Search Search


A person can reject food, abstain from sex and control his or her thirst, but cannot keep from falling asleep. What is the genetic and neuronal basis for this insistent bodily need, why is it observed in essentially all multicellular animals, and how is it regulated?

Despite years of study, answers to these questions remain elusive. We are using zebrafish as a new model to discover and understand genetic and neuronal circuits that regulate sleep.



Yes, fish sleep.

We and others have shown that larval and adult zebrafish exhibit behavioral criteria that define sleep-like states:

  • rapidly reversible behavioral quiescence
  • elevated arousal threshold
  • homeostatic regulation

Sleep has been studied in a variety of model organisms, including roundwormsfruit flies and rodents. Zebrafish larvae offer several advantages

  • optical transparency facilitates monitoring and manipulating neurons in intact animals
  • small size enables large-scale behavioral assays that can be used to identify genes, drugs and neurons that regulate sleep
  • brain is anatomically and molecularly similar to mammalian brain but contains far fewer neurons, providing a simpler system to study neuronal circuits that regulate sleep
  • diurnal sleep/wake pattern makes zebrafish more appropriate than nocturnal rodents as a model for circadian and light-dependent regulation of human sleep

We developed high-throughput assays to measure locomotor activity and arousal threshold in zebrafish larvae, which we study using genetic, pharmacological and neuronal approaches.


Hcrt signaling promotes locomotor activity and inhibits sleep

As proof-of-principle, we studied the zebrafish hypocretin/orexin (Hcrt) ortholog, whose loss causes the sleep disorder narcolepsy. We found that Hcrt overexpression impairs both the initiation and maintenance of sleep, consolidates wakefulness and induces hyperarousal, as in mammals. The movie below shows that most Hcrt-overexpressing larvae are active at night, while most wild type larvae are inactive.


Identifying novel sleep regulators using a genetic screen

Genetic screens have been a powerful approach to identify genes that regulate sleep in invertebrates, but some genes identified in these screens lack clear vertebrate orthologs, and most genes known to regulate vertebrate sleep lack clear invertebrate orthologs. To identify genes that regulate vertebrate sleep, we developed a genetic overexpression screening strategy and used it to perform the first large-scale screen for genes that affect vertebrate sleep. Overexpression of one gene identified in the screen, neuromedin U (Nmu), promotes locomotor activity (left) and inhibits sleep (middle), whereas nmu mutants are hypoactive. We found that Nmu-induced arousal requires Nmu receptor 2 and signaling via corticotropin releasing hormone (Crh) receptor 1, and likely acts via brainstem Crh neurons (right, circled regions). These results revealed an unexpected interaction between Nmu and a brainstem arousal system that represents a novel wake-promoting pathway. We are further exploring this system and characterizing other genes identified in the screen.

Melatonin is required for circadian regulation of sleep

classical model postulates that sleep is regulated by homeostatic (S) and circadian (C) processes. Adenosine and other factors are implicated in process S and the circadian clock is understood at the molecular level, but little is known about how the circadian clock regulates sleep. Melatonin is a good candidate to mediate this process because its production is regulated by the circadian clock and it can induce sleep in diurnal animals. However, melatonin's role in sleep is controversial because most nocturnal lab mouse strains do not synthesize it, and pinealectomy studies have produced different phenotypes in different species, suggesting that melatonin has species-specific functions. However, melatonin levels peak at night in both diurnal and nocturnal animals, and exogenous melatonin only promotes sleep in diurnal animals. The discrepant pinealectomy results may be due to differences in the imprecise pinealectomy procedure in different species and labs. To address the function of melatonin in a diurnal vertebrate in a clean and reproducible manner, we generated zebrafish that lack melatonin due to mutation of arylalkylamine N-acetyltransferase 2 (aanat2). These mutants sleep 50% less at night in light:dark conditions (left), and circadian regulation of sleep is abolished in free-running conditions (middle). In contrast to some experiments using pinealectomy or exogenous melatonin, we found that melatonin is not required for normal circadian rhythms. Finally, we found that an adenosine receptor agonist rescues the aanat2 mutant sleep defect (right), suggesting that melatonin promotes sleep in part by promoting adenosine signaling, providing a simple mechanism to integrate circadian and homeostatic control of sleep. These results help to explain how the circadian clock regulates sleep in a diurnal vertebrate, and provide a basis to explore how the circadian clock and melatonin regulate sleep.
 aanat2 graphs_no bkgd.png

Prokineticin 2 regulates the direct effects of light and dark on sleep

Light affects sleep/wake behaviors indirectly by entraining circadian rhythms and also directly and rapidly via a phenomenon known as the direct or masking effect of light and dark on behavior, in which light and dark rapidly lead to arousal and sedation, respectively, in diurnal animals. While this phenomenon is widely observed in the animal kingdom, including in humans, the underlying mechanism is largely unknown. We found that both overexpression and mutation of the hypothalamic neuropeptide prokineticin 2 (Prok2) affects sleep/wake behaviors in a light-dependent but circadian-independent manner, suggesting that Prok2 regulates the direct effects of light and dark on behavior. For example, in the graphs shown below, larvae were entrained in LD, heat shocked during the day, and then exposed to alternating one hour periods of light and dark. Compared to their WT siblings, Prok2-overexpressing animals were less active and slept more in light, and were more active and slept less in dark, showing that Prok2 overexpression antagonizes the direct effects of light and dark on behavior. These effects were evident during both subjective day (white bars) and subjective night (grey bars). We found that Prok2-induced sedation is associated with increased expression of galanin, a sleep-inducing peptide, in neurons that express both galanin and the prok2 receptor in the anterior hypothalamus, a likely homolog of the mammalian sleep-promoting ventrolateral preoptic nucleus. Prok2 thus provides an entry point to explore genetic and neuronal mechanisms that underlie the direct effects of light and dark on sleep and wake states.

hs-Prok2 masking.png

Candidate gene studies

We are using candidate gene gain- and loss-of-function approaches to identify and explore additional genes that regulate sleep. For example, we found that zebrafish that lack noradrenaline due to mutation of dopamine beta hydroxylase (dbh) exhibit dramatically less locomotor activity (left) and more sleep (middle). We also found that overexpression of the hypothalamic neuropeptide QRFP (also known as 26RFa and P518) inhibits locomotor activity (right), while mutation of qrfp or its receptors gpr103a and gpr103b results in more locomotor activity (right) and less sleep. The effect of QRFP overexpression is blocked in gpr103a; gpr103b double mutants (right), confirming their role as QRFP receptors in vivo. We are using similar methods to characterize other known and novel sleep regulatory pathways.

dbh gpr103.png 


The transparency and relatively simple but conserved vertebrate brain of zebrafish larvae make it a powerful system to study neuronal circuits that regulate sleep. As proof-of-principle, we and others showed that zebrafish larvae only have ~10 Hcrt neurons, compared to thousands of these neurons in mammals, providing a simpler system to study Hcrt neuron development and function. We showed that zebrafish Hcrt neurons project to wake-promoting brain regions and are active during periods of consolidated wakefulness, as they are in mammals. We are exploiting the transparency and relatively simple neuronal circuits of zebrafish larvae to study the development and function of Hcrt neurons, and to discover additional neuronal populations that regulate sleep. We are characterizing these neural circuits at the single neuron level using Brainbow, testing the effects of activating and inhibiting specific neurons on behavior using optogenetics and chemogenetics, and monitoring effects on neuronal activity throughout the brain using GCaMP6.

Hcrt-Brainbow entire larva.jpg                            Zebrafish larva whose Hcrt neurons are labeled with different colors using Brainbow (credit: C. Chiu).

Stimulation of Hcrt neurons promotes locomotor activity

Because zebrafish larvae are transparent, all neurons are accessible to ambient light. We exploited this feature to develop a non-invasive assay that allows optogenetic manipulation of genetically specified neurons in 96 freely-behaving larvae while monitoring the behavior of each animal. Using this assay, we found that stimulating Hcrt neurons promotes locomotor activity and that this effect requires noradrenaline (left). We also combined optogenetics and GCaMP6s imaging in intact larvae to show that activation of Hcrt neurons stimulates the locus coeruleus (right), the main source of noradrenaline in the brain.


As an alternative method to manipulate neurons that does not require light, we found that heterologously expressed TRPV1, TRPM8 and TRPA1 can be used to stimulate or ablate genetically specified neurons in response to their chemical or thermal agonists in freely behaving larvae. Using this approach, we found that stimulating Hcrt neurons using TRPV1 and a low concentration of its small molecule agonist capsaicin (left) results in increased locomotor activity and decreased sleep. Conversely, ablation of Hcrt neurons using a higher concentration of capsaicin (right) results in decreased locomotor activity and increased sleep. These tools may be particularly useful for activating specific neural populations while simultaneously monitoring whole-brain neural activity using GCaMP6.