Studying Cerebellar Circuits by Remote Control of Selected Neuronal Types with GABA(A) Receptors

Abstract

Although GABA(A) receptor-mediated inhibition of cerebellar Purkinje cells by molecular layer interneurons (MLIs) has been studied intensely at the cellular level, it has remained unclear how this inhibition regulates cerebellum-dependent behaviour. We have implemented two complementary approaches to investigate the function of the MLI-Purkinje cell synapse on the behavioural level. In the first approach we permanently disrupted inhibitory fast synaptic transmission at the synapse by genetically removing the postsynaptic GABA(A) receptors from Purkinje cells (PC-Deltagamma2 mice). We found that chronic disruption of the MLI-Purkinje cell synapse strongly impaired cerebellar learning of the vestibular occular reflex (VOR), presumably by disrupting the temporal patterns of Purkinje cell activity. However, in PC-Deltagamma2 mice the baseline VOR reflex was only mildly affected; indeed PC-Deltagamma2 mice show no ataxia or gait abnormalities, suggesting that MLI control of Purkinje cell activity is either not involved in ongoing motor tasks or that the system compensates for its loss. To investigate the latter possibility we developed an alternative genetic technique; we made the MLI-Purkinje cell synapse selectively sensitive to rapid manipulation with the GABA(A) receptor modulator zolpidem (PC-gamma2-swap mice). Minutes after intraperitoneal zolpidem injection, these PC-gamma2-swap mice developed severe motor abnormalities, revealing a substantial contribution of the MLI-Purkinje cell synapses to real time motor control. The cell-type selective permanent knockout of synaptic GABAergic input and the fast reversible modulation of GABAergic input at the same synapse illustrate how pursuing both strategies gives a fuller view.

Keywords: memory consolidation; purkinje cell; zolpidem; β-carboline; γ-aminobutyric acid type A receptor.

Figure 1

The cerebellar circuitry and the interventions at the MLI-Purkinje cell synapses. Excitatory cells are shown in red, inhibitory cells are shown in blue. (A) All cells in the cortex are inhibitory apart from granule cells (Gr) which give rise to parallel fibres (PF), and unipolar brush cells (not shown). The cerebellar cortex receives excitatory input via mossy fibres (MF) and climbing fibres (CF). The only output of the cortex is via Purkinje cells (PC), which project to the deep cerebellar nuclei (DCN) and vestibular nuclei (VN). The activity of Purkinje cells is under inhibitory control from molecular layer basket and stellate cells (BC/SC) through α1βγ2 subunit containing GABA_A receptors. Basket/stellate cells mutually inhibit each other and are coupled by gap junctions (zig-zag line). Diagram adapted from Grillner et al. (2005). (B) and (C) show the MLI-Purkinje cell synapse after chronic disruption in PC-Δγ2 mice (B) and during rapid enhancement with zolpidem in PC-γ2-swap mice (C). GC, Golgi cell; IO, inferior olive; PN/VG, pontine nuclei and vestibular ganglion.

Figure 2

Compensatory changes in an abstract biological network. Blue lines indicate inhibitory interactions, red lines indicate excitatory interactions. Shown are an intact network (left side) and the same network after ablation of one of its nodes. Lesioning one element causes the rest of the system to compensate for the loss. Some interactions are lost (orange dashed lines) and some new interactions emerge (green lines), so that the final output is unchanged. Adapted from Greenspan (2001).

Figure 3

The strategy for bi-directional modulation of selected GABAergic synapses. (A) shows the wild-type situation with a phenylalanine (F) at position 77 in the γ2 subunit of the GABA_A receptor. Zolpidem/DMCM binding occurs at the interface of the α and the γ2 subunits and is widely distributed throughout the brain (indicated in grey in the cartoon of a horizontal brain section). In a first step (B) we have changed the phenylalanine at position 77 to isoleucine (I) by homologous recombination in mouse embryonic stem cells. In these mice zolpidem/DMCM binding is abolished. In the last step (C) we have reintroduced the drug-sensitive wild-type γ2 subunit under the control of the Purkinje cell-specific L7 promoter to selectively restore zolpidem/DMCM-sensitivity by a Cre-mediated Purkinje cell-specific swap of γ2 subunits (PC-γ2-swap mice). In these mice expression of zolpidem/DMCM-sensitive GABA_A receptors is restricted to Purkinje cells of the cerebellum (grey). Note the difference to wild-type mice in (A), which show drug binding throughout the cerebellar cortex. In Purkinje cells zolpidem (red trace) and DMCM (blue trace) now selectively enhance or reduce, respectively, GABA_A receptor-mediated inhibitory post synaptic currents, illustrated by the artificial traces on the right (GABA alone: black trace). All other neurons in the brain are insensitive to these drugs (left traces).