Sleep Homeostasis

Sleep Homeostasis and Sleep Drive
Image Unavailable
Sleep Homeostasis is implicated in driving the need to sleep, particularly after all those late-nighters

Sleep, despite the elusiveness of its function, is a heavily regulated periodic behaviour in animals. Sleep homeostasis is believed to be one of two complementary systems that work to regulate sleep patterns in mammals. The Two Process system [1] consists of a Circadian (Process C) and Homeostatic (Process S) component. The former regulates the timing and onset of sleep, while the latter determines the amount of sleep, depth of sleep, and is implicated in memory consolidation. As of the present, these two processes are differentially described. The Suprachiasmatic Nucleus, its projections, and associated target areas are well defined components of the circadian process, with a body of literature in anatomy, physiology as well as lesion and knockout studies [2]. In contrast, the homeostatic process is less well defined, but is currently being elucidated by a variety of molecular, imaging, and stimulation studies.

The current model of the homeostatic process dictates that a somnogenic substance accumulates in the brain with wakefulness, effectively producing a “drive” to sleep. The best candidate endogenous signalling molecules are PGD2 and Adenosine, both of which produce a strong drive to sleep independent of the Sleep-wake switch [3]. In addition, recent research suggests that sleep induces homeostatic regulation of activity at the synaptic level through global synaptic downscaling [4].

The Two Process Model

SCN regulation of sleep
Image Unavailable
Process C is controlled by gene expression in the SCN[5]

Process C

The Circadian Process is a 24 hour clock mediated by protein transcription in the cells of the Suprachiasmatic Nucleus (SCN)[5]. The SCN is situated superior to the optic chasma in the hypothalamus and has projections to a variety of peripheral pacemakers. The clock is maintained by the alternating expression of Clock:Bmal1 and Per:Cry hetermodimers. Clock:Bmal1 prmotes expression of Per and Cry, while Per:Cry promotes expression of Clock and Bmal1. Per:Cry also inhibits Clock:Bmal1, thus slowing its own expression. As a result, Per:Cry levels peak at the end of the subjective Day, while Clock:Bmal1 levels peak during the subjective night. The SCN provides information as to the timing of the onset of sleep by output to a variety of regulatory systems[5]. Process C is entrained in mammals by input from retinal ganglion cells by the retinohypothalamic tract[5]. This allows sleep to be synchronized with the time of day thus mediating nocturnal and diurnal behaviors.

Process S

The Two Process Model
Image Unavailable
Mediation of the onset and duration of sleep

Sleep quantity is also shown to be homeostatically controlled. A variety of sleep deprivation experiments have shown that lack of sleep has severe cognitive and physiological consequences. Long episodes of Total Sleep Deprivation (TSD) yield longer sleep episodes in "sleep rebound", but the duration of sleep rebound does not always correspond to the duration of sleep lost. Sleep latency, the time to fall asleep, decreases significantly with increased TSD[6]. TSD is also implicated in Slow Wave Activity (SWA). This suggests that increased wakefulness is directly tied to a subsequent increased drive to sleep and often an increased duration of sleep. Process S must therefore act as "a homeostatic process mediating the rise of sleep propensity during waking and its dissipation during sleep"[1]. Since wakefulness is tied to an increase in global cerebral blood flow, metabolism and activity in ascending diffuse arousal systems, this has led to the search to identify a somnogen that accumulates with increased wakefulness and dissipates with sleep.

Mechanisms and Effects of Homeostatic Regulation

Slow Wave Activity

SWA in Sleep
Image Unavailable

As early as 1970, the homeostatic regulation of sleep was known to be tied to EEG activity during Slow Wave (NREM) sleep. Most early research, particularly by Gillberg and colleagues, focused on using SWA as a measure of Process S. In normal sleep, EEG waves have high amplitude at the beginning of the sleep period and steadily decline with increased time spent sleeping. Individuals who undergo Total Sleep Deprivation (TSD) showed increased EEG activity the longer TSD lasted[7]. When sleep is reduced to 4hr intervals, subsequent EEG activity increases by 20%, indicating an increased sleep drive and subsequent "depth" of sleep[8]. Borbely and colleagues developed and cataloged a set of equations that accurately simulate SWA as a function of time in a variety of scenarios[8]. Thus Process S, a key factor in maintaining sleep homeostasis, can be indirectly measured through its effect on synchronized NREM sleep.

The VLPO Sleep-Wake Switch

VLPO Projections
Image Unavailable

The Ventrolateral Preoptic Area (VLPO) is the main sleep switch that induces a sleeping state from an awake state. the VLPO contains inhibitory projections to the Histaminergic system of the Tuberomamillary Nucleus, as well as to the ascending Aminergic system in the Locus Coeruleus[5]. Sleep-promoting neurons of the VLPO have been shown to be directly targeted by anesthetics like isofluorane, which act to block postsynaptic potassium conductance[9]. VLPO neurons are shown to express both kappa and mu opioid receptors, which increase time asleep during the night and time awake during the day respectively [10]. Finally, it is important to note that the VLPO is proportionally active with the time spent asleep. Taken together, this demonstrates that the VLPO is able to respond to substances that alter sleep-wake activity. These neurons are therefore ideal candidates to gauge Sleep Homeostasis, exhibit Process S, and as a result directly induce sleep.

Adenosine, Cytokines, PGD2 and somnogen-induced Sleep Drive

Adenosine and Caffeine
Image Unavailable
Structure of Adenosine and its inhibitor Caffeine

In order to gauge the amount of activity during wakefulness as per Process S, the ideal somnogenic substance would be a byproduct of neuronal activity. Given the brain's enormous energy demand, a likely candidate would be one of the metabolic byproducts of glucose metabolism. Despite the role of glucose in energy homeostasis and its respective feeding behaviours, the main sleep drive brought on by food intake is orexin-mediated by the TMH. Another candidate is Adenosine, the concentration of which is demonstrated to increase due to ATP breakdown during wakefulness. As such, early research looked mainly at Adenosine as the primary somnogenic mediator of Process S. Adenosine concentrations in the extracellular Basal Forebrain are shown to increase with prolonged wakefulness[11]. The Basal Forebrain is a major component of the ascending cholinergic arousal system. Adenosine is also known to tonically inhibit the cholinergic neurons of the Basal forebrain through Adenosine channels[11]. Caffeine, notably, is an antagonist for the Adenosine receptor and as such is believed to mask Process S[12]. Indeed, caffeine effectively promotes arousal while, critically, suppressing recovery sleep after deprivation[11]. In this model, sleep is triggered by the disinhibition of the VLPO by suppressing the inhibitory projections of the basal forebrain[12]. Recently, however, Adenosine A2A receptors have been identified on sleep-promoting neurons in the VLPO[13]. Accumulations of adenosine can therefore directly induce sleep by stimulating the VLPO via A2A receptors. Both of these mechanisms work in concert: A1 receptors in the basal forebrain inhibit the cholinergic arousal system, while A2A receptors stimulate the sleep neurons of the VLPO.

Image Unavailable
IL-1 structure

Recently, the Adenosine's role on the somnogenic pedestal has been challenged by other potential candidates. Disease, for instance, is known to disrupt sleep. Many infections introduce cytokines into the CNS, which has been shown to induce episodic NREM sleep [14]. Ilk-1 and TNF are the most somnogenically potent of these substances, and are shown to fluctuate with a circadian rhythm. However, IL-1 and TNF produce fragmented NREM sleep, and antagonizing them in healthy individuals yields insignificant changes in NREM sleep duration. Thus, although these cytokines have somnogenic properties, their role in sleep regulation is generally restricted to pathological conditions[14].

Prostaglandin D2
Image Unavailable
PGD2 structure

The Prostaglandins are 20-carbon fatty acid derivatives with a constitutive 5-carbon ring structure. Prostaglandin D2 (PGD2) is produced from arachidonic acid by the action of cyclooxygenase followed by PGD Synthase. Since 1980, PGD2 was known to be the most abundant prostaglandin in the brain, exhibiting a circadian variation, and the concentration of which becomes elevated by sleep deprivation[3]. It is also crucially implicated in mastocytosis, a disorder characterized by long sleep episodes due to PGD2 elevation and symptomatic of African Sleeping Sickness[3]. Clearly this molecule is an important endogenous somnogen. Recent work by Urade et al. is looking into the molecular mechanism of this prostaglandin's interactions.

PGD2 receptors include the DP1 receptor (localized to the leptomeninges of the basal forebrain) and the DP2 receptor (whose cellular distribution is not well known). The current model describes the synthesis of PGD2 by L-PGDS (localized to the leptomeninges, choroid plexus, and oligodendrocytes in the brain), after which PGD2 interacts with DP1 receptors on the ventral surface of the leptomeninges of the rostral basal forebrain. This stimulates an increase in extracellular adenosine concentrations, which act on the sleep centers as described above. It is shown that DP1 knockout mice do not exhibit PGD2-induced sleep, nor does adenosine concentration increase, indicating that adenosine-mediated sleep induction is PGD2-dependent.[3]. This is highly indicative of a molecular mechanism for process S.

PGD2-Induced Sleep
Image Unavailable
PGD2 is synthesized by L-PGDS on the meninges. It circulates in the CSF to interact with DP1 receptors on the ventral surface. This stimulates a increase in Adenosine, which acts on A1 receptors to inhibit the Basal Forebrain and TMH, and on A2A receptors to stimulate the VLPO[3]

Homeostatic Regulation at the Synaptic Level

Research has made tremendous headway into demonstrating that global sleep is homeostatically controlled. Recently, however, studies have also pointed towards homeostatic control at a smaller level: the synapse. Since 2006, research into the homeostatic regulation of synaptic strength has aimed to address the Synaptic Homeostasis Hypothesis, as proposed by Tononi in 2006[15].

Synaptic Homeostasis Hypothesis

Slow wave activity has long been associated with memory consolidation. Slow EEG activity during SWA correlates to increased memory retention[17]. This process is presumed to occur through LTP in the hippocampus and neocortex. This is a synapse-dependent process which is initiated by novel experience during wakefulness. Tononi recognized the problem with unregulated increased potentiation. As such, he proposed a model to provide a homeostatic control to synaptic strength:

  • Wakefulness is associated with synaptic potentiation in several cortical circuits
  • Synaptic potentiation is tied to the homeostatic regulation of slow wave activity
  • Slow wave activity is associated with synaptic downscaling
  • Synaptic downscaling is tied to the beneficial effects of sleep on neural function and, indirectly, on performance.

Synaptic Homeostasis
Image Unavailable
The brain favours a global synaptic weight of 200 arbitrary units. Wakefulness and experience preferrentially increases synaptic weight at certain synapses leading to global synaptic weight greater than 200. Global synaptic downscaling occurs during sleep, eliminating weak synapses. After sleep global synaptic weight returns to 200, with a bias towards the potentiated neuron[15]

The hypothesis relies on the principle that synaptic homeostasis follows Process S, in that there is an increases in synaptic strength through LTP that correlates with the time spent awake. Once sleep ensues, synaptic strength decreases until it reaches a threshold level. This is supported by studies in rats, which show that stimulation of whiskers results in localized increased synaptic density at the associated barrel field[16]. Although the concept of globally down-regulating synapses may seem contradictory to memory consolidation, it is important to take into account the body's energy need. Increased potentiation during wakefulness has a cost in terms of enregy, space and potentially poses the risk of saturating the brain's ability to learn[15]. Thus, the brain proportionally downregulates all synapses thereby eliminating weak connections and preserving the signal-to-noise ratio in those that are critical for memory.

Synaptic Downscaling vs. Plasticity

There is some debate on whether or not Synaptic Homeostasis accurately describes the activity of the brain during sleep. In particular the downscaling effect predicted by Tononi has been challenged by recent rsearch. In a study by Chauvette et al, somatosensory cells that had been stimulated showed no decrease in activity over the first or subsequent SWS episodes[17]. In addition, there was no evidence of increased cortical responsiveness during wakefulness. On the other hand, it has been observed that the amplitude and frequency of miniature EPSCs in the rat cortex increase during wakefulness and decrease during sleep[19]. Clearly, the exact behavior of the cortex during SWS is under extensive debate. There is evidence that synaptic plasticity increases with sleep yet evidence also shows that downscaling does occur in the cortex; given that both processes occur, they must work in concert to achieve memory consolidation[18]. Grosmark et al. were able to demonstrate precisely this. Their study was centered around an episode featuring 2 NREM and 1 REM periods. It was observed that discharge activity steadily increased during NREM sleep as a result of potentiation, while a dramatic decrease was observed during NREM sleep, characteristic of downscaling. Over the whole period, there was a net decrease in firing rate consistent with Tononi's hypothesis[20].

Synaptic Potentiation and Downscaling in Sleep
Image Unavailable
During a REM-NREM cycle, potentiation occurs as a result of SWA. REM sleep then initiates a downscaling effect that ultimately lowers to activity of the synapse to a level below its initial value[20]

Grosmark's study demonstrates that both downscaling and potentiation occur during sleep, but they may be temporally separated such that downscaling is characteristic of REM sleep. Overall, it is conclusive that the body regulates the strength of synapses as a result of memory and novel experience, and this regulation is characteristic of sleep. Although the exact mechanism is still being investigated, it is safe to conclude that the brain homeostatically regulates synaptic activity while carefully selecting and potentiating synapses for memory consolidation.

1. Borbély, A. A two process model of sleep regulation. Human Neurobiology 1982;1:195-204.
2. Saper CB, Cano G, Scammell TE. Homeostatic, circadian, and emotional regulation of sleep. Journal of Comparitive Neurology. 2005;493(1):92–98.
3. Urade Y. Prostaglandin D2 and adenosine as endogenous somnogens. Sleep and Biological Rhythms. 2011;9:10-17.
4. Tononi G, Cirelli C. Sleep function and synaptic homeostasis. Sleep Medicine Reviews. 2006;10:49-62.
5. Pace-Schott E, Hobson J. The Neurobiology of Sleep: Genetics, Cellular Physiology and Subcortical Networks. Nature Reviews. 2002 Aug;3:591-605.
6. Åkersted T, Gillberg M. Effects of Sleep Deprivation on Memory and Sleep Latencies in Connection with Repeated Awakenings From Sleep. Psychophysiology. 1979;16(1):49-52.
7. Jan Dijk D, Beersma D. Effects of SWS deprivation on subsequent EEG power density and spontaneous sleep duration. Electroencephalography and Clinical Neurophysiology. 1989 Apr;72(4):312-320.
8. Borbély A, Achermann P. Sleep Homeostasis and Models of Sleep Regulation. Journal of Biological Rhythms. 1999 Dec;14(6):559-568.
9. Moore J, Chen J, Han B, Meng Q, Veasey S, Beck S, Kelz M. Direct Activation of Sleep-Promoting VLPO Neurons by Volatile Anesthetics Contributes to Anesthetic Hypnosis. Current Biology. 2012 Nov;22(21):2008-2016.
10. Greco M, Fuller P, Jhou T, Martin-Schild S, Zadina J, Hu Z, Shiromani P, Lu J. Opioidergic projections to sleep-active neurons in the ventrolateral preoptic nucleus. Brain Research 2008;1245:96-107.
11. Porkka-Heiskanen T, Strecker R, Thakkar M, Bjørkum A, Greene R, McCarley R. Adenosine: A Mediator of the Sleep-Inducing Effects of Prolonged Wakefulness. Science. 1997 May;5316(276):1265-1268.
12. Puckeridge M, Fulcher B, Phillips A, Robinson P. Incorporation of caffeine into a quantitative model of fatigue and sleep. Journal of Theoretical Biology. 2011;273:44-54.
13. Gallopin T, Luppi P, Cauli B, Urade Y, Rossier J, Hayaishi O, Lambolez B, Fort P. The endogenous somnogen adenosine excites a subset of sleep-promoting neurons via A2A receptors in the ventrolateral preoptic nucleus. Neuroscience. 2005 Jan;134(4):1377-1390.
14. Imeri L, Opp M. How (and why) the immune system makes us sleep. Nature Reviews Neuroscience. 2009 Mar;10:199-210.
15. Tononi G, Cirelli C. Sleep function and synaptic homeostasis. Sleep Medicine Reviews. 2006;10:49-62.
16. Knott GW, Quairiaux C, Genoud C, Welker E. Formation of dendritic spines with GABAergic synapses induced by whisker stimulation in adult mice. Neuron 2002;34(2):265–73.
17. Chauvette S. Sleep Oscillations in the Thalamocortical System Induce Long-Term Neuronal Plasticity. Neuron 2012;75(6);1105-1113.
18. Born J, Feld G. Sleep to Upscale, Sleep to Downscale: Balancing Homeostasis and Plasticity. Neuron 2012;75(6);933-935.
19. Liu, Z.W., Faraguna, U., Cirelli, C., Tononi, G., and Gao, X.B. Direct evidence for wake-related increases and sleep-related decreases in synaptic strength in rodent cortex. Journal of Neuroscience 2010;30:8671–8675.
20. Grosmark A, Mizuseki K, Pastalkova E, Diba K, Buzsaki G. REM Sleep Reorganizes Hippocampal Excitability. Neuron 2012;75:1001-1007.

Add a New Comment
Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License