Distinct effects of mGlu4 receptor positive allosteric modulators at corticostriatal vs. striatopallidal synapses may differentially contribute to their antiparkinsonian action

Abstract Metabotropic glutamate 4 (mGlu4) receptor is a promising target for the treatment of motor deficits in Parkinson's disease (PD). This is due in part to its localization at key basal ganglia (BG) synapses that become hyperactive in this pathology, particularly striatopallidal synapses. In this context, mGlu4 receptor activation using either orthosteric agonists or positive allosteric modulators (PAMs) improves motor symptoms in rodent PD models in certain conditions. However, literature data show that mGlu4 receptor PAMs have no effect at striatopallidal GABAergic synapses (unless combined with an orthosteric agonist) and on the firing activity of pallidal neurons, and fail to provide significant motor improvement in relevant PD models. This questions the mechanistic hypothesis that mGlu4 receptor PAMs should act at striatopallidal synapses to alleviate PD motor symptoms. To shed light on this issue, we performed brain slice electrophysiology experiments. We show that Lu AF21934, an mGlu4 PAM small-molecule probe-compound, was ineffective at striatopallidal synapses at all concentrations tested, while it significantly inhibited corticostriatal synaptic transmission. Similarly, Lu AF21934 did not affect electrophysiology readouts at striatopallidal synapses in the presence of haloperidol or in 6-hydroxydopamine-lesioned rats. Interestingly, co-application of Lu AF21934 with a glutamate transporter inhibitor revealed a significant inhibitory action at striatopallidal synapses. Possibly, this effect could rely on increased level/permanence of glutamate in the synaptic cleft. Such differential efficacy of mGlu4 receptor PAMs at corticostriatal vs . striatopallidal synapses raises several issues regarding the synaptic target(s) of these drugs in the BG, and challenges the mechanisms by which they alleviate motor deficits in experimental PD models.