|Home | About | Journals | Submit | Contact Us | Français|
Essential tremor is the most common movement disorder and has an unknown etiology. Here we report that γ-aminobutyric acidA (GABAA) receptor α1–/– mice exhibit postural and kinetic tremor and motor incoordination that is characteristic of essential tremor disease. We tested mice with essential-like tremor using current drug therapies that alleviate symptoms in essential tremor patients (primidone, propranolol, and gabapentin) and several candidates hypothesized to reduce tremor, including ethanol; the noncompetitive N-methyl-D-aspartate receptor antagonist MK-801; the adenosine A1 receptor agonist 2-chloro-N6-cyclopentyladenosine (CCPA); the GABAA receptor modulators diazepam, allopregnanolone, and Ro15-4513; and the L-type Ca2+ channel antagonist nitrendipine. Primidone, propranolol, and gabapentin reduced the amplitude (power) of the pathologic tremor. Nonsedative doses of ethanol eliminated tremor in mice. Diazepam, allopregnanolone, Ro15-4513, and nitrendipine had no effect or enhanced tremor, whereas MK-801 and CCPA reduced tremor. To understand the etiology of tremor in these mice, we studied the electrophysiological properties of cerebellar Purkinje cells. Cerebellar Purkinje cells in GABAA receptor α1–/– mice exhibited a profound loss of all responses to synaptic or exogenous GABA, but no differences in abundance, gross morphology, or spontaneous synaptic activity were observed. This genetic animal model elucidates a mechanism of GABAergic dysfunction in the major motor pathway and potential targets for pharmacotherapy of essential tremor.
As the most common movement disorder, with higher prevalence than Parkinson disease, essential tremor can be both socially and physically debilitating, rendering patients unable to write, clothe themselves, or feed themselves (1, 2). In contrast to resting tremor found in Parkinson disease, essential tremor is characterized by postural and kinetic components (3). Postural tremor is an involuntary oscillation that occurs when a posture is maintained against gravity. Essential tremor worsens during movement, adding a kinetic component to the disorder that results in disturbances in tandem gait and ataxia in approximately 50% of patients (4). First-line treatments for essential tremor include the anticonvulsant primidone and the β-adrenergic blocker propranolol, although their mechanisms of action are unknown (5). Like primidone, gabapentin is an anticonvulsant serendipitously found to be effective in the treatment of essential tremor (6, 7). The oldest treatment for essential tremor may be ethanol, which temporarily ameliorates tremor; however, there are serious drawbacks to chronic use of ethanol for tremor management (3).
The etiology of essential tremor has not been elucidated (2). A major obstacle in the search for a mechanism is the lack of an adequate animal model for essential tremor. Current tremor models utilize drug- or lesion-induced tremors or arise from mutant strains. The only model sharing some features of essential tremor uses the γ-aminobutyric acidA (GABAA) receptor inverse agonist harmaline to induce a temporary tremor in animals (8) through its action at the inferior olive. However, animals develop rapid tolerance to harmaline, enhanced physiological tremor, and locomotor deficits due to Purkinje cell degeneration and do not respond to propranolol (9–11). Although the harmaline model of tremor has made a valuable contribution to the understanding of human essential tremor, inconsistencies with essential tremor demonstrate the need for additional models.
Several lines of evidence suggest that the GABAergic system is involved in the etiology of essential tremor because this major inhibitory system has a central role in motor control pathways. GABAA receptors are the primary inhibitory receptors in brain that regulate motor function (12–14). Tremor induction by harmaline has been attributed to inhibition of GABAA receptors resulting in enhanced electrical coupling of cerebellar afferents in the inferior olive (15). Patients with essential tremor have reduced cerebrospinal fluid concentrations of GABA (16), and thalamic microinjection of the GABA agonist muscimol results in tremor arrest in patients (17). In addition, ethanol is particularly effective for treating essential tremor and has GABAergic effects that are believed to be related to its pharmacologic action (18–20).
Deletion of GABAA receptor α1 subunits results in the loss of 50% of all GABAA receptors in brain (21), including the motor pathways in brainstem, cerebellum, thalamus, and basal ganglia. Mice carrying this deletion exhibit an essential-like tremor and motor incoordination that mimics human essential tremor. To test the relevance of this model and to explore potential mechanisms underlying essential tremor, we examined the properties of the tremor and responses to ethanol and drugs currently used to treat essential tremor. Purkinje cell morphology and function were examined in light of their involvement in the etiology of essential tremor. The results indicate that GABAA receptor α1–/– mice represent a useful genetic animal model of essential tremor that predicts new therapeutic targets and a potential etiology of the disease.
GABAA receptor α1–/– mice develop normally but exhibit a tremor as they move freely about their cages. The tremor was observed in the first litters tested at the University of North Carolina at Chapel Hill (21) and has persisted through 9 generations of heterozygote breeding over more than 3 years. The tremor is most evident in the limbs, becomes significantly more pronounced upon tail suspension or movement, and is absent when the mouse is relaxed. The presence of postural and kinetic tremor is characteristic of essential tremor disorder. To determine the tremor frequency and amplitude (power) in GABAA receptor α1–/– mice, the animals were suspended by the tail from a force transducer and the displacement of the whole animal was measured over time. Physiologic tremor was observed in wild-type (α1+/+) and heterozygous (α1+/–) mice with mean frequencies between 32.2 and 35.0 hertz (Hz) and maximal power ranging between 2.6 × 1011 and 6.00 × 1011 N (Figure (Figure1,1, A and B). Homozygous (α1–/–) mice exhibited a pronounced pathologic tremor at a lower mean frequency of 19.3 ± 0.9 Hz and a greater mean maximal power of 24.3 × 1011 ± 4.2 × 1011 N and a physiologic tremor similar to that of α1+/+ and α1+/– mice (Figure (Figure1C).1C). In addition, motor coordination was determined by performance on the accelerating rotorod. Compared with α1+/+ mice, α1–/– mice demonstrated significant impairment in their ability to remain on the rotating rod (Table (Table1).1). Deficits in motor coordination of the forearms and tandem gait are commonly observed in patients with essential tremor (4, 22, 23).
Patients with essential tremor respond positively to a small number of drugs, including primidone, propranolol, and gabapentin (3). The efficacy of these drugs was investigated to determine whether α1–/– mice respond like patients with essential tremor. Primidone, gabapentin, and propranolol were effective in α1–/– mice, significantly reducing tremor by 45–70% (Table (Table2).2). The effect of propranolol is similar to that in essential tremor patients, where 50–60% reductions in tremor amplitude are observed (24).
Although rarely prescribed to patients with essential tremor, alcohol (ethanol) markedly reduces tremor amplitude through action on a central component (20). Ethanol completely blocked the pathologic tremor in α1–/– mice in a dose-dependent manner (Figure (Figure2A).2A). Ethanol was remarkably potent with an apparent IC50 of 0.35 g/kg ethanol, a nonsedating dose in mice (24). Ethanol had no effect overall on the frequency or power of physiologic tremor in α1+/+ mice, although an activating effect was observed at the lowest dose.
Ethanol acts on several neurotransmitter systems that could contribute to reduction of tremor, including inhibition of N-methyl-D-aspartate (NMDA) receptors (25), potentiation of GABAA receptor–mediated neurotransmission (18), and enhancement of adenosinergic transmission (26). The NMDA receptor antagonist MK-801 significantly reduced the power of the pathologic tremor in a dose-dependent manner, suggesting that ethanol may exert its effects by inhibiting the NMDA subtype of glutamate receptors (Figure (Figure2B).2B). MK-801 exhibited an IC50 of 0.024 mg/kg and a maximal effect of 65–80% inhibition. The adenosine agonist 2-chloro-N6-cyclopentyladenosine (CCPA) reduced tremor severity by approximately 60% in α1–/– mice (Table (Table2),2), although significant motor incoordination was observed. Therefore, the contribution of the peripheral effect of CCPA on reduction of the tremor is currently unknown. Potentiation of GABAergic neurotransmission by the benzodiazepine diazepam and the neuroactive steroid allopregnanolone exacerbated the power of the pathologic tremor in α1–/– mice, whereas the GABAA receptor inverse agonist Ro15-4513 had no effect (Table (Table2).2). Because primidone, gabapentin, and ethanol block voltage-gated Ca2+ channels, the L-type Ca2+ channel antagonist nitrendipine was tested at a dose that blocks ethanol withdrawal seizures; however, it did not have a measurable effect on tremor amplitude (Table (Table22).
A functional disturbance of the olivocerebellar circuit is postulated to mediate human essential tremor. To understand the etiology of the tremor and motor incoordination in α1–/– mice, we focused on cerebellar Purkinje cells that integrate excitatory input from the inferior olive and cerebellar granule cells and inhibitory input from cerebellar interneurons to control the primary output of the cerebellar cortex, playing a central role in motor coordination. Because Purkinje cells primarily express the α1 subunit–containing subtype (type 1) of GABAA receptors (27) and Purkinje cell degeneration produces tremor, we first determined whether deletion of GABAA receptor α1 subunits caused a loss of Purkinje cells. No difference in the number (46.0 ± 2.7 cells/mm3 in α1+/+ mice) or morphology of calbindin-stained Purkinje cells was observed between genotypes at 4, 8, or 16 months of age, demonstrating that the tremor is not caused by Purkinje cell degeneration (Figure (Figure3A).3A). Furthermore, the spontaneous firing rate of Purkinje cells in α1+/+ and α1–/– mice did not differ (Figure (Figure33B).
Spontaneous GABAergic inhibitory postsynaptic potentials were absent in Purkinje cells recorded from dissociated neurons (Figure (Figure3C).3C). Similar results were found in cerebellar slices (data not shown). In addition, GABA-gated currents in dissociated Purkinje cells were undetectable at concentrations up to 100 μM GABA in α1–/– mice compared with a concentration-dependent response observed in α1+/+ mice (Figure (Figure3,3, D and E).
The principal finding of this study is that deletion of the GABAA receptor α1 subunit results in a mouse with a persistent postural and kinetic tremor, characteristic of human essential tremor disease. Tremor was observed in the first generation of GABAA receptor α1–/– mice (21) and has persisted through more than 9 generations of mice. Our first report on these mice (28) failed to recognize the phenotype, probably owing to the focus on electrophysiological analysis rather than behavior. The tremor is easily observed in all 3 laboratories that currently breed the mice. Current drug therapy for essential tremor was efficacious in reducing tremor in α1–/– mice. Furthermore, the tremor in α1–/– mice responded to ethanol, MK-801, and CCPA, suggesting that glutamatergic and adenosinergic mechanisms are involved in the inhibitory effects of ethanol and are potential targets for novel pharmacotherapy. GABAA α1–/– mice exhibit a complete loss of both endogenous and exogenous GABA inhibition in cerebellar Purkinje cells. In addition, these mice exhibit the loss of 50% of all GABAA receptor binding sites throughout the brain, including brain regions in the major motor pathways from the brainstem to the thalamus and the basal ganglia (21). It is likely that loss of GABA inhibition in these motor pathways underlies tremor in α1–/– mice.
There are many similarities between tremor in GABAA receptor α1–/– mice and human essential tremor disease (Table (Table3).3). Deletion of the GABAA receptor α1 subunit produces a pathologic tremor with postural and kinetic components similar to essential tremor disorder. Furthermore, the tremor is genetic and persistent when compared with existing models that produce a chemical-induced tremor that is short-lived. Human physiologic and essential tremor occur at frequency ranges of 8–12 Hz and 4–8 Hz, respectively (29). The higher frequency of both physiologic and pathologic tremor in rodents is probably due to the smaller size of the animal. The appearance of a pathologic tremor at 19 Hz, relative to the physiologic tremor frequency of 25–40 Hz in α1–/– mice, is indicative of essential tremor rather than enhanced physiologic tremor.
The efficacy of drugs used in the treatment of human essential tremor was observed in α1–/– mice, lending further support to the model. The ability of the first-line treatments, primidone and propranolol, to alleviate essential tremor in α1–/– mice supports common mechanisms with human essential tremor. Furthermore, successful novel application of drugs such as the anticonvulsant gabapentin in both α1–/– mice and patients demonstrates the necessity for animal models in which new and existing drugs can be screened for clinical utility (7). GABAA receptor α1–/– mice provide a useful animal model with which to investigate the pathophysiology of essential-like tremor and the mechanisms by which current and future therapies exert their effects.
Ethanol is effective in the suppression of essential tremor symptoms in humans and α1–/– mice. Ethanol was the most effective treatment of tremor in α1–/– mice at doses that preclude the contribution of its sedative effects on tremor (30). Ethanol is believed to act centrally in the alleviation of tremor and, in human studies, has been shown to involve a central component (20). Furthermore, positron emission tomography studies have shown that ethanol consumption increases cerebral blood flow in the inferior olivary nucleus, a central oscillator, which coincides with alleviation of tremor (31).
The mechanism responsible for ethanol’s effects on tremor appears to involve inhibition of excitatory glutamatergic transmission but not activation of inhibitory GABAergic receptors. The efficacy of the glutamatergic antagonist MK-801 but not of GABAergic agonists diazepam and allopregnanolone suggest that the tremor may result from inactivation of inhibitory neurons. Although the lack of tremor inhibition by diazepam in α1–/– mice is consistent with low responsiveness in patients with essential tremor (20), the augmentation of tremor by diazepam was unexpected. Ethanol has several other targets in the brain that could account for its effects. Ethanol-sensitive adenosine receptors may be involved; however, studies with the adenosine agonist CCPA did not eliminate the possibility for peripheral effects of the drug, because significant motor incoordination was also observed in the mice.
Disruption of Purkinje cell function following deletion of GABAA receptor α1 subunits likely contributes, in part, to the etiology of the tremor in α1–/– mice. The loss of GABAergic inhibition on Purkinje cells suggests that α1 subunit expression is critical for normal function of motor circuits and likely contributes to tremor and motor incoordination observed in α1–/– mice. Tremor induced by alcohol withdrawal is clinically similar to essential tremor and also results in decreased α1 subunit expression in rat cerebellum (32, 33). The lack of change in spontaneous activity of Purkinje cells is consistent with the lack of resting tremor and implies some adaptation to the loss of α1 subunits in these mice. Although compensatory changes in other GABAA receptor subunits have been observed in some brain regions of α1–/– mice (21), no GABAergic compensation in cerebellar Purkinje cells was found (data not shown). Because α1 subunits are normally expressed in other cells of this motor pathway, including the deep cerebellar nuclei and thalamus (but not the inferior olive), it is likely that communication among these regions is also disturbed by α1 subunit deletion and subsequent adaptations in receptor subunit expression. Although we have described altered GABA responses in the cerebellum, impairment of GABAergic inhibition is likely throughout the motor pathways that could contribute to the observed tremor. Additional studies will be required to determine if loss of α1 subunits in any single brain area is sufficient to mimic the clinical manifestations of human essential tremor disease. Furthermore, a common target of the major drugs used to treat essential tremor (primidone and propranolol) has not been elucidated. Therefore, treatment of essential tremor may target several pathways in the brain.
The similarities between human essential tremor and pathologic tremor in α1–/– mice suggest that expression of α1 subunits in the brains of patients with essential tremor may be abnormal. Further studies are needed to test this possibility. The GABAA receptor α1–/– mouse model of essential tremor provides a behavioral paradigm with predictive validity for testing novel potential treatments for this prevalent disorder.
Production of GABAA receptor α1+/+, α1+/–, and α1–/– mice was previously described (28). Wild-type controls were homozygous for a floxed α1 allele that was functionally equivalent to the wild-type allele. GABAA receptor α1–/– mice were homozygous for a cre-mediated global deletion of exon 8 of the α1 gene. All mice were of a mixed C57BL/6J, 129Sv/SvJ, and FVB/N genetic background of F8+ generations and genotyped by Southern blot analysis. Mice were group-housed, provided ad libitum access to food and water, and maintained under a 12-hour light/12-hour dark schedule with lights on at 0700. All animal experiments were conducted in accordance with the Guide for the Care and Use of Laboratory Animals by the NIH and were approved by the Institutional Animal Care and Use Committees at the University of North Carolina and the University of Memphis.
Vehicles were saline for ethanol (AAEPER Alcohol), MK-801 (Sigma-Aldrich), CCPA (Sigma-Aldrich), primidone (Sigma-Aldrich), propranolol (MP Biochemicals), nitrendipine (Axxora), and gabapentin (Sigma-Aldrich); 20% 2-hydroxypropyl-β-cylcodextrin (Sigma-Aldrich)/saline for allopregnanolone (provided by R. Purdy, The Scripps Research Institute, La Jolla, California, USA); and 1% Tween-80 (MP Biochemicals)/saline for diazepam (Biomol) and Ro-154513 (Sigma-Aldrich).
Tremor was quantified by determination of amplitude and frequency using a tremor monitor built in the Bowles Center for Alcohol Studies at the University of North Carolina. Tremor was measured in mice suspended by the tail for 22 seconds so that the animals held a posture against gravity and the tremor was easily observed and measured. The mice exhibited no signs of stress (defecation, vocalization, biting, or jumping) during the measurement as the mice were accustomed to handling by tail suspension in the laboratory. Each mouse was suspended by its tail from a cord attached to the center of a stereo speaker (Archer 3” 8Ωmax2W). Tremor resulted in vibration of the speaker surface and the resulting signal was amplified before passing through an analog-to-digital converter that assigned voltage equivalents to the vibrations transmitted to the speaker. These voltage measurements were analyzed via a Pascal program developed in house. Voltage recordings were made over a 22-second period at a rate of 454 recordings per second. The data obtained from the program was analyzed using MatLab (Mathworks) by Fourier transformation to determine the peak amplitude and frequency of the tremor for each animal.
For all drugs tested (except ethanol), each animal served as its own control in which a baseline measurement of tremor was obtained following administration of vehicle. Dose-response curves for MK-801 were generated by administration of each dose to a group of animals in a χ2 design. Animals were tested at 3-day intervals to allow clearance of drug. Each animal was used to test 1 drug. For ethanol, every dose was tested in a separate group of mice. Drugs and vehicles were administered intraperitoneally in a 10 ml/kg injection volume 1 hour (45 minutes for CCPA) before measurement of tremor.
Mice were tested on an accelerating rotorod (Columbus Instruments) with a rod diameter of 3 inches that accelerated from 5 to 45 rpm over 2 minutes. Mice were placed on the rotating rod (5 rpm) and each trial lasted until the mouse fell off the rod. Each mouse was given 10 trials on the rotorod. For each trial, the length of time the mouse stayed on the rotorod was recorded. The mean was calculated from the last 3 trials for each mouse.
Mice were sacrificed at 4, 8, and 16 months of age, perfused, and processed as previously described (34), with cerebella serially sectioned (10 μm) in the sagittal plane. Purkinje cells were labeled using an anti-calbindin rabbit polyclonal antibody (Chemicon). Calbindin-positive cells were profile-counted throughout the entire cerebellum. The area of each cerebellum tissue section was determined using the Bioquant system and data expressed as the number of Purkinje cells per mm2.
Action potentials from cerebellar Purkinje cells in a slice preparation were recorded extracellularly using an Axon Instruments Axopatch-200A in track mode. Cerebellar slices (250 μm) from 20- to 40-day-old mice were cut in cold artificial cerebrospinal fluid (aCSF) (124 mM NaCl, 3.25 mM KCl, 1.25 mM KH2PO4, 1 mM CaCl2, 20 mM NaHCO3, 2 mM MgSO4, and 10 mM glucose). Slices were maintained with continuous 95% O2 and 5% CO2 for at least 1 hour before use. Slices were transferred to a recording chamber under an upright microscope where they were continuously bathed in aCSF. Purkinje cells were identified using infrared illumination and differential interference contrast optics. Recording pipettes were fabricated from N51A capillary glass (Drummond Scientific) and filled with HEPES-buffered aCSF (described below). The electrode was apposed to a Purkinje cell, and if no spontaneous activity was observed, it was lowered through the Purkinje layer for 100 μm or until an active neuron was found. Probability of finding a neuron during the 100-μm penetration and firing rate of each neuron were recorded.
Spontaneous GABA-gated miniature inhibitory postsynaptic currents (mIPSCs) and currents generated by exogenously applied GABA were measured in the slice preparation and in mechanically dissociated cerebellar Purkinje cells using standard whole-cell voltage-clamp recording. For dissociation, slices (as described above) were transferred into standard HEPES-buffered recording medium (145 mM NaCl, 5 mM KCl, 10 mM HEPES, 2 mM CaCl2, and 10 mM glucose, pH 7.4; 340 milliosmoles per liter) with room air, and a vibrating probe (60 Hz; 0.5–1 mm amplitude) was touched to the surface of the slice for 30–60 seconds. The slice was removed and the dissociated neurons allowed to settle onto the floor of the recording chamber. This procedure dissociates Purkinje cells with intact presynaptic terminals, allowing recording of mIPSCs.
Voltage-clamp (–60 mV) recording in the whole-cell configuration was accomplished using an Axopatch-1D amplifier. Recording pipettes were fabricated from N51A capillary glass (Drummond Scientific). The internal solution used for measuring mIPSCs and ion currents induced by GABA included 150 mM KCl, 3.1 mM MgCl2, 15 mM HEPES, 2 mM K-ATP, 5 mM EGTA, 15 mM phosphocreatine, 500 nM tetrodotoxin, and 50 U/ml creatine phosphokinase. This solution was adjusted to pH 7.4 with KOH and osmolality of 310 milliosmoles per liter with sucrose. A U-tube was used to apply varying concentrations of GABA to neurons. Counts of action potentials and mIPSCs were detected off-line using Mini Analysis software (Synaptosoft).
Data were analyzed using Prism (GraphPad). All data were presented as mean (or mean percent control) ± SEM and subjected to Student’s t test or ANOVA with an appropriate post hoc test. Dose responses were analyzed by nonlinear regression. The IC50 was defined as the drug dose that inhibited 50% of the pathologic tremor and maximal effect as the best-fit bottom value estimated by nonlinear regression.
This research was supported by NIH grants AA11605 (to A.L. Morrow and G.R. Breese), AA09013 (to A.L. Morrow), MH61971 (to D.B. Matthews and K. Hamre), AA10422, and GM52035 (to G.E. Homanics). We thank Jean-Marc Fritschy and Kirk Wilhelmson for discussion of the manuscript.
See the related Commentary beginning on page 584.
Jason E. Kralic’s present address is: Institute of Pharmacology and Toxicology, University of Zurich, Zurich, Switzerland.
Jason E. Kralic, Hugh E. Criswell, Jessica L. Osterman, and A. Leslie Morrow contributed equally to this work.
Nonstandard abbreviations used: aCSF, artificial cerebral spinal fluid; CCPA, 2-chloro-N6-cyclopentyladenosine; GABA, γ-aminobutyric acid; Hz, hertz; mIPSC, miniature inhibitory postsynaptic current; NMDA, N-methyl-d-aspartate.
Conflict of interest: The authors have declared that no conflict of interest exists.