Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The behavioral effects of ⁹-tetrahydrocannabinol (THC) and other cannabinoids have been studied extensively during the past four decades, and include pronounced motor deficits such as catalepsy, hypolocomotion, and static ataxia (Dewey, 1986; Chaperon and Thiebot, 1999). Both the basal ganglia and the cerebellum have long been acknowledged as probable sites of action for these cannabinoid-induced motor deficits (Gough and Olley, 1977; Dewey, 1986). It has been proposed that cannabinoid-induced catalepsy and other manifestations of extrapyramidal deficits are due primarily to inhibition of nigrostriatal dopaminergic transmission (Sanudo-Pena et al., 1999). There is less known about the contribution of cannabinoid effects on the cerebellum to motor deficits seen after their administration. In this study, we used behavioral tests specific for cerebellar function to characterize cannabinoid-induced cerebellar deficits in mice.

Physiological symptoms of cerebellar dysfunction include ataxia, side-to-side truncal tremor, and an abnormal wide stance base in the gait. To evaluate the ability of various cannabinoids to produce cerebellar dysfunction in mice, we used two specific physiological measures: gait analysis and the bar cross test. Gait analysis includes quantification of various parameters of gait such as step width and length, alternation coefficient, and linear movement. Alterations of these parameters, especially gait width, are hallmarks of cerebellar dysfunction (Gilman et al., 1981). In the bar cross test, the mouse crosses a narrow bar with several obstacles. Crossing the bar without falling requires smooth, accurate, and coordinated movement of several muscle groups; an animal with cerebellar ataxia will slip or fall multiple times while completing the task. Since both of these tests are analyses of animal movement, it is unlikely that the extrapyramidal effects of cannabinoids (in which difficulty in initiating rather than accurately completing movements is the predominant symptom) will report false positive results.

The cerebellum influences descending motors systems by evaluating disparities between intended and actual movement. The Purkinje cell of the cerebellum receives excitatory inputs from the inferior olivary nucleus via climbing fibers, and cortical and peripheral sensory information via mossy fibers. Mossy fiber input is disynaptic; mossy fibers synapse on granule cells whose axon terminals (parallel fibers) activate Purkinje cell dendrites in the molecular layer of the cerebellar cortex. A third major input to the Purkinje cell is the basket cell, an inhibitory GABAergic interneuron. Basket cell terminals synapse with several Purkinje cells, forming a peri-cellular basket surrounding the axon hillocks of these cells. Purkinje cells send inhibitory projections to the deep cerebellar nuclei, which in turn project to postural and motor structures such as the vestibular nucleus and motor cortex.

It has long been hypothesized that symptoms of cerebellar dysfunction arise when the excitatory output of the deep cerebellar nuclei is reduced or interrupted. For example, lesions of the deep cerebellar nuclei result in symptoms of cerebellar dysfunction in primates (Poirier et al., 1974). In addition, abnormal Purkinje cell activation, which results in inhibition of the deep cerebellar nuclei, induces symptoms of cerebellar dysfunction. For example, indole alkaloids such as harmaline and ibogaine, increase Purkinje cell activation subsequent to an increase in synchronous firing of the inferior olive and produce profound symptoms of cerebellar dysfunction (O'Hearn and Molivar, 1997). We hypothesize that cannabinoids produce cerebellar deficits by a similar mechanism, i.e., cannabinoids induce abnormal Purkinje cell activation and subsequent inhibition of the deep cerebellar nuclei and, therefore disruption of normal posture and movement.

Recent anatomical and functional studies have identified two populations of CB₁ receptors within the rodent cerebellum. Immunohistochemistry has revealed a high density of CB₁ receptors on the axon terminals of GABAergic basket cells (Tsou et al., 1998). Activation of this CB₁ receptor population in cerebellar slice preparations results in inhibition of GABA release at the basket cell-Purkinje cell synapse (Takahashi and Linden, 2000). High CB₁ receptor density has also been demonstrated in the molecular layer of the cerebellum (Herkenham et al., 1991; Tsou et al., 1998). It is likely that the CB₁ receptor of the cerebellar molecular layer is present on granule cell axons (Herkenham et al., 1991). Electrophysiological experiments have demonstrated that cannabinoids also inhibit transmitter release at the parallel fiber-Purkinje cell synapse (Levenes et al., 1998; Takahashi and Linden, 2000). These data suggest that cannabinoids may alter cerebellar function via inhibition of synaptic transmission at either the parallel fiber-Purkinje cell or basket cell-Purkinje cell synapse, or both. It is the purpose of this study to quantify the effects of various cannabinoids on cerebellar function in mice using standard behavioral tests and to determine the most likely cellular mechanism for these effects.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals and Drugs. Male ICR albino mice (Harlan, Madison, WI) weighing between 21 and 24 g were used for all experiments. Animals were housed five to a cage with ad libitum access to food and water under a 12-h light/dark cycle with lights on at 6:00 AM. All experiments were carried out in accordance with the Guide for the Care and Use of Laboratory Animals as promulgated by the National Institutes of Health.

Behavioral Tests. All behavioral tests were conducted between 12:00 PM and 3:00 PM, with animals given several hours to acclimate to the testing room before the beginning of each experiment. The room was quiet with no distractions and maintained at a temperature of 21°C. The observer was not blinded to the treatment.

View larger version (32K): [in this window] [in a new window]   Fig. 1.   A, representative footprints and gait diagrams from vehicle- (left) and CP55940 (3 mg/kg; right)-treated animals. X refers to the measurement that was taken for gait width. B, mean step length, alternation coefficient, and linear movement in vehicle and CP55940-treated mice. There were no significant differences between treatments for these measures.

Behavioral Observations. We observed mice for the specific cerebellar symptoms of postural hypotonia and hyperreflexia. Postural hypotonia is characterized by animals lying flat on their abdomens with fore and hind limbs splayed laterally (Modianos and Pfaff, 1976). This effect is known to be caused by inhibition of postural muscle tone (Gilman, 1969). Hyperreflexia is manifested as jumping or popcorn behavior (Dewey, 1986). This involved exaggerated and synchronous contraction of fore and hind limb extensors, which could be elicited by auditory stimulus or disturbing the animals cage. In addition, animals suspended by their tails were observed for exaggerated bilateral limb extension.

Data Analysis. For the bar cross test, the mean number of slips was recorded for each group at each time point tested. A one-way ANOVA factoring number of slips and treatment group, or a one-way repeated measures ANOVA factoring slips and treatment were used to determine significance. Post hoc tests for individual groups included Dunnett's, paired, and unpaired t tests as indicated.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effects of CP55940 on Cerebellar Function. Mice were injected intraperitoneally with the CB₁ receptor agonist CP55940 at 0.3, 1, or 3 mg/kg. Mice treated with 1 or 3 mg/kg CP55940 slipped significantly more on the bar cross test than mice treated with vehicle (p < 0.05 and p < 0.01, respectively) at all time points tested (Fig. 2A). When placed on the horizontal bar for testing, CP55940-treated mice exhibited symptoms of truncal ataxia, including oscillations of the trunk when attempting to balance and walk on the bar. Doses of CP55940 higher than 3 mg/kg elicited significant catalepsy, which prevented the mice from completing the bar cross.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Mean number of slips on the bar cross (A) and average gait width (B) after administration of vehicle () or CP55940 at 0.3 (), 1 (), and 3 mg/kg (). Statistical significance determined by repeated measures ANOVA (A) or one-way ANOVA (B) followed by post hoc Dunnett's test (n = 4-7/group). *Significantly different from vehicle-treated group with p < 0.05.

Effects of SR141716 on CP55940-Induced Cerebellar Dysfunction. To evaluate the contribution of the CB₁ receptor to the effects of CP55940 on cerebellar function, animals were pretreated with the CB₁ receptor antagonist SR141716 30 min before the administration of 3 mg/kg CP55940. Treatment of mice with SR141716 alone at 1 mg/kg did not significantly alter performance on the bar cross or gait width (Fig. 3). However, SR141716 at a dose of 5 mg/kg significantly increased gait width compared with vehicle-treated animals (Fig. 3A). Pretreatment of mice with SR141716 at both doses significantly attenuated the diminished performance on the bar cross (p < 0.01) and increased gait width (p < 0.01) induced by administration of 3 mg/kg CP55940 (Fig. 3). The exaggerated bilateral limb extension produced by CP55940 was also inhibited by pretreatment with SR141716 (Table 1).

View larger version (23K): [in this window] [in a new window]   Fig. 3.   Mean number of slips on the bar cross (A) and average gait width (B) after administration of vehicle, SR141716 (SR; 1 or 5 mg/kg), CP55940 (CP; 3 mg/kg), or SR141716 plus CP55940. In A, vehicle (), 3 mg/kg CP (), 1 mg/kg SR (), 1 mg/kg SR + CP (), 5 mg/kg SR (), 5 mg/kg SR + CP (). Statistical significance determined by repeated measures ANOVA (A) or one-way ANOVA (B) followed by unpaired t tests to compare treatment groups. *Significantly different from vehicle-treated group with p < 0.05. Significantly different from CP55940-treated group with p < 0.01.

Effects of Other Cannabinoids on Cerebellar Function. Administration of THC at 100 mg/kg significantly increased the number of slips on the bar cross (p < 0.01) and significantly increased gait width (p < 0.01) compared with measurements taken before administration of THC (Table 2). THC did not significantly affect either measure at 30 mg/kg i.p. Animals treated with 100 mg/kg THC showed some symptoms of truncal ataxia as indicated by mild oscillatory motions when attempting to balance or cross the bar. These effects were much less dramatic than those seen with CP55940. Mice also displayed postural hypotonia, but only a few mice exhibited mild hyperreflexive movements.

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Mean number of slips (A) and average gait width (B) before drug administration, after administration of SR141716A (30 min), and immediately after administration of both SR141716A and AEA (i.v.). Significance determined by one-way ANOVA and paired t test. *Significantly different from vehicle-treated group with p < 0.05.

Effects of Dopamine Agonists on CP55940-Induced Cerebellar Dysfunction. To evaluate the contribution of cannabinoid-induced inhibition of nigrostriatal dopaminergic transmission to the diminished performance on the bar cross and abnormal gait, we pretreated animals with apomorphine (15 min) and bromocriptine (3 h) before the administration of CP55940. Neither administration of apomorphine alone at 4 mg/kg, nor bromocriptine alone at 8 mg/kg, significantly altered any parameters of gait or performance on the bar cross, although there was a trend toward motor incoordination after administration of apomorphine at 4 mg/kg. Pretreatment of mice with apomorphine or bromocriptine did not attenuate the effects of 3 mg/kg CP55940 on either the bar cross or gait width (Fig. 5). In fact, mice pretreated with both dopamine agonists showed a significant increase in the number of slips on the bar cross test at 30- and 50-min time points for apomorphine and 10-, 30-, and 50-min time points for bromocriptine compared with mice treated with CP55940 alone (Fig. 5A). Bromocriptine also potentiated the effects CP55940 on gait width (Fig. 5B).

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Mean number of slips on the bar cross (A) and average gait width (B) after administration of vehicle, apomorphine (Apo; 4 mg/kg), bromocriptine (Brom; 8 mg/kg), CP55940 (CP; 3 mg/kg), apomorphine plus CP55940, or bromocriptine plus CP55940. In A, vehicle (), 3 mg/kg CP55940 (), 8 mg/kg Brom (), 8 mg/kg Brom + CP (), 4 mg/kg Apo (), 4 mg/kg Apo + CP (). All slip data obtained from the CP55940-treated groups are significantly different from the respective control groups (p < 0.01). *Significantly different from vehicle-treated group with p < 0.05. Significantly different from CP55940-treated group with p < 0.05.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Early research into the behavioral effects of cannabinoids noted cerebellar deficits such as static ataxia in dogs and incoordination and hyperreflexia in primates (Conrad et al., 1972; Ho et al., 1972; Stark and Dews, 1980; Dewey, 1986). In spite of these early observations, little is known about the cellular mechanism(s) underlying these cannabinoid effects. The purpose of this study was to use well characterized tests of cerebellar function to determine the role of the CB₁ receptor in, and to begin to understand the neuroanatomical bases for, these cannabinoid effects.

All of the CB₁ receptor agonists tested were found to induce cerebellar motor incoordination as measured by the bar cross test. These results are in agreement with the recent report of decreased latency to fall from a roto-rod apparatus in mice given intracerebellar injections of THC (Dar, 2000). In addition to cerebellar motor incoordination, all of the CB₁ receptor agonists tested induced symptoms of profound truncal ataxia, including increased base of support while walking (revealed as an increase in gait width) and truncal oscillations when attempting to balance or walk on the bar (Poirier et al., 1974). Postural hypotonia was also induced by all of the CB₁ receptor agonists, although the duration of the deficit was variable among the various compounds. Symptoms of postural hypotonia have been described previously (Dewey, 1986).

The CB₁ receptor agonists also induced alterations in reflexes that are consistent with a cerebellar locus of action. The CB₁ receptor agonists produced hyperreflexia as manifest by pronounced jumping behavior as a result of synchronous bilateral contraction of both fore and hind limb extensors. This effect is consistent with alterations in the vestibulo-limb reflex pathway that is under the control of cerebellar circuits (Lindsay et al., 1976; Lindsay and Rosenberg, 1977). In addition, mice exhibited exaggerated bilateral extension of fore and hind limbs when suspended by their tails. During normal postural adjustments, animals exhibit ipsilateral shortening and contralateral lengthening of limb extensors; however, animals with cerebellar/vestibular lesions exhibit bilateral shortening of limb extensors under similar circumstances (Lindsay et al., 1976; Lindsay and Rosenberg, 1977).

The motor incoordination, truncal ataxia, and exaggerated bilateral limb extension induced by administration of CP55940 and AEA were attenuated by the CB₁ receptor-selective antagonist SR141716 (Rinaldi-Carmona et al., 1995). SR141716 has also been shown to attenuate THC- and AEA-induced static ataxia in dogs (Lichtman et al., 1998). Similarly, Dar (2000) reported that the motor deficits seen after intracerebellar THC infusion were eliminated by pretreatment with an antisense deoxyoligonucleotide sequence for the CB₁ receptor. The negative controls used in the present study, two compounds structurally similar to active cannabinoids, without affinity for the CB₁ receptor [cannabidiol (Devane et al., 1988) and Win 55212-3 (Kuster et al., 1993)], did not produce any cerebellar motor deficits. These data further support a CB₁ receptor-mediated mechanism.

The structure-activity profile for the CB₁ receptor agonists tested is similar between the two cerebellar tests with the exception of THC, which has significantly lower efficacy in the bar slip assay than the other compounds. In vitro studies demonstrate that THC is a low-efficacy, partial agonist of the CB₁ receptor compared with CP55940 and Win 55212-2 (Sim et al., 1996). We suggest that the lower efficacy of THC is apparent in the bar slip assay because there is no upper limit of impairment (i.e., there is theoretically no limit to the number of times an animal can slip). On the other hand, the gait width assay has a ceiling, the mice can only spread their limbs a certain distance, such that even a partial agonist is fully effective. The ratio of equieffective doses of CP55940, Win 55212-2, and THC in the gait width assay is approximately 1:3:10, which agrees well with the relative potencies of these compounds to produce other physiological and behavioral effects in mice, including hypothermia, hypolocomotion, and antinociception (Little et al., 1988; Compton et al., 1993; Wiley et al., 1998).

We have also compared the potencies of three ethanolamides of arachidonic acid: the endocannabinoid AEA and two high-affinity, CB₁ receptor-selective derivatives, ACPA and ACEA (Hillard et al., 1999). These CB₁ agonists were administered by i.v. injection because of their rapid metabolism and short half-lives (Willoughby et al., 1997; Hillard et al., 1999). In spite of the fact that ACPA and ACEA bind to the CB₁ receptor with K_D values 25- to 50-fold lower than AEA, all three arachidonamides affected cerebellar function in a similar dose range. This is in agreement with the effects of these agonists on body temperature in mice (Hillard et al., 1999) and likely reflects the low bioavailability of these compounds.

Increased GABAergic tone and subsequent inhibition of nigrostriatal dopaminergic projection neurons have been proposed to mediate the extrapyramidal deficits seen after administration of CB₁ receptor agonists (for review, see Ameri, 1999). It has been demonstrated that dopamine agonists attenuate cannabinoid-induced catalepsy (Ghosh et al., 1980). The dopamine agonist apomorphine has also been used to rule out extrapyramidal effects in the cerebellar neurotoxicity of 3-acetyl pyridine (De Michelle et al., 1980). To eliminate cannabinoid-induced inhibition of nigrostriatal dopaminergic transmission as a possible cause of the motor incoordination and ataxia, mice were pretreated with D1/D2 receptor agonist apomorphine or the D2 receptor agonist bromocriptine before challenge with CP55940. Neither apomorphine nor bromocriptine attenuated the effects of CP55940 on either bar slips or increased gait width. Since CP55940 was able to produce significant cerebellar deficits as measured by gait analysis and the bar cross test in the presence of dopamine agonists, it is unlikely that the extrapyramidal deficits induced by cannabinoids contribute to the motor deficits observed in this study. These results are consistent with our hypothesis that the CB₁ receptor agonists produce direct effects on cerebellar function, and the behavioral deficits measured by these tests are not a consequence of inhibition of dopaminergic function in the extrapyramidal system.

Interestingly, both dopamine agonists tested produced a potentiation of CP55940-induced motor incoordination as measured by the bar cross; bromocriptine also potentiated the effects of CP55940 on gait width. The mechanism by which dopamine agonists interact with cannabinoids to potentiate the cerebellar deficits remains unclear. However, both dopamine agonists increase spontaneous locomotion (Dogrul and Yesilyurt, 1999), which could increase the number of slips in the bar cross test.

The CB₁ receptor is localized at two synapses within the cerebellum: the granule cell/Purkinje cell synapse and the basket cell/Purkinje cell synapse (Tsou et al., 1998). Activation of the CB₁ receptor at either of these synapses results in an inhibition of neurotransmitter release (Takahashi and Linden, 2000). Since the granule cell input is excitatory, CB₁ agonists would be expected to decrease Purkinje cell activity through effects on this synapse. Conversely, the basket cell synapse is inhibitory, so CB₁ agonists acting here would increase Purkinje cell excitability. Based upon data discussed below, we hypothesize that the primary mechanism by which cannabinoids induce cerebellar deficits is via inhibition of GABAergic neurotransmission at the basket cell-Purkinje cell synapse (Fig. 6). Inhibition of GABAergic transmission at this synapse would lower the threshold for Purkinje cell firing and result in inhibition of the deep cerebellar nuclei.

View larger version (22K): [in this window] [in a new window]   Fig. 6.   Functional diagram of normal cerebellar circuitry (A) and alterations we hypothesize are induced by cannabinoids (B). Line thickness represents strength of synaptic transmission. CB1 receptor agonists inhibit GABAergic transmission at the basket cell-Purkinje cell synapse, which leads to an increased Purkinje cell output in response to climbing fiber excitation, and inhibition of the deep cerebellar nuclei. The reduced facilitatory output of these nuclei to projection sites such as the vestibular nucleus or motor cortex may underlie the cerebellar deficits seen after cannabinoid administration. BC, basket cell; CF, climbing fibers; DCN, deep cerebellar nuclei; LVN, lateral vestibular nucleus; PC, Purkinje cell; inf. olive., inferior olivary nucleus.

Acute treatment of mice with indole alkaloids produces cerebellar effects that are similar to the effects of the cannabinoids (O'Hearn and Molivar, 1997). In particular, these agents induce ataxia and extensor limb movements that propel the animals into the air. The proposed mechanism of action of these agents is increased Purkinje cell excitation via synchronous activation of climbing fibers, which leads to inhibition of the deep cerebellar nuclei and manifestation of motor deficits. It is our working hypothesis that CB₁ receptor agonists also increase Purkinje cell excitation; however, the mechanism is a reduction in the threshold for Purkinje cell firing in response to normal levels of climbing fiber input. This model is further supported by data indicating lesions of the deep cerebellar nuclei or the vestibular nucleus produce cerebellar symptoms strikingly similar to those seen after cannabinoid administration (Poirier et al., 1974; Modianos and Pfaff, 1976).

Other evidence that supports this hypothesis includes the demonstration, using functional magnetic resonance imaging, that THC decreases the activity of neurons in the deep cerebellar nuclei in humans (Bloom et al., 2000). Furthermore, cannabinoids increase glucose metabolism (Volkow et al., 1991) and blood flow (Mathew et al., 1998) in the cerebellar cortex, which is also predicted by the model since the Purkinje cell is disinhibited as a result of the activation of CB₁ receptors on the basket cell (Fig. 6).

In summary, we have demonstrated that CB₁ agonists produce cerebellar deficits such as motor incoordination, truncal ataxia, postural hypotonia, and hyperreflexia via a receptor-mediated mechanism. Additionally, these effects were not attenuated by pretreatment with dopamine agonists, indicating that cannabinoids produce motor deficits that are not dependent on cannabinoid-induced inhibition of nigrostriatal dopaminergic transmission. A mechanism of action that is consistent with the localization of the CB₁ receptors within the functional anatomy of the cerebellum is that activation of the CB₁ receptor of the basket cell results in inhibition of GABAergic transmission at the basket cell-Purkinje cell synapse. This leads to an increased Purkinje cell output and inhibition of neurons in the deep cerebellar nuclei. In addition, we suggest that the assays used in this study to quantify the effects of cannabinoids on cerebellar function be added to the classic behavioral measures currently being used to evaluate cannabimimetic activity.