Neurotransmitter Systems

Based on the above account of the neuronal pathways thought to be responsible for the basic sleep-wake cycle, the neurotransmitters that are most likely to be involved in the cycle are those which:

(1) Are released either in the cortex or the non-specific thalamic nuclei.

(2) Augment, or more probably, break up thalamic-cortico synchrony and its tendency to promote slow-wave EEG activity and non-REM sleep. Whether this results in full arousal, or merely a temporary disruption of sleep to give REM periods without full awaking, will depend on the balance of inputs and the overall state of cortical activity.

Some of these inputs come from cholinergic, histaminergic, noradrenergic and 5-HT neurons. These neurons innervate the cortex more than the thalamus and their possible roles will be considered in the following sections. This material draws on studies designed to show: which neurotransmitters are associated with those brain structures concerned with sleep and waking; how their function may change during the cycle; to what extent pharmacological manipulation of their activity influences the cycle; and how drugs which modify our state of arousal affect neurotransmitters.


Studies of several animal species, ranging from rats to sheep, have shown that the release of acetylcholine (ACh) into cortical cups (see Chapter 4 and 6) is increased in proportion to cortical (EEG) activity, being maximal during convulsions and lowest under deep anaesthesia. These findings are consistent with evidence that cortical arousal (EEG desynchronisation) is increased by injection of ACh into the carotid artery of animals, or by direct stimulation of the ascending reticular system (ARAS), and that both these actions are blocked by the muscarinic receptor antagonist, atropine. It has even been shown in humans that REM sleep is induced by intravenous infusion of centrally-acting cholinomimetic agents, such as arecoline or physostigmine (an acetylcholinesterase inhibitor), and, again, the effects of these treatments are inhibited by atropine. Yet antimuscarinic drugs do not have any marked sedative effects on behavioural arousal. This could mean that sedation requires recruitment of the 'sleep' system, as well as blockade of arousal.

As outlined previously (Chapter 6), cholinergic neurons are located in two broad groups of nuclei, both of which are linked to the ARAS and thalamus (Fig. 22.6). One group lies rostrally in the basal forebrain, within the nucleus basalis, medial septum and diagonal band. This system is more active during the waking state than during sleep and blocking its effects could well explain how antimuscarinic drugs inhibit EEG desynchronisation. The nucleus basalis, which sends diffuse projections to the cortex and hippocampus, has also been linked with memory function (Chapter 18).

The second cluster of neurons lies more caudally, near the pons, in the pedunculo-pontine (PPT) and laterodorsal tegmental (LDT) nuclei (see Fig. 22.6) and could be regarded as part of the ARAS (see McCormick 1992). It innervates the non-specific thalamic nuclei as well as some more specific ones like the lateral geniculate nucleus (visual pathway), the pontine reticular formation and occipital cortex. Because long

Reticular Formation

Figure 22.6 Cholinergic influences on sleep and arousal. Cholinergic neurons are found primarily either rostral to the ascending reticular activating system (ARAS) in the nucleus basalis (NcB) caudally in the pedunculo pontine tegmentum (PPT) nucleus. The former, which innervate much of the cortex, receive inputs from the ARAS and appear to be partly responsible for maintaining the EEG and behavioural arousal. The latter innervate non-specific (NspThNu) and specific (SpThN) thalamic nuclei, including the lateral geniculate nucleus as well as the pontine reticular formation (PRF) and occipital cortex (OC). The high-voltage pontine-geniculo-occipital (PGO) waves they initiate in all three areas are characteristic of REM sleep, which is reduced by their destruction

Figure 22.6 Cholinergic influences on sleep and arousal. Cholinergic neurons are found primarily either rostral to the ascending reticular activating system (ARAS) in the nucleus basalis (NcB) caudally in the pedunculo pontine tegmentum (PPT) nucleus. The former, which innervate much of the cortex, receive inputs from the ARAS and appear to be partly responsible for maintaining the EEG and behavioural arousal. The latter innervate non-specific (NspThNu) and specific (SpThN) thalamic nuclei, including the lateral geniculate nucleus as well as the pontine reticular formation (PRF) and occipital cortex (OC). The high-voltage pontine-geniculo-occipital (PGO) waves they initiate in all three areas are characteristic of REM sleep, which is reduced by their destruction bursts of high-voltage waves occur in all these three terminal areas during REM sleep, forming the pontine-geniculo-occipital (PGO) waves described above, they could derive from the PPT (see Hobson 1992).

In fact, there is a good deal of evidence to support this suggestion. First, more than half the neurons in the PPT fire rhythmically only when PGO waves are evident and their firing starts immediately before the PGO waves appear. Second, in cats, REM sleep is augmented by direct injection of either carbachol, or more selective muscarinic agonists, or the anticholinesterase, neostigmine, into the pontine reticular formation (one of the projection sites for PPT). Third, REM sleep is abolished by lesion of the PPT nucleus but, interestingly, not by lesion of the LDT.

Overall, there are compelling reasons to believe that cholinergic pathways not only play a part in arousal but also contribute to the induction of the 'arousal-like' features of REM sleep.


Although histamine has mixed excitatory and inhibitory effects on central neurons, those antihistamines (H1-receptor antagonists) that enter the brain produce sedation; this indicates that the predominant overall effect of histamine is excitatory. The preferred explanation for this rests on evidence that histaminergic neurons in the posterior hypothalamus are active in waking and silent in deep SWS and REM sleep.

The histamine neurons in the tuberomammillary nucleus, in the posterior hypothalamus, project to the cortex and thalamus and receive an afferent input from

Ips Drug Screening
Figure 22.7 Histamine influences on sleep and arousal. The activity of histamine-releasing neurons increases with arousal and diminishes during sleep. Both H antagonists and agonists of H3-autoreceptors depress release of histamine and reduce arousal (see text for details)

GABAergic neurons in the ventrolateral preoptic area (VLPO). Since the VLPO is more active in SWS sleep, this phase of the sleep cycle could depend in part on GABAergic inhibition of histamine-releasing neurons that project to the cortex (Fig. 22.7). What activates the VLPO is not clear, however. There also seems to be some feedback control of histamine release because H3-receptor agonists, that activate the autoreceptors on histamine-releasing neurons and reduce release of this transmitter, augment SWS while H3-receptor antagonists have the opposite effect. Finally, other effects of histamine that could contribute to increased arousal are increasing the activity of excitatory cholinergic neurons in the basal forebrain and inhibition of neurons in the hypothalamic preoptic area which promote sleep.

A much higher profile has recently been claimed for histamine in the control of circadian rhythm (see Jacobs, Yamatodani and Timmerman 2000). When injected intracerebroventricularly in rats it appears to alter locomotor and drinking rhythms in a somewhat complex manner depending on when it is given in the light-dark cycle, being most active when the animals are in constant darkness. Some of the effects can also be mimicked by increasing the amount of endogenous histamine released with the H3 autoreceptor antagonist thioperamine. Certainly histamine has both excitatory (Hi) and inhibitory (H2) effects on SCN neuron firing and autoradiography has revealed the presence there of Hi receptors. Since glutamate and 5-HT have been shown to increase histamine release in the SCN and GABA to inhibit it, the above authors consider histamine to be the final mediator of their effects. Whether this is so remains to be seen for, despite the sedative effects of some Hi antagonists, rhythm changes have not been reported with their long-term clinical use.


Although some studies show that noradrenaline inhibits neuronal firing it is generally considered to increase behavioural activity and arousal. This impression is borne out to the extent that CNS stimulants, like amphetamine, increase release of noradrenaline and produce behavioural and EEG arousal, while reserpine, which reduces noradrena-line storage and hence release, causes psychomotor retardation. It is also supported by evidence that the firing rate of neurons projecting from the locus coeruleus is greater during waking (1-2 Hz) than during SWS (0.2-0.5 Hz) and is increased even more as behaviour progresses from vegetative or consummatory activities (e.g. grooming or feeding) to vigilance. Furthermore, stimulation of the locus coeruleus in cats causes EEG desynchronisation and increases arousal, while a neurotoxic lesion of these neurons leads to EEG synchrony, increases SWS and reduces REM sleep. In fact, some ('REM-off') cells in the locus coeruleus stop firing altogether during REM sleep. Because a reduction in the activity of noradrenergic neurons precedes the onset of sleep, this change in activity is thought to have a permissive role in sleep induction.

How all these actions of noradrenaline are manifest is not clear and, unfortunately, most experiments in this area have been carried out on anaesthetised animals which, arguably, are not ideal for investigating mechanisms underlying arousal! One of the few investigations to have been carried out in unanaesthetised rats has shown that infusions of noradrenaline into the nucleus basalis of the medial septum increases waking (and the y-wave activity of the waking phase), but reduces the y-waves of SWS.

These changes, which are thought to be mediated by activation of ^-adrenoceptors, suggest that noradrenaline increases cholinergic influences on arousal, in the nucleus basalis, at least (Cape and Jones 1998). However, a fairly common side-effect of ^-adrenoceptor antagonists, used clinically to relieve hypertension, is sleep disturbance which is expressed as nightmares, insomnia and increased waking. Clearly, these drugs must have additional actions either in other brain centres, or non-selective effects on other (possibly 5-HT1A) receptors that have quite different effects on arousal. It has even been suggested that ^-blockers disrupt sleep patterns by inhibiting melatonin synthesis and release, but this is controversial.

In contrast, a2-adrenoceptor agonists are well-known for their sedative effects. Since their activation of presynaptic a2-autoreceptors will reduce noradrenergic transmission, by depressing the firing of neurons in the locus coeruleus and release of noradrenaline from their terminals, this action is entirely consistent with the proposal that increased noradrenergic transmission increases arousal. Although this presynaptic action of a2-agonists would explain their sedative effects it must be borne in mind that many ^-adrenoceptors in the brain are in fact postsynaptic. Their role (if any) in sedation is unclear but it must be inferred that, if they make any contribution to sedation, then either a specific brain region or a specific a2-adrenoceptor subtype is involved. Another possible confounding factor is that many a2-adrenoceptor ligands have an imidazoline structure (see Chapter 8) and the recently discovered imidazoline receptors are also thought to influence the sleep cycle and arousal. Even less is known about the role of ^-adrenoceptors on arousal partly because most drugs acting at these receptors do not readily cross the blood-brain barrier.

The role of noradrenergic neurons from the locus coeruleus on behaviour during the waking phase is rather controversial. It is doubtful that noradrenaline release is actually required for waking because animals with more than a 90% lesion of these neurons are still capable of staying awake, although they are rather subdued. Nevertheless, the single-unit activity of these neurons is increased by sensory stimuli ranging from those that cause physical discomfort (e.g. tailpinch) to environmental stimuli (e.g. tones and light flashes), especially those that provoke orientation to the stimulus (e.g. approach of the experimenter). The evoked neuronal response typically shows a brief (phasic) burst of activity followed by a quiescent period of post-stimulus inhibition but this response, along with behavioural arousal, habituates on successive presentation of the stimulus.

Such findings have led to suggestions that neurons in the locus coeruleus complex serve as a central 'alarm' system while others have argued that their increased neuronal firing during the waking period mediates changes in 'selective attention'. It has even been suggested that the tonic activity of these neurons could determine overall arousal, whereas the more transient, phasic, response determines 'attentiveness'. In fact, these neurons could serve all these purposes, thereby helping to protect the individual from threatening stimuli as well as directing attention to interesting, or salient environmental features (see also Chapter 8).

Few studies have investigated the role in behaviour of noradrenergic neurons originating in the nuclei of the lateral tegmental area (see Chapter 8). However, what little evidence there is suggests that they respond primarily to unconditioned environmental stimuli but are capable of adaptive changes in their activity on repeated presentation of the stimulus. Because noradrenergic neurons arising in the lateral tegmental nuclei have numerous reciprocal connections with other brainstem nuclei involved in homeostasis (e.g. regulating blood pressure and heart rate), it is likely that they make an important contribution to the adjustments in the activity of the peripheral autonomic system during the various states of sleep and waking (see Goldstein 1995).


The role of this neurotransmitter in the sleep-waking cycle has not received as much attention as that devoted to noradrenaline and interpretation of existing evidence is not straightforward. On the one hand, the firing rate of neurons projecting from the dopaminergic neurons in the ventral tegmental area does not vary across the sleep-waking cycle and, in any case, the dopaminergic innervation of the cortex is much more restricted than that of noradrenaline or 5-HT. On the other hand, drugs that modify dopaminergic transmission do affect arousal albeit in complex ways (see Gottesmann 1999).

Low doses of the dopamine agonist, apomorphine, induce SWS and, in humans, dopamine agonists can induce somnolence which is a problem when treating Parkinson's disease. This action is thought to be due to activation of presynaptic D2-autoreceptors and some antagonists of this receptor increase waking state and reduce both non-REM and REM sleep. That a reduction in firing of dopaminergic neurons is associated with reduced arousal is consistent with evidence that local infusion of GABA into the dopaminergic ventral tegmental area also reduces waking. However, others have suggested that activation of postsynaptic D2-receptors in the dorsal striatum is responsible.

By contrast, high doses of dopamine agonists increase arousal and cortical desynchronisation, possibly by activating postsynaptic D2-receptors. Indeed, local infusion of dopamine into the nucleus accumbens increases waking, an effect blocked by the D2-receptor antagonist, haloperidol. Such an action is consistent with the general improvement in sleep (especially sleep continuity) in patients treated with neuroleptics, such as haloperidol and clozapine, which share D2-receptor antagonism as a common target. However, the various changes seen in the different phases of the EEG seem to depend on the actual compound tested.


This neurotransmitter presents something of a paradox in respect of its role in sleep and waking behaviour, although its importance to both is undoubted. Early experiments suggested that an increase in 5-HT transmission actually helps to induce sleep (see Jouvet 1974). Thus ^CPA, which blocks the synthesis of 5-HT, causes insomnia in cats and reduces SWS; this insomnia is reversed by giving the 5-HT precursor, 5-hydroxytryptophan (5-HTP), which bypasses the ^CPA block. Also, a lesion of the dorsal Raphe nucleus (DRN) produces insomnia, the degree of which is proportional to the loss of 5-HT neurons and the decrease of 5-HT turnover in their projection areas. Despite such lesions, sleep patterns return to normal after some days and, if they are made in new-born rats, sleep patterns normalise after a few weeks, suggesting that they are not solely dependent on 5-HT.

In contrast to this evidence that 5-HT activity decreases arousal, antidepressants are generally thought to increase serotonergic transmission while the central depressant, reserpine, reduces it, although it must be remembered that both these treatments affect central noradrenergic transmission as well. Nevertheless, direct stimulation of Raphe neurons, or systemic administration of a 5-HT precursor, actually increases waking. This suggests that 5-HT has either an excitatory influence on behaviour and/or an inhibitory effect on sleep. This view is supported by electrophysiological recordings of the activity (firing frequency) of neurons in the cat DRN. Insofar as it can be certain that it is serotonergic neurons that are being monitored in this nucleus, these studies have shown that, during quiet waking, their activity is about 2-3 spikes/s, but that this rate decreases progressively and becomes less regular as sleep progresses to SWS. In fact, these neurons become virtually totally quiescent during REM sleep and this reduction in activity is probably effected by GABAergic inputs to the DRN. For a review of all this evidence, see Jacobs and Azmitia (1992).

Assigning a particular role for changes in 5-HT transmission in sleep is confounded by the existence of 5-HT neurons in several distinct Raphe nuclei (Fig. 22.8; see also Chapter 9). These project to different regions of the brain but the differences in their functional influences are, as yet, poorly understood. Most studies have in fact investigated the DRN, which innervates forebrain areas, but it does seem that other serotonergic nuclei in the medulla show a similar pattern of responses. Thus, neurons in the 'inferior' Raphe nuclei (the Raphe magnus (NRM), the nucleus Raphe obscurus (NRO) and the nucleus Raphe pallidus (NRP)) (see Fig. 22.8) which project to the lower brainstem and spinal cord, all show a reduced discharge during SWS when compared with that in the awake subject. However, their firing rate is generally higher than in the DRN. Moreover, unlike DRN neurons, those in the NRO and NRP continue to fire, albeit at a reduced frequency, during REM sleep. The implications of these differences in the regulation of the sleep cycle are unclear.

The role of 5-HT transmission in waking behaviour is even less clear. The tonic activity of DRN neurons during 'active' waking is certainly greater than during 'quiet' waking but it is not increased further by arousing or threatening stimuli. However, environmental stimuli that provoke behavioural orientation induce a marked phasic increase in serotonergic neuronal activity (see Chapter 9) suggesting that they do have some role in the response to stimuli requiring attention.

A link between 5-HT release and increased waking is supported by evidence from in vivo microdialysis of cats and rats. This has confirmed that the extracellular concentration of 5-HT in all brain regions studied to date is lower during both SWS and REM sleep than in the awake state (see Portas, Bjorvatn and Ursin 2000). Interestingly, if behaviour is maintained at a constant level, the activity of 5-HT neurons does not show circadian variation although 5-HT turnover in the brain areas to which they project

Ecosistema Dibujos Faciles

Figure 22.8 The distribution of brainstem Raphe nuclei. Neurons that release 5-HT are clustered in two groups of nuclei in the pons and upper brainstem. The 'superior' group, which projects to forebrain areas, includes the dorsal Raphe nucleus (DRN) and the median Raphe nucleus (MRN). The 'inferior' group projects to the medulla and spinal cord and includes the nucleus Raphe pallidus (NRP), the nucleus Raphe obscurus (NRO) and the nucleus Raphe magnus (NRM)

Figure 22.8 The distribution of brainstem Raphe nuclei. Neurons that release 5-HT are clustered in two groups of nuclei in the pons and upper brainstem. The 'superior' group, which projects to forebrain areas, includes the dorsal Raphe nucleus (DRN) and the median Raphe nucleus (MRN). The 'inferior' group projects to the medulla and spinal cord and includes the nucleus Raphe pallidus (NRP), the nucleus Raphe obscurus (NRO) and the nucleus Raphe magnus (NRM)

does show such a rhythm. The reasons for this apparent dissociation between firing rate and transmitter release are not clear but it does suggest that neuronal firing rate is not necessarily a reliable indicator of transmitter release in the terminal field.

One specific theory for the role of 5-HT in arousal suggests that serotonergic transmission serves to coordinate target cell responses by adjusting their excitability to match the subjects' general level of arousal. In so doing, they are responsible for gating motor output and coordinating this with homeostatic and sensory function (Jacobs and Azmitia 1992; Jacobs and Fornal 1999). This would be consistent with evidence that, like the noradrenergic system, increases in the firing rate of neurons in the DRN precede an increase in arousal. The frequency of discharge would code the state of arousal and prime target cells for forthcoming changes in the response to sensory inputs.

Apart from the problem of trying to associate the effects of 5-HT with specific nuclei, there is also no clear picture of which 5-HT receptors mediate any of these changes in sleep and waking. This is not least because of the large number of receptor subtypes, the limited receptor selectivity of most test drugs, species differences in the response, as well as time- and dose-related differences in the response to any given agent. 5-HT is also known to affect noradrenaline and dopamine release in the brain (see Stanford 1999) and such interactions undoubtedly explain some of the inconsistencies between the early findings and recent studies of the role of these different 5-HT neurons in sleep.

Nevertheless, it is evident that activation of many different receptor subtypes affect the sleep-waking cycle. For instance, recent evidence suggests that activation of 5-HT1a, 5-HT1B, 5-HT2a/c and 5-HT7 receptors in the SCN all affect circadian rhythms. Activation of 5-HT1B (presynaptic) receptors in the retinohypothalamic tract is thought to attenuate 5-HT release and so blunt light inputs to the SCN and reduce its phototic regulation. In contrast, postsynaptic 5-HT7 receptors, 5-HT2C, and possibly postsynaptic 5-HT1a receptors, are thought to have an important role in phototic entrainment and to mediate phase-shifts in circadian rhythms (reviewed by Barnes and Sharp 1999). In addition to these effects on circadian rhythms, it is clear that 5-HT receptors affect sleep more directly. A detailed review of this subject is to be found in Portas, Bjorvatn and Ursin (2000) but key findings are summarised here.

Table 22.1 Effects of activation of 5-HT receptors on sleep-waking cycle

Receptor REM Waking Location Effect on 5-HT transmission

5-HT1A T I Presynaptic |

5-HT1A I I Postsynaptic T

5-HT1B I T Not known ?

5-HT2A/2C | T ?Postsynaptic T

5-HT3 T Postsynaptic T

The actions of 5-HT1A receptor agonists in rats depend on their route of administration (Bjorvatn and Ursin 1998). When they are given systemically they cause a transient increase in waking time and a reduction in SWS and REM sleep which is followed by a delayed increase in SWS. This latter response is possibly mediated by activation of inhibitory postsynaptic 5-HT1A receptors in the nucleus basalis (Table 22.1). Certainly, local infusion of 5-HT1A agonists into this area increases SWS. Another contributory factor is suggested by the reduction in waking and increase in SWS following intrathecal infusion of 8-OH-DPAT. This is thought to reflect inhibition of primary sensory afferents, by activation of presynaptic 5-HT1A receptors, an action which would be conducive with induction of sleep. However, infusion of low concentrations of the 5-HT1A agonist, 8-OH-DPAT, into the DRN to activate autoreceptors induces a type of REM sleep which is explained by a reduction in the firing rate of 5-HT neurons. In turn, this is presumed to result in disinhibition of mesopontine cholinergic neurons in the PPT and LTD nuclei which are responsible for REM sleep. Such a scheme is supported by evidence that local infusion of a 5-HT1A agonist into these areas reduces REM sleep, presumably by inhibition of mesopontine cholinergic neurons by postsynaptic 5-HT1A receptors.

Administration of 5-HT1B receptor agonists increases waking time and reduces REM sleep. This is consistent with recent evidence gathered from 5-HTiB-receptor knockout mice which exhibit more REM sleep and less SWS than the wild-type. Moreover, 5-HT1B agonists reduce, while antagonists increase, REM sleep in the wild-type mouse, but neither type of compound has any effect in the knock-outs (Boutrel et al. 1999). Unfortunately, it is not known whether these actions are mediated by presynaptic, postsynaptic or heteroceptors and therefore whether 5-HT activity is increased or decreased. It is also not helped by the limited selectivity of test agents.

5-HT2A/2C agonists increase waking and reduce SWS and REM sleep in humans and rats, possibly through an action in the thalamus. Conversely, blockade of 5-HT2A receptors, e.g. by ritanserin, increases SWS, an action that might contribute to the beneficial effects of antidepressants that share this action. However, these findings are confounded by evidence that activation of 5-HT2C receptors increases SWS.

Infusion of 5-HT3 receptor agonists into the nucleus accumbens increases waking and reduces SWS, although REM sleep is unchanged. These effects of 5-HT3 receptor activation are prevented by co-administration of a D2-receptor antagonist. This is consistent with evidence that activation of 5-HT3 receptors can increase dopamine release and points to functional interactions between these two groups of neurons that affect the sleep-waking cycle. Such interactions will certainly confound any attempts to define the specific role of 5-HT in the regulation of sleep and arousal.

Overall, 5-HT transmission seems to increase during waking and to decline in sleep although it may only reach its minimal level, in some neurons anyway, during REM

sleep. Whether its role is simply to prime target cells to enable an increase in the motor activity associated with waking, as has been suggested, remains to be seen.


It is perhaps not surprising that, since adenosine has been presented as an endogenous inhibitor of neuronal function with its antagonists, like theophylline, being stimulants (see Chapter 13), it should have been implicated in sleep induction.

In fact EEG studies have shown that administration of an adenosine A1 agonist increases SWS in humans and induces it in sleep-deprived rats while adenosine also inhibits the important cortical activating brainstem cholinergic neurons. Of more physiological significance is the finding from microdialysis in rats that the extracellular concentration of adenosine progressively increases in the hippocampus, reaching a maximum at the end of the animal's active (lights off) period. After that, it falls sharply within an hour as the animal enters the quiet (lights on) sleepy period (see Huston et al. 1996). Of course, the hippocampus is not generally associated with sleep patterns and whether these studies establish adenosine as a potential sleep inducer, or merely as an 'activity-restrictor' that facilitates sleep, is unclear.


Since most excitatory transmission is mediated by glutamate this must be involved in the sleep-waking cycle. It certainly mediates the input of the retinohypothalamic tract to the SCN, apart from afferent inputs more generally to the ARAS, etc. So far, specific in vivo manipulation of the direct glutamate input to the SCN has not been possible.

The fact that SCN neurons contain GABA, and that this appears to be the neurotransmitter released by the geniculohypothalamic tract onto the SCN, clearly puts it in a prime position for regulation of sleep rhythms. However, its precise role is unclear, not least because it can act as an excitatory, as well as an inhibitory, neurotransmitter in this nucleus and that these varied responses appear to follow a circadian rhythm (see Chapter 11). Again, specific manipulation of this pathway is difficult although GABA enhancement generally (e.g. by benzodiazepines) is, of course, sedative (see later section on drug-induced sleep).


In classical times, sleep was thought to be induced by sleep factors (vapours) emanating from food in the stomach. To this day, and despite the encyclopedic evidence that neurotransmitters have discrete effects on sleep and arousal, the idea still lingers that there are sleep-inducing ('somnogenic') factors. These are thought to have a pervading influence on sleep throughout the brain, although the stomach is no longer regarded as their source! This view was strongly encouraged by experiments, carried out in the early twentieth century, by Pieron in Paris, who showed that the CSF of sleep-deprived dogs contained a substance that had a somnogenic effect when infused into non-sleep-deprived animals. Since then, many candidate sleep substances have emerged, some of which are more convincing than others.

The first serious attempts to identify and characterise an endogenous somnogenic agent was carried out by Pappenheimer and colleagues (see Pappenheimer 1983) who found that transferring samples of CSF from sleep-deprived goats into normal rabbits increased the latter's REM sleep. Chemical extraction from thousands of rabbit brains and many gallons of human urine yielded a sleep factor and established it as a muramyl peptide. Unfortunately muramyl peptides are not synthesised by mammalian cells but are components of bacterial cell walls. Apart from the obvious possibility of mere contamination, it is not clear how the substance turned up in the CSF and brain tissue. Despite this setback, and some scepticism about whether somnogenic peptides exist at all, research still continues in this area and many candidates have been suggested. These include well-known peptides such as prolactin, CCK-8, VIP and somatostatin as well as some novel ones such as ¿-sleep-inducing peptide. (For a full review of this subject, see Garcia-Garcia and Drucker-Colin 1999.)

Another line of research has produced convincing evidence that the pro-inflammatory cytokines, interleukin IL-1^ and TNFa modify the sleep cycle: these agents generally increase non-REM sleep and suppress REM sleep. IL-6 also reduces REM sleep and SWS in the first half of the sleep cycle but subsequently increases SWS. However, all these responses vary with dose, test species and even time of day. These factors are produced by T-cell lymphocytes but their receptors are associated with neurons, astrocytes, microglia and endothelial cells. Because these agents induce nitric oxide synthase, and there is some evidence that nitric oxide increases waking, possibly through modulation of ACh release in the medial pontine reticular formation, there is no need for them to cross the blood-brain barrier (although there is evidence that they do). Nevertheless, how these factors actually cause changes in the sleep cycle is as yet unclear. An indirect effect via changes in the rate of prostaglandin synthesis (see below) is one possibility but others include modulation of 5-HT2A-mediated serotonergic transmission and suppression of glutama-tergic neuronal activity through an adenosine-dependent process.

Prostaglandins, in particular PGD2, have also been shown to act as sleep-promoting substances. PGD2 is synthesised in the arachnoid membrane and choroid plexus and its receptors are prevalent in the basal forebrain. Moreover, its concentration in the CSF shows a circadian rhythm and increases during sleep deprivation. It is not yet known how PGD2 influences sleep but when it is infused locally, it changes the firing rate of neurons in the preoptic and basal forebrain areas in ways suggesting that it promotes sleep. Like the interleukins, and TNFa, mechanisms proposed to explain these actions include modification of monoaminergic or adenosinergic transmission.

Finally, the endogenous fatty acid amide, oleamide, has somnogenic effects. This compound is chemically related to the endogenous ligand for cannabinoid receptors, anandamide. Although oleamide has even been reported to augment anandamide binding to cannabinoid (CB1) receptors, it is still not known whether this action is relevant to its somnogenic effects. Oleamide has been shown to potentiate (benzodiazepine-sensitive) GABAA receptor responses through a mechanism that seems to involve the y-subunit. However, modifications of 5-HT2, muscarinic, metabotropic glutamate and NMDA receptor function have all been suggested as possible mechanisms.

Continue reading here: Druginduced Sleep

Was this article helpful?

0 0