Unravelling the role of GABA_A receptor subtypes in distinct neurons and behaviour

Recently, a variety of optogenetic or pharmacogenetic techniques have been established that allow a modulation of the electrical activity of genetically modified neuronal populations in the brains of living animals. These techniques rely on the transgenic or viral introduction of exogenous ion channels or receptors into neurons, whose membrane potential can then be selectively changed in vivo by optical or pharmacological stimuli, respectively. The subsequently observed behavioural changes in these animals allow conclusions to be made on the function of the manipulated neurons in behaviour. These methods have already contributed and will further contribute enormously to our understanding of the role of defined cell populations in certain behaviours (Callaway, 2005). A common disadvantage of all these methods is the genetic introduction of exogenous proteins into neurons. Whether these proteins interfere with the translation, assembly, or processing of endogenous proteins or receptors, or whether they interfere with the intrinsic excitability of neurons, is not currently known. In addition, experimental excitation or inhibition of neurons not directly involved in eliciting a certain behaviour might generate long range effects in the brain that overall could modify the investigated behaviour, thus leading to erroneous conclusions on the function of the respective neurons.

A fundamentally different pharmaco-genetic approach was developed by Wulff et al. (2007). They introduced a point mutation (γ2^F77I) into the benzodiazepine binding site of GABA_A receptors that renders these receptors insensitive towards the positive allosteric modulator zolpidem, the negative allosteric modulator DMCM, and the benzodiazepine site antagonist (null modulator) flumazenil (Buhr et al. 1997). The respective knock-in mouse line (GABA_ARγ2^77Ilox), in which the point-mutated γ2 subunit was flanked by loxP sites that allowed its subsequent removal by Cre-recombinase, exhibited an unaltered distribution and function of GABA_A receptors in the brain. These receptors, however, could no longer be modulated by zolpidem, DMCM, or flumazenil, but could still be modulated by diazepam and other benzodiazepine site ligands. Using the Purkinje cell-specific promoter L7, Wulff et al. (2007) then generated two additional transgenic mouse lines expressing either Cre-recombinase or the wild-type γ2 subunit coupled to enhanced green fluorescent protein in Purkinje cells of cerebellum only. By crossing these two mouse lines with the GABA_ARγ2^77Ilox line, they selectively swapped zolpidem-insensitive with zolpidem-sensitive GABA_A receptors in cerebellar Purkinje cells and demonstrated that GABAergic inhibition was potentiated exclusively in this cell type by systemically applied zolpidem. In addition the motor-impairing effects of zolpidem in the rotarod test, that were eliminated by the point mutation γ2^F77I in the GABA_ARγ2^77Ilox mouse, were restored in the triple-crossed mouse in which the wild-type γ2 subunit was introduced in Purkinje cells only, thus, providing the proof of principle for this method. However, the necessity of generating different transgenic mouse lines as well as backcrossing to the GABA_ARγ2^77Ilox mouse for each cell type to be investigated, and the unavailability of viruses suitable for a rapid swap of the γ2^77I to γ2^77F in the GABA_ARγ2^77Ilox mouse, so far delayed the further use of this technique.

In a recent issue of The Journal of PhysiologySumegi et al. (2012) described how they overcame these problems by generating lenti- and adeno-associated viruses expressing both Cre-recombinase and N-terminal AU1-tagged wild-type γ2^77F subunits. Using light- and electron microscopic immunolocalization as well as in vitro electrophysiological methods, they verified the brain region-specific swapping of the γ2^77I with the ^AU1γ2^77F in GABA_ARγ2^77Ilox mice. The resulting GABA_A receptors could be located by the efficient immunohistochemical detection of the AU1-tag, had normal synaptic enrichment, unaltered kinetic properties and restored sensitivity to zolpidem. The availability of these lenti- and adeno-associated viruses thus represents a breakthrough in the applicability of this technique. Injection of these viruses into any brain region of GABA_ARγ2^77Ilox mice will render virus-infected cells in this brain region selectively zolpidem sensitive, if these cells endogenously express GABA_A receptors containing α1, α2, or α3, and any one of the β subunits. Since GABA_A receptors are ubiquitously distributed in the brain and since the large majority of these receptors contain these subunits, it can be expected that most if not all neurons in the brain infected by these viruses will become zolpidem sensitive. Zolpidem, like other benzodiazepine site ligands, cannot directly activate GABA_A receptors, but only modulates ongoing GABAergic transmission. Systemically applied zolpidem rapidly penetrates into the brain and only enhances those GABAergic inputs in the virus-infected neurons that are currently active and thus required for the respective behavioural task. Alternatively, a systemic application of DMCM, a negative allosteric modulator at the benzodiazepine site of GABA_A receptors, would cause a reduction of the on-going GABAergic transmission in these neurons. And both actions could be rapidly terminated at any time during the behavioural task by the application of flumazenil.

The applicability of this technique, however, could be even further enhanced by using multiple kinds of viruses carrying different promoters or elements of alternative conditional recombination systems. By these means, it will be possible to express the AU1-tagged wild-type γ2^77F subunit into specific cell types of different brain regions and study their function in behaviour. Finally, combined with currently available and newly developed receptor subtype-selective benzodiazepine site ligands from the structural classes of zolpidem, DMCM, flumazenil, and others (Ramerstorfer et al. 2010), it will be possible to address the effects of specific GABA_A receptor subtypes on specific cell types in brain function and behaviour. This, certainly, will be a major breakthrough in the investigation of neuronal circuits as well as unravelling the function of individual GABA_A receptor subtypes in different behavioural tasks.