We screened for novel circuits in the mouse brain that promote wakefulness. Chemogenetic activation experiments and EEG recordings pointed to glutamatergic/nitrergic (NOS1) and GABAergic neurons in the VTA. Activating glutamatergic/NOS1 neurons, which were wake- and REM-sleep-active, produced wakefulness through projections to the nucleus accumbens and the lateral hypothalamus. Lesioning the glutamate cells impaired the consolidation of wakefulness. By contrast, activation of GABAergic VTA neurons elicited long-lasting NREM-like sleep resembling sedation. Lesioning these neurons produced an increase in wakefulness that persisted for at least 4 months. Surprisingly, these VTA GABAergic neurons were wake-and REM-sleep-active. We suggest that GABAergic VTA neurons may limit wakefulness by inhibiting the arousal-promoting VTA glutamatergic and/or dopaminergic neurons and through projections to the lateral hypothalamus. Thus, in addition to its contribution to goal- and reward-directed behaviours, the VTA has a role in regulating sleep and wakefulness.

SummaryMammals, including humans, prepare for sleep by nesting and/or curling up, creating microclimates of skin warmth. To address whether external warmth induces sleep through defined circuitry, we used c-Fos-dependent activity tagging, which captures populations of activated cells and allows them to be reactivated to test their physiological role. External warming tagged two principal groups of neurons in the median preoptic (MnPO)/medial preoptic (MPO) hypothalamic area. GABA neurons located mainly in MPO produced non-rapid eye movement (NREM) sleep but no body temperature decrease. Nitrergic-glutamatergic neurons in MnPO-MPO induced both body cooling and NREM sleep. This circuitry explains how skin warming induces sleep and why the maximal rate of core body cooling positively correlates with sleep onset. Thus, the pathways that promote NREM sleep, reduced energy expenditure, and body cooling are inextricably linked, commanded by the same neurons. This implies that one function of NREM sleep is to lower brain temperature and/or conserve energy.

Recently, we marked TRIO for the first time as a candidate gene for intellectual disability (ID). Across diverse vertebrate species, TRIO is a well-conserved Rho GTPase regulator that is highly expressed in the developing brain. However, little is known about the specific events regulated by TRIO during brain development and its clinical impact in humans when mutated. Routine clinical diagnostic testing identified an intragenic de novo deletion of TRIO in a boy with ID. Targeted sequencing of this gene in over 2300 individuals with ID, identified three additional truncating mutations. All index cases had mild to borderline ID combined with behavioral problems consisting of autistic, hyperactive and/or aggressive behavior. Studies in dissociated rat hippocampal neurons demonstrated the enhancement of dendritic formation by suppressing endogenous TRIO, and similarly decreasing endogenous TRIO in organotypic hippocampal brain slices significantly increased synaptic strength by increasing functional synapses. Together, our findings provide new mechanistic insight into how genetic deficits in TRIO can lead to early neuronal network formation by directly affecting both neurite outgrowth and synapse development.

Homeostatic plasticity, a form of synaptic plasticity, maintains the fine balance between overall excitation and inhibition in developing and mature neuronal networks. Although the synaptic mechanisms of homeostatic plasticity are well characterized, the associated transcriptional program remains poorly understood. We show that the Kleefstra-syndrome-associated protein EHMT1 plays a critical and cell-autonomous role in synaptic scaling by responding to attenuated neuronal firing or sensory drive. Chronic activity deprivation increased the amount of neuronal dimethylated H3 at lysine 9 (H3K9me2), the catalytic product of EHMT1 and an epigenetic marker for gene repression. Genetic knockdown and pharmacological blockade of EHMT1 or EHMT2 prevented the increase of H3K9me2 and synaptic scaling up. Furthermore, BDNF repression was preceded by EHMT1/2-mediated H3K9me2 deposition at the Bdnf promoter during synaptic scaling up, both in vitro and in vivo. Our findings suggest that H3K9me2-mediated changes in chromatin structure govern a repressive program that controls synaptic scaling.