Flumazenil as Potential Therapy for Essential Tremor

IETF Funded Research

Principal Investigator: Adrian Handforth, MD, assistant chief of neurology at Veterans Affairs Greater Los Angeles Healthcare System


A barrier to finding new medications for essential tremor (ET) has been the identification of molecular targets to which drugs could be designed to act on with selectivity and good tolerability. In previous IETF-funded work, we explored the role of inhibitory receptors that respond to GABA. GABAA receptors are not uniform but differ according to their subunit composition. They all contain 2 α subunits (alpha, 6 types), 2 β (beta, 3 types), and either a δ (delta, only 1 type) or a γ (gamma, 3 types) subunit. We have been interested in GABAA receptors containing both δ and α6 subunits, as such receptors are found mainly on cerebellar granule cells, the most abundant type of cell in the cerebellum. These cells are especially interesting in ET, because brain imaging studies suggest these cells are too active in ET patients. One imaging study found that low-dose alcohol reduces the excessive activity of the cerebellum in such a way that it is best explained by less cerebellar granule cell firing. Interestingly, GABAA receptors of the type located on cerebellar granule cells that contain both δ and α6 subunits have been found to respond to low levels of alcohol, so that alcohol could easily reduce firing in these cells by activating α6βδ GABAA receptors.

In previous work we found that low-dose alcohol, ganaxolone, and gaboxadol each suppresses tremor in a mouse model of ET in which tremor is induced with the drug harmaline, but not if the mice lack the δ or the α6 GABAA receptor subunit. Because the cerebellar granule cell is the main site where GABAA receptors express both these subunits, this provides an explanation how alcohol suppresses tremor. From this work, it may be inferred that drugs acting on only these receptors could be good candidates for ET therapy. Unfortunately, alcohol, gaboxadol, and ganaxolone not only activate δ GABAA receptors containing α6, but also those containing α4, found in widespread brain areas and which mediate adverse symptoms. So far, no one has devised a drug that activates only α6βδ GABAA receptors.

We now consider flumazenil, a medication that is on the market for treating benzodiazepine overdose and reversing conscious sedation. In addition, flumazenil activates α6βγ2 receptors that, like α6βδreceptors, are expressed intensely on cerebellar granule cells, where their activation should suppress tremor. Flumazenil also activates α2βγ2 and α3βγ2 receptors (involved in alleviating anxiety and depression), but not α1βγ2 receptors (involved in causing sedation). Interestingly, another group has reported that activation of α2βγ2 and α3βγ2 receptors by other drugs suppresses tremor in the harmaline model. Thus, flumazenil’s profile suggests that it may be effective and tolerated for ET.

In initial experiments, we will use the straight wire test, a sensitive index of psychomotor ability, to assess what doses of flumazenil do or do not cause impairment. Wild-type (WT) mice and mice lacking the α6 subunit (knockout, KO) will be tested. Only doses at which all mice pass this test will be used in subsequent harmaline experiments.

In the harmaline experiment, mice will receive harmaline to induce tremor, then receive either vehicle, a low, medium, or high dose of flumazenil. WT and littermate α6 KO mice will be tested.

We anticipate that flumazenil will likely be found to suppress harmaline-induced tremor in WT mice, but not in KO mice, indicating that flumazenil is likely suppressing tremor by inhibiting cerebellar granule cells, where most brain α6βγ2 GABAA receptors are located. If flumazenil suppresses tremor in both WT and KO mice that will suggest it works on other receptors, such as α2βγ2 and α3βγ2 GABAA receptors, which have already been implicated in tremor suppression. More importantly, because flumazenil has few if any sedative effects, we anticipate that flumazenil will suppress tremor in doses much less than those that cause impairment. This work is expected to provide insights to foster new anti-tremor therapy for ET that will be effective and well-tolerated, including potential consideration of flumazenil itself as an ET therapy.

Specific Aims: The development of new, high-efficacy, well-tolerated medications for essential tremor (ET) will require the identification of molecular targets that selectively affect tremor circuitry. To date, no marketed therapy for ET has emerged from laboratory research. Many persons with ET report that low doses of alcohol effectively suppress tremor. Metabolic imaging has shown that the cerebellar cortex is hyperactive in ET. This finding suggests that cerebellar granule cells (CGCs) and Purkinje cells (PCs) are over-active in ET. When low-dose alcohol that reduces tremor is given, cerebellar hyper-metabolism is reduced, consistent with reduction of CGC and PC firing.1High density EEG also indicates that alcohol suppresses tremor by acting on the cerebellum.2PC firing is dominated by the massive excitatory granule cell population, whose activity is controlled by Golgi cells, which release the inhibitory transmitter GABA. GABAA receptors are made of α and β subunits and either a δ or γ subunit. The Golgi-granule cell synaptic and extra-synaptic interface is doubly unique among brain areas in intensely expressing α6βγ2 and α6βδ GABAA receptors respectively. There is relatively little expression of these receptors outside the cerebellum.

Like synaptic GABAA receptors, extra-synaptic GABAA receptors are pentameric, but contain a δ rather than γ subunit. In view of their abundance on cerebellar granule cells, and their sensitivity to low levels of alcohol, as well as to neurosteroids, and to THIP (gaboxadol; 4,5,6,7-tetrahydroisoxazolo[5,4-c]pyridin-3-ol), we investigated the ability of these agents to suppress tremor in the harmaline model of ET. As we reported, THIP suppresses tremor in wild-type (WT) mice, but not in knockout (KO) mice lacking the α6 or the δ GABAA receptor subunit, indicating that this drug was very likely suppressing tremor via its action on CGCs, where α6βδ GABAA receptors are mainly expressed. Similar observations were made with low-dose alcohol and ganaxolone (unpublished). The effect of THIP on harmaline tremor was partial, however, as higher doses could not be administered due to psychomotor effects that are very likely mediated by α4βδ receptors expressed in brain regions outside the cerebellum.

There is no agent available that selectively activates α6βδ GABAA receptors, thus compromising for now the potential utility of this therapeutic target. However, CGCs also intensely express α6βγ2 receptors, found only in low levels elsewhere in the brain. A compound that activates α6βγ2 receptors, as a partial agonist, is flumazenil, which offers the advantages of good brain penetration and being already on the market for other indications. Flumazenil also activates α3βγ2 and α2βγ2 GABAA receptors less potently but of interest as these actions may contribute to therapeutic effects against anxiety and depression (common in ET) and against tremor. In contrast, flumazenil does not exhibit the undesirable property of activating α1βγ2 GABAA receptors, which promote sedation. In view of this highly favorable profile, we thus now want to assess flumazenil as potential therapy in the harmaline mouse model of essential tremor.

Aim 1: Effect of genotype on straight wire test response to flumazenil
Flumazenil’s effect on psychomotor performance will first be studied in the straight wire test, a highly sensitive test for impairment in mice. The purpose of this experiment is to determine the highest dose of flumazenil at which 6/6 α6+/+ (WT) and 6/6 littermate α6-/-(KO) mice pass all straight wire tests. Only doses at which all mice pass, or lower doses, will be used in subsequent harmaline experiments. If the KO mice are less sensitive than WT mice, that will indicate that doses that impair WT mice do so via the α6βγ2 GABAA receptor. If comparable doses impair WT and KO mice, that will suggest high doses have a non-α6βγ2 effect.

Aim 2: Effect of flumazenil on harmaline tremor in mice with or without GABAA receptor α6 subunit
Mice, either WT or littermate α6 KO, will receive vehicle, low, moderate, or high doses of flumazenil in harmaline experiments, all doses within the range that did not cause impairment in straight-wire testing. Pilot experiments will guide dose selection. It is anticipated that flumazenil will suppress harmaline-induced tremor in WT mice, but not in mice lacking the α6 subunit, in dose-dependent fashion, in doses much less than those that cause straight wire impairment. Such findings will support flumazenil as potential therapy for ET and the notion that α6βγ2 GABAA receptors represent a promising therapeutic molecular target for ET. Alternatively, if α6 KO mice still show tremor suppression by flumazenil, that would support a focus on other receptors, such as α3βγ2 and α2βγ2 GABAA receptors for future therapies.

Impact: We anticipate that, based on its pharmacologic profile, flumazenil may constitute an effective and well-tolerated therapy for ET. In addition, this project will provide evidence whether α6βγ2 GABAA receptors, located mainly on cerebellar granule cells, are a viable molecular target for developing future therapies for ET.