Dopamine Control of REM Sleep and Cataplexy

The brain circuits underlying sleep-wake regulation are not fully understood, although significant progress has been made in the past few decades [1–3]. Different sleep-wake states are controlled by distinct neural circuits: Wakefulness is mainly driven by the ascending reticular activating system (ARAS) that originates from the brainstem and activates the cortex via dorsal and ventral pathways; The circuits essential for rapid eye-movement (REM) sleep are within the pons and medulla in the brainstem; non-REM (NREM) sleep is achieved through inhibiting wake-promoting circuits by widely distributed subcortical nuclei. The dopamine (DA) neurons in the ventral tegmental area (VTA) are part of the ARAS and are known to promote wakefulness: DA-stimulating drugs such as cocaine and amphetamine cause potent wakefulness [3]; manipulating the activity of VTA DA neurons bidirectionally modulates the sleep-wake cycle [4]. However, unlike other monoamine systems such as norepinephrine, serotonin, or histamine neurons whose activity substantially decreases during REM sleep, VTA DA neurons exhibit phasic activation during REM [3, 4]. The role of this phasic activity is unclear. Hasegawa et al. tackled this question and uncovered a surprising role for DA signaling in the basolateral amygdala (BLA) in initiating REM sleep [5].

VTA DA neurons are heterogeneous with multiple downstream targets [6]. Therefore, the authors first used a GPCR activation-based DA sensor [7] to measure the dynamic of DA levels in multiple brain regions. They found that the DA signals showed a transient increase immediately before the NREM-to-REM transition in both the BLA and the nucleus accumbens. The authors next used optogenetic methods to selectively stimulate DA axons to determine the role of this transient DA increase. Surprisingly, stimulation in the BLA during NREM potently increased the total time of REM (from ~10% to ~30%) and decreased NREM. Furthermore, optogenetic inhibition of DA release in the BLA reduced the amount of REM sleep. These results suggested that the transient DA increase in the BLA plays a critical role in controlling the NREM-to-REM transition.

Because the DA receptor D2 (DRD2) has been implicated in REM sleep regulation [8], the authors next focused on the DRD2-expressing neurons in the BLA (BLA^DRD2+). In vitro brain slice recording showed that optogenetic stimulation-evoked DA release hyperpolarized BLA^DRD2+ neurons, and optogenetic or chemogenetic inhibition of BLA^DRD2+ neurons in intact mice prominently increased REM and decreased NREM, indicating that DA release in the BLA promotes REM sleep by suppressing BLA^DRD2+ neurons.

A key feature of REM sleep is muscle atonia, or loss of skeletal muscle tone [9]. This feature is also shared by cataplexy (a sudden and transient loss of skeletal muscle tone during wakefulness), a symptom of narcolepsy [10]. Hasegawa et al. next asked whether the transient DA increase in the BLA during wakefulness plays a role in cataplexy. To address this question, the authors first measured the dynamics of DA levels in the BLA in narcoleptic orexin-ataxin 3 mice [5], in which orexin (also known as hypocretin) neurons in the hypothalamus are genetically ablated, and eating chocolate frequently induces cataplexy-like episodes (CLEs) in these mice [5]. They found a transient increase in DA levels in the BLA, followed by CLEs, while the orexin-ataxin 3 mice were eating chocolate. Interestingly, they also found that DA release in the narcoleptic mice appeared to be greatly upregulated, which may underlie the frequent CLEs in these mice. The authors next transiently activated VTA DA fibers in the BLA or directly inhibited the BLA^DRD2+ neurons using optogenetic tools during wakefulness, and both manipulations induced cataplexy in non-narcoleptic mice that rarely show cataplexy. Consistent with these results, chemogenetic inhibition of BLA^DRD2+ neurons significantly increased CLEs in the orexin-ataxin 3 mice. Together, these results strongly suggest that a pathological increase of DA levels in the BLA causes cataplexy by inhibiting BLA^DRD2+ neurons in narcoleptic animals (Fig. 1).

Fig. 1.

The role of transient dopamine release in the BLA under physiological and pathological conditions. A Schematic illustrating projection from dopamine neurons in the ventral tegmental area (green) and three key nuclei for REM regulation (magenta). B The BLA neural circuits revealed by Hasegawa et al. Transient dopamine release in the BLA inhibits DRD2+ neurons and promotes REM sleep or cataplexy. C Transient dopamine release occurs during the NREM-to-REM transition; optogenetic stimulation-induced transient dopamine release during NREM promotes REM sleep (left). Dopamine transmission in the BLA is upregulated in narcoleptic animals; Optogenetic stimulation-induced transient dopamine release during wakefulness can cause cataplexy (right)

The current study provided multiple lines of evidence supporting the hypothesis that a unique subset of the VTA neurons is associated with REM sleep—transient DA release from these neurons and its inhibition of DRD2+ neurons in the BLA is essential to initiate REM, revealing an unexpected role of the VTA DA neurons in sleep-wake regulation beyond their documented wake-promoting effects. This study also demonstrated a causal link between the transient DA increase in the BLA and cataplexy, offering a new explanation of why cataplexy attacks are often associated with strong positive emotions [10] because the latter is known to activate VTA DA neurons.

This study also provides critical new insights into our understanding of several aspects of sleep-wake regulation, including REM sleep regulation and the role of orexin neurons. Although different models have been proposed for REM sleep regulation, it is generally agreed that the core circuits in the brainstem consist of two antagonizing neural populations that promote or suppress REM, and various brain regions in the brainstem and hypothalamus modulate the core circuits to regulate REM sleep [1–3, 9]. The ventrolateral peri-aqueductal grey/lateral pontine tegmentum, the sublaterodorsal nucleus, and the ventral medulla play a powerful role in controlling REM sleep [9]. However, this study showed that BLA^DRD2+ neurons are also essential in gating REM sleep, demonstrating that REM is regulated by widely distributed networks. As for the role of orexin neurons in sleep-wake regulation, it is hypothesized that a primary function of the orexin system is to stabilize sleep and wakefulness based on the fact that narcoleptic animals show fragmented sleep-wake states [11]. Much of the orexin regulation of the sleep-wake cycle has been attributed to the innervation of several neuromodulator systems, including the norepinephrine, histamine, serotonin, and cholinergic systems [11]. The current findings point to the importance of the change in the VTA DA system in narcoleptic animals, strongly suggesting a close relationship between the increase of DA release and cataplexy attacks. Given the critical role of VTA DA neurons in promoting and maintaining wakefulness, it is thus possible that the altered VTA DA system may play a role in producing the sleep-wake state fragmentation after the loss of orexin neurons. However, a remaining mystery related to this is how the loss of orexinergic neurons cause the upregulation of DA levels in the BLA. This is likely caused by changes at the network level rather than the loss of direct orexinergic input to the VTA DA neurons. Because orexin has been reported to excite the VTA DA neurons [11], the loss of such input would downregulate but not upregulate DA release. These results demonstrate how the complex interactions between different regulatory systems control the sleep-wake cycle.

Besides these new insights, the current findings have also raised several important questions. First, while the authors showed that inhibiting BLA^DRD2+ neurons increased the total time of REM sleep and cataplexy, it is unclear how these neurons influence neural circuits in the brainstem that are critical in controlling REM sleep or the loss of muscle tone during cataplexy, because it seems that BLA^DRD2+ neurons are largely local interneurons that do not project outside of the BLA. It is possible that the central amygdala (CEA) mediates part of the effect, as CEA activation promotes cataplexy but not REM sleep [9]. In addition, this question is particularly interesting because the authors reported a considerably long delay (>100 s) between optogenetic stimulation and REM sleep initiation. The long delay indicates a gradual buildup of the REM-promoting effects and is different from manipulating other REM-regulating circuits, suggesting that the BLA^DRD2+ neurons may not belong to the NREM-REM switching circuits. Therefore, future studies are needed to determine how the BLA^DRD2+ neurons contribute to the homeostatic regulation of REM sleep. Second, the authors showed that transient DA release in the BLA is required for REM sleep initiation under physiological conditions. Further study is needed to dissect the neural circuits that trigger this phasic activation of the VTA DA neurons and how this phasic DA release is controlled by the REM-regulating circuits in the brainstem. Answering these questions will help to understand how the widely distributed neural circuits coordinate and control REM sleep.

In conclusion, Hasegawa et al. revealed a surprising role for a subpopulation of VTA DA neurons in gating REM sleep and producing cataplexy in narcoleptic animals. These findings greatly expand our current understanding of sleep-wake regulation under physiological and pathological conditions, and provide potential drug targets for treating narcolepsy and REM-related sleep disorders.