Sleep and dreaming:
induction
and mediation
by cholinergic
of REM sleep
mechanisms
J. Allan Hobson
The
Harvard Medical School, Boston, Massachusetts,
USA
most
sleep
focused eye
important
on the
movement
recent
precise
work
cellular
sleep mediation.
acetylcholine-containing triggering
and
maintenance
new
studies
provide
support
generator
the and
in the the
indirect eye
pons that
influences.
A growing new
thinking
consensus about
regarding the
waking
Current
Opinion
brain
the
implicates
as critical
movement
hypothesis
has
of rapid
evidence
system is gated during waking by serotonergic
stimulated
of
mechanisms
peribrachial
of rapid for
neurobiology
biochemical
Direct
neurons
the
on
and
sleep. the
in
Other
cholinergic
and noradrenergic
basic
neurobiology
basis of consciousness
has during
and dreaming.
in Neurobiology
Introduction
1992,
2:759-763
REM sleep mediation cholinoceptive
The role of acetylcholine (ACh) in promoting sleep, and especially rapid eye movement (REM) sleep, has been suspected for many years, but only recently has the cellular and molecular basis of its mediating role begun to be clarified. The critical cholinergic neurons are found in the peribrachial region of the pons. They project to the paramedian pontine reticular formation and to the thalamus where they excite other cells which directly mediate the rostral signs of REM sleep, including the cortical electroencephalogram activation, the ocular saccades (or REMs), and the ponto-geniculo-occipital (PGO) waves (Fig.la,b,c).
via cholinergic
neurons
and
in the pons
The pedunculo pontine tegmental nucleus (PPT), which contains the Ch5 group of cholinergic neurons, has been of interest to sleep scientists since it was discovered that the cells of this region fire in bursts in advance of ipsiversive REMs and the larger PGO waves in the ipsilateral geniculate body (see Fig.la,b). Together with the Ch6 group of the dorsolateral tegmental nucleus, these neurons constitute the entire brainstem compliment of ACh-containing cells. Lesion studies now concur in showing that the selective suppression of REM and PGO waves [4] is proportional to the loss of cholinacety-transferase activity [ 21. Brainstem transection at the prebulbar or caudal pontine levels eliminates the atonia, but not the PGO waves or eye movements that are triggered by carbachol [ 51, indicating again a necessary link to the medullary inhibitory reticular function.
An indirect effect of the rostral pontine trigger zone upon magnocellular medullary elements (which execute the muscle atonia of REM) may be mediated by gluta mate. The cholinergic system and its postsynaptic REM sleep executive population are modulated by serotonergic, noradrenergic, and dopaminergic inputs which are, in general, inhibitory. The REM sleep bursting pattern of cholinergic neuronal discharge is thought to be mediated by local y-aminobutync acid (GABA)ergic interneurons via membrane hyperpolarization. The details of these effects are the main focus of this review. An emerging theoretical model reconciling previously controversial theory can be seen in three recent reviews [ 1,2,3-l.
Because the muscle atonia of REM is not as dramatically altered by PPT lesions as when the lesions are in the locus subcoeruleus, a non-cholinergic mediation of this REM sleep sign is suggested. The magnocellular medullary neurons (which are thought to inhibit the anterior horn cells via a glycinergic mechanism) may be excited by glutamate [6]. Interestingly, most of these neu-
Abbreviations ACLacetylcholine;
AChE-acetylcholinesterase;
5-HTP-5-hydroxytryptophan;
GABA--y-aminobutyric
PGO-ponto-geniculo-occipital; REM-rapid
@
Current
Biology
acid; 5-HT-5-hydroxytryptamine;
PPT-pedunculo
pontine
tegmental
nucleus;
eye movement.
Ltd ISSN 09594388
759
760
Neural
control
(a) Location of PC0
(c) Neurophysiological and schematic
cell
Burst
(b) Long (0) and short REM Induction sites
cells
(0)
term
recording
Percentage of REM sleep signs
LGB
LGBL
LGB
LGBW
EEG
EEG
60
1
L
R
v
Post carbachol Injection days
5 Sec.
Fig.1. Relationship
of PC0 burst cells to cholinergic REM induction sites. (a) Filled circle indicates site of injection of carbachoi into peribrachial pons. Small dots indicate location of cholinergic cells and crosses location of PC0 burst cells. (b) Location of peribrachial long-term (filled circle) and paramedian short-term (open circle with dots) REM induction sites. k) On the left is an extracellular unit recording of a PC0 burst cell and PC0 waves in the ipsilateral LGB. On the right is a hypothetical wiring diagram of the PC0 trigger zone (PB) elements (PC0 and ACh) with modulatory aminergic raphe (R) and locus coeruleus (LC), thalamic (LCB), and paramedian pontine reticular cells (FTC). (d) Injections of carbachol at the PB site shown in (a) and (b) produce immediate PC0 waves in the ipsilateral LGB. (e) After 24 h REM triples and remains elevated for 6 days, this is the long-term REM effect. Reticular tegmental nuclei (FTC, FTP, FTL, FTC). Aminergic nuclei (R, LC). Cholinergic nuclei (C5, C6). Reference structures shown are red nucleus (RN), trigeminal motor nucleus (MN), and brachium conjuctivum (BC). (a,b and c) reproduced with permission from S Datta, J Quattrochi and JA Hobson (unpublished data). (d and e) are reproduced with permission from [14*1.
rons are not activated in canine narcolepsy, suggesting a Werent mechanism for the atonia of that disease [7]. An important series of electrical stimulation studies indicates that the cortical desynchronization of REM sleep is mediated via a muscat-ink cholinergic depolarization of neurons in the ST, VS-VL, and VL nuclei of the thala-
mus [@I. The cholinergic excitation appears to disrupt the sequences of bursting discharge and afterhyperpolarizations that underlie the spindling pattern of slow wave sleep, and which are due to GABAergic inhibition of nucleus reticular-is and local circuit inhibitory neurons [9*]. The possibility that ACh inhibits reticularis neurons [ 101
Sleep and dreaming
is consistent with the presumed cholinergic mediation of cortical synchronization. The 20-40Hz oscillations that can be recorded in thalamocortical neurons also seem to be mediated in part by the brainstem cholinergic system (PPT/ChS), as they are enhanced by stimulation of that structure, blocked by scopolamine, and unaffected by lesions of the Chl-Ch4 cell groups, which constitute the forebrain cholinergic ensemble in the nucleus basalis of Meynert [ 11.1. In a related study a role for the cholinergic system in the mediation of the cat mid-latency audition-evoked potential has been suggested [ 121. A major impetus for the recent increase in experimental work on the cholinergic mediation of REM sleep is the continuing refinement and exploitation of the chemical microstimulation technique using carbachol and other muscarinic agonists to trigger REM sleep. Direct evidence of cholinergic effects upon cholinoceptive mPRF neurons of the pons has recently been obtained via intracellular recording in rat brainstem slices (131. The cholinergic microstimulation approach has also been used to show that ipsiversive eye movements and ipsilateral thalamic LGB cell PGO waves are immediately excited by carbachol (see Fig.ld) when it is microinjected into the caudo-lateral PPT (or PB; see Fig.la,b). Following a delay of several hours to one day (during which waking behavior is actually enhanced) REM sleep triples in amount and thereafter remains significantly elevated for 6 days (see Fig.le) [ 14.1. Such long-term increases in sleep are unprecedented and imply the activation, perhaps via the intense bursting of the cholinergic neurons themselves, of an intracellular metabolic process (such as the nuclear synthesis of cholinacetyltransferase). Both the cholinergic (PB) and cholinoceptive (mPRF) regions of the pons are among the structures shown by increases in glucose use to be metabolically active during natural REM sleep [15]. When carbachol is microinjected into the paramedian pons (see Fig.lb) it produces a more immediate and much briefer period of REM enhancement, as if a cholinoceptive executive population of neurons were being activated without affecting long-term regulatory processes. In such cases an increase in ACh release can be measured in the contralateral pons [ 161. That the carbachol-induced REM-like state is a physiologically valid model is suggested by its association with the same kind of respiratory depression that is seen in natural REM sleep [ 171.
Functional
studies of REM sleep and PC0
waves Pm-like waves can be elicited by tones in all behavioral states and tend to be associated with the orienting response when they are triggered during waking. The amplitude, frequency and latency of the waves are all dependent upon stimulus intensity [18] and can still be recorded in waking when the orienting response has
habituated, indicating ciable [ 191.
that the two processes
Hobson
are disso-
When auditory stimuli are delivered to cats they produce first an excitation and then an inhibition of dorsal pontine neurons; the inhibitory component is reduced by 24-48 h of REM sleep deprivation [20], indicating the dramatic state dependency of this system and its intimate relationship to REM sleep. That this state-dependency may be cholinergically mediated is suggested by an increase in acetylcholinesterase (AChE) activity in the brainstem during REM deprivation [ 211. A leading theory about REM sleep function is that it enables leaming and memory. Following exposure to maze learning, rats show a delayed vulnerability to REM sleep loss which results in deficits only when experienced during a critical time period (called the REM sleep window). It has now been demonstrated that both anisomycin and scopolamine given during the REM sleep window produce similar learning deficits and decreases in both ACh and AChE activity [22].
Serotonin
as a wake state enhancer
and REM
sleep inhibitor Early lesion and pharmacological studies indicating a positive role for the serotonergic raphe nucleus in slow wave sleep mediation were strongly countered by the consistent finding of a selective decrease in raphe firing and 5-hydroxytryptamine (5-HT) release during that state (for a review, see [ 11; for further confirmations, see [23,24]). Because the raphe neurons showed an almost complete arrest of firing during REM it was instead proposed that serotonin might actually inhibit the cholinergic/cholinoceptive REM sleep generator. It has now been determined that the reversal of parachlorophenylalanineinduced insomnia by 5-hydroxytryptophan (5-HTP) occurs without any increase in 5-HT [25], suggesting that 5-HTP is sedative but not via serotonergic enhancement. The 5-HT REM inhibition hypothesis has received a spate of affirmation during the past year. The most direct evidence for anticholinergic mediation of serotonergic REM suppression comes from intracellular and whole-cell patch-clamp recordings of identified cholinergic neurons in brainstem slices. Bursting cells show a hyperpolarization and a decrease in input resistance in response to iontophoretic application of the 5-HT agonist carboxyamidotryptamine [ 26.1. In behavioral studies, the 5-HTI agonist eltoprazine exerts a dose-dependent REM/PGO suppression in cats, which is followed by a rebound [27]. The surplus of PGO waves that is observed when the drug is withdrawn indicates that the PGO rebound is more than a recovery, and implicates a potentiation of PGO mediating mechanisms. A rebound likewise follows the REM suppression induced by electrical stimulation of the nucleus raphe dorsalis [28]. Human subjects also show reductions in REM sleep with serotonergic enhancers such as the 5-
761
762
Neural control
HT agonist m-CCP [ 291, and the 5-HT reuptake
paraoxetine
[30].
Aminergic
restraint
of the cholinergic
blocker
REM
tenrion syndrome that follows right-sided cortical lesions [ 37’1. Such patients make only rightward eye movements in REM sleep, indicating a powerful ‘top-down’ effect upon the pontine saccade generator, which one would expect in waking, but which is also surprisingly effective in FEM. It is natural to wonder whether these subjects neglect the left half of visual space in their dreams.
generator Noradrenergic neurons have been known for some time to be selectively active in waking and inactive in REM sleep. This neurophysiological fact runs directly counter to the early theory that noradrenaline mediates REM, and suggests, instead, an inhibitory role for noradrenaline as well as for serotonin. This concept fits well with other behavioral evidence of noradrenergic enhancement of waking brain functions, such as attention, learning and memory (all of which are decidedly deficient in REM). The a2-adrenergic agonist clonidine suppresses REM when it is microinjected into the pontine reticular formation of cats [31], or when it is given parenterally to human subjects [32]. The failure to enhance REM with c12-antagonists in man [32], or with P-adrenergic blockers in rats [33], may be a function of the parenteral route of drug delivery and the ensuing disruption of multiple peripheral physiological systems. As with the cholinergic REM sleep enhancement story, it is now clear that other transmitters are also involved in striking the excitatory-inhibitory balance in the brainstem that is expressed as sleep or waking behavior. When injected into the pontine reticular formation of rats, the GABA agonist muscimol produces an increase in waking and a decrease in sleep, especially REM 1341, but the finding of even greater sleep suppression with the GABA antagonist bicuculline at the same sites confounds any easy interpretation of the role of GABA in sleep.
Neuropsychology
of REM sleep and dreaming
Stimulated by the strong data base reviewed above, new empirical and theoretical work in neuropsychology has focused on the similarities and differences between REM sleep (with dreaming) and waking. Basing their assertions on elegant and detailed studies of the thalamocortical system, Llinas and Pare [35] emphasize the greater weight assigned to external inputs as the critical factor of the waking brain. The fact that in waking so relatively little of the neuronal connectivity is actually devoted to the transfer of sensory input, suggests that consciousness in both waking and dreaming is a closed-loop process in which the intrinsic activity of cells is critical. A similar position is taken by Antrobus [36] in his connectionist model, which is viewed as the result of an inhibition of external input, which leaves the perceptual and cognitive modules with their own outputs as their sole inputs. The importance to REM sleep of cortical areas classically considered critical for attention has recently been examined in a study of patients with the left visual hemi-inat-
Conclusion Intrinsic cholinergic neurons of the brainstem directly mediate the PGO waves of each REM period, and indirectly prime other mediating circuits of the brain so as to regulate the amount of REM sleep. This cholinergic REM induction and maintenance system is gated by serotonergic and noradrenergic inhibition. The cellular and molecular specificity of this experimental model provides the theoretical and operational framework for a molecular biological approach to sleep mechanism and function in both in uiz~oand in vitro preparations.
References
and recommended
Papers of particular interest, published view, have been highlighted as: . of special interest .. of outstanding interest 1.
2.
within the annual period of re-
JACOBSBL, FORNAL CA: V. Activity
Neurons in 43:563-578.
the
Behaving
reading
Animal.
of Brain Serotonergic &Y~~~rmacol Rez) 1991,
JONES BE: Paradoxical
Substrates
Sleep and its Chemical/Structural in the Brain. Neuroscience 1991, 401637-656.
3.
MAMELAK M: A Model for Narcolepsy.
4.
SHOUSEMN, SIEGELJM: Pontine Regulation of REM Sleep Components in Cats. Bruin Res 1992, 571:5@63.
5.
VANNI-MERCIER G,
6.
LAIYY, SIEGEL JM: Pontomedullary Glutamate Receptors Mediating Locomotion and Muscle Tone Suppression. J Neurosci 1992, 11:2931-2937.
7.
SIEGELJM, NIENHLJIS R, FAHRINGER HM, PAUL R, SHIROMANI P, DEMENTWC, MIGNOT E, CHKJ C: Neuronal Activity in
Can J Psychol 1991, 45:194-220. :he first attempt to integrate data regarding brain neurotransmitter levels in human narcolepsy with cellular and molecular findings regarding REM sleep triggering in animals. Mamelak’s model incorporates new findings concerning the possible role of dopaminergic substantia nigra neurons in REM sleep control.
SAKAIK, LINJS, JOUVETM: Carbachol Microinjections in the Mediodorsal Pontine Tegmentum are Unable to Induce Paradoxical Sleep After Caudal Pontine and Prebulbar Transections in the Cat. Neurosci Lett 1991, 130:41-45.
Narcolepsy: Identification of Cataplexy-Related Cells in the Medial Medulla. Science 1991, 252:1315-1318.
CURRODOSSI R, Pa D, STE~UA~EM: Short-Lasting Nicotinic and Long-Lasting Muscarinic Depolarizing Response of Thaiamocortical Neurons to Stimulation of Mesopontine Cholinergic Nuclei. J Neuvopbysiol 1991, 65:3931tO6. The first clear cellular and molecular demonstrations (see also [90,11*]) that the cortical desynchronization (including the newty discovered 40Hz oscillations) may be mediated by a muscarinic choliner8. .
Sleep and dreaming
gic influence of mesopontine elements.
cholinergic
neurons
STER~ADE M, CURRO DOSSI R, NUNEZ A: Network Modulation of a Slow Intrinsic Oscillation of Cat Thalamocortical Neurons Implicated in Sleep Delta Waves: Cortically Induced Synchronization and Brainstem Cholinergic Suppression. J Neurosci 1991, 11:320&3217. See 18.1.
10.
MARKSG, ROFFWARGH: Cholinergic Modulation of Responses to Glutamate in the Thalamic Reticular Nucleus of the Anesthetized Rat. Brain Res 1991, 557:4%56.
11. .
STER~A~EM, Doss1 RC, PARE D, OAKSON G: Fast Oscillations (2tiOHz) in the ThalamocorticaI Systems and Their Potentiation by Mesopontine Cholinergic Nuclei in the Cat. Proc Nat1 Acud Sci USA 1991, 88:43964400. See [8-l. 12.
25.
L, BUCHWALDJ: Midlatency Auditory-Evoked Responses: Effect of Scopolamine in the Cat and Implications for BraInstem Cholinergic Mechanisms. EJCP Neural 1991, 112:229239.
GERBERU, S’~E~ENSDR, MCCARLEY RW, GREENERW: Muscarinic Agonists Activated an Inwardly Rectifying Potassium Conductance in Medial Pontine Reticular Formation Neurons of the Rat In Vitro. J Neurosci 1991, 11:386-3867.
14. .
DANA
S, CALVO J, QUATIROCHI
J, HOBSON
JA: Long-Term EnCholinergic Stlmula-
hancement of REM Sleep Following tion. NeuroReport 1991, 2:619622. Evidence that direct triggering of PGO waves by cholinergic chemical stimulation is followed by a subsequent week long enhancement of REM. 15.
of Cat Pons LYLXCR, BAGHWYAN HA, ILXINC 2: Microdialysis Reveals Enhanced Acetylcholine Release During StateDependent Respiratory Depression. Am J Pkysiol 1991, 261:766770.
16.
LYDIC R, BAGHWYAN HA, HLBBARDL, BONYAK EV, DEJOS~PH MR, HAWKINSRA: Regional Brain Glucose Metabolism is Altered During Rapid Eye Movement Sleep in the Cat: a Preliminary Study. J Comp Neural 1991, 304:517-529.
17.
LYDIC R, BAGHDOYAN HA, WERTZ R, WHITE DP: Cholinerglc Reticular Mechanisms InIluence State-Dependent Ventilatory Response to Hypercapnia. Am J Pbysiol 1991, 261~738746.
18.
Bw WA HUNT WH, SANFORD LD, Ross RJ, MORRISIONAR: Effects of Stimulus Intensity on Elicited Ponto-GeniculoOccipital Waves. Electroencephalogr Clin Neurophysiol 191, 78:35-39.
19.
BALL WA, SANFORD LD, MORRISONAR, Ross RJ, HUNT WH, MANNGL The Effects of Changing State on Elicited PontoGeniculo-Occipital Waves, Electroencephalogr Clin Neuro physiol 1991, 79:42@429.
20.
MALLICKBN, FAHRINGERHM, Wu MF, SIEGELJM: REM Sleep Deprivation Reduces Auditory Evoked Inhibition of Dorsolateral Pontine Neurons. Brain Res 1991, 552:333-337. THAKKARM, MALuc BN: Effect of REM Sleep Deprivation on Rat Brain Acetylcholinesterase. Pharmacol Biocbem Bebar) 1991, 39:211-214.
22.
SMITH C, TENN C, ANNErr R: Some Biochemical and Behavioral Aspects of the Paradoxical Sleep Window. Can J Psy dwl 1991, 45:115-124.
23.
WILKINSONLo, AUERBACHAB, JACOBS BL: Extracellular Serotordn Levels Change with Behavioral State but not with Pyrogen-Induced Hyperthermia. J Neurosci 1991, 11:2732-2741.
24.
HOUDoUtN F, CESPUGUOR, JOUVET M: Effects Induced by the Electrical Stimulation of the Nucleus Raphe Dorsalis Upon
TOURE’~M, SARDAN, GHARIB A, GEFFARD M, JOWET M: Tbe Role of 5-Hydroxytryptophan (5HTP) in the Regulation of the Sleep/Wake Cycle in Paracblorophenylalanine (p-CPA) Pretreated Rat: a Multiple Approach Study Exp Brain Res 1991, 86:117-124.
LUEBKE JL, GREENE RW, SEMBA K, WONDI A, MCCARLEY RW, REINERPB: Serotonin Hyperpolarizes Cholinergic LowThreshold Burst Neurons in the Rat Laterodorsal Tegmental Nucleus In Vitro. Neurobiology 1992, 89~743-747. The first membrane potential level evidence that serotonin may exert its REM sleep suppression effects via hyperpolarization of intrinsic cholinergic neurons. 26. .
27.
QUAT~ROCHIJ, MAMEL%A, BINDERD, WILUAMSJ, RITTENHOUSE C, HOBSON JA Dynamic Suppression of REM Sleep by Parental Administration of the Serotonin-1 Agonist Eltoprazine. Sleep 1991, 15:125-132.
28.
HOUDOUIN F, CESPUCLIO R, CHAIUB A, SARDA N, JOWET M: Detection of the Release of 5-Hydroxylndole Compounds in the Hypothalamus and the n. Raphe Dorsal& Throughout the Sleep-Waking Cycle and During Stressful Situations in the Rat. Exp Brain Res 1991, 85:153-162.
29.
IAWLOHBA, NEWHOUSE PA, BALKINTH, MOIXHAN SE, MELLOW AM, MUR~HYDL, SUNDERLAND T: A Preliminary Study of the Effects of Night-Time Administration of the Serotonin Agonist, m-CPP, on Sleep Architecture and Behavior in Healthy Volunteers. Biol Pq%atr 1991, 29:281-286.
30.
SALUTU B, FREY R, KRUPKAM, ANDEKER P, GRUNBERGER J, SEE WR: Sleep Laboratory Studies on the Single-Dose Effects of Serotonin Reuptake Inhibitors Paroxetine and Fluoxetine on Human Sleep and Awakening Qualities. Sleep 1991, 14:439447.
31.
TONONI G, POMPEIANOM, CIRELLIC: Suppression of Desynchronized Sleep Through Microinjection of the cL2-Adrenergic Agonist Clonidine in the Dorsal Pontine Tegmentum of the Cat. Pjtigers Arch 1991, 418:512-518.
32.
NICHOLSONAN, PASCOE PA: Presynaptic Alpha-Adrenoceptor Function and Sleep in Man: Studies with Clonldine and Idazoxan. Neurophamacology 1991, 301367-372.
DICKERSON
13.
21.
Hypothalamic Release of 5-Hydroxyindole Compounds and Sleep Parameters in the Rat. Brain Res 1991, 556:48-56.
upon thalamocortical
9. .
Hobson
33.
ONCINI E, MILANIS, MARZANATT~ M, TRAMPUS M, MONOPOLI Effects of Selected Beta-Adrenerglc Blocking Agents on Sleep Stages in Spontaneously Hypersensitive Rats. J Phar~ macol Exp Ther 1991, 257:114-119.
A:
34.
CAMACHO-ARROYO I, ALVARADO R, MANJARIUJ, TAP~AR: Microinjections of Muscimol and Bicuculllne into the Pontine Reticular Formation Modify the Sleep-Waking in the Rat. Neurosci Len 1991, 129~95-96.
35.
LIJNAS RR, PARI? D: Of Dreaming science 1991, 44:521-535.
36.
ANTROB~S
Activation 98:96121. 37 .
and Wakefulness.
Neuro
J: Dreaming: Cognitive Processes During Cortical and High Afferent Thresholds. Psycho1 Rev 1991,
DOFXCHI
F, G~JAIUGUAC, PAOLUCCI S, P!ZZAMIGLIO L: Disappearance of Leftward Rapid Eye Movements During Sleep in Left Visual Hemi-Inattention. NeuroReport 1991, 2:285-288. The first evidence of a relationship between a cortical lesion (causing hemi-neglect) and a selective redirection of eye movement during sleep.
JA Hobson, Department of Psychiatry, Halvard Medical School, Boston, Massachusetts 02115, USA.
763
induction
and mediation
by cholinergic
of REM sleep
mechanisms
J. Allan Hobson
The
Harvard Medical School, Boston, Massachusetts,
USA
most
sleep
focused eye
important
on the
movement
recent
precise
work
cellular
sleep mediation.
acetylcholine-containing triggering
and
maintenance
new
studies
provide
support
generator
the and
in the the
indirect eye
pons that
influences.
A growing new
thinking
consensus about
regarding the
waking
Current
Opinion
brain
the
implicates
as critical
movement
hypothesis
has
of rapid
evidence
system is gated during waking by serotonergic
stimulated
of
mechanisms
peribrachial
of rapid for
neurobiology
biochemical
Direct
neurons
the
on
and
sleep. the
in
Other
cholinergic
and noradrenergic
basic
neurobiology
basis of consciousness
has during
and dreaming.
in Neurobiology
Introduction
1992,
2:759-763
REM sleep mediation cholinoceptive
The role of acetylcholine (ACh) in promoting sleep, and especially rapid eye movement (REM) sleep, has been suspected for many years, but only recently has the cellular and molecular basis of its mediating role begun to be clarified. The critical cholinergic neurons are found in the peribrachial region of the pons. They project to the paramedian pontine reticular formation and to the thalamus where they excite other cells which directly mediate the rostral signs of REM sleep, including the cortical electroencephalogram activation, the ocular saccades (or REMs), and the ponto-geniculo-occipital (PGO) waves (Fig.la,b,c).
via cholinergic
neurons
and
in the pons
The pedunculo pontine tegmental nucleus (PPT), which contains the Ch5 group of cholinergic neurons, has been of interest to sleep scientists since it was discovered that the cells of this region fire in bursts in advance of ipsiversive REMs and the larger PGO waves in the ipsilateral geniculate body (see Fig.la,b). Together with the Ch6 group of the dorsolateral tegmental nucleus, these neurons constitute the entire brainstem compliment of ACh-containing cells. Lesion studies now concur in showing that the selective suppression of REM and PGO waves [4] is proportional to the loss of cholinacety-transferase activity [ 21. Brainstem transection at the prebulbar or caudal pontine levels eliminates the atonia, but not the PGO waves or eye movements that are triggered by carbachol [ 51, indicating again a necessary link to the medullary inhibitory reticular function.
An indirect effect of the rostral pontine trigger zone upon magnocellular medullary elements (which execute the muscle atonia of REM) may be mediated by gluta mate. The cholinergic system and its postsynaptic REM sleep executive population are modulated by serotonergic, noradrenergic, and dopaminergic inputs which are, in general, inhibitory. The REM sleep bursting pattern of cholinergic neuronal discharge is thought to be mediated by local y-aminobutync acid (GABA)ergic interneurons via membrane hyperpolarization. The details of these effects are the main focus of this review. An emerging theoretical model reconciling previously controversial theory can be seen in three recent reviews [ 1,2,3-l.
Because the muscle atonia of REM is not as dramatically altered by PPT lesions as when the lesions are in the locus subcoeruleus, a non-cholinergic mediation of this REM sleep sign is suggested. The magnocellular medullary neurons (which are thought to inhibit the anterior horn cells via a glycinergic mechanism) may be excited by glutamate [6]. Interestingly, most of these neu-
Abbreviations ACLacetylcholine;
AChE-acetylcholinesterase;
5-HTP-5-hydroxytryptophan;
GABA--y-aminobutyric
PGO-ponto-geniculo-occipital; REM-rapid
@
Current
Biology
acid; 5-HT-5-hydroxytryptamine;
PPT-pedunculo
pontine
tegmental
nucleus;
eye movement.
Ltd ISSN 09594388
759
760
Neural
control
(a) Location of PC0
(c) Neurophysiological and schematic
cell
Burst
(b) Long (0) and short REM Induction sites
cells
(0)
term
recording
Percentage of REM sleep signs
LGB
LGBL
LGB
LGBW
EEG
EEG
60
1
L
R
v
Post carbachol Injection days
5 Sec.
Fig.1. Relationship
of PC0 burst cells to cholinergic REM induction sites. (a) Filled circle indicates site of injection of carbachoi into peribrachial pons. Small dots indicate location of cholinergic cells and crosses location of PC0 burst cells. (b) Location of peribrachial long-term (filled circle) and paramedian short-term (open circle with dots) REM induction sites. k) On the left is an extracellular unit recording of a PC0 burst cell and PC0 waves in the ipsilateral LGB. On the right is a hypothetical wiring diagram of the PC0 trigger zone (PB) elements (PC0 and ACh) with modulatory aminergic raphe (R) and locus coeruleus (LC), thalamic (LCB), and paramedian pontine reticular cells (FTC). (d) Injections of carbachol at the PB site shown in (a) and (b) produce immediate PC0 waves in the ipsilateral LGB. (e) After 24 h REM triples and remains elevated for 6 days, this is the long-term REM effect. Reticular tegmental nuclei (FTC, FTP, FTL, FTC). Aminergic nuclei (R, LC). Cholinergic nuclei (C5, C6). Reference structures shown are red nucleus (RN), trigeminal motor nucleus (MN), and brachium conjuctivum (BC). (a,b and c) reproduced with permission from S Datta, J Quattrochi and JA Hobson (unpublished data). (d and e) are reproduced with permission from [14*1.
rons are not activated in canine narcolepsy, suggesting a Werent mechanism for the atonia of that disease [7]. An important series of electrical stimulation studies indicates that the cortical desynchronization of REM sleep is mediated via a muscat-ink cholinergic depolarization of neurons in the ST, VS-VL, and VL nuclei of the thala-
mus [@I. The cholinergic excitation appears to disrupt the sequences of bursting discharge and afterhyperpolarizations that underlie the spindling pattern of slow wave sleep, and which are due to GABAergic inhibition of nucleus reticular-is and local circuit inhibitory neurons [9*]. The possibility that ACh inhibits reticularis neurons [ 101
Sleep and dreaming
is consistent with the presumed cholinergic mediation of cortical synchronization. The 20-40Hz oscillations that can be recorded in thalamocortical neurons also seem to be mediated in part by the brainstem cholinergic system (PPT/ChS), as they are enhanced by stimulation of that structure, blocked by scopolamine, and unaffected by lesions of the Chl-Ch4 cell groups, which constitute the forebrain cholinergic ensemble in the nucleus basalis of Meynert [ 11.1. In a related study a role for the cholinergic system in the mediation of the cat mid-latency audition-evoked potential has been suggested [ 121. A major impetus for the recent increase in experimental work on the cholinergic mediation of REM sleep is the continuing refinement and exploitation of the chemical microstimulation technique using carbachol and other muscarinic agonists to trigger REM sleep. Direct evidence of cholinergic effects upon cholinoceptive mPRF neurons of the pons has recently been obtained via intracellular recording in rat brainstem slices (131. The cholinergic microstimulation approach has also been used to show that ipsiversive eye movements and ipsilateral thalamic LGB cell PGO waves are immediately excited by carbachol (see Fig.ld) when it is microinjected into the caudo-lateral PPT (or PB; see Fig.la,b). Following a delay of several hours to one day (during which waking behavior is actually enhanced) REM sleep triples in amount and thereafter remains significantly elevated for 6 days (see Fig.le) [ 14.1. Such long-term increases in sleep are unprecedented and imply the activation, perhaps via the intense bursting of the cholinergic neurons themselves, of an intracellular metabolic process (such as the nuclear synthesis of cholinacetyltransferase). Both the cholinergic (PB) and cholinoceptive (mPRF) regions of the pons are among the structures shown by increases in glucose use to be metabolically active during natural REM sleep [15]. When carbachol is microinjected into the paramedian pons (see Fig.lb) it produces a more immediate and much briefer period of REM enhancement, as if a cholinoceptive executive population of neurons were being activated without affecting long-term regulatory processes. In such cases an increase in ACh release can be measured in the contralateral pons [ 161. That the carbachol-induced REM-like state is a physiologically valid model is suggested by its association with the same kind of respiratory depression that is seen in natural REM sleep [ 171.
Functional
studies of REM sleep and PC0
waves Pm-like waves can be elicited by tones in all behavioral states and tend to be associated with the orienting response when they are triggered during waking. The amplitude, frequency and latency of the waves are all dependent upon stimulus intensity [18] and can still be recorded in waking when the orienting response has
habituated, indicating ciable [ 191.
that the two processes
Hobson
are disso-
When auditory stimuli are delivered to cats they produce first an excitation and then an inhibition of dorsal pontine neurons; the inhibitory component is reduced by 24-48 h of REM sleep deprivation [20], indicating the dramatic state dependency of this system and its intimate relationship to REM sleep. That this state-dependency may be cholinergically mediated is suggested by an increase in acetylcholinesterase (AChE) activity in the brainstem during REM deprivation [ 211. A leading theory about REM sleep function is that it enables leaming and memory. Following exposure to maze learning, rats show a delayed vulnerability to REM sleep loss which results in deficits only when experienced during a critical time period (called the REM sleep window). It has now been demonstrated that both anisomycin and scopolamine given during the REM sleep window produce similar learning deficits and decreases in both ACh and AChE activity [22].
Serotonin
as a wake state enhancer
and REM
sleep inhibitor Early lesion and pharmacological studies indicating a positive role for the serotonergic raphe nucleus in slow wave sleep mediation were strongly countered by the consistent finding of a selective decrease in raphe firing and 5-hydroxytryptamine (5-HT) release during that state (for a review, see [ 11; for further confirmations, see [23,24]). Because the raphe neurons showed an almost complete arrest of firing during REM it was instead proposed that serotonin might actually inhibit the cholinergic/cholinoceptive REM sleep generator. It has now been determined that the reversal of parachlorophenylalanineinduced insomnia by 5-hydroxytryptophan (5-HTP) occurs without any increase in 5-HT [25], suggesting that 5-HTP is sedative but not via serotonergic enhancement. The 5-HT REM inhibition hypothesis has received a spate of affirmation during the past year. The most direct evidence for anticholinergic mediation of serotonergic REM suppression comes from intracellular and whole-cell patch-clamp recordings of identified cholinergic neurons in brainstem slices. Bursting cells show a hyperpolarization and a decrease in input resistance in response to iontophoretic application of the 5-HT agonist carboxyamidotryptamine [ 26.1. In behavioral studies, the 5-HTI agonist eltoprazine exerts a dose-dependent REM/PGO suppression in cats, which is followed by a rebound [27]. The surplus of PGO waves that is observed when the drug is withdrawn indicates that the PGO rebound is more than a recovery, and implicates a potentiation of PGO mediating mechanisms. A rebound likewise follows the REM suppression induced by electrical stimulation of the nucleus raphe dorsalis [28]. Human subjects also show reductions in REM sleep with serotonergic enhancers such as the 5-
761
762
Neural control
HT agonist m-CCP [ 291, and the 5-HT reuptake
paraoxetine
[30].
Aminergic
restraint
of the cholinergic
blocker
REM
tenrion syndrome that follows right-sided cortical lesions [ 37’1. Such patients make only rightward eye movements in REM sleep, indicating a powerful ‘top-down’ effect upon the pontine saccade generator, which one would expect in waking, but which is also surprisingly effective in FEM. It is natural to wonder whether these subjects neglect the left half of visual space in their dreams.
generator Noradrenergic neurons have been known for some time to be selectively active in waking and inactive in REM sleep. This neurophysiological fact runs directly counter to the early theory that noradrenaline mediates REM, and suggests, instead, an inhibitory role for noradrenaline as well as for serotonin. This concept fits well with other behavioral evidence of noradrenergic enhancement of waking brain functions, such as attention, learning and memory (all of which are decidedly deficient in REM). The a2-adrenergic agonist clonidine suppresses REM when it is microinjected into the pontine reticular formation of cats [31], or when it is given parenterally to human subjects [32]. The failure to enhance REM with c12-antagonists in man [32], or with P-adrenergic blockers in rats [33], may be a function of the parenteral route of drug delivery and the ensuing disruption of multiple peripheral physiological systems. As with the cholinergic REM sleep enhancement story, it is now clear that other transmitters are also involved in striking the excitatory-inhibitory balance in the brainstem that is expressed as sleep or waking behavior. When injected into the pontine reticular formation of rats, the GABA agonist muscimol produces an increase in waking and a decrease in sleep, especially REM 1341, but the finding of even greater sleep suppression with the GABA antagonist bicuculline at the same sites confounds any easy interpretation of the role of GABA in sleep.
Neuropsychology
of REM sleep and dreaming
Stimulated by the strong data base reviewed above, new empirical and theoretical work in neuropsychology has focused on the similarities and differences between REM sleep (with dreaming) and waking. Basing their assertions on elegant and detailed studies of the thalamocortical system, Llinas and Pare [35] emphasize the greater weight assigned to external inputs as the critical factor of the waking brain. The fact that in waking so relatively little of the neuronal connectivity is actually devoted to the transfer of sensory input, suggests that consciousness in both waking and dreaming is a closed-loop process in which the intrinsic activity of cells is critical. A similar position is taken by Antrobus [36] in his connectionist model, which is viewed as the result of an inhibition of external input, which leaves the perceptual and cognitive modules with their own outputs as their sole inputs. The importance to REM sleep of cortical areas classically considered critical for attention has recently been examined in a study of patients with the left visual hemi-inat-
Conclusion Intrinsic cholinergic neurons of the brainstem directly mediate the PGO waves of each REM period, and indirectly prime other mediating circuits of the brain so as to regulate the amount of REM sleep. This cholinergic REM induction and maintenance system is gated by serotonergic and noradrenergic inhibition. The cellular and molecular specificity of this experimental model provides the theoretical and operational framework for a molecular biological approach to sleep mechanism and function in both in uiz~oand in vitro preparations.
References
and recommended
Papers of particular interest, published view, have been highlighted as: . of special interest .. of outstanding interest 1.
2.
within the annual period of re-
JACOBSBL, FORNAL CA: V. Activity
Neurons in 43:563-578.
the
Behaving
reading
Animal.
of Brain Serotonergic &Y~~~rmacol Rez) 1991,
JONES BE: Paradoxical
Substrates
Sleep and its Chemical/Structural in the Brain. Neuroscience 1991, 401637-656.
3.
MAMELAK M: A Model for Narcolepsy.
4.
SHOUSEMN, SIEGELJM: Pontine Regulation of REM Sleep Components in Cats. Bruin Res 1992, 571:5@63.
5.
VANNI-MERCIER G,
6.
LAIYY, SIEGEL JM: Pontomedullary Glutamate Receptors Mediating Locomotion and Muscle Tone Suppression. J Neurosci 1992, 11:2931-2937.
7.
SIEGELJM, NIENHLJIS R, FAHRINGER HM, PAUL R, SHIROMANI P, DEMENTWC, MIGNOT E, CHKJ C: Neuronal Activity in
Can J Psychol 1991, 45:194-220. :he first attempt to integrate data regarding brain neurotransmitter levels in human narcolepsy with cellular and molecular findings regarding REM sleep triggering in animals. Mamelak’s model incorporates new findings concerning the possible role of dopaminergic substantia nigra neurons in REM sleep control.
SAKAIK, LINJS, JOUVETM: Carbachol Microinjections in the Mediodorsal Pontine Tegmentum are Unable to Induce Paradoxical Sleep After Caudal Pontine and Prebulbar Transections in the Cat. Neurosci Lett 1991, 130:41-45.
Narcolepsy: Identification of Cataplexy-Related Cells in the Medial Medulla. Science 1991, 252:1315-1318.
CURRODOSSI R, Pa D, STE~UA~EM: Short-Lasting Nicotinic and Long-Lasting Muscarinic Depolarizing Response of Thaiamocortical Neurons to Stimulation of Mesopontine Cholinergic Nuclei. J Neuvopbysiol 1991, 65:3931tO6. The first clear cellular and molecular demonstrations (see also [90,11*]) that the cortical desynchronization (including the newty discovered 40Hz oscillations) may be mediated by a muscarinic choliner8. .
Sleep and dreaming
gic influence of mesopontine elements.
cholinergic
neurons
STER~ADE M, CURRO DOSSI R, NUNEZ A: Network Modulation of a Slow Intrinsic Oscillation of Cat Thalamocortical Neurons Implicated in Sleep Delta Waves: Cortically Induced Synchronization and Brainstem Cholinergic Suppression. J Neurosci 1991, 11:320&3217. See 18.1.
10.
MARKSG, ROFFWARGH: Cholinergic Modulation of Responses to Glutamate in the Thalamic Reticular Nucleus of the Anesthetized Rat. Brain Res 1991, 557:4%56.
11. .
STER~A~EM, Doss1 RC, PARE D, OAKSON G: Fast Oscillations (2tiOHz) in the ThalamocorticaI Systems and Their Potentiation by Mesopontine Cholinergic Nuclei in the Cat. Proc Nat1 Acud Sci USA 1991, 88:43964400. See [8-l. 12.
25.
L, BUCHWALDJ: Midlatency Auditory-Evoked Responses: Effect of Scopolamine in the Cat and Implications for BraInstem Cholinergic Mechanisms. EJCP Neural 1991, 112:229239.
GERBERU, S’~E~ENSDR, MCCARLEY RW, GREENERW: Muscarinic Agonists Activated an Inwardly Rectifying Potassium Conductance in Medial Pontine Reticular Formation Neurons of the Rat In Vitro. J Neurosci 1991, 11:386-3867.
14. .
DANA
S, CALVO J, QUATIROCHI
J, HOBSON
JA: Long-Term EnCholinergic Stlmula-
hancement of REM Sleep Following tion. NeuroReport 1991, 2:619622. Evidence that direct triggering of PGO waves by cholinergic chemical stimulation is followed by a subsequent week long enhancement of REM. 15.
of Cat Pons LYLXCR, BAGHWYAN HA, ILXINC 2: Microdialysis Reveals Enhanced Acetylcholine Release During StateDependent Respiratory Depression. Am J Pkysiol 1991, 261:766770.
16.
LYDIC R, BAGHWYAN HA, HLBBARDL, BONYAK EV, DEJOS~PH MR, HAWKINSRA: Regional Brain Glucose Metabolism is Altered During Rapid Eye Movement Sleep in the Cat: a Preliminary Study. J Comp Neural 1991, 304:517-529.
17.
LYDIC R, BAGHDOYAN HA, WERTZ R, WHITE DP: Cholinerglc Reticular Mechanisms InIluence State-Dependent Ventilatory Response to Hypercapnia. Am J Pbysiol 1991, 261~738746.
18.
Bw WA HUNT WH, SANFORD LD, Ross RJ, MORRISIONAR: Effects of Stimulus Intensity on Elicited Ponto-GeniculoOccipital Waves. Electroencephalogr Clin Neurophysiol 191, 78:35-39.
19.
BALL WA, SANFORD LD, MORRISONAR, Ross RJ, HUNT WH, MANNGL The Effects of Changing State on Elicited PontoGeniculo-Occipital Waves, Electroencephalogr Clin Neuro physiol 1991, 79:42@429.
20.
MALLICKBN, FAHRINGERHM, Wu MF, SIEGELJM: REM Sleep Deprivation Reduces Auditory Evoked Inhibition of Dorsolateral Pontine Neurons. Brain Res 1991, 552:333-337. THAKKARM, MALuc BN: Effect of REM Sleep Deprivation on Rat Brain Acetylcholinesterase. Pharmacol Biocbem Bebar) 1991, 39:211-214.
22.
SMITH C, TENN C, ANNErr R: Some Biochemical and Behavioral Aspects of the Paradoxical Sleep Window. Can J Psy dwl 1991, 45:115-124.
23.
WILKINSONLo, AUERBACHAB, JACOBS BL: Extracellular Serotordn Levels Change with Behavioral State but not with Pyrogen-Induced Hyperthermia. J Neurosci 1991, 11:2732-2741.
24.
HOUDoUtN F, CESPUGUOR, JOUVET M: Effects Induced by the Electrical Stimulation of the Nucleus Raphe Dorsalis Upon
TOURE’~M, SARDAN, GHARIB A, GEFFARD M, JOWET M: Tbe Role of 5-Hydroxytryptophan (5HTP) in the Regulation of the Sleep/Wake Cycle in Paracblorophenylalanine (p-CPA) Pretreated Rat: a Multiple Approach Study Exp Brain Res 1991, 86:117-124.
LUEBKE JL, GREENE RW, SEMBA K, WONDI A, MCCARLEY RW, REINERPB: Serotonin Hyperpolarizes Cholinergic LowThreshold Burst Neurons in the Rat Laterodorsal Tegmental Nucleus In Vitro. Neurobiology 1992, 89~743-747. The first membrane potential level evidence that serotonin may exert its REM sleep suppression effects via hyperpolarization of intrinsic cholinergic neurons. 26. .
27.
QUAT~ROCHIJ, MAMEL%A, BINDERD, WILUAMSJ, RITTENHOUSE C, HOBSON JA Dynamic Suppression of REM Sleep by Parental Administration of the Serotonin-1 Agonist Eltoprazine. Sleep 1991, 15:125-132.
28.
HOUDOUIN F, CESPUCLIO R, CHAIUB A, SARDA N, JOWET M: Detection of the Release of 5-Hydroxylndole Compounds in the Hypothalamus and the n. Raphe Dorsal& Throughout the Sleep-Waking Cycle and During Stressful Situations in the Rat. Exp Brain Res 1991, 85:153-162.
29.
IAWLOHBA, NEWHOUSE PA, BALKINTH, MOIXHAN SE, MELLOW AM, MUR~HYDL, SUNDERLAND T: A Preliminary Study of the Effects of Night-Time Administration of the Serotonin Agonist, m-CPP, on Sleep Architecture and Behavior in Healthy Volunteers. Biol Pq%atr 1991, 29:281-286.
30.
SALUTU B, FREY R, KRUPKAM, ANDEKER P, GRUNBERGER J, SEE WR: Sleep Laboratory Studies on the Single-Dose Effects of Serotonin Reuptake Inhibitors Paroxetine and Fluoxetine on Human Sleep and Awakening Qualities. Sleep 1991, 14:439447.
31.
TONONI G, POMPEIANOM, CIRELLIC: Suppression of Desynchronized Sleep Through Microinjection of the cL2-Adrenergic Agonist Clonidine in the Dorsal Pontine Tegmentum of the Cat. Pjtigers Arch 1991, 418:512-518.
32.
NICHOLSONAN, PASCOE PA: Presynaptic Alpha-Adrenoceptor Function and Sleep in Man: Studies with Clonldine and Idazoxan. Neurophamacology 1991, 301367-372.
DICKERSON
13.
21.
Hypothalamic Release of 5-Hydroxyindole Compounds and Sleep Parameters in the Rat. Brain Res 1991, 556:48-56.
upon thalamocortical
9. .
Hobson
33.
ONCINI E, MILANIS, MARZANATT~ M, TRAMPUS M, MONOPOLI Effects of Selected Beta-Adrenerglc Blocking Agents on Sleep Stages in Spontaneously Hypersensitive Rats. J Phar~ macol Exp Ther 1991, 257:114-119.
A:
34.
CAMACHO-ARROYO I, ALVARADO R, MANJARIUJ, TAP~AR: Microinjections of Muscimol and Bicuculllne into the Pontine Reticular Formation Modify the Sleep-Waking in the Rat. Neurosci Len 1991, 129~95-96.
35.
LIJNAS RR, PARI? D: Of Dreaming science 1991, 44:521-535.
36.
ANTROB~S
Activation 98:96121. 37 .
and Wakefulness.
Neuro
J: Dreaming: Cognitive Processes During Cortical and High Afferent Thresholds. Psycho1 Rev 1991,
DOFXCHI
F, G~JAIUGUAC, PAOLUCCI S, P!ZZAMIGLIO L: Disappearance of Leftward Rapid Eye Movements During Sleep in Left Visual Hemi-Inattention. NeuroReport 1991, 2:285-288. The first evidence of a relationship between a cortical lesion (causing hemi-neglect) and a selective redirection of eye movement during sleep.
JA Hobson, Department of Psychiatry, Halvard Medical School, Boston, Massachusetts 02115, USA.
763
