Galanin neurons unite sleep homeostasis and α2-adrenergic sedation

Abstract

Our urge to sleep increases with time spent awake, until sleep becomes inescapable. The sleep following sleep deprivation is longer and deeper, with an increased power of delta (0.5 - 4 Hz) oscillations, a phenomenon termed sleep homeostasis [1-4]. Although widely-expressed genes regulate sleep homeostasis [1, 4-10], and the process is tracked by somnogens and phosphorylation [1, 3, 7, 11-14], at the circuit level sleep homeostasis has remained mysterious. Previously we found that sedation induced with 2 adrenergic agonists (e.g. dexmedetomidine) and sleep homeostasis both depend on the preoptic (PO) hypothalamus [15, 16]. Dexmedetomidine, increasingly used for long-term sedation in intensive care units [17], induces a NREM-like sleep but with undesirable hypothermia [18, 19]. Within the PO, various neuronal subtypes (e.g. GABA/galanin and glutamate/NOS1) induce NREM sleep [20-22] and concomitant body cooling [21, 22]. This could be because NREM sleep’s restorative effects depend on lower body temperature [23, 24]. Here, we show that mice with lesioned PO galanin neurons have reduced sleep homeostasis: in the recovery sleep following sleep deprivation, there is a diminished increase in delta power, and the mice catch up little on lost sleep. Furthermore, dexmedetomidine cannot induce high-power delta oscillations or sustained hypothermia. Some hours after dexmedetomidine administration to wild-type mice there is a rebound in delta power when they enter normal NREM sleep, reminiscent of emergence from torpor. This delta rebound is reduced in mice lacking PO galanin neurons. Thus, sleep homeostasis and dexmedetomidine-induced sedation require PO galanin neurons and likely share common mechanisms.