Despite decades of research, there is a persistent debate regarding the localization of GABA/glycine neurons responsible for hyperpolarizing somatic motoneurons during paradoxical (or REM) sleep (PS), resulting in the loss of muscle tone during this sleep state. Combining complementary neuroanatomical approaches in rats, we first show that these inhibitory neurons are localized within the ventromedial medulla (vmM) rather than within the spinal cord. We then demonstrate their functional role in PS expression through local injections of adeno-associated virus carrying specific short-hairpin RNA in order to chronically impair inhibitory neurotransmission from vmM. After such selective genetic inactivation, rats display PS without atonia associated with abnormal and violent motor activity, concomitant with a small reduction of daily PS quantity. These symptoms closely mimic human REM sleep behavior disorder (RBD), a prodromal parasomnia of synucleinopathies. Our findings demonstrate the crucial role of GABA/glycine inhibitory vmM neurons in muscle atonia during PS and highlight a candidate brain region that can be susceptible to α-synuclein-dependent degeneration in RBD patients.

SEE SCHENCK AND MAHOWALD DOI101093/AWW329 FOR A SCIENTIFIC COMMENTARY ON THIS ARTICLE: Idiopathic REM sleep behaviour disorder is characterized by the enactment of violent dreams during paradoxical (REM) sleep in the absence of normal muscle atonia. Accumulating clinical and experimental data suggest that REM sleep behaviour disorder might be due to the neurodegeneration of glutamate neurons involved in paradoxical sleep and located within the pontine sublaterodorsal tegmental nucleus. The purpose of the present work was thus to functionally determine first, the role of glutamate sublaterodorsal tegmental nucleus neurons in paradoxical sleep and second, whether their genetic inactivation is sufficient for recapitulating REM sleep behaviour disorder in rats. For this goal, we first injected two retrograde tracers in the intralaminar thalamus and ventral medulla to disentangle neuronal circuits in which sublaterodorsal tegmental nucleus is involved; second we infused bilaterally in sublaterodorsal tegmental nucleus adeno-associated viruses carrying short hairpin RNAs targeting Slc17a6 mRNA [which encodes vesicular glutamate transporter 2 (vGluT2)] to chronically impair glutamate synaptic transmission in sublaterodorsal tegmental nucleus neurons. At the neuroanatomical level, sublaterodorsal tegmental nucleus neurons specifically activated during paradoxical sleep hypersomnia send descending efferents to glycine/GABA neurons within the ventral medulla, but not ascending projections to the intralaminar thalamus. These data suggest a crucial role of sublaterodorsal tegmental nucleus neurons rather in muscle atonia than in paradoxical sleep generation. In line with this hypothesis, 30 days after adeno-associated virus injections into sublaterodorsal tegmental nucleus rats display a decrease of 30% of paradoxical sleep daily quantities, and a significant increase of muscle tone during paradoxical sleep concomitant to a tremendous increase of abnormal motor dream-enacting behaviours. These animals display symptoms and behaviours during paradoxical sleep that closely mimic human REM sleep behaviour disorder. Altogether, our data demonstrate that glutamate sublaterodorsal tegmental nucleus neurons generate muscle atonia during paradoxical sleep likely through descending projections to glycine/GABA premotor neurons in the ventral medulla. Although playing a role in paradoxical sleep regulation, they are, however, not necessary for inducing the state itself. The present work further validates a potent new preclinical REM sleep behaviour disorder model that opens avenues for studying and treating this disabling sleep disorder, and advances potential regions implicated in prodromal stages of synucleinopathies such as Parkinson's disease.