rats.
Male Sprague-Dawley rats weighing between 300 and 350 g at the beginning of the behavioral experiments (Holtzman, Madison, WI) were housed in pairs. The colony room was maintained at 20°C and 60% relative humidity. The room was illuminated 12 h/day (7:00 AM to 7:00 PM). All rats had free access to laboratory chow (Teklad 4% rat diet; Harlan Teklad, Madison, WI) except during experimental sessions. Water was available for a 20-min period after the daily behavioral session. Sixteen operant-conditioning chambers (MED Associates, St. Albans, VT) were used. The lever in these chambers was mounted on one wall with the water access next to the lever in the middle of the wall. A reinforced response caused the dipper (0.02-ml cup) to be lifted from a water trough to an opening in the floor of the access port for 4 s. The house light, which was mounted on the opposite wall, was turned on when the session began, remained on throughout the entire session, and was turned off at the end of the session. Each experimental chamber was enclosed in a separate sound-attenuating cubicle equipped with a white noise generator to provide masking noise. Rats were water deprived for 22.5 h before each session. Each rat was initially trained under an alternative fixed ratio 1, fixed time 1-min schedule for water reinforcement where each response was reinforced, and water was also provided every minute if a response did not occur. The few rats that did not acquire lever-pressing behavior after three daily 1-h sessions under this schedule were trained by the experimenter using the method of successive approximation. After the rats had acquired lever-pressing behavior, they were trained during daily DRL 18-s sessions for 2 weeks before moving directly to DRL 72-s sessions. Under the DRL schedules, a response only produced water if it occurred at least 18 or 72 s after the last water delivery; responses before the designated time period reset the timer. Responding on these sessions became stable after 8 weeks. Experimental sessions lasted for 1 h and were conducted 5 days/week during light hours. The data analyses were conducted on the raw response and reinforcement data. All behavioral data are expressed as the mean ± S.E.M. normalized to the vehicle or vehicle/vehicle condition. The effects of THIIC and imipramine were analyzed with a one-way repeated-measures ANOVA. The level of significance was set for p < 0.05 for all analyses.
Dominant-Submissive Test in Rats.
Experimentally naive male Sprague-Dawley rats (Harlan), weighing 160 to 180 g at the start of the study, were assigned randomly to pairs, and the paired animals remained separated through the course of the experiment, except during the 5-min testing periods. Rats were maintained on a 12 h light/dark cycle (6:00 AM/6:00 PM), and all procedures were performed between 12:00 PM and 4:00 PM in ambient room light. The testing apparatus was constructed from transparent plastic material and consisted of two identical chambers (24 × 17 × 14 cm) connected by a round tunnel (4.5 cm diameter × 52 cm long). A container (10-ml beaker) of sweetened milk (9% sucrose) was placed in an opening in the floor at the midpoint of the tunnel. Animals were food-deprived overnight before the first test session. During the testing period, each member of a pair was placed in the different chambers of the testing apparatus, the gates were opened, and the time spent drinking was recorded for each animal for a 5-min period. At the end of the 5-min testing period, animals were returned to their home cages and given free access to food for 1 h. The animals were also given free access to food from Friday afternoon (after the test session) to Sunday morning, when they were once again food-deprived. During the first week of the testing (acclimation week), drinking time was not scored. During the second week of testing (selection week), the time spent drinking was recorded. Pairs of animals that passed the following criteria continued into the drug administration phase of the experiment: 1) the difference between the average daily drinking scores of the two animals was significant (two-tailed t test, p < 0.05), and 2) the dominant animal's score was at least 25% greater than the submissive animal's score. Any pairs not passing the selection criteria were dropped from the study. The animal of a pair with the higher drinking score was labeled as “dominant,” and the animal of a pair with lower drinking score was labeled as “submissive”. The submissive rat of each pair was treated with the THIIC (3 or 30 mg/kg) or imipramine (10 mg/kg), whereas the dominant rat of each pair was treated with vehicle for the next 21 days. The 5-min test was repeated once daily, except weekends, for 21 days after the beginning of the drug treatment. Dominance levels were calculated as the difference in daily drinking scores between paired rats. The daily dominance level values were averaged over each week, and statistical comparisons were made by analysis of repeated measurements coupled Tukey's post hoc test.
Assays for Anxiolytic-Like Activity
Marble Burying and Rotorod Performances.
Male CD1 mice were used as described previously (Li et al.,2006). In brief, mice were dosed with either THIIC (intraperitoneally) or chlordiazepoxide (intraperitoneally) and tested on a rotating rod without prior training. Mice were scored as failing if they fell off the rotorod on two occasions during a 2-min test. Marble-burying behavior was assessed in the same mice immediately after rotorod testing. The number of marbles buried out of 20 by sawdust bedding by at least 2/3 was measured. The number of marbles buried was evaluated by ANOVA followed by Dunnett's test; rotorod performances were evaluated by Fisher's exact probability test.
Rat Stress-Induced Hyperthermia.
Male Fischer F-344 rats (Harlan), weighing between 275 and 350 g, were individually housed with food and water available ad libitum and maintained on a 12-h light/dark cycle (lights on at 6:00 AM). The rats were fasted for approximately 12 to 18 h before the experiment. Rats were transported from the colony room in groups of 10 to a procedure room for dosing. Rats were dosed orally in a dose volume of 2 ml/kg with THIIC (1, 3, 10, and 30 mg/kg). The mGlu5 receptor antagonist MTEP, which has demonstrated robust anxiolytic-like activity in preclinical models, was used as a positive control (10 mg/kg p.o.). Immediately after dosing, rats were returned to their home cage, and the dosing room (room A) was darkened for the remainder of the 4-h pretreatment period. After the pretreatment period, rats were taken individually to a brightly lit adjacent room (room B) where baseline body temperatures were determined by insertion of a rectal probe lubricated with mineral oil. Core body temperature was assessed using a Physitemp BAT-12 Microprobe Thermometer with a Physitemp RET-2 rat rectal probe (Physitemp Instruments Inc., Clifton, NJ) inserted approximately 2 cm into the rectum and is designated as the baseline body temperature, T1, in °C. The rat was then placed back in the home cage and remained in room B. Ten minutes later, a second body temperature measurement was recorded (T2). The difference between the first and second body temperature measurements (T2 − T1) was used as an index of stress-induced hyperthermia. Data represent means and S.E.M. and were subjected to one-way ANOVA followed by Dunnett's test.
Reversal of Stress-Induced Increase in Cerebellar cGMP.
Male CF-1 mice (18–20 g) (Harlan) were housed in an enriched environment in groups of eight mice per cage with food and automated water available ad libitum. Animals were kept under a 12-h light/dark cycle with lights on at 6:00 AM in a temperature- and humidity-controlled environment for 5 to 7 days before their use. The mice were kept in their home cages to avoid stress and received an intraperitoneal injection of either vehicle, THIIC (3, 10, and 30 mg/kg), or alprazolam (1 mg/kg) 30 min before stressor (foot shock). To evaluate the effects of the test compounds on basal cGMP levels, vehicle control “no stress” mice were sacrificed without being subjected to a foot shock. Vehicle stress, THIIC-treated, and alprazolam-treated mice were subjected to a stressor of inescapable electric foot shock (1-mA intensity for 10 s) in a metal cage with a stainless-steel grid floor (Coulbourn Instruments, Allentown, PA), and immediately sacrificed for estimation of basal cGMP levels. All mice were euthanized by using a beam of microwave radiation focused on the skull for 0.5 s (Thermax Thermatron, Louisville, KY). The cerebellum was quickly removed from the skull, and the weight was recorded. The tissue was then placed in a polystyrene tube containing 2.0 ml of 1% perchloric acid on ice. Tissues were then homogenized and incubated on ice for 30 min after which the samples were boiled for 5 min. The tubes were then centrifuged at 11,700g for 20 min at room temperature. The supernatant was acetylated using a 2:1 ratio of triethylamine (EM Science, Gibbstown, NJ) and acetic anhydride. The samples were immediately vortexed and then centrifuged at 13,000g for 20 min at 4°C. The cGMP assay buffer included 0.05 M sodium acetate, trihydrate (6.8 g/liter), and 0.1% sodium azide adjusted to pH 6.2 with acetic acid. For the radioligand tracer, a separate tracer buffer was prepared using the cGMP assay buffer and adding 0.002 M EDTA disodium salt dihydrate (0.7 g/liter), with 1% normal rabbit serum (PerkinElmer Life and Analytical Sciences) added just before use.
[125I]cGMP, obtained from PerkinElmer Life and Analytical Sciences (SA 2200 Ci/mmol, 32.3 μCi/ml), was prepared in the tracer diluent, pH 6.2, at a concentration of 0.75 μCi/10-ml tracer buffer per FlashPlate (PerkinElmer Life and Analytical Sciences). A cGMP standard curve was prepared by using guanosine-3′:5′-cyclic monophosphate lyophilized standard (PerkinElmer Life and Analytical Sciences) reconstituted with 2 ml of distilled water to make a 2000 pm/ml solution (stored at 4°C). This stock was diluted with assay buffer to generate a nine-point standard curve. A 1-ml aliquot of each standard, including a blank, were acetylated and then vortexed immediately. One hundred microliters of each standard was added, in duplicate, to flashplate wells, including two 100-μl aliquots directly from the stock cGMP standard bottle for nonspecific binding determination. Forty microliters of each sample, in duplicate, was added to flashplate wells containing 60 μl of cGMP assay buffer for a total sample volume of 100 μl (yields a 1:50 final dilution). One hundred microliters of [125I]cGMP was added to all flashplate wells, and plates were covered before being shaken overnight (16–24 h) at 2 to 8°C using a titer plate shaker. Contents of the wells were aspirated and the plates were resealed. Plates were then counted for 1 min/well in a Packard TopCount radioactive counter (PerkinElmer Life and Analytical Sciences).
cGMP data calculations were analyzed by using Excel (Microsoft, Redmond, WA) and GraphPad Prism. Mean cpm from duplicate samples were averaged, and the average nonspecific binding cpm was subtracted to get specific cpm. Specific cpm were used to determine the normalized percentage bound. GraphPad Prism was used to determine the nonlinear regression curve fit from the normalized percentage bound for the standard curve and interpolated the values of the unknown samples as log picomole of cGMP. Results were normalized to tissue weights for each sample to calculate picomoles of cGMP/g tissue. Group averages were calculated as percentage of vehicle control and percentage of vehicle stress control, and the S.E.M. was determined. Statistical analyses were carried out by GraphPad Prism using one-factor (dose group) ANOVA followed by Dunnett's multiple comparison test versus vehicle and vehicle stress groups indicating a p < 0.05 value for significance.
Sleep EEG in the Rat.
Male Wistar rats were surgically prepared with a cranial implant that permitted chronic EEG and electromyogram (EMG) recording, in addition to an abdominally implanted telemeter to measure temperature and locomotor activity. After recovery, animals were individually housed and connected to low torque commutator via a flexible tether allowing access to all areas of the cage. Animals were maintained on a 24-h light/dark schedule (12:12) with ad libitum food and water. Animals were undisturbed for 48 h before and after each treatment. Sleep and wakefulness were determined by using SCORE2004 (Hypnion, Inc., Lexington, MA), a microcomputer-based sleep-wake and physiological monitoring system. In the present study, the system monitored amplified EEG [×10,000, bandpass 1–30 Hz; initial digitization rate 400 Hz (Grass Instruments Corp., Quincy, MA)], integrated EMG (bandpass 10–100 Hz, RMS integration), telemetered body temperature, nonspecific locomotor activity, and drinking activity simultaneously. Arousal states were classified on-line as NREM sleep, REM sleep, wake, or θ-dominated wake every 10 s by using EEG period and amplitude feature extraction and ranked membership algorithms. Individually taught EEG-arousal-state templates and EMG criteria differentiated states of arousal. Drug was administered 6 h after lights off (circadian time denoted CT-18) to look for sleep-promoting effects of the drug at 10 or 30 mg/kg or vehicle. Experiments were run as simultaneous parallel groups, and all doses including vehicle were tested in parallel on the same day in approximately equal-sized dose groups of 10 to 14 rats. Total NREM sleep and total REM sleep were computed for each animal. The primary outcomes were the total NREM sleep and REM sleep during the first 12 h and during the first 18 h after treatment. To assess sleep continuity, the single longest bout of the six hourly longest sleep bouts during the first 6 h after treatment was calculated, as was the average sleep bout over this period. Each outcome was analyzed by analysis of covariance using treatment group as the factor and the pretreatment period as the covariate. Sleep bouts were analyzed on the log scale to stabilize the variation. Adjusted means and the change from vehicle means and their corresponding standard errors were summarized for each treatment group. For longest sleep bouts results were transformed back to the linear scale and represent fold changes over vehicle. Unadjusted and Dunnett's multiple-comparison adjusted P values are shown for each outcome in each period.
In Vivo Neurochemistry
In Vivo Microdialysis in the Rat.
Implantation of a BAS guide cannula into the mPFC [anterior (A) 3.2 mm; lateral (L) 0.8 mm; and ventral (V), -2 mm] was carried out by Taconic Farms (Germantown, NY) 5 to 7 days before the experiment as described previously (Fell et al., 2010). A concentric type probe (BR-4) from BAS Bioanalytical Systems (West Lafayette, IN) to match the implanted cannula with a 4-mm membrane tip extending below the cannula was flushed with water and carefully inserted through the cannula 16 h before the experiment began. Immediately after insertion of the probe the rat was returned to the home cage. On the morning of the experiment the rat was moved to the cage where the experiment was to be conducted and allowed to acclimate for a period of 4 h. After the acclimation period the rat was connected to a fraction collection system for freely moving animals (BAS Bioanalytical Systems). The input tube of the dialysis probe was connected to a syringe pump (BeeHive and BabyBee; BAS Bioanalytical Systems), which delivered an artificial cerebrospinal fluid containing 150 mM NaCl, 3 mM KCl, 1.7 mM CaCl2, and 0.9 mM MgCl2, pH 6.0, to the probe at a rate of 1 μl/min. The output tubes from the rats were attached to a refrigerated fraction collector (BAS Bioanalytical Systems). After a period of 2 h for equilibration of the probe and establishment of stable histamine baseline levels, collection of 60-min fractions was started. The flow from the output lines was collected in 400-μl plastic tubes that contained 15 μl of an antioxidant solution (3 mM l-cysteine, 10 nM EDTA, and 50 nM isoproterenol as an internal standard, in 0.5 M acetic acid). The antioxidant serves to acidify the dialysate and prevents degradation of the amines and metabolites during the 24-h period that it takes to collect the samples and complete the HPLC assays. Typically the total sample volume was 60 μl. Three baseline samples were collected before injection of any drugs. At 6:00 PM the lights were turned out in the experimental room and not turned back on until 6:00 AM the next morning. THIIC (10 and 30 mg/kg) or vehicle were dosed orally 60 min before the onset of the dark phase (6:00 PM). All microdialysis data were calculated as percentage change from dialysate basal concentrations with 100% defined as the average of the three drug preinjection values. Each group had five to eight rats.
Histamine was separated on an HPLC column and quantified by fluorimetric detection after a postcolumn derivatization with an o-phthaldialdehyde (OPA)-containing reagent as follows. The mobile phase consisted of 0.16 M KH2PO4, 0.1 mM sodium octanesulfonic acid, and 0.1 mM EDTA; the pH was adjusted to 4.6. The flow rate of the mobile phase was 0.7 ml/min. The eluent line was connected by a T-piece for mixing with a reagent line through which a 0.02% solution of OPA in 0.15 M NaOH was delivered at 0.6 ml/min. The OPA reagent was mixed with the eluent in a mixing coil of Teflon tubing (o.d. 1.1 mm; i.d. 0.55 mm; length 1 m), which was insulated to allow the derivatization reaction to proceed at ambient temperature. The OPA reagent solution was prepared fresh daily, protected from light, and kept cooled in ice. Histamine was separated on a reverse-phase column, a BDS Hypersil, 3 μm, C18 analytical column (4.6 × 100 mm from ThermoFisher Scientific). The fluorescence of the reaction product was measured by a fluorimeter (Jasco, Tokyo, Japan; excitation: 350 nm; emission: 450 nm) set at the highest sensitivity. The sensitivity for histamine was approximately 20 fmol per sample (20 μl). All dialysis probe placements were checked by perfusing a tetra red solution through the dialysis probe for subsequent histological examination. Only results derived from rats with correctly positioned probes were included in the data analysis. Time course data represent the percentage change from baseline and were analyzed by two-way ANOVA followed by post hoc Bonferroni corrected t test. The level of significance was set at p < 0.05. All analyses were performed using the GraphPad PRISM statistical program.
Analysis of Tele-Methylhistamine in Rat CSF.
Male Harlan-Sprague-Dawley rats (200 g) were dosed acutely or for 5 days with THIIC at doses of 1 to 10 mg/kg orally. Six hours after dosing the rats were sacrificed with CO2 and CSF samples were collected from the cisterna magna. Concentrations of histamine and tele-methylhistamine (t-MeHA) were determined by using HPLC separation and detected by tandem mass spectrometry. Cerebrospinal fluid (50–100 μl) samples were spiked with stable label internal standards for both analytes. The CSF was then passed though an anion exchange solid-phase extraction cartridge to remove protein and interferences. The extracts for CSF were injected onto a Varian Inc. (Palo Alto, CA) monochrome 100 × 2 mm HPLC column for separation. The separation was achieved by using a mobile phase of 0.5% heptafluorobutyric acid and 10% acetic acid/0.5% heptafluorobutyric acid in acetonitrile. A linear gradient was used for separation of histamine and its metabolite. Histamine, its tele-methyl metabolite, and the stable label internal standards were detected by electrospray ionization with the specific tandem mass spectrometry transitions. Standard curves were analyzed from 0.150 to 50 ng/ml, and unknown concentrations were calculated by using an internal standard method. Detection limit for t-MeHA was 0.3 ng/ml. Data are expressed as mean ± S.E.M. and were subjected to a one-way ANOVA followed by post hoc analyses using Dunnett's corrected t test.
Results
THIIC Induces a Selective Allosteric Potentiation In Vitro of mGluR2 in Recombinant Systems.
Activation of mGluR2 was assessed by measuring glutamate-induced increases in intracellular calcium in AV12 cells expressing human or rat mGluR2 coexpressed with the G protein (Gqi5). Consistent with the effects of other allosteric potentiators of mGluR, THIIC positively modulated the mGluR2 as shown by the parallel leftward shift of a glutamate dose-response curve in the presence of increasing concentrations of THIIC with no increase in the maximal response to glutamate (Fig. 2). This leftward shift of the glutamate dose response represents a positive cooperativity between the allosteric binding site for THIIC and the orthosteric glutamate binding site. In the presence of an EC10 concentration of glutamate, the EC50 value for potentiation by THIIC was 22.5 ± 3.2 nM (mean ± S.E.M., n = 7) for the human mGluR2 and EC50 = 12.8 ± 3.8 nM (mean ± S.E.M., n = 4) for the rat mGluR2 homolog (data not shown).
Concentration-dependent potentiation of glutamate-stimulated mobilization of intracellular Ca2+ by THIIC using the FLIPR assay in clonal AV12 cells stably expressing the human mGluR2 receptor (A) and typical FLIPR traces showing minimal impact from agonist activity of compound alone (B). The effects of different concentrations of THIIC were measured in the presence of increasing concentrations of the agonist glutamate. EC50 values were calculated using a four-parameter logistic curve-fitting program. RLU, relative light units.
The ability of THIIC to shift glutamate-induced [35S]GTPγS binding in membranes expressing human mGluR2 is shown in Fig. 3. Increasing concentrations of THIIC produced an approximately 25-fold leftward shift in the glutamate-induced [35S]GTPγS in the human mGluR2 (positive cooperativity factor = 23.6, Kb = 824 nM). It is noteworthy that in the absence of glutamate THIIC also increased [35S]GTPγS binding in membranes expressing mGluR2 (EC50 1.5 μM). Agonist activity is also evident in the increasing baselines and the progressively flattening Hill slopes of the glutamate dose-response curves as the concentration of THIIC is increased. Together, these data indicate allosteric agonist and potentiation effects of THIIC in this particular assay.
Concentration-dependent stimulation of GTPγ35S binding by THIIC in clonal AV12 cells stably expressing the human mGlu2 receptor.
THIIC showed no significant activity across a panel of mGlu receptors (mGluR1, mGluR3, mGluR4, mGluR5, mGluR6, mGluR7, mGluR8) and GABAB receptors in antagonist, agonist, and potentiator modalities (see Table 1). In addition, THIIC was profiled further against a diverse panel of CNS receptors, showing no significant radioligand displacement at concentrations of 1 μM. However, THIIC did show signs of radioligand displacement at the following receptors and concentrations: hNET, (71% at 10 μM and 7% at 1 μM, rCl− channel, 80% at 10 μM and 12% at 1 μM, rNa+ channel, 78% at 10 μM and 25% at 1 μM, h5-HT2B agonist site 97% at 10 μM and 41% at 1 μM with a Ki = 3.7 μM). A follow-up study for agonist activity at the h5-HT2B using GTPγS binding revealed no significant stimulation of binding at concentrations up to 10 μM.
Summary of potency values (EC50 or IC50 in nM) for THIIC tested in a panel of human cloned mGlu and GABAB receptors using the calcium mobilization assay
Overall mean IC50 or EC50 ± S.E.M. as shown.
THIIC Engenders Antidepressant-Like Effect in the Mouse Forced-Swim Test that Depends on mGlu2 Receptors.
The forced-swim (behavioral despair) test in rodents in often used as a screen assay for drugs with potential antidepressant activity. Orally administered THIIC decreased immobility in the mouse forced-swim test in a dose-dependent manner after both acute and 5-day dosing (Fig. 4). However, somewhat greater potency and efficacy was achieved after repeat dosing than after acute dosing despite the lack of significant changes in the plasma kinetics of THIIC after repeat dosing (data not shown).
Top, comparison of acute (open bars) and 5-day repeated dosing (filled bars) of THIIC administered orally 60 min before testing on immobility in the mouse forced-swim test. Bars represent mean ± S.E.M. of eight mice/group. Imipramine (15 mg/kg) was dosed intraperitoneally 30 min before (n = 4/group). Data were analyzed by a one-way ANOVA followed by Dunnett's test. *, p < 0.05 compared with vehicle-treated group. Bottom, comparison of effects of THIIC (30 mg/kg i.p., 30 min) in the forced-swim test in wild-type (WT) and mGlu2 receptor KO mice. *, p < 0.05 compared with the WT vehicle control mice.
To determine whether mGlu2 receptors are responsible for the antidepressant-like effects of THIIC, mGlu2 receptor knockout (KO) mice were studied. In this study, mGlu2 receptor KO mice showed a small antidepressant-like phenotype on their own. When dosed acutely with 30 mg/kg THIIC (intraperitoneally, 30 min before), the antidepressant-like efficacy observed in the wild-type mice was eliminated in the mGlu2 receptor KO mice (Fig. 4). In contrast, the antidepressant-like effects of imipramine occurred in both mGlu2(−/−) and mGlu2(+/+) mice (data not shown).
THIIC Produces Antidepressant-Like Activity in DRL 72-s Assay.
The DRL 72-s assay in rats is an operant responding paradigm that is sensitive to clinically active antidepressant drugs including SSRIs, NRIs, and SNRIs (Marek et al., 1989, 2005). Under this assay, THIIC increased the number of water reinforcements received and reduced responses emitted in a dose-dependent manner (Fig. 5A). Although animals receiving doses of 1 and 3 mg/kg were not significantly different from vehicle-treated animals, a dose of 10 mg/kg produced a significant decrease in responses emitted (p < 0.05), while producing an increase in reinforcers received (though not statistically significant). The results on responses and reinforcers matched well with the data produced by the positive control, imipramine, which also produced a significant decrease in responses emitted (p < 0.05), while producing a nonsignificant increase in reinforcers received. In these animals, the magnitude of effect on responses and reinforcers was comparable. At 10 mg/kg, THIIC also produced a statistically significant increase in efficiency of responding by the animals (p < 0.05) as with 10 mg/kg imipramine. After treatment of 10 mg/kg THIIC for 5 days, there was a significant increase in reinforcers received and a significant decrease in responses emitted (Fig. 5B) compared with vehicle. These results showed a trend for a larger magnitude of effect in the DRL 72-s assay with THIIC after repeated treatment compared with acute treatment.
The effect of acute (A) and 5-days (B) repeated dosing of THIIC on responses (●) and reinforcers (○) on a DRL 72-s schedule. Each point represents the mean (± S.E.M.) of 10 rats at each dose tested. Imipramine (10 mg/kg i.p.) results are also shown. Data were analyzed by one-way ANOVA followed by Dunnett's test. *, p < 0.05 compared with the vehicle-treated group.
THIIC Produces Antidepressant-Like Activity in the Dominant-Submissive Behavior Assay.
The dominant-submissive test is a chronic dosing assay that is sensitive to clinically active antidepressant drugs including SSRIs and SNRIs. In the dominant-submissive test, the dominance level was significantly lower than baseline after oral treatment with THIIC at 3 and 30 mg/kg (p < 0.05) with statistical significant differences at weeks 2 and 3 (p < 0.05 in all cases). In contrast, 10 mg/kg of the tricyclic antidepressant imipramine required 3 weeks of dosing for statistical significant alleviation of submissive behavior (p < 0.05; Fig. 6).
Dose-dependent decrease in difference in dominance level (i.e., decrease in difference in feeding time) between the dominant and submissive rat after treatment with THIIC (3 and 30 mg/kg p.o. daily dosing) and imipramine (10 mg/kg/day). Bars represent mean ± S.E.M. in 8 to 13 rat pairs. *, p < 0.05 compared with baseline with Tukey's post hoc test.
Effects of THIIC on Sleep EEG in the Rat.
As shown in Fig. 7A, THIIC (10 and 30 mg/kg p.o.) dosed at CT18 (circadian time 18 = 18 h after lights on), i.e., during the dark active phase produced a significant increase in NREM sleep at 30 mg/kg over the first 12 h after dose, and an increase at both 10 and 30 mg/kg over the first 18 h, eliciting an extra 36.0 ± 10.9 and 57.0 ± 10.9 min, respectively, relative to vehicle control. A significant increase in NREM sleep of 30.8 ± 7.7 min was also seen after 10 mg/kg p.o. after dosing at CT5 (5 h after lights on, i.e., during the light inactive phase) (data not shown). A significant decrease in REM sleep was also observed over the same time course with 30 mg/kg THIIC, resulting in a reduction of 11.2 ± 2.7 min over the first 12 h and 15.1 ± 3.6 min over the first 18 h compared with vehicle (Fig. 7B). Analysis of sleep bouts over the first 6 h after dosing resulted in a dose-dependent increase, more than doubling at 30 mg/kg, for both average sleep bout length (2.2- ± 0.3-fold) and longest sleep bout (2.1- ± 0.3-fold) (Fig. 7C), indicating the ability of the compound to consolidate sleep at 10 and 30 mg/kg p.o.. A lower dose of 3.0 mg/kg p.o. did not produce statistically significant changes in sleep parameters (data not shown).
Effect of THIIC on NREM sleep (A), REM sleep (B), and sleep bout length (C) in rats. A, dose-dependent increase in NREM sleep over the first 12 h and first 18-h period after dosing orally at 10 and 30 mg/kg. B, dose-dependent decrease in REM sleep over the first 12 h and first 18-h period after dosing orally at 10 and 30 mg/kg. C, dose-dependent increase in longest sleep bout length over the first 6 h after dosing at 10 and 30 mg/kg. Bars represent mean ± S.E.M. of 10 to 13 rats. NREM and REM sleep values are cumulative minutes over the stated period. The longest sleep bout was analyzed on the log scale and back-transformed to the linear scale to stabilize the variation. Each outcome was analyzed by analysis of covariance using treatment group as the factor and the pretreatment period as the covariate. *, p < 0.01 compared with vehicle group; **, p < 0.0001 compared with vehicle group.
THIIC Suppresses Marble Burying in a Genotype-Dependent Manner.
Marble burying behavior in the mouse can detect both GABA and serotonin-based anxiolytic agents (Li et al., 2006). THIIC produced dose-dependent decreases in marble burying as did the anxiolytic agent chlordiazepoxide (Fig. 8). THIIC did not affect rotorod performance of mice, whereas chlordiazepoxide did (Fig. 8). The effect of THIIC was eliminated in mGlu2 receptor null mice (Fig. 8).
THIIC suppresses marble burying of mice that depends on mGlu2 receptors. Chlordiazepoxide HCl was studied as an anxiolytic comparator. Each point represents effects in eight mice/dose group with 12 to 16 mice/group in the vehicle control animals. Significant effects by ANOVA were followed by Dunnett's post hoc test. **, p < 0.01 versus control. Inset, the effects of 30 mg/kg THIIC and 30 mg/kg chlordiazepoxide on rotorod performance. *, p < 0.05 by Fisher's exact probability test.
THIIC Inhibits Stress-Induced Hyperthermia in Rats.
Hyperthermia is a general phenomenon that has been reliably demonstrated in many species in response to stress and is a component of the well characterized fight-or-flight response. Stress-induced hyperthermia is attenuated by clinical anxiolytics and is widely used preclinically to predict anxiolytic efficacy of novel compounds (Spooren et al., 2002). As is shown in Fig. 9, THIIC (1–30 mg/kg) produced a dose-dependent reduction in stress-induced hyperthermia (F5,53 = 7.06, p < 0.001), with significant reductions compared with vehicle observed at 3, 10, and 30 mg/kg (P < 0.05 in all cases). The effect of THIIC was comparable with the response of the mGlu5 antagonist MTEP (10 mg/kg). THIIC did not affect baseline core body temperature (data not shown).
Dose-dependent inhibition of stress-induced hyperthermia in Fischer F-344 rats after oral administration of THIIC. Data represent means and SEM for 9 to 10 rats/group. Data were analyzed by a one-way ANOVA (F5,53 = 7.06, p < 0.001) followed by Dunnett's test. *, p < 0.05 compared with the vehicle-treated group. Percentage values in each bar represent the percentage inhibition relative to the vehicle control group.
THIIC Prevents Stress-Induced Increases in Cerebellar cGMP.
Another animal model that is often used to assess the potential anxiolytic/antistress effects of novel compounds is reversal of stress-induced levels of cGMP in the cerebellum. As shown in Fig. 10, exposure to a brief foot shock resulted in a significant increase in cGMP (∼80%) compared with the vehicle-treated group, which received no foot shock (F5,41 = 6.227; p < 0.001). This stress-induced increase in cGMP was completely reversed by THIIC at the highest dose tested of 30 mg/kg (p < 0.05), although lower doses of 3 and 10 mg/kg were not significantly different from the vehicle + foot shock-treated group. In the same experiment, alprazolam (1 mg/kg) produced a similar antagonism of the stress effect (p < 0.05).
Inhibition of stress-induced increase in cerebellar cGMP in male CF-1 mice after intraperitoneal administration of THIIC. Data represent means and S.E.M. for 7 to 10 mice/group and were analyzed by one-way ANOVA followed by Dunnett's test. #, P < 0.05 compared with the vehicle-no stress group. *, p < 0.05 and **, p < 0.01 compared with the vehicle + stress group.
THIIC Decreases the Dark-Phase Increase in HA Efflux in the mPFC and Reduces t-MeHA Levels in Rat CSF after Acute and 5-Day Dosing.
As shown in Fig. 11, THIIC at 10 and 30 mg/kg significantly attenuated the dark-phase increase in mPFC histamine efflux as measured by in vivo microdialysis. A two-way ANOVA revealed a significant effect of drug (F2,208 = 34.52; P < 0.001), time (F15,208 = 10.73; p < 0.001), and a significant drug × time interaction (F30,208 = 16.11; P < 0.05). THIIC showed a dose-dependent inhibition of t-MeHA in the CSF with approximately 50% inhibition after 3 mg/kg (p < 0.05) and 10 mg/kg (p < 0.05; Fig. 12A). After 5 days of repeated dosing, levels of t-MeHA were completely suppressed and were below the level of detection after both the 3 and 10 mg/kg p.o. doses (Fig. 12B).
Effect of THIIC on the dark-phase increase in histamine efflux in the mPFC. THIIC (10 and 30 mg/kg) was dosed 60 min before the onset of the dark phase (6:00 PM and time 0 on graph). Data are the mean ± S.E.M. of the dialysate levels, expressed as a percentage of the baseline, and were analyzed by two-way ANOVA followed by a post hoc Bonferroni corrected t test (n = 5–8 per group). *, p < 0.05 compared with the vehicle-treated group.
Effects of acute (A) and 5 days repeated dosing (B) with THIIC on CSF levels of tele-methylhistamine. Animals were administered THIIC (1–10 mg/kg p.o.), and CSF samples were collected 6 h after dosing. Data were analyzed by a one-way ANOVA followed by Dunnett's test (n = 4–8 per group). *, p < 0.05 and **, p < 0.01 compared with the vehicle-treated group.
Discussion
The current article details the pharmacological properties of a structurally novel positive allosteric modulator of mGlu2 receptors, THIIC. In rat and human recombinant systems, THIIC is a potent potentiator of the response of mGlu2 receptors to glutamate and is highly selective for mGlu2 receptors, having no direct effect on the glutamate response of other mGlu receptors, ionotropic glutamate receptors, glutamate transporters, or receptors implicated in the actions of clinically effective psychiatric medications including dopaminergic, serotonergic, muscarinic, adrenergic, GABAergic, and histaminergic. The selectivity of this molecule for mGlu2 receptors thus enables an evaluation of the biological consequences of augmenting mGlu2 receptor function in vivo. THIIC showed robust activity in behavioral models used to predict anxiolytic- and antidepressant-like activity. It is noteworthy that in several assays the effects of THIIC were negated in mice without mGlu2 receptors, providing in vivo support for the molecular target of these biological impacts from THIIC. Encouragingly, these data also provide evidence that doses of THIIC with anxiolytic/antidepressant-like activity also produced significant electrophysiological and neurochemical changes, including changes on sleep EEG parameters (NREM and REM sleep), and decreased histamine release in the prefrontal cortex and t-MeHA in the CSF, which may represent potential translational biomarkers for this drug target.
THIIC demonstrated efficacy in three animal models used to predict antidepressant-like activity. Such antidepressant activity of orthosteric mGlu2/3 receptor agonists have not been reported although modification of antidepressant effects have been shown by one group (Matrisciano et al., 2008; see Witkin et al., 2007 for an overview). Systemically administered THIIC dose-dependently decreased immobility time in the mouse forced-swim test (an effect lost in mice deficient of mGlu2 receptors) and was also efficacious in the rat DRL 72-s assay, an operant responding paradigm that is sensitive to clinically active antidepressant drugs, including SSRIs, NRIs, and SNRIs. In both assays the degree of efficacy with THIIC was similar to that of the clinically efficacious tricyclic antidepressant imipramine. Moreover, the antidepressant-like effects of THIIC were retained after 5 days of dosing, suggesting that there is a low potential for tolerance with this mechanism. These findings are further supported by our finding that THIIC was efficacious in the rat dominant-submissive test, a chronic assay in which THIIC was administered for a 3-week treatment period. Although neurogenesis has been suggested to occur with repeat dosing of antidepressants (c.f., Witkin et al., 2007), we did not study this phenomenon here. Together, these data suggests that THIIC might produce antidepressant-like efficacy in humans.
To the best of our knowledge these are the first data showing that the potentiation of mGlu2 receptor activity and a subsequent reduction in glutamatergic tone has efficacy in preclinical models that detect acute or subchronic dosing with antidepressants in vivo. However, increasing glutamate tone also produces antidepressant-like effects in a host of rodent models (Witkin et al., 2007). The seeming disparity of mechanisms producing a common behavioral outcome is potentially puzzling. Resolution to such potential mechanistic convergence could come from several lines of argument. First, mGlu2 receptor potentiation engages mGlu2 in a physiological-relevant manner whereas the orthosteric antagonists do not. Second, mGlu2 potentiators influence mGlu2 in exclusion of mGlu3. Third, only one mGlu2 potentiator was exemplified here and might not be representative of a class effect. Fourth, in contrast to THIIC, mGlu2/3 receptor antagonists do not generally produce anxiolytic-like effects (Witkin and Eiler, 2006), mGlu2/3 antagonists do not display antipsychotic activity as reported for mGlu2 potentiators (Galici et al., 2005), and mGlu2/3 antagonists are not active in some antidepressant-detecting assays that showed positive results with THIIC and vice versa (Witkin et al., 2007). It is noteworthy that major depressive disorder is a composed of a heterogeneous set of symptoms and symptom clusters that probably have distinct implications for therapeutics (Witkin et al., 2007).
Consistent with previous reports with mGlu2/3 receptor agonists and other mGlu2 receptor potentiators, THIIC produced anxiolytic-like effects in pharmacological models predicative of anxiolytic efficacy. THIIC produced a dose-related reduction in the stress-induced increase in cerebellar cGMP and reduced the number of marbles buried in the mouse marble burying assay, an effect that was lost in mGlu2 receptor-deficient mice. THIIC also prevented the development of stress-induced hypothermia in rats, and again the magnitude of the observed decrease in body temperature was similar to effects observed previously with structurally unrelated allosteric potentiators of mGlu2 receptor (CBiPES, BINA) and mGlu2/3 receptor agonists (LY354740) (Galici et al., 2005; Johnson et al., 2005; Rorick-Kehn et al., 2006).
THIIC was also characterized in a series of in vivo and ex vivo neurochemical experiments to shed further light on the mechanism of action of mGlu2 receptor potentiation and to identify potential neurochemical markers of behavioral efficacy. Although not reported here, we did not observe any effects of THIIC on extracellular (or brain tissue) monoamine levels (dopamine, serotonin, and norepinephrine) or their metabolites. Based on the finding that THIIC has significant sleep-promoting effects and that we observed significant interactions for CBiPES with the histamine system (Fell et al., 2010), we assessed the effects of THIIC on extracellular histamine levels. THIIC clearly attenuated the increase in extracellular histamine in the mPFC, which occurs during the dark phase, suggesting that potentiation of mGlu2 receptor activity can modulate central histaminergic neurotransmission in the brain. Although this mechanism is not fully understood, previous studies have shown a stimulatory action of glutamate on histaminergic neuronal activity (Okakura et al., 1992). Thus the ability of mGlu2 receptors to dampen presynaptic glutamate release may account for the effect of THIIC on HA efflux in the mPFC. Further studies are currently ongoing in our laboratories to elucidate this mechanism in more detail. Nonetheless, that THIIC reduced HA efflux is intriguing and may contribute to the behavioral effects of THIIC. Neuronal histamine is known to play a key modulatory role in a number of CNS functions, including arousal, and has also been implicated in stress-related disorders such as anxiety (Frisch et al., 1998; Zarrindast et al., 2005). Furthermore, the ability to modulate histamergic neurotransmission is a feature shared by clinically effective anxiolytic drugs, diazepam and buspirone (Oishi et al., 1986, 1992). In addition, the relationship between extracellular histamine levels and sleep-wake activity is well established, i.e., histamine H3 antagonists that increase histamine release are wake-promoting in several species, including humans (Witkin and Nelson, 2004).
In recent years, an increasing emphasis has been placed on the alignment of preclinical and clinical research in drug discovery to facilitate the development of novel therapeutic agents. Indeed, a central aim of the present study was to identify potential biomarkers that could facilitate the translational development of THIIC. This drive for a biomarker combined with the finding that THIIC modulates histaminergic neurotransmission led us to evaluate the effects of THIIC on histamine and its metabolite t-MeHA in the CSF of rats. The analysis of CSF neurotransmitter levels is a relatively underexploited but potentially translatable assay that can be performed in both animals and human subjects. Moreover, because CSF is in direct contact with the CNS, brain processes are more likely to be reflected in the CSF than other in other body fluids. Indeed, CSF histamine levels in rats have been shown to reflect central histamine neurotransmission (Soya et al., 2008). Consistent with the in vivo microdialysis data, we observed a significant reduction in CSF t-MeHA levels after acute and repeated dosing with THIIC that probably reflects the effect of THIIC on central histaminergic activity. Although further studies are needed to evaluate the role of mGlu2 receptors in the regulation of histaminergic activity, these findings suggest that the analysis of the histamine metabolite t-MeHA in the CSF could represent a viable translational biomarker approach for the clinical development of mGlu2 potentiators.
Another preclinical assay that has the potential to translate to the clinical setting is sleep EEG. Abnormalities of REM sleep (shortening of REM latency, lengthening of the duration of the first REM period) are frequently observed in patients with major depressive disorder and have been considered specific of depressive disease (Feige et al., 2002; Chalon et al., 2005). In the present study, THIIC decreased REM sleep and increased NREM sleep. These effects on REM sleep are consistent with a previous report showing that the mGlu2 receptor potentiator BINA similarly inhibited REM sleep and increased its onset latency (Ahnaou et al., 2009). Moreover, these changes are also similar to the effects of many clinically effective antidepressants that are known to produce changes in sleep pattern, including reduced REM sleep and increased REM sleep-onset latency (Mayers and Baldwin, 2005; Thase, 2006).
In summary, convergent data from the present series of experiments demonstrates that allosteric potentiation of the mGlu2 receptor with THIIC suggests a novel and potentially promising target for the treatment of anxiety and/or depression. Our findings also provide evidence that sleep EEG and CSF t-MeHA may represent viable translational biomarker approaches to the clinical development of mGlu2 potentiators.
Authorship Contributions
Participated in research design: Fell, Witkin, Perry, Rorick-Kehn, Rasmussen, Benvenga, Leuke, Edgar, Wafford, Seidel, Quets, Felder, Heinz, Kuo, Ebert, Eckstein, Ackermann, Swanson, Catlow, Dean, Jackson, Tauscher-Wisniewski, Marek, Schkeryantz, and Svensson.
Conducted experiments: Fell, Falcone, Katner, Hart, Overshiner, Chaney, Li, Marlow, Thompson, Wafford, Quets, Wang, Nikolayev, Eckstein, and Ebert.
Contributed new reagents or analytic tools: Mayhugh, Khilevich, Zhang, and Schkeryantz.
Performed data analysis: Fell, Witkin, Falcone, Rorick-Kehn, Rasmussen, Benvenga, Li, Marlow, Thompson, Seidel, Wafford, Wang, Heinz, and Quets.
Wrote or contributed to the writing of the manuscript: Fell, Witkin, Rorick-Kehn, Rasmussen, Benvenga, Wafford, Quets, Heinz, Schkeryantz, and Svensson.
Footnotes
These studies were funded by Eli Lilly and Company, USA.
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
doi:10.1124/jpet.110.172957.
-
ABBREVIATIONS:
- mGlu
- metabotropic glutamate
- mGluR
- mGlu receptor
- THIIC
- N-(4-((2-(trifluoromethyl)-3-hydroxy-4-(isobutyryl)phenoxy)methyl)benzyl)-1-methyl-1H-imidazole-4-carboxamide
- LY404039
- (−)-(1R,4S,5S,6S)-4-amino-2-sulfonylbicyclo [3.1.0]hexane-4,6-dicarboxylic acid
- LY379268
- (−)-2-oxa-4-aminobicyclo[3.1.0]hexane-4,6-dicarboxylic acid
- LY354740
- 1S,2S,5R,6S-2 aminobicyclo[3.1.0]hexane-2,6-bicaroxylate monohydrate
- CBiPES
- N-4′-cyano-biphenyl-3-yl)-N-(3-pyridinylmethyl)-ethanesulfonamide hydrochloride
- BINA
- biphenyl-indanone A
- LY487379
- N-(4-(2-methoxyphenoxy)-phenyl-N-(2,2,2-trifluoroethylsulfonyl)-pyrid-3-ylmethylamine
- mPFC
- medial prefrontal cortex
- HA
- histamine
- t-MeHA
- tele-methylhistamine
- REM
- rapid eye movement
- NREM
- non-rapid eye movement
- ANOVA
- analysis of variance
- CNS
- central nervous system
- EEG
- electroencephalogram
- DRL
- differential reinforcement of low rate
- CSF
- cerebrospinal fluid
- FLIPR
- fluorometric imaging plate reading
- DMSO
- dimethyl sulfoxide
- MTEP
- 3-[(2-methyl-1,3-thiazol-4-yl)ethynyl]pyridine
- EMG
- electromyogram
- HPLC
- high-performance liquid chromatography
- OPA
- o-phthaldialdehyde
- GTPγS
- guanosine 5′-O-(3-thio)triphosphate
- KO
- knockout
- WT
- wild type
- SSRI
- selective serotonin reuptake inhibitor
- NRI
- norepinephrine reuptake inhibitor
- SNRI
- serotonin-norepinephrine reuptake inhibitor
- CT
- circadian time.
- Received July 15, 2010.
- Accepted October 12, 2010.
- Copyright © 2011 by The American Society for Pharmacology and Experimental Therapeutics