Abstract
Positive allosteric modulators (PAMs) of metabotropic glutamate receptor 4 (mGluR4) have been proposed as a novel therapeutic approach for the treatment of Parkinson's disease. However, evaluation of this proposal has been limited by the availability of appropriate pharmacological tools to interrogate the target. In this study, we describe the properties of a novel mGluR4 PAM. 5-Methyl-N-(4-methylpyrimidin-2-yl)-4-(1H-pyrazol-4-yl)thiazol-2-amine (ADX88178) enhances glutamate-mediated activation of human and rat mGluR4 with EC50 values of 4 and 9 nM, respectively. The compound is highly selective for mGluR4 with minimal activities at other mGluRs. Oral administration of ADX88178 in rats is associated with high bioavailability and results in cerebrospinal fluid exposure of >50-fold the in vitro EC50 value. ADX88178 reverses haloperidol-induced catalepsy in rats at 3 and 10 mg/kg. It is noteworthy that this compound alone has no impact on forelimb akinesia resulting from a bilateral 6-hydroxydopamine lesion in rats. However, coadministration of a low dose of l-DOPA (6 mg/kg) enabled a robust, dose-dependent reversal of the forelimb akinesia deficit. ADX88178 also increased the effects of quinpirole in lesioned rats and enhanced the effects of l-DOPA in MitoPark mice. It is noteworthy that the enhancement of the actions of l-DOPA was not associated with an exacerbation of l-DOPA-induced dyskinesias in rats. ADX88178 is a novel, potent, and selective mGluR4 PAM that is a valuable tool for exploring the therapeutic potential of mGluR4 modulation. The use of this novel tool molecule supports the proposal that activation of mGluR4 may be therapeutically useful in Parkinson's disease.
Introduction
Parkinson's disease (PD) is a progressive neurodegenerative disease that results in the loss of dopaminergic neurons in the substantia nigra. One consequence of the depletion of dopamine in this disease is a series of movement disorders, including bradykinesia, akinesia, tremor, gait disorders, and problems with balance (Schapira, 2009). These motor disturbances form the hallmark of PD, although there are many other nonmotor symptoms that are associated with the disease (Chaudhuri et al., 2006). Early in the course of the disease PD symptoms are effectively treated by dopamine replacement or augmentation with the use of dopamine D2 receptor agonists, l-DOPA, or monoamine oxidase B inhibitors. However, as the disease progresses these agents become less effective in controlling motor symptoms. In addition, their use is limited by the emergence of adverse effects including dopamine agonist-induced dyskinesias (Schapira, 2009). Consequently, there remains a need for new approaches to the treatment of PD that improve the effectiveness of the control of motor symptoms (Meissner et al., 2011).
Activation of metabotropic glutamate receptor 4 (mGluR4) has been proposed as a potential therapeutic approach to Parkinson's disease (Marino et al., 2003; Conn et al., 2005; Duty, 2010). A member of the group III mGluRs, mGluR4 is predominantly a presynaptic glutamate receptor that is expressed in several key locations in the basal ganglia circuits that control movement (Bradley et al., 1999; Conn et al., 2005). Activation of mGluR4 with group III-preferring agonists decreases inhibitory and excitatory postsynaptic potentials, presumably by decreasing the release of GABA and glutamate, respectively (Marino et al., 2003; Valenti et al., 2003, 2005). Because of the strategic location of mGluR4 on output neurons from the globus pallidus and subthalamic nucleus, it has been suggested that increasing mGluR4 activity will normalize aberrant synaptic transmission resulting from dopamine depletion and thus restore normal activity in the basal ganglia circuits that control movement. It is noteworthy that the highest density of mGluR4 is found in the cerebellum in granule cells (Bradley et al., 1999). The contribution of these receptors to the normal control of movement is not fully appreciated, although mice lacking in mGluR4 show impairments in motor learning (Pekhletski et al., 1996). The similarity in the ligand binding domains of group III mGluRs creates a challenge for identifying selective orthosteric agonists of this receptor, although some progress has been made in this area (Beurrier et al., 2009; Goudet et al., 2012). However, targeting positive allosteric modulators (PAMs) rather than orthosteric agonists provides a broader opportunity for identifying molecules that are exclusively selective between mGluRs.
In this study, we report for the first time the properties of 5-methyl-N-(4-methylpyrimidin-2-yl)-4-(1H-pyrazol-4-yl)thiazol-2-amine (ADX88178) (Fig. 1), a novel mGluR4 PAM. This molecule is selective for mGluR4, potent, orally available, and brain penetrant. Using this molecule we have tested the hypothesis that mGluR4 activation will ameliorate the parkinsonian symptoms in two rodent models of dopamine depletion.
Materials and Methods
Drugs
ADX88178 was synthesized at Addex Pharmaceuticals, Geneva, Switzerland. l-DOPA, benserazide, and quinpirole dihydrochloride were purchased from Sigma-Aldrich (St. Louis, MO). Haloperidol (D2 antagonist) was purchased from Janssen-Cilag (Boulogne, France; Haldol, injectable solution 2 mg/ml). For in vivo studies, ADX88178 was suspended in 1% carboxymethylcellulose (CMC) sodium salt. Haloperidol and l-DOPA in combination with benserazide were dissolved in saline. All solutions and suspensions were prepared fresh daily. Doses referred to are the free base.
In Vitro Studies
Stable Cell Lines.
The cDNAs encoding the human or the rat mGluR4 were subcloned into an expression vector also containing the hygromycin resistance gene. In parallel, the cDNA encoding a Gα16 protein allowing redirection of the activation signal to intracellular calcium flux was subcloned into a different expression vector also containing the puromycin resistance gene. Transfection of both of these vectors into HEK293 cells with PolyFect reagent (QIAGEN, Valencia, CA), according to the supplier's protocol, and hygromycin and puromycin treatment allowed selection of antibiotic-resistant cells that had integrated stably one or more copies of the plasmids. Positive cellular clones expressing mGluR4 were identified in a functional assay measuring changes in calcium fluxes in response to glutamate or selective known mGluR4 orthosteric agonists and antagonists as well as (−)-N-phenyl-7-(hydroxyimino)cyclopropa[b]chromen-1a-carboxamide (PHCCC), a known positive allosteric modulator of group III mGluRs (Maj et al., 2003). HEK293 cells expressing mGluR4 were maintained in media containing Dulbecco's modified Eagle's medium, 10% dialyzed fetal calf serum, 2 mM Glutamax, 100 units/ml penicillin, 100 μg/ml streptomycin, 100 μg/ml G418, 40 μg/ml hygromycin B, and 1 μg/ml puromycin at 37°C with 5% CO2 in a humidified atmosphere.
Fluorescent Cell Based-Ca2+ Mobilization Assay.
All assays were 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 experiment, human or rat mGluR4-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 a glutamine/glutamate-free Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, 100 units/ml penicillin, and 100 μg/ml streptomycin. On the day of the assay, the cells were loaded with a 3 μM solution of Fluo4-AM (LuBioScience, Lucerne, Switzerland) in assay buffer containing 0.03% Pluronic F-127 in DMSO (Invitrogen, Carlsbad, CA). After 1 h at 37°C with 5% CO2 in a humidified atmosphere, the extracellular dye was removed by washing the cell plate with the assay buffer, and calcium flux was measured with a fluorometric imaging plate reader (Molecular Devices, Sunnyvale, CA). After 10 s of basal fluorescence recording, various concentrations of the compounds to be tested were added to the cells. Changes in fluorescence levels were first monitored for 180 s to detect any agonist activity of the compounds. Then the cells were stimulated by glutamate at an EC20 concentration (concentration giving 20% of the maximal glutamate response) for an additional 110 s to measure enhancement activity of the compounds. In Schild-plot experiments, different concentrations of PAMs were preincubated for 3 min on the cells before addition of glutamate to generate concentration-response curves.
Selectivity Analysis.
ADX88178 was functionally tested up to 30 μM as agonist, antagonist, and positive or negative allosteric modulator at rat and/or human family III G protein-coupled receptors (mGluR1, mGluR2, mGluR3, mGluR5, mGluR6, mGluR7, and mGluR8) in fluorometric imaging plate reader experiments using the technique described in Fluorescent Cell-Based-Ca2+ Mobilization Assay. Those clones were prepared following the same method described for mGluR4 cell lines. In addition, ADX88178 was tested up to 10 μM in competition radioligand binding assays on membranes expressing 71 different receptors, transporters, enzymes, and ion channels (binding profile at Cerep, Poitiers, France).
Radioligand Binding Assay.
Binding to mGluR4 was assessed with a novel mGluR4-selective radioligand, [3H]PAM2 (F. Hess, E. Le Poul, N. Liverton, J. McAuley, T. McDonald, and I. J. Reynolds, unpublished work). Membranes were prepared from Chinese hamster ovary cells expressing the human mGluR4 receptor. Cells were harvested by using trypsin and then washed in assay buffer containing 50 mM HEPES, 1 mM EDTA, and protease inhibitors (Roche Diagnostics, Indianapolis, IN). Harvested cells were pelleted by centrifugation and stored at −80°C for a minimum of 12 h. Cells were resuspended in assay buffer and lysed by a polytron (Kinematica AG, Luzern, Switzerland). A low-speed spin, 2000g for 10 min, was used to recover unbroken cells for additional treatment by using the polytron. The cleared supernatant from the low-speed spins was subjected to a high-speed spin at 96,000g for 1 h. The membrane pellet from the high-speed spin was resuspended in assay buffer and used in the mGluR4 binding assay.
The binding assay followed a standard filtration binding paradigm. The binding buffer consisted of 20 mM HEPES, 100 mM NaCl, and 3 mM MgCl2, pH 7.4. The assay was performed in a 96-well format with each well containing test compound, 25 to 50 μg of membrane protein, 20 μM (2S)-2-amino-4-phosphonobutanoic acid, and 7 nM radiotracer. The reaction was incubated at room temperature for 1 h and passed through a GF/B filter that was presoaked with 0.5% polyethyleneimine. The cell membranes were collected on the filter and washed five times with ∼5 ml of ice-cold binding buffer containing 0.01% bovine serum albumin. The washed filter plate was dried for 30 min at 37°C, scintillation fluid was added, and the amount of filter-bound radioactivity was determined by using a Packard TopCount-NXT (PerkinElmer Life and Analytical Sciences, Waltham, MA). Total binding was determined in the absence of any test compound, but in the presence of the DMSO concentration that would result from the addition of test compound. The level of nonspecific binding for the radiotracer [3H]PAM2 was determined in the presence of 5 μM PAM1, a structurally distinct mGluR4 PAM known to bind to the same binding site (F. Hess, E. Le Poul, N. Liverton, J. McAuley, T. McDonald, and I. J. Reynolds, unpublished work).
Pharmacokinetic Studies
In Vivo Studies.
For intravenous administration, ADX88178 was dissolved in 80% polyethyleneglycol-400 and 20% saline (1 ml/kg). For oral administration ADX88178 was suspended in 1% CMC (5 ml/kg). All doses refer to free base of compound. For blood sampling, rats were surgically implanted under isoflurane-oxygen anesthesia (1–2%) with jugular vein catheters that ran subcutaneously from the jugular vein and exited in the midscapular region. Animals were allowed to recover for 2 to 6 days after surgery before administration of ADX88178. CSF was sampled as described previously (El Mouedden et al., 2005). Rats received a single intravenous injection in the tail vein (0.3 mg/kg) or oral gavage of 3, 10, or 30 mg/kg of ADX88178. Serial blood and CSF samples were collected from the catheter at 5 (intravenous only), 15, and 30 min and 1, 2, 4, 6, 8, and 24 h after administration. The 24-h time point was collected for intravenous and 30 mg/kg p.o. only. Blood samples were collected in 1.5-ml polyethylene Eppendorf tubes containing 4 μl of 15% EDTA solution and immediately placed on ice. Samples were centrifuged at 4°C for 12 min at 5900g for 3 min. Plasma was transferred to 1.5-ml Eppendorf tubes and stored at −20°C until analysis. CSF was collected into 20-μl heparin-coated capillary tubes. For sample preparation, CSF was transferred into Eppendorf tubes containing 20 μl of control rat plasma. These samples were checked carefully for blood contamination. 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 next section).
Plasma Sample Analysis.
To precipitate proteins, 150 μl of acetonitrile was added to 50 μl of plasma spiked with 10 μl of DMSO for unknown samples or 10 μl of ADX88178 in DMSO for calibration and quality-control samples. After vortexing and centrifugation (15 min; 4°C; 13,200 rpm), a portion (100 μl) was transferred into the 384-well analytical plate. Five microliters of the supernatant was injected to an ultra performance liquid chromatography system (Waters, Milford, MA) coupled with a mass spectrometer (API 3200; Applied Biosystems, Foster City, CA). Samples were injected onto an Acquity BEH C18 reverse-phase column (Waters; 1.7 μm; 2.1 × 50 mm). Elution was performed with high-pressure linear gradient from 25 to 100% acetonitrile in 10 mM ammonium formate at pH 3.5. Run time was 1.5 min (retention time = 0.43 min). The electrospray positive ionization was used in multiple reaction monitoring mode (transition 273.04/132.2). Two calibration curves and three quality-control levels in duplicate were used to quantify and validate the run (Quadratic 1/x).
Cytochrome P450 Inhibition.
Cytochrome P450 inhibition was determined by using recombinant human cytochromes as described previously (Crespi, 1999).
Intrinsic Clearance in Microsomes.
ADX88178 was incubated at 37°C with rat and human liver microsomes (BD Gentest, Woburn, MA) at a concentration of 0.3 μM in the absence and presence of an NADPH-regenerating system. Aliquots of the incubations were taken at 0, 2, 5, 10, 15, and 20 min and added to three volumes of cold acetonitrile. Samples were then centrifuged for 15 min at 3000 rpm and 4°C. After centrifugation, a portion of the supernatant was transferred into a 384-well plate for analysis of the remaining ADX88178 by liquid chromatography/tandem mass spectrometry. The first-order rate constant for consumption of the substrates (k) was obtained from the slope of the linear regression from log percentage remaining versus incubation time plots. In vitro t1/2 values were calculated from: t1/2 = −0.693/k. Verapamil was used as a control in all experiments. Intrinsic clearance values were calculated from the obtained half-life as described previously (Obach, 1999).
Plasma Protein Binding.
Plasma 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 was assembled so that each well was divided vertically in two parts by a dialysis membrane. Molecular weight cutoff-regenerated dialysis cellulose membranes (Dialysis Membrane Strips; HT Dialysis) with a molecular mass cutoff of 12 to 14,000 Da were conditioned according to the manufacturer and used for all experiments. One microliter of a 1 mg/ml DMSO solution of ADX88178 was added to rat, mouse, or human plasma to reach the final concentration of 1000 ng/ml. Portions (150 μl) of the plasma solution were added to one side of the membrane, and the pH 7.4 phosphate buffer solution was added to the other side of the well simultaneously. Experiments were carried out in duplicate for each species and 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 plasma and buffer compartments at the start of the experiments and after 6 and 7 h, and they were immediately stored at 4°C.
The plasma and buffer samples were subsequently analyzed by a specific liquid chromatography/mass spectrometry method to determine ADX88178 concentrations. The portion of bound ADX88178 in plasma was calculated according to the following equation: percentage of compound bound = (Cplasma − Cbuffer) × 100/Cplasma, where Cplasma is the total (bound + free) concentration of ADX88178 in plasma after equilibrium is reached and Cbuffer is the concentration of ADX88178 measured in the buffer solution at the same time. This equation is valid only when the equilibrium between the plasma and buffer solutions is completed. Previous experiments demonstrated that, for the reference substances, the equilibrium between the plasma and buffer solution is reached after 5 to 6 h.
In Vivo Studies
Animals.
Rat experiments used adult male Sprague-Dawley (SD) rats (260–450 g) supplied by Charles River Laboratories (Les Oncins, France) (commercial name OFA) or Taconic Farms (Germantown, NY). Mouse studies used MitoPark (MP) mice (Ekstrand et al., 2007) originally obtained from Kampavata Inc. (Karolinska Institute, Stockholm, Sweden) and rederived on a C57BL/6(Tac) background (Taconic Farms). MitoPark mice express a homozygous disruption of Tfam selectively in midbrain dopamine neurons [genotype TfamloxP/TfamloxP, DAT-cre/(+)]. Littermate wild-type (WT) mice [Tfam/(+), (+/+)] served as controls. Rats and mice were housed under a 12-h light/dark cycle in constant temperature (22 ± 2°C) and humidity (>45%) with food and water available ad libitum. Animals were acclimated in-house for at least 5 days before use, and experiments were conducted during the light phase between 7:00 AM and 1:00 PM. All experiments were carried out in compliance with American, French, and European legislation governing the protection of vertebrate animals used in scientific research and approved by the Addex Pharmaceuticals Ethical Committee and the Merck Research Laboratories (West Point) Institutional Animal Care and Use Committee.
Haloperidol-Induced Catalepsy in Rats.
Catalepsy was induced by administration of haloperidol (1.5 mg/kg i.p.). Thirty minutes later, rats received ADX88178 (0.1, 0.3, 1, 3, and 10 mg/kg p.o.), vehicle (1% CMC), or l-DOPA (150 mg/kg i.p. in saline, in combination with 50 mg/kg i.p. benserazide) as a reference control. The degree of catalepsy was assessed 60 min after test compound administration by using a grid test. Individual rats were gently placed on a vertical wire grid with their heads pointing toward the ceiling and all paws gripping the grid. Latency to move both forepaws to relocate the body was measured, with a maximum latency cutoff time of 120 s. The mean latency over three trials was recorded.
Forelimb Akinesia in 6-Hydroxydopamine-Lesioned Rats.
Male SD rats (225–300 g) underwent stereotactic 6-hydroxydopamine (6-OHDA) lesions of bilateral striata by using a convection-enhanced delivery approach to minimize mechanical and nonspecific damage to striatal neurons (Oiwa et al., 2003). In brief, under isoflurane anesthesia (1–2% in O2 for maintenance), the rats' heads were placed in a stereotactic frame (David Kopf Instruments, Tujunga, CA), the scalp was exposed by a midline incision, and burr holes were drilled through the skull at anterior-posterior + 0.7 mm; lateral ± 2.8 mm relative to Bregma (Paxinos and Watson, 1997). Fine-tipped cannulae, generated by inserting fine silica tubing into a 27-g needle tip, were inserted through each burr hole to 5 mm below the dura. Twenty microliters of 6-OHDA-hydrogen bromide was delivered over 100 min into each striata via a connected pump (1.48 mg/ml in 2% ascorbic acid/saline). Cannulae were left in place for an additional 5 min before removal, and then the scalp was closed with tissue adhesive. Rats were maintained at 37°C throughout by an isothermic heating pad. During the first week post-6-OHDA-lesion rats were supplemented b.i.d. with 5 ml of lactated Ringer's saline intraperitoneally and 5 ml of goats milk esbilac orally and given access to hydra-gel electrolytes and powdered rat chow. Rats were allowed 3 to 4 weeks recovery postsurgery before behavioral testing, with softened chow and/or subcutaneous injections of saline administered if needed.
Forelimb akinesia was assessed as follows. The rat was held by its torso with its hindquarters and one forelimb lifted above a bench top, so that its body weight was supported by the other forelimb alone. The head and forequarters were oriented forward by the experimenter's thumb and index finger, with the supporting forelimb free to move and elicit spontaneous steps. The number of voluntary steps taken over 30 s was recorded for four trials per forelimb. For each forelimb, the number of steps from four trials were averaged, and these scores combined to give the final stepping score (average for left + right forelimb).
The forelimb akinesia test was performed 1 day before 6-OHDA lesions to measure baseline activity in the rats and again 4 weeks after the 6-OHDA lesions to evaluate the extent of the lesion in terms of the forelimb stepping deficit induced. We set inclusion criteria for study animals at a threshold of <25% baseline stepping ability at 4 weeks postlesion.
To assess the effects of ADX88178 on forelimb akinesia, in the presence or absence of l-DOPA, we selected 14 of 16 lesioned rats that met the stepping criterion (prelesion, steps = 35.5 ± 0.6; 4 weeks postlesion, steps = 3.1 ± 0.52; mean ± S.E.M.). We first examined stepping ability after increasing doses of l-DOPA (6, 20, 60 and 120 mg/kg i.p. in saline plus 15 mg/kg benserazide) alone and in the presence of a fixed dose of ADX88178 (10 mg/kg i.p. in CMC, a dose that had no effect on forelimb akinesia, as determined in a pilot study) (see Fig. 5A). In all akinesia studies ADX88178 was administered 20 min pre-l-DOPA, and the step test commenced 60 min post-ADX88178. Because of the nature of the forelimb akinesia assay, a maximum of 16 animals could be tested per day. With the possibility of high doses of l-DOPA causing sensitization to future doses, we applied a dose escalation paradigm to the study design. In all studies, the evaluator was blinded to dose. After assaying vehicle effects in the rats, study groups were tested in the following order (one group, n = 14, per day, with at least 1 day of washout between successive l-DOPA doses): 10 mg/kg ADX (+ l-DOPA vehicle, saline); 6 mg/kg l-DOPA (+ ADX vehicle, 1% CMC); 6 mg/kg l-DOPA + ADX; 20 mg/kg l-DOPA + ADX; 20 mg/kg l-DOPA; 60 mg/kg l-DOPA; 60 mg/kg l-DOPA + ADX; and 120 mg/kg l-DOPA. Assays of vehicle effects were randomly interspersed to check whether basal activity shifted during the study duration, and no significant changes were detected (data not shown).
At 8 weeks postlesion, we reassessed activity after vehicle administration to the animals to accommodate any plasticity in animals' stepping abilities with lesion age (see V in Fig. 5B). We then commenced a second study evaluating the effects of ADX88178 alone (3, 10, and 30 mg/kg p.o. in 1% CMC) and ADX88178 doses in the presence of a fixed submaximal dose of l-DOPA (6 mg/kg i.p. in saline with 15 mg/kg benserazide) (see Fig. 5B). Groups were tested in the order shown in Fig. 5B to minimize the impact of any potential sensitivity to l-DOPA that might develop. The observation that the stepping response size after 6 mg/kg l-DOPA alone was almost identical in both studies (see Fig. 5), despite the ADX88178 dose escalation study being run after animals had been exposed to l-DOPA doses of up to 120 mg/kg, suggests that animals' stepping responses were not showing sensitization to l-DOPA over time.
In a separate set of animals, we examined whether ADX88178 (30 mg/kg p.o. in 1% CMC) interacted with a submaximal dose of quinpirole (0.1 mg/kg i.p in saline). In this study 16 animals were selected from 20 lesioned rats and divided into two groups of eight balanced for baseline deficits measured at 10 weeks postlesion (<25% of prelesion). On the subsequent test days, compounds were tested as follows: test day 1, group 1, vehicles, and group 2, ADX88178 alone; test day 2, group 1, quinpirole alone, and group 2, quinpirole + ADX88178; with a day for washout in between tests. ADX88178 was administered 30 min prequinpirole, and the step test commenced 60 min post-ADX88178.
l-DOPA-Induced Dyskinesias in Rats.
Male SD rats (250–330 g) with a unilateral 6-OHDA lesion of the medial forebrain bundle were imported from Taconic Farms. Animals were prescreened for lesion by an apomorphine rotation assay, conducted at Taconic Farms. After acclimation in-house, rats underwent daily l-DOPA administration (6 mg/kg l-DOPA plus 15 mg/kg i.p. benserazide q.i.d.) for a minimum of 21 days, to generate l-DOPA-induced adnormal involuntary movements (AIMs) (Cenci et al., 1998). Thereafter, AIMs were assessed by rater observation, blinded to test dose, immediately after administration of the daily dose of l-DOPA. Movement abnormalities were scored in 1-min periods, at 20-min intervals, over 180 min. The rater scored three dyskinesia parameters: axial (dystonic or choreiform torsion of the trunk and neck toward the side contralateral to the lesion), limb (jerky and/or dystonic movements of the forelimb contralateral to the lesion), and orolingual (twitching of orofacial muscles, and bursts of empty masticatory movements with protrusion of the tongue toward the side contralateral to the lesion). Each subtype was scored on a severity scale from 0 to 4, where 0 = absent, 1 = present during less than half of the observation time, 2 = present for more than half of the observation time, 3 = present throughout but suppressible by external stimuli, and 4 = present throughout and not suppressible by external stimuli.
Immediately before compound testing, rats underwent baseline AIMs rating on 3 successive days. Animals were selected based on the inclusion criteria of demonstrating a cumulative AIM score of >9 of a maximum of 12 for at least 1 h on each of the three test days (n = 8). For compound tests, ADX88178 (0.1, 1, and 10 mg/kg p.o. in 1% CMC) was administered 30 min before the daily l-DOPA (6 mg/kg) dose, and AIMs were scored for 3 h beginning immediately after l-DOPA administration. Each dose was tested in a separate crossover design study over 2 days [day 1, group, ADX88178 (n = 4), group 2, vehicle (n = 4); day 2, group 1, vehicle (n = 4), group 2, ADX88178 (n = 4), to achieve n = 8/group)].
MitoPark Mouse Locomotor Activity.
Spontaneous locomotor activity was assessed in 17-week-old symptomatic MP mice and littermate WT mice, using an infrared beam-break mouse cage apparatus (Hamilton-Kinder, Poway, CA) under dark room conditions. The study used 12 MP and 12 WT mice (female), using a crossover design over 3 test days to achieve group sizes of 8 MP mice and 12 WT mice. MP mice received ADX88178 (30 mg/kg p.o.) or vehicle (1% CMC) 10 min before challenge with l-DOPA (15 mg/kg l-DOPA + 3.75 mg/kg i.p. benserazide) or vehicle (saline). Littermate wild-type mice (n = 12) received only vehicle (1% CMC orally and saline intraperitoneally). Mice were placed in the beam-break boxes immediately after administration of l-DOPA or its vehicle and allowed to explore freely. Beam breaks in X, Y, and Z dimensions were recorded over the next 180 min.
Data and Statistical Analysis
For in vitro studies, the concentration-response curves of glutamate or compounds were generated by using Prism software (GraphPad Software Inc., San Diego, CA). The curves were fitted to a four-parameter logistic equation allowing the determination of EC50 and IC50 values: (Y = bottom + (top − bottom)/(1 + 10̂((LogE(I)C50 − X) × Hill slope).
For in vivo studies, catalepsy duration (timed up to a maximum of 120 s) was analyzed by the Kruskal-Wallis test, followed by Dunn's multiple comparison test. Rat forelimb stepping ability was analyzed by one-way ANOVA, with Dunnett's post hoc t test or Bonferroni correction for multiple comparisons. MitoPark mouse locomotor activity changes within subjects were analyzed by using repeated-measures ANOVA and post hoc multiple comparison test, whereas total distances traveled over the assay were compared between groups by one-way ANOVA with Bonferroni correction for multiple analyses. Rat dyskinesia assays were analyzed by repeated-measures ANOVA. Analyses used GraphPad Prism 4.01, Statistica 6.1 (StatSoft, Tulsa, OK), or SPSS 20 (IBM, White Plains, NY). The alpha level chosen was p < 0.05.
Results
In Vitro Properties of ADX88178.
We first evaluated the in vitro properties of ADX88178 by monitoring intracellular calcium mobilization in HEK293 cell lines modified to express human and rat mGluR4 as described under Materials and Methods. The addition of ADX88178 alone had no effect on calcium levels measured with Fluo4. However, the PAM was effective in enhancing calcium responses resulting from the addition of an EC20 concentration of glutamate (Fig. 2A). In the cell lines expressing human mGluR4, ADX88178 potentiated the response to glutamate with an EC50 value of 3.5 ± 0.3 nM and amplified the effects of the EC20 glutamate concentration to 92% of the maximal glutamate response. As shown in Fig. 2A, ADX88178 produced a leftward shift in the glutamate concentration response relationship and increased the apparent affinity for glutamate by approximately 100-fold at 3 μM, the highest concentration tested. In cell lines expressing the rat receptor the compound was slightly less potent at 9.1 ± 1.0 nM, but the responses were enhanced to 195% of the glutamate maximum. We also evaluated the effects of ADX88178 by using a radioligand binding assay with human mGluR4 receptors obtained from membranes from the HEK293 cell line (Fig. 2B). The compound completely displaced the specific binding of [3H]PAM2, a structurally distinct compound, in this assay, with a Ki value of 39 nM. Taken together, these results confirm that the allosteric properties of ADX88178 are mediated by a site distinct from the orthosteric binding site.
The selectivity for mGluR4 was investigated by using cell lines expressing each of the other mGluRs as well as GABAB receptors. Compound activity was investigated to enable the detection of agonists and positive allosteric modulation as well as receptor antagonism. ADX88178 had no effect on mGluR1, mGluR2, mGluR3, mGluR5, mGluR7, or GABAB at concentrations up to 30 μM. Agonist activity at mGluR6 was detected with an EC50 value exceeding 10 μM, whereas the compound acted as a PAM at mGluR8 with an EC50 of 2.2 μM. We further evaluated the selectivity of ADX88178 by screening against 71 receptors, transporters, enzymes, and ion channels in the Cerep panel. The compound was inactive at 10 μM in all assays except human adenosine A1 and human adenosine A3 where 10 μM resulted in 75 and 80% inhibition of binding, respectively. The Ki for inhibition of human adenosine A3 receptor binding was determined to be 2.2 μM. These results demonstrate that ADX88178 is both potent and highly selective for mGluR4.
Pharmacokinetic Profile of ADX88178.
We next evaluated the pharmacokinetic properties of ADX88178 as a prolegomenon to subsequent in vivo studies (Table 1). Intrinsic clearance was determined in rat and human liver microsomes. ADX88178 showed relatively high intrinsic clearance with values of 194 and 282 μl/min/mg protein in rat and human microsomes, respectively. Cytochrome P450 interactions were evaluated in vitro by using recombinant enzymes purified from insect cells. ADX88178 showed limited interactions with cytochrome enzymes. CYP1A2 was inhibited with a half-maximal effect at 0.5 μM. The compound was less potent at CYP3A4 (IC50 = 7 μM), CYP2C9 (14 μM), and 2D6 (>50 μM). Plasma protein binding was limited, and the fraction unbound was determined to be 8, 21, and 11% in rat, mouse, and human plasma proteins, respectively.
The pharmacokinetic profile of ADX88178 was determined in rats after either intravenous administration of 0.3 mg/kg or oral administration of 30 mg/kg in carboxymethylcellulose. Measurements made after intravenous administration revealed a relatively short half-life of 0.2 h and a VDSS of 0.97 l/kg (Table 2) as predicted from the intrinsic clearance. ADX88178 was very effectively absorbed after oral administration in rats (%F > 100) and reached a Cmax in plasma and CSF 1 to 2 h after dosing (Fig. 3, 30 mg/kg p.o.). Plasma exposures at 3, 10 and 30 mg/kg increased dose proportionally. The CSF concentration time course followed the plasma concentration (Fig. 3), and the CSF exposure also increased dose proportionally after oral administration (Table 3). The CSF/plasma percentage value correlates well with the free fraction measured in plasma in independent experiments, showing that ADX88178 freely permeates through the blood-brain barrier to reach the central compartment. Thus, although ADX88178 has a relatively short plasma half-life after intravenous administration, Cmax brain exposures after oral administration are greater than the in vitro EC50 values, and the tight correlation observed between plasma and CSF concentration curves over time, suggest that after dosing 3, 10, and 30 mg/kg p.o. exposures are maintained >EC50 for the duration of the in vivo studies to assess the behavioral effects of mGluR4 activation (up to 3.5 h postdose).
In Vivo Properties of ADX88178.
Several rodent assays assess the potential symptomatic benefit of putative antiparkinsonian compounds. Acute administration of haloperidol induces a catalepsy in rats that can be reversed by several classes of compound that improve the motor symptoms of Parkinson's disease, including dopamine agonists, l-DOPA, and adenosine A2A antagonists. In the present study, administration of haloperidol resulted in catalepsy that lasted approximately 80 to 90 s. ADX88178 produced a dose-dependent reversal of catalepsy, with 10 mg/kg producing a significant reversal that was similar to the effects of l-DOPA (150 mg/kg; Fig. 4). In these studies we measured the compound exposures in plasma at the end of the experiment to determine the in vivo EC50. These data are shown in Tables 4 and 5 and indicate that the median effective dose for ADX88178 in catalepsy reversal is 3 mg/kg, the ED50 is 1.1 mg/kg, and in vivo EC50 in plasma is 92 ng/ml, equivalent to 338 nM. Using the CSF/plasma ratio previously determined in the pharmacokinetics experiments, we extrapolated the concentrations in CSF from the measured plasma concentrations, and then compared them to the rat in vitro EC50 (9.1 nM). At the median effective dose, the extrapolated concentration in CSF was well above the in vitro EC50, and at the in vivo EC50 in plasma the extrapolated CSF level corresponded well with the in vitro EC50 at rat mGluR4, inferring that the effect seen in this test is mGluR4 related.
We next evaluated ADX88178 in rats that had previously received a bilateral lesion of the striatum induced by 6-OHDA (B. Luo, T. Lis, L. M. Hodge, L. Yao, S. Villarreal, A. G. DiLella, and S. E. Browne, unpublished work). The forelimb akinesia assay, which quantifies the ability of rats to take spontaneous steps, was used to assess motor function in these animals. Before lesioning, rats typically took approximately 35 steps in the allotted time. After the lesion and a recovery period, the total number of steps was reduced by >80% (Fig. 5). The effects of the lesion were fairly stable over 3 to 4 months postlesion, with the number of steps measured in untreated animals typically not exceeding 10 at 4 months postlesion (still less than 30% of the prelesion state). As expected, l-DOPA was effective in improving motor performance (Fig. 5A). Dose-response curves performed with l-DOPA showed that 6 mg/kg represented a threshold dose that was the minimum sufficient to reliably produce an increase in the number of steps taken. Administration of ADX88178 (3–30 mg/kg p.o.) alone had no significant effect on the number of steps made in the lesioned animals (Fig. 5B). It is noteworthy that when ADX88178 was coadministered with the threshold dose of l-DOPA we observed a significant improvement in motor performance that showed dose dependence for the 4-PAM (Fig. 5B). Thus, the combination of 6 mg/kg l-DOPA and 30 mg/kg ADX88178 was able to fully reverse the motor deficit. Similar l-DOPA sparing effects were observed when a dose-response curve to l-DOPA was established in combination with a fixed dose of ADX88178 (10 mg/kg; Fig. 5A). Low doses of l-DOPA were more effective in the presence of ADX88178, with full reversal observed at 60 mg/kg l-DOPA, which is less than the 120 mg/kg required in the absence of the 4-PAM. These effects were observed in the absence of any detectable effect of ADX88178 on plasma concentrations of l-DOPA (data not shown), thus ruling out the possibility that the apparent synergy is caused by a peripheral pharmacokinetic interaction with l-DOPA rather than a central pharmacodynamics interaction.
In addition to l-DOPA, dopamine agonists represent a mainstay of therapy for PD, especially in earlier stages of the disease. Dopamine agonists directly stimulate postsynaptic D2 receptors and thus work differently from l-DOPA. To ensure that the effects of mGluR4 activation were not specific to the presynaptic mechanism associated with l-DOPA we examined the interactions between ADX88178 and quinpirole, a dopamine agonist currently in clinical use (Fig. 6). In this experiment, a low dose of quinpirole (0.1 mg/kg) produced a partial reversal of forelimb akinesia. As observed in the previous experiment, ADX88178 alone had no effect on forelimb akinesia. However, we observed a significant enhancement of the quinpirole response when the two compounds were coadministered. This suggests that mGluR4 activation may potentiate multiple therapeutic modalities in PD.
We also evaluated the effects of ADX88178 in MitoPark mice. In contrast to the rat lesion models, these mice provide an in vivo model of a progressive dopaminergic deficit, as the result of a loss of the mitochondrial transcription factor Tfam in dopamine transporter-expressing neurons (Ekstrand et al., 2007). Consequently, MP mice show an age-dependent decrease in ambulatory movements (Ekstrand et al., 2007; S. E. Browne, K. Smith, L. Hodge, S. Villarreal, L. Yao, T. Rosahl, and I. J. Reynolds, unpublished work). In the experiments reported here, mice were 17 weeks old, approximately 5 weeks after the motor deficit was first detected. As shown in Fig. 7, MitoPark mice show a significant deficit in ambulatory movements compared with wild-type animals, as shown by distance traveled in a beam-break box. ADX88178 alone (30 mg/kg) had limited effects in MitoPark mice. However, l-DOPA (15 mg/kg) was effective in reversing the decrease in ambulation. The combination of l-DOPA and ADX88178 produced a significant increase in ambulatory activity over l-DOPA effects alone, consistent with the findings in 6-OHDA-lesioned rats.
The treatment of PD with l-DOPA is limited by motor adverse effects, with l-DOPA-induced dyskinesias (LIDs) being a particular problem. Although the enhancement of the effects of l-DOPA might be considered therapeutically useful, this would be of limited clinical utility if the dopamine-related adverse effects were increased, whereas an agent that reduced or delayed the onset of LIDs would have greater benefit. We examined whether ADX88178 treatment modulated LIDs in rats, by monitoring AIMs induced by repeated administration of l-DOPA to animals with 6-OHDA lesions of the medial forebrain bundle (Cenci et al., 1998). In such primed animals the administration of l-DOPA results in AIMs that persist for the duration of the l-DOPA treatment. We evaluated the effects of ADX88178 by dosing primed animals with both l-DOPA (6 mg/kg) and ADX88178 at 0.1, 1, and 10 mg/kg and found no evidence for modulation of AIMs in these animals compared with animals that received l-DOPA alone (Fig. 8).
Discussion
In this study we have characterized the properties of a novel mGluR4 PAM. ADX88178 is a potent allosteric modulator of mGluR4 that shows excellent selectivity against other metabotropic glutamate receptors. This compound is orally available and brain penetrant. Although it is cleared relatively quickly, oral dosing produces exposures of sufficient duration to evaluate its properties in vivo. Efficacy of ADX88178 was seen in a number of rodent models of Parkinson's disease, although in dopamine depletion models (rat bilateral 6-OHDA lesion and MitoPark mice), efficacy was contingent on the coadministration of a low dose of l-DOPA. These findings that coadministration of the 4-PAM with l-DOPA results in robust efficacy in reversing movement deficits are consistent with an l-DOPA sparing action that may prove to be therapeutically useful for the management of the symptoms of Parkinson's disease.
There is interest in developing 4-PAMs for a number of potential therapeutic applications (Lavreysen and Dautzenberg, 2008; Duty, 2010; Niswender and Conn, 2010; Célanire and Campo, 2012). The demonstration that PHCCC acted as a 4-PAM was important because it provided an early tool to manipulate mGluR4 receptors both in vivo and in vitro (Maj et al., 2003; Marino et al., 2003). However, PHCCC is also an mGluR1 antagonist, and its physicochemical and pharmacokinetic properties limit its usefulness (Williams et al., 2010). More recently, a series of papers have described useful advances based on several different chemical scaffolds. This includes molecules with improved selectivity for mGluR4 (Niswender et al., 2008b; Williams et al., 2010) and improved central nervous system penetration (Williams et al., 2009, 2010; East and Gerlach, 2010). This approach has also recently generated mGluR4 receptor antagonists (Utley et al., 2011). A very recent study reported the effects of a systemically active mGluR4 PAM, N-(3-chlorophenyl)picolinamide (VU0364770), in rat models of PD (Jones et al., 2012). This compound was active after subcutaneous administration and, like ADX88178, reversed haldol-induced catalepsy and improved motor performance in 6-OHDA-lesioned animals. VU0364770 is (or seems) less potent than ADX88178 (EC50 values of 290 versus 3.5 nM against human mGluR4, respectively). VU0364770 is also a reasonably potent inhibitor of MAO-B (720 nM; Jones et al., 2012). Although MAO-B inhibition would not interfere with the treatment of PD in patients, this property makes it considerably more difficult to interpret preclinical antiparkinsonian effects of VU0364770 in terms of activity at mGluR4. It is noteworthy that there is monotherapy activity described for VU0364770 in 6-OHDA-lesioned rats, which is notably devoid in ADX88178. It will be interesting to determine whether monotherapeutic activity (i.e., 4-PAM alone without l-DOPA) is an intrinsic property of a subset of 4-PAMs, whether this reflects some amount of orthosteric agonist activity, or whether it is a consequence of both 4-PAM and MAO-B inhibition activity. The 4-PAM mechanism may be broadly useful for enhancing the efficacy of symptomatic PD treatments, because the data obtained with both ADX88178 and VU0364770 demonstrate their abilities to potentiate the effects of l-DOPA, dopamine agonists, adenosine A2A antagonists, and potentially MAO-B inhibitors. Nevertheless, ADX88178 is notable in that it seems to be the most potent of the molecules reported to date and also has a greater window of selectivity against mGluRs, as well as other non-G protein-coupled receptor targets.
Combined with the opportunity to achieve peak CSF exposures in vivo of more than 100-fold greater than its EC50 in vitro while maintaining good pharmacological selectivity, this 4-PAM should prove to be a valuable tool for exploring the biological role of mGluR4.
There have been a number of proposed uses for mGluR4 PAMs. The location of the mGluR4 receptor presynaptically in several key locations in the basal ganglia (Conn et al., 2005) suggested a potential utility in treating the motor dysfunction in PD (Marino et al., 2003; Valenti et al., 2003, 2005; Niswender et al., 2008a). PHCCC also protects mice from the neurotoxic effects of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine, which raises the interesting possibility of neuroprotection as well as symptomatic benefit in PD (Battaglia et al., 2006). The potential therapeutic utility may extend beyond PD. Neuroprotection with PHCCC has been reported in cerebral ischemia models in mice (Moyanova et al., 2011). Studies with PHCCC have suggested an anxiolytic effect of 4-PAMs (Stachowicz et al., 2004, 2006), an effect that we have also observed with ADX88178 (B. Campo, S. E. Browne, F. Girard, F. Hess, J. Uslaner, E. Le Poul, and I. J. Reynolds, unpublished work). Iacovelli et al. (2006) reported that PHCCC could inhibit the growth of medulloblastomas. In addition, mGluR4 is found on human dendritic cells in the immune system, and mice lacking mGluR4 show an enhanced susceptibility to experimental autoimmune encephalomyelitis, a model of multiple sclerosis (Fallarino et al., 2010). Also outside the central nervous system, Uehara et al. (2004) described the presence of mGluR4 on α cells in the pancreas and demonstrated regulation of glucagon release that suggests a role for glucose control in diabetes. Several other reports have suggested a role for mGluR4 in affective disorders, epilepsy, and pain (Lavreysen and Dautzenberg, 2008; Célanire and Campo, 2012), and it will be interesting to determine which, if any, of these potential therapeutic effects are still evident with a more selective molecule such as ADX88178.
Perhaps the most surprising observation in the present study is the finding that the effects on forelimb akinesia require the coadministration of l-DOPA. Although the effects of mGluR4 activation have been evaluated at several different synapses that may be relevant for this behavioral effect, including in the substantia nigra and the striatopallidal synapse (Valenti et al., 2003, 2005), the key site of action of 4-PAMs in regulating motor function is not clear. It may be appropriate to infer that the key target in these experiments is downstream of D2 receptor activation after l-DOPA administration. However, this hypothesis remains to be critically evaluated.
Authorship Contributions
Participated in research design: Le Poul, Boléa, Girard, Poli, Charvin, Campo, Bortoli, Bessif, Luo, DiLella, Liverton, Hess, Browne, and Reynolds.
Conducted experiments: Girard, Charvin, Bortoli, Bessif, Luo, Koser, Hodge, and Smith.
Contributed new reagents or analytic tools: Boléa.
Performed data analysis: Girard, Poli, Charvin, Campo, Bortoli, Bessif, Luo, Koser, Hodge, Smith, DiLella, and Browne.
Wrote or contributed to the writing of the manuscript: Le Poul, Girard, Poli, Campo, Koser, Hodge, Smith, Liverton, Hess, Browne, and Reynolds.
Footnotes
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
ABBREVIATIONS:
- PD
- Parkinson's disease
- ADX88178
- 5-methyl-N-(4-methylpyrimidin-2-yl)-4-(1H-pyrazol-4-yl)thiazol-2-amine
- AIM
- abnormal involuntary movement
- ANOVA
- analysis of variance
- BLQ
- below level of quantitation
- CMC
- carboxymethylcellulose
- CSF
- cerebrospinal fluid
- DMSO
- dimethyl sulfoxide
- GABA
- γ-aminobutyric acid
- HEK
- human embryonic kidney
- LD
- l-DOPA
- LID
- l-DOPA-induced dyskinesia
- MAO-B
- monoamine oxidase-B
- mGluR
- metabotropic glutamate receptor
- MP
- MitoPark
- 6-OHDA
- 6-hydroxydopamine
- PAM
- positive allosteric modulator
- PHCCC
- (−)-N-phenyl-7-(hydroxyimino)cyclopropa[b]chromen-1a-carboxamide
- SD
- Sprague-Dawley
- V
- vehicle
- VU0364770
- N-(3-chlorophenyl)picolinamide
- WT
- wild type.
- Received May 1, 2012.
- Accepted July 9, 2012.
- Copyright © 2012 by The American Society for Pharmacology and Experimental Therapeutics