Abstract
GABAA receptors containing the α6 subunit are highly expressed in cerebellar granule cells and less abundantly in many other neuronal and peripheral tissues. Here, we for the first time summarize their importance for the functions of the cerebellum and the nervous system. The cerebellum is not only involved in motor control but also in cognitive, emotional, and social behaviors. α6βγ2 GABAA receptors located at cerebellar Golgi cell/granule cell synapses enhance the precision of inputs required for cerebellar timing of motor activity and are thus involved in cognitive processing and adequate responses to our environment. Extrasynaptic α6βδ GABAA receptors regulate the amount of information entering the cerebellum by their tonic inhibition of granule cells, and their optimal functioning enhances input filtering or contrast. The complex roles of the cerebellum in multiple brain functions can be compromised by genetic or neurodevelopmental causes that lead to a hypofunction of cerebellar α6-containing GABAA receptors. Animal models mimicking neuropsychiatric phenotypes suggest that compounds selectively activating or positively modulating cerebellar α6-containing GABAA receptors can alleviate essential tremor and motor disturbances in Angelman and Down syndrome as well as impaired prepulse inhibition in neuropsychiatric disorders and reduce migraine and trigeminal-related pain via α6-containing GABAA receptors in trigeminal ganglia. Genetic studies in humans suggest an association of the human GABAA receptor α6 subunit gene with stress-associated disorders. Animal studies support this conclusion. Neuroimaging and post-mortem studies in humans further support an involvement of α6-containing GABAA receptors in various neuropsychiatric disorders, pointing to a broad therapeutic potential of drugs modulating α6-containing GABAA receptors.
Significance Statement α6-Containing GABAA receptors are abundantly expressed in cerebellar granule cells, but their pathophysiological roles are widely unknown, and they are thus out of the mainstream of GABAA receptor research. Anatomical and electrophysiological evidence indicates that these receptors have a crucial function in neuronal circuits of the cerebellum and the nervous system, and experimental, genetic, post-mortem, and pharmacological studies indicate that selective modulation of these receptors offers therapeutic prospects for a variety of neuropsychiatric disorders and for stress and its consequences.
I. GABAA Receptors
GABAA receptors (GABAARs) are the major inhibitory neurotransmitter receptors in the brain and are the site of action of a variety of pharmacologically and clinically important drugs, such as benzodiazepines, barbiturates, anesthetics, neuroactive steroids, and convulsants (Sieghart, 1995). These drugs allosterically modulate the function of GABAARs via distinct binding sites (Sieghart, 2015; Puthenkalam et al., 2016), and from their effects on animal and human behavior it can be concluded that GABAARs are involved in modulating anxiety, cognition, vigilance, memory, and motor functions. GABAARs are ligand-gated chloride channels that are formed by five subunits. A total of 6α, 3β, 3γ, δ, ε, π, θ, and 3ρ subunits as well as several alternatively spliced isoforms of some of these subunits have been identified in the mammalian nervous system (Olsen and Sieghart, 2008). Each one of these subunits exhibits a distinct regional and cellular distribution in the brain (Wisden et al., 1992; Fritschy and Mohler, 1995; Pirker et al., 2000). Varying the combination of up to five different subunits in a single GABAAR subtype results in an enormous possible heterogeneity of GABAAR subtypes (Barnard et al., 1998).
The majority of GABAARs are composed of two α subunits, two β subunits, and one γ subunit (αβγ receptors). In these receptors, alternating α and β subunits are connected by a γ subunit (Tretter et al., 1997; Baumann et al., 2002; Laverty et al., 2019) (Fig. 1). In contrast to the αβγ receptors, the subunit stoichiometry and arrangement of receptors containing δ, ε, π, or θ subunits have not been unequivocally identified (Minier and Sigel, 2004; Bollan et al., 2008; Has and Chebib, 2018).
GABAAR subunit composition and arrangement. The majority of GABAARs is composed of two α subunits, two β subunits, and one γ subunit that form a central anion channel with the indicated subunit arrangement when viewed from above. The two α and β subunits can be identical or different. The upper left image depicts the experimentally verified subunit arrangement for the major GABAAR subtype composed of α1βγ2 subunits, and the other three panels depict α6-containing arrangements that have been shown to exist in cerebellar granule cells (Scholze et al., 2020). Each subunit has a principal (+) and a complementary (−) side. The + side contains the large loop C in the extracellular domain that is part of the pharmacological binding sites formed by the interfaces of two subunits. This loop C is schematically depicted as the pointed end of each subunit. The two GABA binding sites are formed by the two β+/α- interfaces. The benzodiazepine binding site (Bz) is formed by the α+/γ2- interface. Receptors containing α1, α2, α3, or α5 subunits form diazepam-sensitive benzodiazepine sites (DS-Bz). Receptors containing α4 or α6 subunits form diazepam-insensitive benzodiazepine binding sites (DI-Bz). Pyrazoloquinolinones (PQs) bind to two distinctive interfaces: to the benzodiazepine binding sites of all αβγ2 receptors with high affinity as silent binders (antagonists, similar to flumazenil) and to the PQ sites of these receptors formed by the α+/β- interfaces as modulators. α6GABAAR-selective pyrazoloquinolinones, such as Compound 6 (PZ-II-029), exert their positive modulatory action via the PQ site at the α6+/β- interface. The δ-containing α6βδ and α1α6βδ receptors presumably have a similar stoichiometry and subunit arrangement as the respective γ2-containing receptors, but this is still a matter of controversy.
II. GABAA Receptors Containing α6 Subunits (α6GABAARs)
The α6 subunit of GABAARs was first cloned from rat cerebellum in 1990 (Luddens et al., 1990). In situ hybridization (Laurie et al., 1992; Persohn et al., 1992) and immunohistochemical studies (Gutierrez et al., 1996; Pirker et al., 2000) revealed that the α6 subunit is located predominantly in cerebellar granule cells (Fig. 2A) and that its location and protein sequence are highly conserved in man, rat, mice, goldfish, and chicken, suggesting that receptors containing this subunit are of fundamental importance for cerebellar function (Bahn et al., 1996). In the meantime, however, α6 subunits were also identified in multiple other neuronal and non-neuronal tissues, although at a much lower abundance (Fig. 2A; Table 1). The function of α6GABAAR subunits in most of these tissues currently is not known.
GABAAR α6 subunit expression in the mouse brain and trigeminal ganglia. (A) In situ hybridization of a parasagittal mouse whole brain section for Gabra6 mRNA, obtained from Allen Brain Atlas Data Portal by Allen Institute for Brain Science (https://mouse.brain-map.org/experiment/show/308055734). Gabra6 expression is high in cerebellar granule cells but low in olfactory areas, isocortex, hippocampal formation, hypothalamus, midbrain, and pons. (B) Immunohistochemistry demonstrating GABAAR α6 subunit expression in neurons and satellite glial cells of TG in mice. Micrographs are representative mouse TG sections triple immunofluorescence-stained with antibodies against the α6 subunit of GABAA receptors (B-i, B-iii, and B-v, green), NeuN (B-ii and B-iii, blue, neuronal marker), and glutamine synthetase (GS, B-iv and B-v, red, satellite glial cells marker). Modified from Tzeng et al. (2021). Neurons and satellite glial cells are organized in bands or clusters within the ganglion. The spaces between the clusters are filled by nerve fibers, microglia, and macrophages (Legradi et al., 2020).
Distribution of α6-subunits in the nervous system and in peripheral tissues
So far, the subunit composition of α6GABAARs has been investigated in the cerebellum, only (Jechlinger et al., 1998). Results indicated that 45% of GABAARs in the cerebellum contained α6 subunits and that α6GABAARs in the rat cerebellum are predominantly composed of α6βγ2 (32%), α1α6βγ2 (37%), α6βδ (14%), or α1α6βδ (15%) combinations (Fig. 1). Other experiments indicated that 10%, 51%, or 21% of α6 receptors contained homogeneous β1, β2, or β3 subunits, respectively, whereas two different β subunits were present in 18% of all α6 receptors (Jechlinger et al., 1998). Overall, these data were confirmed in a second study that compared the total GABAAR composition in the cerebellum of mice and rats and investigated the abundance of minor GABAAR subtypes in the cerebellum (Poltl et al., 2003).
Cerebellar granule cells express α1, α6, β2, β3, γ2, and δ subunits abundantly, and the α1, α6, β2/3, and γ2 subunits are concentrated at postsynaptic sites of many GABAergic Golgi synapses, demonstrating that both α1 and α6 subunits are involved in synaptic transmission in the same synapse (Nusser et al., 1998). These could be pentameric α1βγ2 and α6βγ2 receptors sitting side by side in the same synapse or pentameric α1α6βγ2 receptors containing two different α subunit types. The actual existence of α1α6βγ2 or α1α6βδ receptors is supported by several lines of evidence, resulting in their classification as “existing with high probability” (Olsen and Sieghart, 2008; Scholze et al., 2020) (Fig. 1).
However, Golgi synapses immunopositive for only one of the α subunits were also found (Nusser et al., 1996). Receptors composed of the α1, α6, β2/3, and γ2 subunits are not only present at synapses but also present extrasynaptically at lower concentrations. δ subunits could not be detected in synaptic junctions, although they were abundantly present in the extrasynaptic dendritic and somatic membranes (Nusser et al., 1998). α6, β2/3, and γ2 but not α1 and δ subunits were also concentrated in some glutamatergic mossy fiber synapses, where they are colocalized with AMPA (α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid) receptors (Nusser et al., 1996, 1998). The extrasynaptic α6β2/3δ and α1α6β2/3δ receptors could thus mediate tonic inhibition, whereas the synaptic α1β2/3γ2, α6β2/3γ2, and α1α6β2/3γ2 receptors could mediate phasic inhibition (Nusser et al., 1998; Stell and Mody, 2002).
The α6 subunit confers unique characteristics to GABAARs. Electrophysiological investigations in Xenopus laevis oocytes indicated that GABA exhibited an up to 14-fold higher potency for activating recombinant α6βγ2 than for α1βγ2 receptors, whereas it exhibited an up to 50-fold further increase in potency for α6βδ compared with α6βγ2 receptors, depending on the type of β subunit within the receptors. For α1β3γ2L: α6β3γ2L: α6β3δ receptors, the potency ratio was 156: 23,3: 0,44 μM (Karim et al., 2013). However, the potency ratios were different in other recombinant expression systems because of possible differences in the receptor subunit composition, regulation by endogenous neurosteroids or Zn2+, or other factors, such as phosphorylation states (Mortensen et al., 2012). Thus, in mouse fibroblast L929 cells, the potency ratio of α1β3γ2L: α6β3γ2L: α6β3δ receptors for submaximal concentrations of GABA was 10: 3: 0,3 μM (Saxena and Macdonald, 1996) (Fig. 3A), whereas in human embryonic kidney 293 cells it was 2.1: 0.17: 0.17 μΜ (Mortensen et al., 2012), indicating that GABA potency was comparable for α6β3γ2L and α6β3δ receptors in the latter expression system. α6βγ2 and α6βδ receptors exhibited a slower rise time and a longer decay time than α1βγ2 receptors (Bianchi et al., 2002) (Fig. 3B). The rise and decay times of α1α6βγ2 were different from that of α1βγ2 and α6βγ2 receptors but closer to those of α1βγ2 receptors (Saxena and Macdonald, 1996; Tia et al., 1996). The unique properties of α6GABAARs that are grossly different from those of other GABAAR subtypes (Mortensen et al., 2012; Karim et al., 2013) support the conclusion that these receptors have important physiologic functions in regulating the activity of neurons especially at low GABA concentrations.
Representative GABA-evoked currents of recombinant GABAARs. (A) Whole-cell currents elicited by respective submaximal concentrations of GABA recorded from mouse fibroblast L929 cells transfected with α1β3γ2L, α6β3γ2L, and α6β3δ GABAA receptor subunits. Modified from Saxena and Macdonald (1996). (B) Concentration jump experiments in the outside-out patch configuration induced by a saturating GABA concentration (1 mM) recorded from human embryonic kidney 293T cells transfected with α6β3γ2L and α6β3δ GABAA receptor subunits. Modified from Bianchi et al. (2002). Horizontal bars above traces indicate the duration of GABA application.
The GABAAR agonist THIP (4,5,6,7-tetrahydroisoxazolo[4,5-c]pyridine-3-ol, gaboxadol, Fig. 4) is less potent than GABA at αβγ receptors but approximately 10 times more potent at δ-containing than at γ2S containing receptors (Brown et al., 2002; Meera et al., 2011). In contrast to GABA that is a partial agonist at δ-containing receptors (Bianchi and Macdonald, 2003; Meera et al., 2011), THIP is a full agonist at these receptors and thus elicits a markedly larger response than GABA [but see also the discussion of this topic in (Sieghart and Savic, 2018)]. Both α6β3γ2 and α6β3δ receptors as well as α1α6βγ2 and α1α6βδ receptors also are modulated by neurosteroids, and this modulation is stronger for α6β3δ than for α6β3γ2 receptors (Bianchi and Macdonald, 2003).
Chemical structures of GABAAR ligands. BZ, benzodiazepine.
α6βγ2 receptors, like α4βγ2 receptors, are diazepam-insensitive αβγ GABAARs (Luddens et al., 1990) (Fig. 1) and are also insensitive to a variety of other benzodiazepine-site ligands (Sieghart, 1995). These receptors can be positively modulated by the imidazobenzodiazepines Ro15-1788 (flumazenil) and Ro15-4513 (Korpi et al., 2007; Ramerstorfer et al., 2010) as well as by some but not all β-carbolines (Sieghart, 1995; Ramerstorfer et al., 2010) and by the flavonoid hispidulin (Fig. 4) (Kavvadias et al., 2004; Chiou et al., 2018). Other nonbenzodiazepine positive allosteric modulators (PAMs) of GABAARs, such as loreclezole or etomidate (Fig. 4), are also able to modulate α6βγ2 receptors, but these compounds also modulate the activity of α1-5βγ2 GABAARs (Wafford et al., 1994; Belelli et al., 1997). However, some pyrazoloquinolinones, such as PZ-II-029 (Compound 6), LAU159 (8-chloro-2-(3-methoxyphenyl)-2H-pyrazolo[4,3-c]quinolin-3(5H)-one), and LAU463 (7-bromo-2-(4-methoxyphenyl)-2,5-dihydro-3H-pyrazolo[4,3-c]-quinolin-3-one)) (Fig. 4) that are high affinity silent modulators (antagonists) at the benzodiazepine site of α1-6βγ2 receptors, are also highly selective PAMs of α6βγ2 and α6βδ GABAARs (Fig. 5) by acting via a second binding site at the extracellular α6+β2/3- interface of these receptors (Fig. 1) (Ramerstorfer et al., 2011; Varagic et al., 2013; Chiou et al., 2018; Treven et al., 2018; Simeone et al., 2019). In addition, the diuretics amiloride (Fisher, 2002) and furosemide (Korpi et al., 1995; Wafford et al., 1996; Korpi and Luddens, 1997) are relatively selective allosteric antagonists at α6βγ2, α1α6βγ2, α6βδ, and α1α6βδ receptors (Fig. 4) (Sigel et al., 2006; Sieghart and Savic, 2018). Additional compounds preferentially modulating these receptors are discussed by Sieghart and Savic (2018).
Concentration-response curves of α6GABAAR-selective ligands. Curves are reproduced based on published (Varagic et al., 2013; Chiou et al., 2018; Knutson et al., 2018; Treven et al., 2018), (PCT/US2016/035761) and unpublished data for PZ-II-029/Compound 6 (top), 8-chloro-2-(3-methoxyphenyl)-2H-pyrazolo[4,3-c]quinolin-3(5H)-one (LAU159) (middle), and 7-bromo-2-(4-methoxyphenyl)-2,5-dihydro-3H-pyrazolo[4,3-c]-quinolin-3-one) (LAU463) (bottom). Data for α6β3γ2 and α6β3δ are presented in blue. At 100% of control current, the response equals that of the GABA-evoked current in the absence of compounds. All measurements were performed with GABA concentrations eliciting 3%–5% and ∼10%–20% of maximum GABA-elicited current for γ2- and δ-containing receptors, respectively. Modulation of α6-containing receptors by the indicated compounds at 100 nM was already stronger than that of receptors containing other α subunits. Deuterated derivatives, such as DK-I-56-1, behave like the respective parent compounds in concentration-response investigations at different receptor subtypes (Knutson et al., 2018).
Α. α6GABAAR Functions in the Cerebellum
The cerebellar cortex consists of a folded sheet of microcircuitry reiterated on a vast scale (Fig. 6). The same circuitry in different areas of the cerebellum probably computes similar operations linked to different parts of the brain (Wisden et al., 2009). After receiving specific sensory and contextual information, this circuitry probably measures or predicts time intervals for motor activity required for controlling movement of the respective body parts.
Simplified schematic representation of cerebellar microcircuitry. Excitatory terminals are indicated by arrowheads and a plus sign, whereas inhibitory terminals are represented by blunt T-ends and a minus sign. The mossy fiber terminal structure that is known as glomerulus is depicted as drop-shaped terminal ending. Granule cells with their schematic processes are emphasized by red color.
The cerebellar cortex has two excitatory inputs, the mossy fibers and the climbing fibers. Both use glutamate as a transmitter. Mossy fibers provide the inputs that elicit the coordinated movement programs, and climbing fibers provide the error signals that lead to an improved motor program and might work as a timer for motor activity (D'Angelo et al., 2013). The only output of the cerebellar cortex is mediated via the GABAergic Purkinje cells (PCs) that provide the results of the cerebellar program calculations as inhibitory signals to the deep cerebellar nuclei (DCN) (Pugh and Raman, 2009) via α1βγ2 GABAARs (Laurie et al., 1992) (Fig. 6).
As discussed above, α6GABAARs are located at Golgi cell–granule cell synapses and extrasynaptic sites. The synaptic α1βγ2 receptors together with synaptic α6βγ2 and α1α6βγ2 receptors mediate phasic inhibition, whereas the extrasynaptic α6βδ or α1α6βδ receptors as well as extrasynaptically located α6βγ2 and α1α6βγ2 receptors mediate tonic inhibition of granule cells (Nusser et al., 1998). Granule cells are excitatory interneurons in the cerebellar cortex that receive excitatory inputs from mossy fibers and are inhibited by Golgi cells. The synaptic contacts of mossy fiber terminals, granule cell dendrites, and Golgi cells axon and dendrites are enwrapped into a glial sheet, forming the cerebellar glomerulus (Mapelli et al., 2014) (Fig. 6). Each glomerulus is characterized by one mossy fiber rosette and dendrites of several granule cells and Golgi cells as well as Golgi cell axons. By confining the space around the granule cell and Golgi cell compartments, the glomerulus provides optimal conditions for GABA released from Golgi cells to also activate extrasynaptic receptors at the granule cell dendrites as well as α6βγ2 receptors at some of the excitatory mossy fiber synapses (Nusser et al., 1996), possibly limiting the excitation mediated by these synapses and thus sharpening their excitatory signal. Each granule cell dendrite contacts different glomeruli, receiving inputs and information from different mossy fibers. The axons of granule cells ascend in the molecular layer, where they originate the parallel fibers that form excitatory synapses on PCs and Golgi cells. Golgi cells can thus directly inhibit granule cells in a feedforward loop when activated by mossy fibers within the glomeruli. They also can inhibit granule cells in a feedback loop after Golgi cell activation via the ascending axons and parallel fibers of granule cells.
While granule cell depolarization begins, the increasing activity of the feedforward and feedback inhibition via Golgi cells modulates their response. GABA release from the presynaptic terminal of Golgi cells activates α1- and in part α6-containing receptors in the postsynaptic density (Nusser et al., 1998). α1βγ2 receptors open and close rapidly in contrast to the α6βγ2 receptors. Whereas α1βγ2 receptors control granule cell responses in a narrow, approximately 10-millisecond band, α6βγ2 receptors show a broader, approximately 50-millisecond tuning (Mapelli et al., 2014; Nieus et al., 2014). Inhibition by GABA reduces the intensity and duration of granule cell responses without significantly changing the precision of the first spike (Nieus et al., 2014). The longer inhibition time elicited by α6βγ2 receptors thus better separates each incoming signal from other incoming signals. In addition, Golgi cells cause a broad lateral inhibition of granule cells that generates dense clusters of granule cell activity organized in center-surround structures, implementing combinatorial operations of multiple mossy fiber inputs (D'Angelo et al., 2013). This allows generation of granular layer coherent oscillations and resonance, induction of mossy fiber–granule cell long-term potentiation and long-term depression, and spatiotemporal reconfiguration of granular layer activity (Nieus et al., 2014). However, long-term potentiation and long-term depression occur at several excitatory and inhibitory synapses in the main cerebellar subcircuits and regulate the function of granule cells, PCs, and DCN cells (Mapelli et al., 2015).
Although the dynamic nature of cerebellar computations and the exceptional temporal sensitivity of cerebellar circuits require high-precision phasic inhibitory mechanism, tonic inhibition at Golgi cell–granule cell synapses provides at least 97% of the inhibition of granule cells (Hamann et al., 2002). Tonic inhibition reduces the fraction of granule cells activated by mossy fiber inputs and can set the level of neuronal excitability at the mossy fiber–granule cell synapse (Hamann et al., 2002; Mapelli et al., 2014). The excitability of granule cells can additionally be regulated by GABA tonically released from glial cells and acting on cell bodies, axons, and parallel fibers of granule cells (Lee et al., 2010). Such processes might be involved in enhancing pattern recognition in Purkinje cells, for instance by distinguishing between important (strong stimulus) and nonimportant (weak stimulus) information. A perfectly balanced function of the extrasynaptic α6GABAARs also is essential because their overactivity (or overstimulation by drugs such as ethanol) can disrupt the synchronous firing of granule cells, causing disturbances in motor programs, whereas an underactivity can provide problems in the cognitive performance and stress-related and social behaviors (Rudolph et al., 2020).
Granule cells, in addition to activating PCs and Golgi cells, also generate feedforward and lateral inhibition of PCs by activating stellate cells (at the dendrites of PCs) and basket cells (at the soma of PCs), respectively (Fig. 6). Feedforward inhibition limits the excitation of PCs, whereas lateral inhibition reduces the activity of those PCs that fire asynchronously. Spike synchrony of PCs requires coactivation of several discrete groups of granule cells to overcome lateral inhibition by basket and stellate cells. This lateral inhibition also narrows the duration of granule cell–mediated excitation to 1–2 milliseconds, an effect that would be expected to increase the precision of PC firing in response to excitation (Person and Raman, 2012).
PCs represent the major output neurons of the cerebellar cortex, providing an inhibitory control on the DCN. DCN neurons spontaneously fire tens of action potentials per second in vivo when animals are not engaged in cerebellar behavior (Pugh and Raman, 2009) possibly to maintain ongoing muscle tone. Due to their specific ion channels, even a short excitation, possibly via mossy fibers that not only activate granule cells but also the respective DCN neurons, can drive DCN neurons into a depolarization block. Their firing usually resumes only after an active hyperpolarization likely induced by the inhibitory GABAergic input from PCs (Pugh and Raman, 2009). This behavior possibly allows DCN neurons to resume firing in synchrony with PCs.
Accumulated evidence indicates that subsets of PCs synchronize their firing during behaviors that require the cerebellum, for instance in a learned motor task in rats, and the increase of synchrony is time-locked to movement (Heck et al., 2007). The synchronous inhibitory input from PCs sets the timing and rate of action potentials produced by DCN cells, thereby relaying information out of the cerebellum (Person and Raman, 2012). Then, changing spatiotemporal patterns of PC activity allows different subsets of inhibitory neurons to control cerebellar output at different times.
B. Role of the Cerebellum in the Function of the Brain
The cerebellum is involved in sensorimotor functions regulating the movement of all muscles in the body, including those controlling emotional facial expressions and speech, and is thus involved in all our communication with and all our reactions to the surrounding world. For that, information from all our senses is analyzed and integrated in the appropriate brain regions, eliciting context-specific motor programs in the motor cortex that are conveyed to the cerebellum via mossy fibers. These motor programs are constantly adapted and optimized by incoming sensory information via climbing fibers to fit to the individual requirements. The cerebellum is thus also involved in attention and cognitive performance (Mannarelli et al., 2019; Schmahmann et al., 2019) as well as in language and emotional processing (Sacchetti et al., 2009 De Smet et al., 2013) and in controlling the reward circuitry and social behaviors (Wagner et al., 2017; Carta et al., 2019).
The cerebellum forms reciprocal, closed-loop circuits with much of the cerebral cortex as well as subcortical structures and contains repeating processing modules, the function of which is driven by the input the module receives (Schmahmann, 1991; D'Mello and Stoodley, 2015). Motor coordination (anterior cerebellum) and cognitive functioning (posterior cerebellum) are mediated by different parts of the cerebellum. Thus, connections between PCs in the posterior lateral cerebellum to DCN neurons, which are distinct from those controlling motor activity, convey the information from the cerebellar cortex to the prefrontal cortex via the thalamus (Caligiore et al., 2016). Lesions in the “limbic cerebellum” (vermis and fastigial nucleus) cause a “cerebellar cognitive affective syndrome” (Schmahmann and Sherman, 1998) that involves disorders of attentional control, emotional control, and social skills as seen in patients with autism spectrum (D'Mello and Stoodley, 2015) and psychosis spectrum disorders (Schmahmann et al., 2007). The cerebellum also has extensive connections with the ventral tegmental area (Carta et al., 2019), which is involved in the reward circuitry, and the basal ganglia, providing the neural basis for a cerebellar involvement in Parkinson disease, dystonia, Tourette syndrome, and addiction as well as in normal basal ganglia function, such as reward-related learning (Caligiore et al., 2016).
Relative to other regions of the brain, the human cerebellum undergoes enormous growth between 24 and 40 weeks postconception, increasing approximately 5-fold in volume and over 30-fold in surface area. Although this rapid cerebellar growth slows postnatally, neural differentiation and growth of axonal inputs and outputs continue throughout the first postnatal year. This substantial prenatal growth continued postnatally renders the cerebellum especially vulnerable to developmental disruptions and damage (D'Mello and Stoodley, 2015). Abnormalities in cerebellar structure, connectivity, and function have been identified in patients with neurodevelopmental disorders like autism spectrum disorder (Shevelkin et al., 2014; D'Mello and Stoodley, 2015), schizophrenia (Parker et al., 2013; Peters et al., 2016), and dyslexia (Person and Raman, 2012) and might thus be causatively related to changes in emotional, cognitive, and social behaviors of these patients.
From the above description of the structure and function of the cerebellum, the involvement of α6GABAARs in neuropsychiatric disorders can easily be delineated. Synaptic and extrasynaptic α6GABAARs are of central importance for the functioning of the cerebellum. A significant disturbance in their expression or function might thus not only result in motor incoordination, ataxia, dystonia, and epilepsy but also contribute to neuropsychiatric disorders.
III. α6GABAARs and Animal Models of Neuropsychiatric Disorders
A. α6GABAARs and Animal Models of Angelman Syndrome
Angelman syndrome (Table 2) is a neurodevelopmental disorder that is most frequently caused by a maternal deletion of the chromosome 15q11-q13 region, which includes the causative UBE3A gene encoding a ubiquitin E3 ligase (Roden et al., 2010). The major clinical manifestations consist of severe developmental delay, speech impairment, epilepsy, movement and balance problems, and characteristic behavior, such as paroxysmal laughter (Egawa et al., 2012). Abnormal movement and balance are largely attributed to cerebellar ataxia. No effective therapeutic strategies have yet been elucidated.
Therapeutic potentials of α6GABAAR PAMs based on patient studies and animal models of neuropsychiatric disorders
Recently, a study on a UBE3A-deficient mouse model of Angelman syndrome (Egawa et al., 2012) demonstrated that the deletion of the ubiquitin E3 ligase, which controls the degradation of GABA transporter 1 (GAT-1), leads to a surplus of GAT-1 and decreased GABA concentrations in the extrasynaptic space of cerebellar granule cells. Reduced tonic inhibition of granule cells causes abnormal firing of PCs and ataxia. Low doses of THIP (gaboxadol), which preferentially activates extrasynaptic α6β3δ receptors, improved the abnormal PC firing and reduced cerebellar ataxia in these mice (Egawa et al., 2012).
B. α6GABAARs and Animal Models of Down Syndrome
Down syndrome (Table 2) is due to the inheritance of an additional copy of all or part of chromosome 21 (trisomy 21) and is characterized by intellectual disability and impaired motor control. Lack of coordinated movement, poor balance, and unclear speech imply dysfunctions of the cerebellum, which is known to be reduced in volume in Down syndrome (Table 2). The principal cause of the smaller cerebellum is a diminished number of granule cells. The Ts65Dn mouse, the most widely investigated animal model of Down syndrome, was generated by triplication of a region of the mouse chromosome 16, which makes it trisomic for approximately half of the orthologous protein-coding genes and a subset of non–protein-coding RNAs located on the long arm of the human chromosome 21 (Hsa21). Ts65Dn mice replicate the reduced number and density of cerebellar granule cells that are characteristic for Down syndrome (Szemes et al., 2013). Using the Ts65Dn mouse model of Down syndrome, it was demonstrated that the tonic GABAA receptor current in granule cells is smaller than in wild-type mice and that Ts65Dn granule cells are thus more excitable (Szemes et al., 2013). Enhancing tonic inhibition of granule cells might thus improve coordinated movement in this mouse model as well as in patients with Down syndrome (Szemes et al., 2013).
C. α6GABAARs and Animal Models of Essential Tremor
Essential tremor (Table 2) is the most common movement disorder. It is characterized by action tremor manifested with involuntary rhythmic contractions and relaxations of certain muscle groups in one or more body parts and involves an oscillatory network that includes the cerebellum, thalamus, motor cortex, and brainstem, as shown by magnetoencephalography (Schnitzler et al., 2009). The cause of essential tremor currently is not known. Post-mortem studies in individuals with essential tremor identified abnormalities in the cerebellum, which primarily affected PCs, basket cells, and climbing fibers (Louis and Faust, 2020). Functional magnetic resonance imaging (fMRI) studies revealed excessive cerebellar activity in patients with essential tremor (Sharifi et al., 2014).
Harmaline is a β-carboline neurotoxin that can induce action tremor in multiple species, including humans, in a frequency range of 8–16 Hz (Martin et al., 2005). Harmaline affects many types of receptors, ion channels, and gap junctions and likely affects the activity of multiple cell populations in the brain (Paterson et al., 2009; Jiménez-Jiménez et al., 2020). Evidence accumulated that harmaline-induced action tremor involves synchronous rhythmic firing in the inferior olive nucleus, which provides a major input to cerebellar PCs via climbing fibers, strongly suggesting an involvement of PCs in tremor generation (Brown et al., 2020; Handforth and Lang, 2021). Experiments in awake behaving mice indicated that the onset of harmaline-induced tremor was coincident with rhythmic PC firing, and harmaline-induced tremor was blocked by genetically silencing of cerebellar PC output. Finally, mimicking harmaline-induced PC firing pattern by optogenetic PC stimulation reproduced the tremors induced by harmaline (Brown et al., 2020). Since Golgi cells control cerebellar activity by releasing GABA onto synaptic and extrasynaptic α6GABAARs on cerebellar granule cells, it was investigated whether THIP (gaboxadol), which preferentially activates extrasynaptic α6βδ GABAARs, would suppress harmaline-induced tremor in mice (Handforth et al., 2018). Whereas THIP reduced harmaline-induced tremor in the respective littermates, it did not in α6−/− or δ−/− mice, suggesting that cerebellar α6βδ GABAARs mediate THIP-induced tremor reduction (Handforth et al., 2018). Consistently, α6GABAAR-selective PAMs significantly suppressed harmaline-induced tremor activity in mice as did propranolol, a clinically used pharmacotherapy for essential tremor. Importantly, intracerebellar microinjection of furosemide, an α6GABAAR antagonist, prevented the antitremor effect of α6GABAAR-selective PAMs but not that of propranolol (Huang et al., 2021).
D. α6GABAARs and Animal Models of Tic Disorders
In a case report, a pediatric patient with intractable motor tic disorders was found to respond well to an herbal remedy consisting of the leaf juice of Clerodendrum inerme (L.) Gaertn (C. Inerme) without apparent hepatotoxicity or renal toxicity (Fan et al., 2009). Tic disorders (Table 2), including Tourette syndrome, an idiopathic spectrum of tic disorders, are attributed to an over-reactive dopaminergic system in the cortico-thalamic-striatal circuit (Singer, 2005). Using methamphetamine-induced hyperlocomotion in mice, a hyperdopaminergic behavior model, the flavonoid hispidulin (6-methoxy-4´,5,7-trihydroxyflavonoid) (Fig. 4) was identified as a constituent of the leaf extract that alleviated methamphetamine-induced hyperlocomotion (Chen et al., 2012; Huang et al., 2015).
Hispidulin exhibits a micromolar affinity for the benzodiazepine site of human GABAARs and is a PAM of recombinant GABAARs composed of the α1β2γ2S, α2β2γ2S, α3β2γ2S, α5β2γ2S, or α6β2γ2S subunits (Kavvadias et al., 2004). The alleviating effect of hispidulin on methamphetamine-induced hyperlocomotion in mice (Liao et al., 2016) was mimicked by compounds modulating α6GABAARs, including Compound 6 (PZ-II-029), the selective α6GABAAR PAM (Fig. 4), but not by diazepam and was prevented by an intracerebellar microinjection of furosemide, a selective allosteric α6GABAARs antagonist, suggesting that hispidulin acts by positively modulating α6GABAARs (Liao et al., 2016).
Compound 6 also significantly reduced stereotypic behaviors in mice with intrastriatal SLIT and NTRK-like family, member 1 (Slitrk1) gene silencing, an animal model of Tourette syndrome [Wu CY, (2016) Master Thesis, Graduate Institute of Pharmacology, College of Medicine, National Taiwan University, Taipei, Taiwan; Du et al., Soc Neurosci 2017, Prog. No. 690.01, November 11-15, 2017, Washington DC, USA]. The Slitrk1 gene is the mouse version of the human SLITRK1 gene that codes for a transmembrane and signaling protein, which is responsible for synapse regulation and presynaptic differentiation in the brain. It has been linked to the formation of excitatory synapses. Among all adult striatal neurons, Slitrk1 was reported to be only expressed in cholinergic interneurons (Stillman et al., 2009) that were lost in post-mortem brains of patients with Tourette syndrome (Kataoka et al., 2010). Mutations in the SLITRK1 gene have been linked to obsessive-compulsive spectrum disorders, such as Tourette syndrome (O’Rourke et al., 2009), obsessive-compulsive disorder (OCD) (Ozomaro et al., 2013), and trichotillomania (Chattopadhyay, 2012).
In addition, an involvement of α6GABAARs in tic-related behaviors was supported by a study indicating that DK-I-56-1 (Fig. 4), the deuterated derivative of Compound 6 (Knutson et al., 2018), significantly reduced tic-like jerks in the D1CT-7 mouse model of Tourette syndrome (Cadeddu et al., 2021). The D1CT-7 mice are a transgenic line generated through the attachment of a neuropotentiating cholera toxin to the D1 dopamine receptor promoter. These mice display tic-like manifestations consisting of brief, sudden axial jerks (Cadeddu et al., 2021).
In addition to tic attacks, patients with Tourette syndrome also manifest sensorimotor gating deficits, which may contribute to their premonitory urges (Leckman et al., 1993) and can be measured by the disruption in the prepulse inhibition (PPI) of the startle response (Swerdlow et al., 2001). Exposing D1CT-7 mice to brief stressors dramatically increased their tic-like manifestations and elicited PPI deficits (Cadeddu et al., 2021). Similarly, striatal Slitrk1-silenced mice also exhibited a haloperidol-sensitive PPI disruption (Du et al., 2017; Soc Neuroscience 2017, Prog. No. 690.01, November 11–15, 2017, Washington DC, USA). This supports the validity of these two animal models and the hyperdopaminergic status in Tourette syndrome, as implicated in patients because of abnormal function of cortico-striatal-thalamo-cortical circuits (Leckman and Riddle, 2000). Currently, it is not known how the striatal dopaminergic overactivity in Tourette syndrome can be modulated by α6GABAAR PAMs via reducing cerebellar granule cell activity.
PPI disruptions are also manifested in patients with several other neuropsychiatric disorders (Table 2), such as attention deficit/hyperactivity (ADHD) disorder (Ornitz et al., 1992), OCD (Swerdlow et al., 1993), Huntington disease (Swerdlow et al., 1995), schizophrenia (Braff et al., 1978), and autism spectrum disorder (Cheng et al., 2018), to name only a few of them.
E. α6GABAARs and Animal Models of Schizophrenia
Especially in patients with schizophrenia, PPI disruption is a well known endophenotype manifestation (Braff et al., 1978). PPI disruption can be induced by methamphetamine or the N-methyl-D-aspartate (NMDA) channel blockers ketamine, phencyclidine (PCP), and MK-801 (dizocilpine) based on hyperdopaminergic and hypoglutamatergic hypotheses of schizophrenia, respectively (Geyer and Moghaddam, 2002). Using these PPI disruption models in mice, it was demonstrated that hispidulin but not diazepam rescued disrupted PPI when applied either via an intraperitoneal injection or a bilateral intracerebellar microinjection (Chiou et al., 2018). The effect of systemic hispidulin on methamphetamine-disrupted PPI could be mimicked by Compound 6, the α6GABAAR-selective PAM, and its analogs (PCT/US2016/035761). Effects of hispidulin and Compound 6 were prevented by an intracerebellar microinjection of furosemide, suggesting that cerebellar α6GABAARs are the targets of hispidulin and Compound 6 (Chiou et al., 2018). Since hispidulin, in contrast to Compound 6, could not modulate α6β3δ receptors, these data suggest that PAMs of cerebellar α6βγ2 receptors can rescue disrupted PPI (Chiou et al., 2018).
Interestingly, maternal separation stress in mice carrying the mutated schizophrenia risk gene, DISC1 (Disrupted in Schizophrenia 1), can induce PPI impairment and social withdrawal, mimicking positive and negative symptoms, respectively, in schizophrenia (Niwa et al., 2013). Hispidulin reduced both behavioral phenotypes in these mice and in chronic PCP-treated mice (Mouri et al., 2020). The α6GABAAR-selective PAMs Compound 6 and DK-I-56-1 also reduced PPI impairment, social withdrawal, and cognitive impairment in chronic PCP-treated mice (Chiou et al., 2019, poster at the Soc Neurosci 2019, No. 645.14/B55, Chicago, IL, USA) and might thus be able to ameliorate the three core symptom dimensions of schizophrenia, positive symptoms, negative symptoms, and cognitive impairment.
F. α6GABAARs and Animal Models of Stress-Associated Disorders
1. α6GABAARs, Stress Response, Anxiety-Like Behaviors, and Social Deficits
A Gabrd knockout in cerebellar granule cells of mice caused a replacement of α6βδ by α6βγ2 and α6β receptors (Tretter et al., 2001). This resulted in a decreased tonic inhibition and enhanced excitability of granule cells and a differential activation of cerebellar output regions as well as of many cortical and subcortical brain areas involved in cognition, anxiety-like behaviors, social deficits, and stress responses, specifically in females (Rudolph et al., 2020). Gabrd knockout also resulted in deficits of maternal behaviors, which is relevant to disorders like postpartum depression. In contrast, motor functions were unaffected in these mice (Rudolph et al., 2020). PAMs of α6βδ GABAARs may thus attenuate all these behaviors elicited by a decreased tonic inhibition of granule cells.
2. α6GABAARs and Stress-Induced Depressive Behavior
An essential role of hippocampal α6GABAARs in maternal separation stress–induced adolescent depressive behaviors was demonstrated in rats (Yang et al., 2016). α6GABAARs are expressed in low abundance on hippocampal interneurons of naive adolescent rats, as indicated by immunohistochemistry. Maternal separation stress from postnatal days 2–15 significantly reduced α6 and α2 but not α1, α3, α4, α5, γ2, or δ expression in the hippocampus, and this was paralleled by an anhedonia-like depressive behavior. Such a depressive behavior was also elicited in young naive control animals by a knockdown of α6 expression in the hippocampus (Yang et al., 2016). These results indicated that α6GABAARs are key modulators of hippocampal synaptic transmission, and malfunctioning of these receptors likely plays a crucial role in the pathophysiology of maternal separation–induced depression in rats.
3. α6GABAARs and Stress-Induced Attention Deficit and Hyperactivity
The maternal separation stress in rats also led to an increased locomotor activity, as shown by a longer travel distance in an open field test (Yang et al., 2016). A similar phenomenon was observed by Kwak et al. (2009), in which maternal separation after P14 caused hyperactivity, as indicated by increased movement in an open field. The hyperactive profile was similar to the findings in an animal model of ADHD, as characterized by poor concentration, increased motor activity, and behavioral impulsivity (Womersley et al., 2015).
G-protein–coupled receptor kinase-interacting protein-1 (GIT1) has been reported to be associated with ADHD in humans (Won et al., 2011), and the genetic deletion of GIT1 resulted in ADHD-like behaviors in mice (Kim et al., 2017) (Table 2). Moreover, a decreased glial GABA intensity was observed in the cerebellum of GIT1-knockout mice, which led to reduced tonic GABA currents in cerebellar granule cells, which are likely mediated by α6βδ receptors (Kim et al., 2017). Overactive cerebellar granule cells might thus contribute to the symptoms of ADHD.
G. α6GABAARs and Animal Models of Adverse Ethanol Effects
Ethanol is a PAM at most GABAAR subtypes, including α6βγ2 and α6βδ receptors, explaining its anxiolytic, sedative-hypnotic, anticonvulsant, and motor incoordinating properties (Kumar et al., 2004). In addition, ethanol exhibits multiple actions at other ion channels, neurotransmitter receptors, and transporters in the brain (Korpi et al., 2015).
1. α6GABAARs and the Ethanol Antagonist Ro15-4513
The imidazobenzodiazepine Ro15-4513 (Fig. 4) is a negative allosteric modulator at α1βγ2, α2βγ2, α3βγ2, and α5βγ2 receptors and a PAM of the diazepam-insensitive α4βγ2 and α6βγ2 receptors (Mohler et al., 1984; Sieghart et al., 1987; Luddens et al., 1990; Hevers and Luddens, 1998). It was demonstrated to be a potent antagonist of the muscimol-induced and ethanol-stimulated 36Cl- uptake into brain vesicles and also blocked the anticonflict activity of low doses of ethanol as well as the behavioral intoxication observed at higher doses of ethanol (Suzdak et al., 1986). The structural analog Ro15-1788 (flumazenil, Fig. 4) could block the antiethanol effects of Ro15-4513 but not the ethanol effects itself. The antiethanol effect of Ro15-4513 could also not be blocked by other negative allosteric modulators acting via the benzodiazepine site of GABAARs, leading to the conclusion that the antiethanol effect of Ro15-4513 might be mediated by a unique interaction with GABAARs (Suzdak et al., 1986).
2. α6GABAARs and Ethanol-Induced Motor Incoordination
This unique interaction site was tentatively identified by the finding that ethanol impairs motor coordination by enhancing tonic inhibition at cerebellar α6β3δ receptors (Hanchar et al., 2006). The ethanol action at α6β3δ receptors was proposed to be mediated by a novel high-affinity binding site for ethanol, Ro15-4513, and Ro15-1788 located at the α4/α6+β3- subunit interface of α4/α6β3δ receptors at a position homologous to the classic benzodiazepine binding site at the α4/α6+γ2- interface (Hanchar et al., 2006; Wallner et al., 2014). Low concentrations of ethanol could inhibit [3H]Ro15-4513 binding to this site, and this [3H]Ro15-4513 binding could also be inhibited by Ro15-1788 but not by diazepam. It was hypothesized that the bulky azido group of Ro15-4513 might overlap with the alcohol binding site, explaining the antiethanol action of Ro15-4513 and the finding that Ro15-1788 cannot inhibit the ethanol effects in rats (Hanchar et al., 2006).
However, only some (Sundstrom-Poromaa et al., 2002; Wei et al., 2004; Fleming et al., 2007; Mody et al., 2007) but not other groups (Borghese et al., 2006; Botta et al., 2007; Korpi et al., 2007; Baur et al., 2009) could confirm the high potency of ethanol for modulation of recombinant α4/α6β3δ receptors. For a critical discussion of this topic, see (Kumar et al., 2004; Krystal et al., 2006; Wallner and Olsen, 2008; Wallner et al., 2014).
3. α6GABAARs and the Alcohol-Nontolerant Rats
The selectively bred alcohol-nontolerant (ANT) rats, in contrast to the alcohol-tolerant rats, carried the point mutation α6R100Q, in which a glutamine replaced the arginine at position 100 of the α6 subunit. The mutated α6(Q100)β2γ2 receptors exhibited a novel high-affinity benzodiazepine binding site, and their function could thus be potentiated by diazepam. The additional modulation by diazepam of a mutated diazepam-insensitive receptor explained the impairments of postural reflexes by diazepam observed in ANT rats (Korpi et al., 1993).
The high ethanol sensitivity of the ANT rats that voluntarily avoid ethanol was then explained by the finding that the already high ethanol sensitivity of recombinant α6β3δ receptors was even further increased in α6(Q100)β3δ receptors (Hanchar et al., 2005; Wallner et al., 2014).
4. α6 Subunit Expression Changes on Chronic Ethanol Administration
Consistent with the modulatory action of ethanol on multiple ion channels, neurotransmitter receptors, and transporters in the brain (Korpi et al., 2015), chronic ethanol administration changed the expression of literally hundreds of different receptors and proteins and also differentially altered the expression of α6 and δ GABAAR subunit mRNAs and peptides in different brain regions of rodents (Kumar et al., 2004; McClintick et al., 2015, 2016).
In the periaqueductal gray, which is involved in processing pain, fear, and anxiety, repeated binge-like alcohol drinking by adolescent alcohol-preferring rats caused an increased expression of α6 and δ and a decrease of γ2 and other GABAAR subunit mRNAs (McClintick et al., 2016).
Serotonin neurons in the dorsal raphe nucleus modulate the effects of ethanol on dopamine release in the shell of the nucleus accumbens, which is directly associated with the rewarding effects of ethanol consumption (Koob, 2013). In the dorsal raphe nucleus, repeated alcohol binges across adolescence of alcohol-preferring rats led to increased α6 and δ mRNAs, whereas γ2 mRNA and seven other GABAAR mRNAs were decreased (McClintick et al., 2015).
In the cerebellum, α6 subunit mRNA and peptide levels were increased, whereas α1 subunit peptide and mRNAs were decreased, after 6 days of ethanol administration in rats (Mhatre and Ticku, 1992; Morrow et al., 1992).
H. α6GABAARs and Animal Models of Trigeminal Nerve-Related Pain
1. α6GABAARs and Trigeminal Ganglia
α6GABAARs are also expressed in some sensory neurons (Gutierrez et al., 1996), including trigeminal ganglia (TG) (Hayasaki et al., 2006; Puri et al., 2012; Fan et al., 2018; Tzeng et al., 2021) (Fig. 2B), but not in dorsal root ganglia (Kogelman et al., 2017). Immunofluorescent studies performed in adult rats demonstrated that the α6 subunit was expressed not only in >74% of small, medium, and large TG neurons but also in satellite glia cells (Puri et al., 2012; Fan et al., 2018). The subunit composition of α6GABAARs within TG neurons and glia is currently unclear, but both α6β2/3γ2 and α6β2/3δ receptors might be present there (Hayasaki et al., 2006).
In TG, GABA is synthesized by glutamate decarboxylase 65 (GAD65) in neurons and accumulated in both neurons and glia (Hayasaki et al., 2006; Tzeng et al., 2021). Cell bodies of mammalian sensory ganglia are generally devoid of synaptic contacts, but exogenous GABA can induce Cl- currents in all TG neurons examined (Hayasaki et al., 2006). It was thus hypothesized that frequently firing sensory neurons can lead to elevated extracellular K+ concentrations during repolarization of neurons between action potentials, which might induce GABA release from satellite glia and neurons (Minchin, 1975; Hayasaki et al., 2006; Fan et al., 2018). The high GABA sensitivity and long open probability of α6GABAARs at TG neurons (Mortensen et al., 2012; Karim et al., 2013) might thus facilitate an efficient feedback inhibition of firing TG neurons (Fan et al., 2018).
2. α6GABAARs and Animal Models of Trigeminal Neuropathic Pain
Trigeminal neuropathic pain is a continuous and burning pain that involves the orofacial region and can be induced by a surgical incision of the facial skin or oral mucosa or the extraction of a tooth. The activation of nociceptive endings in the orofacial tissues by a noxious stimulus results in activation of A-δ and C-fibers, which have their cell bodies in the TG and conduct the information on quality, location, intensity, and duration of pain into the central nervous system (Sessle, 2011). Since only 11% of patients with painful traumatic trigeminal neuropathy had significant pain reduction in response to the available pharmacotherapy by off-label drugs (Haviv et al., 2014; Benoliel et al., 2016), new treatment options are required for this disorder.
It has been demonstrated that a 30% reduction of TG α6 subunits caused hyperalgesia to temporomandibular joint inflammation (Puri et al., 2012; Kramer and Bellinger, 2013). Positive modulation of α6GABAARs may thus reduce pain elicited by TG activation, such as trigeminal neuropathic pain. This was confirmed in a rat model by the finding that the highly α6GABAAR-selective pyrazoloquinolinone DK-I-56-1 but not a related pyrazoloquinolinone that did not modulate α6GABAARs inhibited the development and reduced the established trigeminal neuropathic pain (Vasovic et al., 2019). These data for the first time demonstrated that PAMs of α6GABAARs might represent a mechanistically novel treatment option for trigeminal neuropathic pain (Knutson et al., 2018; Vasovic et al., 2019).
3. α6GABAARs and Animal Models of Migraine
TG are also involved in the pathogenesis of migraine. The ophthalmic branch of trigeminal sensory nerves receives peripheral inputs from dural vessels and sends central projections to the trigeminal cervical complex, which subsequently projects to higher-order pain centers (Bae et al., 2004; Goadsby, 2007). Activation of this trigemino-vascular system (TGVS) via both peripheral and central sensitization and releasing calcitonin gene-related peptide (CGRP) is considered an essential neuropathogenic mechanism of migraine (Goadsby et al., 1988).
Using intracisternally injected capsaicin, an established rat model of TGVS activation mimicking pathologic changes in migraine, it was investigated whether α6GABAAR PAMs can counteract both peripheral and central effects of capsaicin in rats (Fan et al., 2018). Capsaicin is an agonist of vanilloid 1 type of transient receptor potential channels that are coexpressed with CGRP and substance P in small- and medium-sized unmyelinated TG neurons (Caterina and Julius, 2001; Bae et al., 2004). Vanilloid 1 type of transient receptor potential channel activation in TG causes the release of CGRP and induces vasodilation and neurogenic inflammation within the meninges in experimental animals, mimicking migraine (Meents et al., 2010). Capsaicin thus elevated CGRP immunoreactivity in TG (indicating its induced synthesis) and depleted CGRP-immunoreactivity (indicating its induced CGRP release) in the dura mater. Centrally, intracisternally applied capsaicin activates the trigeminal cervical complex by increasing the number of c-Fos immunoreactivity in these neurons. The α6GABAAR-selective PAM, Compound 6 (PZ-II-029), significantly attenuated both central and peripheral capsaicin-induced changes in the TGVS (Fan et al., 2018). All effects of Compound 6 were mimicked by topiramate, a clinically effective antimigraine agent, but not by diazepam. The brain-impermeable furosemide antagonized the peripheral but not central effects of Compound 6, suggesting that α6GABAARs in TG, which are located outside the blood-brain barrier (Eftekhari et al., 2015), are the action target of Compound 6.
This was the first study indicating that positive modulation of α6GABAARs in TG can inhibit nociceptive activation of the TGVS (Fan et al., 2018). In another migraine-mimicking model, TGVS activation was induced in mice via repeated injections of nitroglycerin (NTG) (Pradhan et al., 2014), a nitric oxide donor known to induce migraine in humans (Demartini et al., 2019). Repeated NTG injections induced a significant elevation in the grimace score in mice, a validated behavioral readout of spontaneous pain in rodents (Langford et al., 2010). Immunofluorescent staining of TG showed a significant downregulation of GAD65 but no change in GAT-1 or α6 subunits in repeated NTG- but not saline-treated mice (Tzeng et al., 2021). The α6GABAAR-selective PAMs Compound 6 and DK-I-56-1 significantly aborted migraine-like grimaces in these mice in a manner prevented by furosemide (Tzeng et al., 2021). This study indicates that hypofunction of GABAergic transmission may be the underlying cause of migraine-like responses in mice and supports the concept that α6GABAARs in TG are a novel target for migraine therapy (Tzeng et al., 2021).
IV. The Human GABAA Receptor α6 Subunit Gene and Neuropsychiatric Phenotypes
The human GABAA receptor α6 subunit gene (GABRA6) is located at the human chromosome 5q34 together with the genes for the α1, β2, γ2, and π subunits (Fig. 7) (Olsen and Sieghart, 2008). A variety of genetic studies indicate an association of GABRA6 variants and neuropsychiatric phenotypes.
5q34 locus of GABAAR genes, GABRA6 gene, and localization of the variants: (A) GABAAR genes in the 5q34 locus are depicted to scale reflecting their respective sizes, with arrows showing the direction of the reading frame. The π subunit encoding gene GABRP is located nearby on the 5q35.1 locus (on the right side of GABRG2). The scale is shown on the right-hand side above the scheme. kbp, kilo base pairs. (B) Enlarged view of GABRA6 with nine exons (E1–E9, boxes), eight introns, and two (5′- and 3′-) flanking regions. The scale is indicated on the right-hand side below the scheme. All variants listed in Table 3 are denoted with black lines. Rs3811995 is part of the 5′-upstream sequence and as such is not transcribed. (C) Mature mRNA and the respective protein precursor are depicted schematically and to a scale wherein one codon equals one amino acid. In the mRNA scheme (upper part of the panel), the relative positions of start and stop codons are marked by yellow lines, exon boundaries by white dashed lines, and the transcribed variants in black lines as in (B). The exon numbers are written in white. In the protein precursor (lower part of the panel), the nonsynonymous variants are indicated by extended black lines. The color code on the protein refers to the localization: gray for the signal peptide (S), purple for all extracellular domains (ECD), dark red for the four transmembrane helices (TM1-4), and green for all intracellular domains (ICD). Note that rs3219151 is part of the 3′-untranslated region on exon 9, and Rs13184586 is a synonymous variant that does not cause a replacement of an amino acid. The 5′- and 3′-untranslated regions of the mRNA are important for many processes pertaining to mRNA structure, fate, and translation. The scale bars above and below the schemes indicate the mRNA and protein sizes, respectively. bp, base pairs; AA, amino acids.
So far, 491 unique variants of the GABRA6 transcription unit have been identified and registered in the Ensembl database derived from the 1000genomes project, but only a few of these variants have been used in genetic studies so far. Each variant has a unique Reference Single Nucleotide Polymorphism (SNP) (rs) number (Holmes et al., 2020), which provides a stable variant notation for mutation and polymorphism analysis (https://www.ncbi.nlm.nih.gov/snp/docs/RefSNP_about/). In Table 3, important information on GABRA6 variants that so far have been reported in genetic studies is summarized, and Fig. 7 provides an overview on the localization of the variants that have been investigated.
GABRA6 variants reported in genetic studies
Nucleotide mutation data are taken from dbSNP database (Sherry et al., 2001). Where the mutation resulted in an alteration of the protein sequence, the change is listed under AA consequence. The numbering of amino acids includes the signal peptide [as suggested in The Universal Protein Resource Knowledgebase (UniProtKB)] (Consortium, 2019). The listed publications are separated according to the alternative names used.
A. GABRA6 Variants and Epilepsy
Experimental and clinical observations suggest that epilepsy may result from an imbalance of excitatory and inhibitory neurotransmitter systems, ultimately leading to a decreased net inhibition. So far, genetic variants of GABAAR subunit genes encoding the α1, α6, β2, β3, γ2, δ, and ε subunits have been reported to be associated with human epilepsy (Hirose, 2014; Markus et al., 2020).
The GABRA6 missense variant, rs373363000 C>T, where the Arg46 of the α6 subunit was mutated to Trp46 [α6(R46W); Table 3], was identified as a susceptibility gene that may contribute to the pathogenesis of childhood absence epilepsy (Dibbens et al., 2009). The α6(R46W) mutation is located at the α6+ interface of GABAARs and thus contributes to the α6+β-, α6+γ-, and α6+δ- subunit interfaces in assembled receptors. The α6(R46W) mutation impaired assembly and gating of both α6(R46W)β2γ2L and α6(R46W)β2δ receptors and substantially reduced the current density of both receptors. The α6(R46W) mutation could thus increase seizure susceptibility through a reduction of α6β2γ2 and α6β2δ receptor expression and function (Hernandez et al., 2011).
Two other genetic studies indicated that the GABRA6 rs3219151 T>C variants at the 3′untranslated region (UTR) of GABRA6 at chromosome 5q33.34 (Fig. 7) can influence epilepsy susceptibility (see Table 3 for divergent amino acid or nucleotide numbering in different publications). In a North Indian population comprising 401 epilepsy patients and 202 healthy controls, the frequencies of the CC genotypes were found to be significantly higher in patients with epilepsy than in control subjects (Kumari et al., 2011). These results were confirmed in a South Indian population of 310 patients with idiopathic generalized epilepsy and 310 healthy controls. The C variant of GABRA6 rs3219151 might thus be a risk factor for the development of idiopathic generalized epilepsy (Prasad et al., 2014).
A direct contribution of α6GABAARs to disorders associated with the GABRA6 rs3219151 C variants is supported by the finding that the 3′UTR of mRNAs is the binding target of microRNAs (miRNAs) and RNA binding proteins that regulate mRNA levels in specific subcellular regions and determine the intensity of gene repression (Koulentaki and Kouroumalis, 2018). Binding of miRNAs to the 3′UTR cause mRNA cleavage or translation repression and thus elicit a reduced gene expression. If the C allele of this polymorphism is a risk factor for generalized epilepsy (Kumari et al., 2011; Prasad et al., 2014), it is reasonable to assume that the C allele might have contributed to a novel miRNA binding site and thus have caused a reduced expression of α6GABAARs in the cerebellum or other brain areas. Impaired motor coordination in the cerebellum could contribute to epileptic seizures elicited by other causes.
This conclusion is supported by the finding that the GABRA6 rs3219151 T>C SNP is targeted by at least four miRNAs that regulate gene expression (Gong et al., 2013; Gonda et al., 2017). Two of the four miRNAs were predicted to bind at either GABRA6 rs3219151 T>C allele, whereas one only bound to allele T and another one only bound to allele C. Three of these miRNAs are expressed in the cerebellum and adrenal gland (Gonda et al., 2017). Since miRNAs can interact with either the T- or the C-allele or with both alleles, depending on the regional and cellular miRNA expression, different symptoms of neuropsychiatric disorders can be associated with carriers of either the T- or C-allele or both alleles.
B. GABRA6 Variants and Stress
Stress activates the hypothalamic-pituitary-adrenal (HPA) axis and the sympathetic adrenomedullary system, which are regulated by neuronal pathways, including the GABAergic system. The GABA system plays a key role in the HPA-axis downregulation in response to stress as demonstrated by the strong inhibitory effect on the HPA axis by alprazolam, an anxiolytic benzodiazepine enhancing GABAAR signaling (Giordano et al., 2006; Gonda et al., 2019). Stress and its glucocorticoid component represent a major vulnerability factor for somatic health problems, alcoholism, depression, anxiety disorders, and cognitive impairment. Approximately 60% of the variance in glucocorticoid levels may be attributable to genetic differences (Lynch et al., 2015).
1. GABRA6 Variants, Stress, and Heart Rate
In a longitudinal study (Lynch et al., 2015), 186 high-risk urban children from New York were recruited at age 4. The presence or absence of child maltreatment was determined, as was the amount of crime that occurred in their neighborhood during the previous year. At age 9, the children were brought to the laboratory where their physiologic responses [i.e., changes in the amplitude of respiratory sinus arrhythmia (RSA) to a video-induced cognitive challenge] was assessed, and their DNA samples were collected for subsequent genotyping. RSA is an index of autonomic cardiac regulation and can be indirectly assessed by measuring the high-frequency component of heart rate variability. High levels of RSA and its reliable suppression or reduction are positive indices of social and emotional regulation. Thus, withdrawal of the vagal break in response to challenge shifts autonomic regulation of the heart from parasympathetic to sympathetic control to provide the cardiac and metabolic outputs necessary for behavioral mobilization. When the challenge has passed, vagal pathways restore parasympathetic control of the heart to promote calm states and allow for social engagement. Low levels of RSA and its unreliable modulation appear to increase the risk for difficulties in social and emotional regulation and in some cases may contribute to psychopathology. Multiple studies support the link between aberrant RSA reactivity patterns and risk for both internalizing (anxiety, doubts, compulsive behavior) and externalizing problems (inadequate behavior, aggression) (Lynch et al., 2015).
In this study, GABRA6 rs3219151 T>C variant carriers exhibited different stress responses (Lynch et al., 2015). TT- or TC-carriers had significantly higher cortisol and blood pressure responses to social stressors than CC-carriers. Among CC-carriers, only nonmaltreated children demonstrated the ability to modulate RSA in response to challenge. In contrast, both nonmaltreated and maltreated children carrying TT- or TC-alleles were able to modulate their RSA levels in response to challenge. It was concluded that homozygosity for the C-allele of GABRA rs3219151 may contribute to risk for psychopathology among neglected and multiply maltreated children due to their inappropriate reactions to stress (Lynch et al., 2015). A reduced expression of α6GABAARs in certain brain regions caused by the miRNA binding site of the C allele of GABRA6 rs3219151 (see above) might thus have resulted in a reduced inhibition via α6GABAARs and could explain this risk for psychopathology.
2. GABRA6 Variants, Stress, and Obesity
In a study on 284 unrelated Swedish men exposed to perinatal stress at the end of the second world war, evidence was provided that cortisol secretion and abdominal obesity are associated with the GABRA6 rs3219151 T>C variants (Rosmond et al., 2002). T-allele carriers had a significantly higher waist-to-hip ratio and abdominal sagittal diameter compared with CC-carriers. TT-carriers had significantly higher mean salivary cortisol levels and had significantly higher diurnal cortisol secretion than TC-carriers. These findings suggest that T-allele carriers exhibit a predisposition to hypercortisolism and perhaps abdominal obesity (Rosmond et al., 2002).
3. GABRA6 Variants, Stress, and Personality Traits
An association of the GABRA6 rs3219151 T-allele and hypercortisolism was also supported by another study (Uhart et al., 2004). In an association study with 56 healthy subjects, the T-allele carriers were found to have greater stress responses, including the elevation of adrenocorticotrophic hormone, cortisol, and blood pressure than CC-carriers in the Trier Social Stress test. T-allele carriers exhibited a personality tending to “extraversion behavior” (Uhart et al., 2004) that takes pleasure in activities involving social interactions and community tasks, whereas individuals with “introversion behavior” prefer to spend time alone rather than engage in social interactions.
4. GABRA6 Variants, Stress, Externalizing Behavior, and Substance Use
Another SNP of GABRA6, the 5′UTR GABRA6 rs3811995 C>T variant (Table 3), was reported to be associated with externalizing behavior and risk of adolescent substance use (Trucco et al., 2016). Subjects were collected from the Michigan longitudinal study, including 487 adolescents from 292 families with high risk of substance use, mostly of European American ancestry. Their temperament traits “behavioral control” and “resiliency” at ages 9–11 were assessed using interviewer ratings, and externalizing behaviors at ages 12–14 were assessed using teacher ratings. Substance uses including maximum alcoholic beverages/24 hours, and frequency of past year cigarette and marijuana use were self-reported at ages 15–17. Results indicated that those with the GABRA6 rs3811995 CC genotype had lower behavioral control and greater externalizing behavior and were associated with a higher liability of substance use (Trucco et al., 2016).
Mutations in the 5′UTR of a gene can potentially affect the process of transcription, whereby RNA polymerase is recruited to the gene. Variation in the DNA sequence may alter the affinities of existing protein-DNA interactions or even recruit new proteins to bind to the DNA, altering the specificity and kinetics of the transcriptional process (Kumari et al., 2011). Thus, mutations in the 5′UTR also can change the expression of a gene and causally contribute to disorders.
5. GABRA6 Variants, Stress, and Perceived Parenting
Early adverse experiences are associated with increased stress, and the way we perceive our parents strongly influences our lifelong relationship with them and our emotional and behavioral functioning. The dopamine D2 receptor is involved in the modulation of social behaviors and might influence the perception of parenting. A possible association of the exon 8 of the dopamine D2 receptor gene DRD2 and the Pro385Ser variant of GABRA6 rs34907804 (Table 3) and perceived parental rearing behavior were studied in 207 unrelated adults (Lucht et al., 2006). Subjects with self-reported rejection of parents and father rejection showed particularly high scores in DRD2 (exon 8) AA genotypes when the α6Ser385-positive genotype was also present. These associations were restricted to the female subgroup. Female GABRA6 Ser385-positive subjects might thus be at particular risk for detrimental effects from early adverse experiences (Lucht et al., 2006).
The point mutation α6Pro385Ser is located in the large intracellular loop of the α6 subunit and might represent a novel phosphorylation site (Iwata et al., 2000). Phosphorylation of this α6Ser385 site could influence the functional properties of α6GABAARs as well as their interaction with the cytoskeleton and their trafficking, internalization, and overall expression. In any case, this mutation presumably causes a significant change in the structure and flexibility of the intracellular loop of α6GABAARs and its interaction with intracellular proteins.
6. GABRA6 Variants, Stress, Anxiety, Depression, and Suicide Risk
In a European sample of 2283 subjects, the effects of the GABRA6 rs3219151 T>C variants in interaction with recent negative life events were investigated on incidences of lifetime and current depression and current anxiety as well as lifetime suicide (Gonda et al., 2017). No main effects of the GABRA6 rs3219151 variants on these disorders were detected. However, in those with recent negative life events, the T-allele carriers showed increased current anxiety and depression as well as specific elements of suicide risk, including suicidal and death-related thoughts, hopelessness, restlessness, agitation, insomnia, and impulsiveness. These results indicate that stress-associated suicide risk is elevated in T-allele carriers (Gonda et al., 2017). Considering that the GABRA6 rs3219151 T-allele might be associated with extraversion behavior that requires high motivation and decisiveness, it might possibly also facilitate the last impulse for suicidal behavior as a reaction to a hopeless situation.
C. GABRA6 Variants and Mood Disorders
1. GABRA6 Variants and Neuroticism
A tendency to experience negative affect, as measured by the neuroticism component of the Neuroticism, Extraversion, and Openness Personality Inventory, is a trait marker for major depression. A serotonin transporter promoter polymorphism (5-HTTLPR) has been extensively studied in neuroticism and several psychiatric disorders. A study with 384 subjects who completed the Neuroticism, Extraversion, and Openness Personality Inventory demonstrated that the 5-HTTLPR short allele and the rs34907804 α6Pro385 allele were associated with higher neuroticism scores (Sen et al., 2004).
2. 5q34 GABAAR Gene Cluster and Depression
After a linkage study identified the chromosomal area 5q34 as a susceptibility region for mood disorders (Edenberg et al., 1997), in the mouse a similar location on chromosomes was demonstrated to be associated with depression-like behaviors based on the forced-swim and tail-suspension tests (Yoshikawa et al., 2002).
Later, a study on 203 Japanese patients with mood disorder (Yamada et al., 2003) demonstrated significant associations of GABRA1 and GABRA6 polymorphisms with mood disorders in females but not males. The female T-allele carriers of the GABRA6 rs3219151 variant were more affected by unipolar or bipolar disorders than female controls (Yamada et al., 2003).
3. GABRA6 Variants and Harm Avoidance
In a study investigating 937 Spanish individuals, the TT-carriers of the GABRA6 rs3219151 T>C polymorphism presented slightly higher scores for harm avoidance than C-allele carriers (Arias et al., 2012). This is consistent with the more active extraversion behavior of T-allele carriers.
4. GABRA6 Variants, Major Depression, and Panic Disorder
Associations of GABRA6 rs3219151 variants with major depression and panic disorder were studied in 280 and 189 Japanese patients, respectively (Inoue et al., 2015). The C-allele was significantly associated with major depression and panic disorder in genotype and allele analyses, which is consistent with the more introversive behavior of the C-allele carriers. However, in panic disorder patients exposed to a dynamic fearful face during fMRI scanning, responses in the bilateral anterior cingulate and medial orbital frontal cortex were stronger in T-allele carriers than in CC-carriers (Inoue et al., 2015). This finding is consistent with the higher stress reactivity of T-allele carriers discussed above.
A study with a Spanish cohort showed an association between nonacrophobic panic disorder and the miRNA miR-138-2. This miRNA is highly enriched in the brain, including the frontal cortex, the hippocampus, and the midbrain, and reduced, among other genes, also GABRA6 expression by 30% (Muinos-Gimeno et al., 2011). Such a reduction could have significantly influenced the amounts of α6GABAARs expressed in the respective brain region quite well, causing disturbances in the behavioral tasks mediated by these receptors.
5. GABRA6 Variants and Bipolar Disorder
A genome-wide association study investigating 200 individuals from 41 families multiply affected with bipolar disorders identified several rare variants in ion channel genes that influence risk for bipolar disorder. Among these, the missense variant in GABRA6 rs3811993 resulting in a threonine-to-methionine substitution at position 187 of the α6 subunit (Table 3) was identified to be significantly associated with bipolar disorder and to be among the top 0.1% most deleterious SNPs genome-wide (Ament et al., 2015).
6. GABRA6 Variants and Schizophrenia
A meta-analysis of 20 genome-wide linkage scans in patients with schizophrenia identified the chromosome 5q23.2-q34 as the second most significant risk locus in the genome (Lewis et al., 2003). A linkage to an overlapping region at chromosome 5q31-q35 was then reported in a genome-wide scan of schizophrenia and schizoaffective depressed type families of Portuguese descent (Sklar et al., 2004). In a subsequent study (Petryshen et al., 2005), significant associations were detected between schizophrenia and 19 SNPs spanning the 5q34 GABAAR gene cluster in a Portuguese patient sample of 321 cases and 242 controls. The majority of associated SNPs were in GABRA1, with some in GABRB2, GABRA6, and GABRP (Petryshen et al., 2005). Twenty-one haplotype blocks (sets of genetic determinants located on a single chromosome) of strong linkage disequilibrium of these SNPs were defined, spanning the whole gene cluster. Significant associations were detected with several haplotypes in the gene cluster. These findings were replicated in the second stage in an independent German family-based sample consisting of 238 parent-proband trios (Petryshen et al., 2005). A comparison of transcript levels of 32 patients with schizophrenia carrying the schizophrenia-associated GABRA1 haplotypes and unaffected siblings of this study in peripheral leukocytes found a significantly lower patient transcript expression of GABRA6 and genes coexpressed in haplotypes, including some synaptic and vesicle-associated genes that have been reported to be altered in schizophrenia prefrontal cortex. These results support the involvement of the chromosome 5q34 GABAAR gene cluster in schizophrenia and suggest that schizophrenia-associated haplotypes may alter the expression of GABRA6 as well as that of GABA-related genes (Petryshen et al., 2005).
In a case control study carried out in 959 control subjects and 598 unrelated schizophrenia cases of Han Chinese origin (Gong et al., 2013), it was demonstrated that GABRA6 rs3219151 C>T was significantly associated with schizophrenia susceptibility. According to Table 2 in Gong et al. (2013), the C-allele was associated with an increased risk for schizophrenia. Assuming that the C-allele of GABRA6 rs3219151 causes a reduced expression of the GABRA6 gene as discussed above, these data are consistent with the (Petryshen et al., 2005) study, indicating that a reduced expression of α6GABAARs might be associated with schizophrenia.
D. GABRA6 Variants and Alcohol-Associated Disorders
1. GABRA6 Variants, Alcohol, and Benzodiazepine Sensitivity
Children of alcoholics are at a high risk of alcoholism. In a study investigating the GABRA6 rs34907804 (α6Pro385Ser) variant in 51 adult Caucasian offspring of alcoholics, the sensitivity to diazepam was assessed by using diazepam-induced changes in two eye movement measures: peak saccadic velocity and average smooth pursuit gain. The α6Ser385 carriers were less sensitive to diazepam and alcohol (Iwata et al., 1999). The same conclusion was supported by another study with 100 patients (Hoffman et al., 2002). The α6Ser385 mutation did not affect baseline sedation and anxiety but significantly decreased the anxiolytic effect of low-dose midazolam (Hoffman et al., 2002). These findings support the assumption that α6GABAARs may modulate anxiety within specific regions of the central nervous system as suggested in (Leggio et al., 2015).
In a pilot study it was investigated whether a low level of response to alcohol in children of alcoholics predicts future alcohol abuse and dependence (Schuckit et al., 1999). Results showed that subjects with the α6Ser385 genotype were significantly more vulnerable to become alcoholics than the α6Pro385 genotype and a genotype of a serotonin transporter polymorphism. All subjects with combined “low level of response to alcohol at age 20” and the α6Ser385 genotype had developed alcoholism and demonstrated the lowest alcohol sensitivity during adolescence overall (Schuckit et al., 1999).
In a subsequent study, the authors strengthened the support for a relationship between the serotonin transporter polymorphism and α6Ser385 alleles to “low level of response to alcohol” and to alcoholism in a prospectively studied cohort (Hu et al., 2005). The α6Ser385 mutation might thus reduce ethanol sensitivity in children of alcoholics (Iwata et al., 1999), but enhance their vulnerability for alcoholism in later life.
2. 5q34 GABAAR Gene Cluster and Alcohol Dependence
Multiple association studies suggest a role for the GABAAR gene cluster on chromosomal 5q34 in alcohol-associated disorders. Thus, in 108 patients and 58 controls from Scotland, it was demonstrated that nearly 70% of alcohol-dependent subjects with Korsakoff psychosis carried the T-allele at GABRA6 rs3219151 T>C and the C-allele at the BanI RFLP of the GABRB2, whereas 62% of patients and 45% of controls carried these alleles (Loh et al., 1999).
This finding could not be confirmed by another study investigating 349 German alcohol-dependent subjects and 182 controls, but the frequency of the T-allele carriers of the GABRA6 rs3219151 T>C variants was significantly increased in dissocial alcoholics compared with the controls (Sander et al., 1999). Nevertheless, in a study with 162 male Korean alcoholics and 172 controls, there was a significant association between genetic polymorphisms of the GABRA1 and GABRA6 genes and alcoholism (Park et al., 2006).
Another study (Radel et al., 2005) conducted genotyping for 6 SNPs at the 5q34 GABAA gene cluster in 433 Southwestern Native American and 511 Finnish alcohol-dependent subjects and unaffected individuals. Results showed an association of alcohol dependence with the GABRA6 rs3219151 T- and the GABRB2 1412 T-alleles in both populations. Linkage disequilibrium mapping with haplotypes yielded evidence for an alcohol-dependence locus at the GABAA gene cluster region in both populations. The most highly significant signals were at 3-locus haplotypes that included one or more GABRA6 variants (Radel et al., 2005).
A study investigating 141 alcohol-dependent cases and 110 volunteers from India (Sahni et al., 2019) showed that GABRA6 rs3219151 variants were significantly associated with age at first use and daily ethanol consumption. TT-carriers were more liable to alcohol dependence, had a lower age at first use, higher daily intake, and longer duration of dependence (Sahni et al., 2019).
3. GABRA6 Variants and Treatment Responses of Alcoholism
In 98 male alcoholic inpatients treated with lorazepam and thiamine, craving for alcohol and food was investigated (Han et al., 2008). Body weight increased in all patients with alcohol dependence. The GABRA6 rs3219151 T-allele carriers had greater changes in alcohol and food craving than CC-carriers. Body weight and craving for food significantly increased in T-allele carriers, and these parameters correlated with a reduced craving for alcohol (Han et al., 2008).
In a randomized double-blind, placebo-controlled study, 126 alcohol-dependent subjects were treated with either naltrexone or acamprosate. Both are effective medications, but the effect sizes are modest (Ooteman et al., 2009). Results indicated that acamprosate might be a better treatment option for CC-carriers of the GABRA6 rs3219151 variant, whereas naltrexone outperformed acamprosate in reducing the urge to drink alcohol in TT-carriers. Both treatments displayed equal efficacy in subjects heterozygous for this polymorphism.
The finding that some genetic indicators have good matching potential in this study may be caused by their close relationship to the neurobiological disease process (Gottesman and Gould, 2003; Ooteman et al., 2009). Thus, patients carrying the C-allele might exhibit a reduced expression of α6GABAARs, as discussed above, and might benefit from reducing the overactive glutamatergic system of the glutamatergic/GABAergic negative reinforcement system by acamprosate, whereas those carrying the T-allele might exhibit standard expression of α6GABAARs and benefit from reducing the dopaminergic/opioidergic positive reinforcement system by the opiate antagonist naltrexone. If these considerations were correct, patients who are alcohol-dependent and carrying the C-allele of the GABRA6 rs3219151 T>C polymorphism might benefit from the administration of α6GABAAR PAMs.
E. Possible Contribution of GABRA6 Variants to Neuropsychiatric Phenotypes
Most of the GABRA6 variants so far used for genetic studies cause a changed expression and altered function of the GABRA6 gene. Mutations in the coding region of the GABRA6 gene (Table 3), such as rs373363000 (α6R46W), that elicit an impaired assembly and gating of α6(R46W)β2γ2L and α6(R46W)β2δ receptors, result in a reduced inhibitory activity in all brain regions where these receptors are expressed and can thus directly increase seizure susceptibility (Hernandez et al., 2011). Mutations causing a replacement of an amino-acid residue of α6GABAARs, such as rs34907804 (α6Pro385Ser) or rs3811993 (α6Thr187Met), can cause changes in the expression or function of receptors and thus directly contribute to the respective disorders. Mutations in the coding regions without change in protein sequence, such as rs13184586 (α6Ala335Ala), or missense mutations that do not cause altered protein function might have more subtle effects or might associate the GABRA6 gene with a disorder through their linkage with another disorder-associated mutation in their proximity.
In addition, mutations in the 5′UTR or 3′UTR can directly cause a reduced expression of the GABRA6 gene. Mutations in the 5′UTR, such as rs3811995 C>T, can interfere with the recruitment of RNA polymerase, thus altering the specificity and kinetics of the transcriptional process (Kumari et al., 2011). Mutations in the 3′UTR region, such as rs3219151 T>C, elicit an altered interaction with specific miRNAs resulting in a reduced expression of the GABRA6 gene in only those brain regions expressing the respective miRNAs (Gonda et al., 2017).
Finally, mutations close to the coding region can also cause changes in the expression of α6GABAARs, although the actual gene associated with the disorder or phenotype might be different from and located in the vicinity of the GABRA6 gene. This is exemplified by studies demonstrating an association of the 5q34 GABAAR gene cluster with schizophrenia in which a reduced expression of the GABRA6 gene as well as that of other genes coexpressed in a 5q34 GABRA1 haplotype were demonstrated (Petryshen et al., 2005).
How could the GABRA6 gene contribute to multiple neuropsychiatric disorders or phenotypes? Genetic studies performed with the rs3219151 T>C variants seem to be most informative for answering this question because they are more abundant than those using other variants (Table 3), and they suggest that both T- and C-alleles are associated with different phenotypes of related neuropsychiatric conditions. This is consistent with the concept that most neuropsychiatric disorders are polygenic, and thus, many genes can contribute to individual symptoms and phenotypes.
Stress is a major vulnerability factor for a variety of neuropsychiatric disorders (Lynch et al., 2015), and genetic studies in humans discussed in Section IV suggest extensive links between GABRA6 and stress response. T-allele carriers of the GABRA6 rs3219151 variants exhibit higher cortisol and higher RSA levels as compared with C-allele carriers. In spite of their greater stress responses, however, high RSA levels are positive indices of social and emotional regulation (Lynch et al., 2015). Thus, T-allele carriers might be able to better compensate various stress situations than C-allele carriers, as indicated by their tendency to harm avoidance (Arias et al., 2012). In addition, their extroversion behavior might allow them to better overcome stress situations by exaggerating eating and drinking behavior in a social context, leading to obesity (Rosmond et al., 2002) or alcohol abuse (Section IV.E.2). In the absence of genetic, epigenetic, or environmental vulnerability factors, mild to moderate stress might not cause the development of neuropsychiatric phenotypes in T-allele carriers. But in the presence of appropriate vulnerability factors and/or depending on the specific types of stress they are exposed to (Gonda et al., 2019), T-allele carriers might develop anxiety or depression (Yamada et al., 2003; Gonda et al., 2017) or tend to bipolar disorders (Yamada et al., 2003). In contrast, C-allele carriers of the GABRA6 rs3219151 variants exhibit lower cortisol and RSA levels than T-allele carriers (Lynch et al., 2015), and their already reduced α6GABAAR expression in certain brain regions might make them more vulnerable to stress, consistent with animal experiments indicating an enhanced stress response under conditions of reduced α6GABAARs expression (Yang et al., 2016; Rudolph et al., 2020). Under appropriate stress conditions they might tend to panic disorder and major depression (Inoue et al., 2015), schizophrenia (Gong et al., 2013), or seizures (Kumari et al., 2011; Prasad et al., 2014). The GABRA6 variant-dependent expression of α6GABAARs together with genetic, epigenetic, and environmental vulnerability factors might thus facilitate the development of different phenotypes of neuropsychiatric disorders. Differences in stress processing and in addiction risk have also been observed for the less studied GABRA6 variants, adding evidence to the emerging picture that GABRA6 variants influence stress processing.
V. Post-Mortem Changes of α6 Subunit Expression in Neuropsychiatric Disorders
Post-mortem changes in the α6 subunit expression in patients with neuropsychiatric disorders could have been elicited by a genetic predisposition or by life events and thus could be the cause or consequence of the disorder. Alternatively, such changes also could have been elicited by the often long-term medication of the patients.
A. α6 Subunit Changes in Schizophrenia and Mood Disorders
Deficits in GABA signaling have been described in the cerebellum, prefrontal cortex, and limbic system in individuals with schizophrenia (Table 2). In a post-mortem study investigating lateral cerebellar hemisphere tissue of 13 males with schizophrenia and 13 matched controls, patients had significantly lower mRNA and protein levels of GAD67, GAD65, or GAT-1 and increases in GABRA6, and GABRD mRNAs (Fatemi et al., 2005; Bullock et al., 2008). These findings suggest a reduced GABAergic transmission at the Golgi-granule cell synapse and a compensatory upregulation of α6GABAARs in individuals with schizophrenia (Bullock et al., 2008).
In another study (Fatemi et al., 2013), the mRNA and protein expression of 12 GABAA subunit proteins (α1, α2, α3, α5, α6, β1, β2, β3, δ, ε, γ2, and γ3) in the lateral cerebella from 9–15 subjects with schizophrenia, 10–15 subjects with bipolar disorder, 12–15 with major depression without psychotic features, and 10–15 healthy controls were investigated. Among other findings, significant increases in protein levels of α1, α6, and γ3 receptors in lateral cerebella of subjects with major depression were demonstrated with no changes in either schizophrenia or bipolar disorder. The absence of changes in α6 subunit expression in schizophrenia was a finding discrepant to that of Bullock et al. (2008) and was potentially explained by a different anatomic location of the cerebellar tissue used for both sets of experiments and regional differences in GABA subunit expression throughout the cerebellum (Fatemi et al., 2013).
The protein expression of 14 GABAA receptor subunits was also investigated in the superior frontal cortex of subjects with schizophrenia, bipolar disorder, and healthy controls (Fatemi et al., 2017). Results indicated a higher expression of GABAA α1, α6, and ε subunits and a lower expression of β1 subunits in subjects with schizophrenia and bipolar disorders versus controls. Thus, impaired GABAergic transmission may negatively contribute to symptoms, such as anxiety or panic as well as impaired learning and information processing, all of which are disrupted in schizophrenia and bipolar disorder (Fatemi et al., 2017).
B. α6 Subunit Changes in Autism
In a similar study, protein and mRNA levels for nine GABAAR subunits were measured in cerebellum, superior frontal cortex, and parietal cortex of subjects with autism versus matched controls. It was demonstrated that there was a downregulation of α6, β2, δ, ε, γ2, θ, and ρ2 protein subunits in the superior frontal cortex of subjects with autism (Fatemi et al., 2014).
C. α6 Subunit Changes in Alcohol Use Disorder
Subjects with excessive alcohol drinking for an average of 35 years had lower mRNA and protein levels of δ subunits but unchanged α1 and α6 subunits in their postmortem cerebella (Gatta et al., 2017) and lower mRNA levels of α6, α2, and δ subunits in the prefrontal cortex as compared with the control group (Gatta et al., 2020).
VI. Discussion
In the present review we for the first time summarize the importance of α6GABAARs for the function of the cerebellum and their possible involvement in various neuropsychiatric disorders. α6βγ2 and α6βδ receptors, because of their synaptic and extrasynaptic location at the GABAergic Golgi cell-granule cell synapse where they mediate phasic and tonic inhibition, respectively, play a crucial role in cerebellar function. Their high GABA sensitivity and long open probability make them ideally suited for sharpening input signals and thus for discriminating between multiple and complex mossy fiber inputs in cerebellum (see Section II.A). Because of the role of the cerebellum in motor control, mutations causing changes in the expression or function of α6GABAARs can thus cause impaired motor control and increased susceptibility to epileptic seizures in humans (Hernandez et al., 2011; Kumari et al., 2011; Prasad et al., 2014).
In addition, the substantial prenatal growth that is continued postnatally and the associated prolonged neurodevelopment render the cerebellum especially vulnerable to developmental disruptions and damage elicited by genetic or environmental causes, such as stress or ethanol exposure (D'Mello and Stoodley, 2015). For instance, alterations in cerebellar structure, connectivity, and function, have been identified by neuroimaging or postmortem studies (for references see Table 2) in patients suffering from Angelman syndrome, Down syndrome, essential tremor, tic disorders, ADHD, OCD, Huntington disease, schizophrenia, autism spectrum disorder, stress-associated disorders, depression, and alcohol use disorder. Cerebellar abnormalities also implicate changes in the expression or functional balance of α6GABAARs. Thus, impaired motor control in animal models of Angelman syndrome (Egawa et al., 2012) or essential tremor (Handforth et al., 2018) could be alleviated by THIP, a potent agonist of extrasynaptic α6βδ GABAARs in the cerebellum (Table 2).
The cerebellum, however, is not only involved in motor regulation but also in cognitive, emotional, and social behaviors (see Section II.B). Depending on the exact nature and location of the genetic or developmental disruptions of the cerebellar circuits and the resulting local dysbalance of α6GABAARs function, different brain regions will be influenced via the closed-circuit connections of the cerebellum with the whole brain, contributing to different symptoms of individual disorders.
Stress represents a major vulnerability factor not only for somatic health problems but also for alcoholism, depression, and other neuropsychiatric disorders (Lynch et al., 2015). In rodents, stress can modulate miRNA expression in pregnant females, leading to altered transcriptomic brain profiles in their offspring (Rinaldi et al., 2010; Zucchi et al., 2013). Similar miRNA alterations are also observed in human psychiatric disorders, and transcriptomic changes in the offspring include genes related to neurodevelopment and axonal guidance. miRNAs can also reduce the expression of α6 subunits (Gong et al., 2013; Gonda et al., 2017). In adolescent rats, chronic maternal separation stress caused a reduced α6 expression in the hippocampus and depressive behaviors (Yang et al., 2016) that also could be elicited in young naive control animals by α6-knockdown in the hippocampus (Yang et al., 2016). Maternal separation stress also elicited ADHD-like behaviors in rodents (Kwak et al., 2009; Womersley et al., 2015; Yang et al., 2016), and a reduced tonic current at cerebellar granule cells was observed in a mouse model of ADHD (Kim et al., 2017).
Stress thus modifies epigenetic signatures linked to disease during critical periods of fetal and postnatal development (Zucchi et al., 2013). In humans, stress during pregnancy, such as family discord, divorce, or stress in early-life, such as disturbances of the infant-parent relationship, social separation, child neglect, or abuse, also can negatively influence brain development of the offspring and increase the risk of behavioral, emotional, and cognitive problems and later-onset disorders, such as ADHD, autism, and schizophrenia (Ronald et al., 2011; Gunn et al., 2015; Beversdorf et al., 2019). Thus, several prospective longitudinal studies in humans reported a significant association of antenatal anxiety or perceived stress during pregnancy with ADHD symptoms as well as autistic traits in school-aged children from community samples (Ronald et al., 2011). Although the underlying pathologic mechanisms causing ADHD currently are not known, a decreased volume of the superior cerebellar vermis was observed in MRI studies in 36 children with ADHD (Mackie et al., 2007), and a meta-analysis of fMRI studies in patients with ADHD indicated a reduced activation in areas of timing, such as the left inferior prefrontal cortex/insula, cerebellum, and left inferior parietal lobe (Hart et al., 2012).
These considerations are consistent with emerging evidence from animal and human studies suggesting that the cerebellum also plays a role in stress responses (Moreno-Rius, 2019) in fear and anxiety-related disorders (Moreno-Rius, 2018), autism spectrum disorder, schizophrenia, and addiction (Carta et al., 2019). Human postmortem studies indicated a decreased GAD65 and GAD67 expression in the cerebellar cortex and parietal cortex of autistic subjects (Fatemi et al., 2002b; Yip et al., 2007) and changes in α6-subunit expression in autism, schizophrenia, and mood disorders (Table 2). Whereas α6 subunits were downregulated in the superior frontal cortex of subjects with autism (Fatemi et al., 2014), an upregulation of α6 subunits was observed in the cerebellum and superior frontal cortex of subjects with schizophrenia and mood disorders (Bullock et al., 2008; Fatemi et al., 2013, 2017). Interestingly, the α6 subunit changes in the cerebellum of subjects with schizophrenia were similar to those observed in rats chronically treated with the N-methyl-D-aspartate (NMDA) channel blocker PCP, which elicits schizophrenia-like symptoms in humans as well as in animal models (Bullock et al., 2009).
Stress-related psychiatric disorders, such as anxiety and major depression are strongly associated with alcohol use disorders (Rompala et al., 2018), and alcohol-dependent patients exhibit an altered stress response (Chen et al., 2020). In addition, extensive ethanol exposure of pregnant rats can elicit neurodevelopmental changes in offspring that are similar to those of adverse early life experiences in humans and can lead to long-lasting alterations suggesting a cerebellar involvement, including deficits in fine motor skills and motor learning (Diaz et al., 2014) as well as ADHD behavior in humans and rats (Wang et al., 2020). Chronic ethanol consumption caused a compensatory upregulation of α6GABAAR subunits in the cerebellum of rodents (Section III.G.4) but no change in humans (Section V.C.), whereas other GABAAR subunits were downregulated in both species, supporting the importance of α6 subunits for the function of the brain.
All these long-lasting pathologic consequences of prenatal and antenatal stress and ethanol consumption in rodents and humans can be enhanced or reduced by genetic predisposition. Most neuropsychiatric disorders can be influenced by multiple genes, and the individual genes usually have only a small contribution to the overall expression of the disorder. Depending on the genetic variants, the phenotypes of the disorders can be different. GABRA6 gene variants can contribute to such genetic predispositions. Multiple genetic studies in humans discussed in this review suggest an association of GABRA6 variants with stress-induced somatic symptoms as well as with personality, temperament, perceived parenting, alcohol-associated disorders, mood disorders, and schizophrenia. Most of these studies were performed with the GABRA6 rs3219151 T>C variant, wherein the C-allele results in a novel miRNA binding site and thus a reduced expression of α6 subunits and hence of α6GABAARs in brain regions expressing the C-allele together with its interacting miRNA. Although α6GABAARs are not the main and only cause of these multigenetic disorders, a changed expression of α6GABAARs could have influenced the susceptibility for these disorders.
Involvement of α6GABAARs and their potential as drug targets for a variety of neuropsychiatric disorders is further supported by the recent identification of PAMs with a high functional selectivity for modulating α6βγ2 (and α6βδ) receptors (Fig. 5) (Varagic et al., 2013; Chiou et al., 2018; Knutson et al., 2018) and their first application in animal models of neuropsychiatric disorders (Chiou et al., 2018; Cadeddu et al., 2021). α6GABAARs expressed in TG are part of an inhibitory feedback system that is mediated by GABA released from neurons and glia cells when the TGVS is activated (Fan et al., 2018). Again, the high GABA-sensitivity of these receptors and their long open probability make them ideally suited for this function. α6GABAAR PAMs enhance the GABAergic feedback inhibition of TG neurons in a furosemide-sensitive way and can ameliorate not only trigeminal neuropathic pain (Vasovic et al., 2019) but also migraine pain (Fan et al., 2018; Tzeng et al., 2021) with an efficacy comparable to clinically used antimigraine agents.
In addition, systemic or intracerebellar application of α6GABAAR PAMs could rescue PPI deficits in animal models of tic disorders [Wu et al., 2016; Master thesis, Graduate Institute of Pharmacology, College of Medicine (Chiou LC, Adviser) p 77, National Taiwan University, Taipei, Taiwan; (Cadeddu et al., 2021)] as well as in animal models of schizophrenia in a furosemide-sensitive way (Chiou et al., 2018), supporting a mechanism of action via α6GABAARs. PPI deficits are sensorimotor gating deficits that are thought to reflect abnormalities in the processing and integration of sensory and motor information (Ahmari et al., 2012). The sensory overload resulting from reduced sensorimotor gating is thought to give rise to cognitive fragmentation, attentional deficits, and some of the complex clinical symptoms associated with these disorders (Giakoumaki et al., 2008). PPI deficits not only are observed in schizophrenia (Braff et al., 1978) and tic disorders, including Tourette syndrome, but also in patients with a variety of other neuropsychiatric disorders (Table 2), such as ADHD (Ornitz et al., 1992), OCD (Swerdlow et al., 1993), Huntington disease (Swerdlow et al., 1995), premenstrual dysphoric disorder (Kask et al., 2008), post-traumatic stress disorder (Pineles et al., 2016; Meteran et al., 2019), panic disorder (Ludewig et al., 2002), nocturnal enuresis (Freitag et al., 2006), mania in bipolar disorder (Giakoumaki et al., 2007), and autism spectrum disorder (Cheng et al., 2018).
Patients suffering from schizophrenia display subtle cognitive abnormalities that may reflect a difficulty in rapidly coordinating the steps that occur in a variety of mental activities. The cerebellum may play a role in coordinating both motor and cognitive performance (Andreasen et al., 1996). In addition, a breakdown of connectivity in a specific dorsolateral prefrontal cortex-to-cerebellum network directly corresponded to negative symptom severity in schizophrenia, and restoration of network connectivity with transcranial magnetic stimulation corresponded to amelioration of negative symptoms (Brady et al., 2019). In mouse models mimicking schizophrenia, α6GABAAR PAMs not only rescued PPI deficits (Chiou et al., 2018) but also reduced social withdrawal and cognitive deficits (Mouri et al., 2020) (Chiou et al., 2019, Abstract and Poster No. 645.14/B55 at the Soc Neurosci Meeting 2019, Chicago, IL, USA). These findings are supported by recent evidence indicating that the reward circuit and social behavior in mice can be directly modulated by an excitatory projection from the cerebellum to the ventral tegmental area (Carta et al., 2019).
Fronto-parieto-cerebellar deficits (Hart et al., 2012) and decreased functional connectivity between the cerebral cortex, striatum, and the cerebellum (Anticevic et al., 2014; Xu et al., 2019) were also observed in ADHD and OCD, respectively. Interestingly, for all disorders with PPI deficits, abnormalities in structure and function of the cerebellum have been reported that might also have caused a dysbalance in the expression or function of α6GABAARs (Table 2). Such a dysbalance can explain a possible benefit of α6GABAAR PAMs in the treatment of a variety of neuropsychiatric disorders. This does not necessarily mean that sensorimotor gating deficits are exclusively caused by the cerebellum. α6GABAARs are also expressed in the prefrontal cortex (Agrawal and Dwivedi, 2020; Gatta et al., 2021), the superior frontal cortex (Fatemi et al., 2014), or the cochlear nerve (Gomez-Nieto et al., 2008) and in many other brain regions (Table 1) that also might contribute to sensorimotor gating deficits or their amelioration by α6GABAAR PAMs.
Thus, experimental and pharmacological studies in animal models of neuropsychiatric disorders as well as genetic, neuroimaging, behavioral (PPI), and post-mortem studies in humans suggest the involvement of α6GABAARs in a variety of neuropsychiatric disorders (Table 2). Synaptic α6βγ2 receptors with their high GABA sensitivity and long open probability are required for enhancing the precision of sensory inputs and adequate timing of motor activity and thus strongly influence cognition and adequate responses to our environment. Tonic inhibition mediated by extrasynaptic α6βδ receptors regulate the amount of information entering the cerebellum, and their optimal function allows the brain to concentrate on really important information. This is essential in stress situations when multiple sensory informations, considerations, and emotions pointing in different directions are coming together, thus impeding a rapid and optimal response. This might be even more of a problem in disorders with suboptimal functioning of α6GABAARs, as indicated by a disrupted PPI, that can be ameliorated by α6GABAAR PAMs. Enhancing the function of α6βγ2 and α6βδ receptors by PAMs might improve the integration of the sensory and emotional input and increase the threshold for incoming information, thus allowing the brain to again concentrate on important information. This conclusion is supported by the recent finding that a granule cell–specific deletion of the GABAAR δ-subunit in the cerebellum of mice resulted in a differential activation of many cortical and subcortical brain areas involved in cognition, anxiety-like, and stress-related behaviors in females (Rudolph et al., 2020). If stress can be elicited by a reduced tonic inhibition of granule cells, α6GABAAR PAMs modulating α6βδ receptors should alleviate stress symptoms.
α6GABAARs thus represent novel targets for the treatment of a large variety of neuropsychiatric disorders. α6GABAARs in the cerebellum but also in other parts of the nervous system because of their much higher GABA sensitivity and longer open probability than other GABAAR subtypes seem to be a major hub that allows to better differentiate between incoming signals and to partially correct various dysfunctions of the circuitries to which they contribute. The overall benefit of α6GABAARs PAMs in these disorders cannot easily be predicted. In some disorders, such as migraine (Fan et al., 2018; Tzeng et al., 2021), tic disorders (Cadeddu et al., 2021) [Wu et al., 2016; Master thesis, Graduate Institute of Pharmacology, College of Medicine (Chiou LC, Adviser) p 77, National Taiwan University, Taipei, Taiwan], schizophrenia (Chiou et al., 2018), and essential tremor (Huang et al., 2021), such compounds in animal models seem to be as effective as currently used standard medications. In other disorders, especially multigenic disorders, these compounds might only improve some symptoms. In contrast to THIP that directly activates all α6βδ and α4βδ (Brown et al., 2002) receptors, α6GABAAR-selective PAMs only modulate α6GABAARs involved in ongoing GABAergic transmission. This property together with their lack of modulation of other GABAAR subtypes explains their absence of unwanted side effects, such as sedation, motor-incoordination, addictive potential, and toxicity (Knutson et al., 2018; Vasovic et al., 2019; Tzeng et al., 2021). α6GABAA R-selective PAMs can be taken orally (Tzeng et al., 2021; Mitrović et al., 2021), and their deuterated derivatives, such as DK-I-56-1, exhibit a rapid onset of action and an optimal half-life (Knutson et al., 2018). These compounds exhibiting satisfactory preclinical pharmacokinetic and safety pharmacology will open new opportunities for the treatment of a variety of neuropsychiatric disorders and might also help to better cope with stress and its consequences.
Acknowledgments
The authors acknowledge the help of Florian Daniel Vogel with literature research and sequence alignments.
Authorship Contributions
Wrote or contributed to the writing of the manuscript: Sieghart, Chiou, Ernst, Fabjan, Savic, Lee.
Footnotes
- Received January 22, 2021.
- Accepted September 8, 2021.
This study was supported by grants to L.-C.C. from the Ministry of Science and Technology (MOST), Taiwan (MOST 104-2923-B-002-006-MY3, MOST 108-2320-B-002-029-MY3, MOST 106-2911-I-002-514, and 109-2320-B-002-042-MY3); the National Health Research Institutes (NHRI), Taiwan (NHRI-EX109-10733NI); and the Ministry of Education, Taiwan (107M4022-3). This study was also supported by grants to M.E. (Austrian Science Fund I2306 and DOC33-B27); by a grant to M.M.S. (Ministry of Education, Science and Technological Development of the Republic of Serbia, contract number 451-03-9/2021-14/200161); and by a grant to M.T.L. (UCSI University Research Excellence and Innovation Grant [Grant Number REIG-FPS-2020/065]).
No author has an actual or perceived conflict of interest with the contents of this article.
Thesis information:
Wu CY (2016) A novel therapeutic target for hyper-dopaminergic neuropsychiatric disorders with stereotypy behaviors: Cerebellar alpha 6 subunit-containing GABAA receptors, in Master thesis, Graduate Institute of Pharmacology, College of Medicine (Chiou LC, Adviser) p 77, National Taiwan University, Taipei, Taiwan.
Meeting abstracts:
Du JC, Tung LW, Lee HJ, Chiou LC (2017), Conditional knockdown of Slitrk-1 gene in striatal cholinergic interneurons impairs sensorimotor function in mice mimicking Tourette syndrome. Soc Neuroscience 2017, Prog. No. 690.01, November 11-15, 2017, Washington DC, USA.
Chiou LC, Lee HJ, Lee MT, Tzeng HR, Fan PC, Kubota H, Mouri A, Knutson DE, Cook JM, Ernst M, Sieghart W, Nabeshima T (2019) The alpha-6 subunit-containing GABAA receptor is a novel therapeutic target for schizophrenia and migraine. Soc Neuroscience 2019, Abstract No. 645.14/B55, October 19-23, 2019, Chicago, IL, USA.
Abbreviations
- ADHD,
- attention deficit hyperactivity disorder;
- ANT,
- alcohol-nontolerant;
- CGRP,
- calcitonin gene-related peptide;
- Compound 6/PZ-II-029,
- 7-methoxy-2-(4-methoxyphenyl)-2H-pyrazolo[4,3-c]quinolin-3(5H)-one;
- DCN,
- deep cerebellar nuclei;
- DK-I-56-1,
- (7-methoxy-2-(4-(methoxy-d3)phenyl)-2,5-dihydro-3H-pyrazolo[4,3-c]quinolin-3-one);
- fMRI,
- functional magnetic resonance imaging;
- furosemide,
- 4-chloro-N-furfuryl-5-sulfamoylanthranilic acid;
- GABAAR,
- GABAA receptor;
- GABRA6,
- human GABAA receptor α6 subunit gene;
- GAD65,
- glutamate decarboxylase 65 kDa;
- GAD67,
- glutamate decarboxylase 67 kDa;
- GAT-1,
- GABA transporter 1;
- GIT1,
- G-protein–coupled receptor kinase-interacting protein-1;
- HPA,
- hypothalamic-pituitary-adrenal;
- miRNA,
- microRNA;
- NTG,
- nitroglycerine;
- OCD,
- obsessive-compulsive disorder;
- PAM,
- positive allosteric modulator;
- PC,
- Purkinje cell;
- PCP,
- phencyclidine;
- PPI,
- prepulse inhibition;
- Ro15-1788/flumazenil,
- ethyl 8-fluoro-5,6-dihydro-5-methyl-6-oxo-4H-imidazo[1,4]benzodiazepine-3-carboxylate;
- Ro15-4513,
- ethyl 8-azido-5,6-dihydro-5-methyl-6-oxo-4H-imidazo[1,5a][1,4]benzodiazepine-3-carboxylate;
- RSA,
- respiratory sinus arrhythmia;
- Slitrk1,
- SLIT and NTRK-like family, member 1;
- SNP,
- single nucleotide polymorphism;
- TG,
- trigeminal ganglia;
- TGVS,
- trigemino-vascular system;
- THIP/gaboxadol,
- 4,5,6,7-tetrahydroisoxazolo[4,5-c]pyridine-3-ol;
- UTR,
- untranslated region.
- Copyright © 2022 by The Author(s)
This is an open access article distributed under the CC BY-NC Attribution 4.0 International license.