www.actaps
Neural mechanism of rapid eye movement sleep generation: Cessation of locus coeruleus neurons is a necessity
Dinesh Pal, Vibha Madan, Birendra Nath Mallick *
School of Life Sciences, Jawaharlal Nehru University, New Delhi 110067, India
Abstract: Two types of neurons are involved in the regulation of rapid eye movement (REM) sleep, the REM-ON and the REM-OFF neurons. As the name suggests, the REM-OFF neurons cease firing during REM sleep and they are norepinephrinergic. It has been shown that cessation of these neurons is a pre-requisite for the generation of REM sleep and GABA shuts them off. Further, if these neurons do not shut off, there is increased levels of norepinephrine in the brain and loss of REM sleep. The REM sleep deprivation induced increase in norepinephrine is responsible for mediating at least REM sleep loss induced increase in Na +-K + A TPase activity,which is likely to be the primary factor for causing REM sleep deprivation induced effects.
Key words: GABA; locus coeruleus; Na-K A TPase; norepinephrine; REM sleep generating mechanism; REM sleep loss
快速眼动睡眠产生的神经机制：蓝斑核神经元停止发放是一个必要的条件
Dinesh Pal, Vibha Madan, Birendra Nath Mallick *
Jawaharlal Nehru 大学生命科学学院，新德里110067，印度
摘 要：两种类型的神经元参与了快速眼动(rapid eye movement, REM)睡眠的调节：快速眼动-发放(REM-ON)神经元和快速眼动-沉寂神经元(REM-OFF)。快速眼动-沉寂神经元属去甲肾上腺素能神经元，正如名字表示的那样——在快速眼动睡眠期间停止发放。已有研究表明，这些神经元放电活动的停止是导致快速眼动睡眠的前提条件，γ-氨基丁酸(γ-aminobutyric acid,GABA)可使它们停止发放。如果这些神经元不停止发放，脑中的去甲肾上腺素水平将升高，不出现快速眼动睡眠。剥夺快速眼动睡眠所引起的去甲肾上腺素增加，至少是快速眼动睡眠丧失引起Na +-K + ATP 酶活性增加的原因，而这可能是导致快速眼动睡眠剥夺所引发的各种效应的主要因素。
关键词：γ-氨基丁酸；蓝斑核；Na +-K + ATP 酶；去甲肾上腺素；快速眼动睡眠产生机制；快速眼动睡眠的丧失中图分类号：Q426；R338.4
Received 2005-01-25    Accepted 2005-07-01
This work was supported by Fund from CSIR, DBT, DST, ICMR and UGC, India.
*Corresponding author. Tel: +91-11-26704522; Fax: +91-11-26717586; E-mail: remsbnm@yahoo
Introduction
Sleep and wakefulness are spontaneous cyclic changes in behavior and associated levels of consciousness in higher living beings. To avoid subjective bias, electrophysiologi-cal signals from the brain, the electroencephalogram (EEG),the muscles, the electromyogram (EMG) and the eye movements, the electrooculogram (EOG) have been used for classification and quantification of sleep and wakeful-ness objectively in higher species including humans. While analyzing sleep using the electrophysiological parameters,
Aserinsky and Kleitman [1] observed that electrophysiologi-cally sleep was not a homogenous state. After a minimum time was spent in deep sleep, the EEG and EOG expressed signs apparently resembling wakefulness, although the EMG did not show signs associated to wakefulness. Since that appeared to be a paradox, i.e. presence of signs of wake-fulness in EEG and EOG during a phase of sleep, it was termed as paradoxical sleep. Also, since it appeared to be an active state of the brain with desynchronization of the EEG within sleep, it was termed as active sleep or desynchronized sleep. Since dreams are associated with
premise behind such studies is that if any normal manifestation, behavioral or otherwise, of a living organ-ism continues to be expressed even after the destruction of certain brain area(s), the damaged area of the brain is possibly not essential for normal manifestation of the func-tion under consideration. Work on cats with spinal transec-tion and in humans with spinal injury showed that the spi-nal cord makes no essential contribution to the brainstem signs of REM sleep. The transection studies suggested that structures caudal to the midbrain and rostral to the spinal cord are necessary for REM sleep regulation. Further, when the pons was connected to midbrain and forebrain structures, most of the defining signs of REM sleep were seen in the rostral structures, whereas, if the pons was connected to the medulla and spinal cord, most of the iden-tifying signs of REM sleep were seen in caudal structures.A transection through the middle of the pontine region abol-ished the major defining characteristic signs of REM sleep.Thus, based on the results of transection studies it was concluded that the pontine region in the brainstem is both necessary and sufficient to generate the basic phenom-enon of REM sleep [12,13].
Pontine region and regulation of REM sleep
The pontine region contains noradrenergic, cholinergic as well as GABA-ergic neurons. The noradrenergic neurons are clustered in the LC, which is the primary site for sup-plying NE in the brai
n. The functional characteristic of these NE-ergic neurons in the LC is that they cease firing during REM sleep [14,15] and hence they have been termed as REM-OFF neurons. On the other hand, the cholinergic neurons in the laterodorsal tegmentum (LDT) and pedunculo pontine tegmentum (PPT) in the brainstem in-crease firing during REM sleep and they have been termed as REM-ON neurons [16]. The present knowledge indicates that interactions between the neurons located in these nu-clei in the pontine region are responsible for the generation and regulation of REM sleep.
Locus coeruleus: An anatomical description
The LC is a small cluster of neurons situated in the pontine region near the wall of the fourth ventricle and is one of the few pigmented structures in the brain. Depending on the size of the cells and their organization, the LC has further been subdivided into LC-principal, LC-α, peri-LC-α and sub-coeruleus by some sleep researchers [17]. The LC or its analage, projecting to the forebrain is not found until rep-tiles [18] and avians [19], though some catecholaminergic neu-
this state of sleep, it has been termed as dream sleep.Additionally, as rapid eye movements during sleep formed a characteristic feature of this state, this state has also been termed as rapid eye movement (REM) sleep. The sleep state was thus classified into non-REM sleep and REM sleep.
The REM sleep is present in species higher in the evolu-tionary ladder, viz. birds and mammals, and has been clas-sically identified by the simultaneous presence of desynchronization (low voltage high frequency waves) of the EEG, frequent eye movements, muscle atonia and hip-pocampal theta rhythm. Several other characteristic features, e.g. ponto-geniculo-occipital waves, irregular res-piration as well as heart rate, body temperature fluctuation,etc . also are associated with this state. Though all the REM sleep signs may not be present in all the species, some of these signs may be expressed in some lower species sug-gesting that REM sleep-like state may be present in lower species as well [2-4]. Hence, it is debatable if REM sleep evolved in lower species or it is of relatively later origin in evolution. REM sleep loss affects several physiological processes necessary for normal routine behavior [5-7] and it is also essential for life to the extent that accumulated ef-fect of its loss may be fatal [8]. REM sleep is regulated by the brainstem though other brain areas may modulate it as well. The role of specific area(s) and group of neurons in the brainstem that play key role in the regulation of REM sleep would be discussed. Briefly, the neurons in the locus coeruleus (LC) cease firing during REM sleep and if an experimental animal was not allowed to have REM sleep,these neurons in the LC continued firing incessantly lead-ing to disturbance or loss of REM sleep. Alternatively, if these neurons were not allowed to cease firing either by continuous electrical stimulation [9] or by applying antago-nist of the neurotransmitter that keep these neurons inhib-ited [
10,11], REM sleep did not continue resulting in its reduction. Thus, the neurons in the LC must cease firing (as if it is a pre-requisite) for the generation of REM sleep and non-cessation of these neurons caused reduction of REM sleep associated with increased levels of norepineph-rine (NE) in the brain which ultimately induces the effects associated to REM sleep deprivation/loss.
Localization of area(s) in brainstem responsible for the regulation of REM sleep
Initial studies to localize anatomical structure(s) in the brain responsible for the regulation of REM sleep generation started with transection and lesion experiments. The
403 Dinesh Pal et al: Neural Mechanism of Rapid Eye Movement Sleep Generation: Cessation of Locus Coeruleus Neurons
rons projecting to the cerebellum and nearby tegmentum has been reported in teleosts and amphibian[20]. Therefore, it was proposed that the development of LC is in tandem with the appearance of its cortical target areas[19,20]. The number of neurons in LC increases from 200 in parakeet to 1 600 in rats and 20 000 in humans[21]. Projections from these neurons divide into ascending and descending branches and innervate almost all the areas in the brain, spinal cord[22,23] and brainstem[24]. The LC receives cholin-ergic[25,26] as well as GABA-ergic projection
s[27,28] from other parts of the brain and it also has GABA-ergic interneurons [29]. Galanin-ergic and GABA-ergic neurons from ventro-lateral preoptic area also project to the LC[30,31].
Locus coeruleus and REM sleep
There is ample evidence that the brain noradrenergic sys-tem plays a significant role in the regulation of REM sleep. Several techniques including electrical as well as chemical lesion, stimulation and microinjection have been extensively used to explore the role of LC in regulating REM sleep. Although these studies gathered a large volume of knowledge, much remains to be known in terms of the relationship of neurons in LC with other nuclei in the brain and the exact role that it plays in the generation and regula-tion of REM sleep. Electrical destruction of the dorsal part of LC did not suppress the occurrence of REM sleep[32]. Similarly, destruction of ventral part of the LC (LCα and peri-LCα), was followed by irreversible disappearance of REM sleep atonia[33]. However, destruction of LCp and LCα along with peri-LCα suppressed REM sleep com-pletely during the two post-lesion months[34]. Electrolytic lesions of the dorsal noradrenergic bundle that ascends from the LCp[35] resulted in increase in both non-REM sleep and REM sleep[36]. The firing rate of the neurons in the LCp is maximum during wakefulness, decreases during non-REM sleep and almost ceases during REM sleep[14,15,37], while that of the neurons located ventrally increase their
firing rate (almost exclusively) during REM sleep[17,37]. The activity of the NE-ergic neurons in the LC has been posi-tively correlated with activation of the sympathetic ner-vous system[38]. Sympathetic activation is normally accom-panied by EEG desynchronization and according to Reiner, the activity of the LC-NE-ergic neurons increases with an increase in discharge in the sympathetic nervous system. Reversible inactivation of the LCp by localized cooling (+10ºC) induced non-REM sleep followed by REM sleep[39]. On the other hand, it was found that continuous activation of the LC neurons inhibited REM sleep by re-ducing the frequency of generation of REM sleep although the duration per episode remained unaffected[9]. Thus, the results suggested that activation of LC neurons did not allow REM sleep occurrence while inactivation of those neurons allowed REM sleep to continue. Norepinephrine in REM sleep
The concentration of NE increased in the brain[40] after REM sleep deprivation. An increase in NE concentration in serum was also reported after REM sleep deprivation[41]. There was an increase in the activity of tyrosine hydroxylase, the enzyme involved in the first rate limiting step of synthesis of NE[42] and mRNA levels[43] of tyrosine hydroxylase, whereas there was a decrease in the activity of the NE degrading enzyme, monoamine oxidase-A[44] af-ter REM sleep deprivation. The above findings may be supported by a recent study that there was increased ty-rosine hydroxylase activity
within the neurons located in the LC[45]. These results suggested that there would be increased NE in the brain after REM sleep deprivation. An inhibitory role of NE on REM sleep may be supported by the fact that NE levels decreased during REM sleep[46]. Since most of the supply of NE in the brain comes from the neurons in LC, it is likely that the activity of those neurons must be getting modulated during normal REM sleep and/or upon REM sleep deprivation. This view may be confirmed by the fact that the REM-OFF neurons in the LC cease firing during REM sleep[14,15], and they con-tinue firing incessantly during REM sleep deprivation[47]. Also, their activation by electrical stimulation[9] or disinhi-bition by GABA-antagonist, picrotoxin[10] did not allow REM sleep to continue whereas GABA in LC caused an increase in REM sleep[48]. A comparative effect on the REM sleep upon electrical stimulation of LC neurons and microinjec-tions of GABA as well as picrotoxin in LC are shown in Fig.1A~E.
The studies mentioned above support the involvement of LC in the regulation of REM sleep. Those studies also suggest that continuous activity of the neurons in the LC possibly prevented generation of REM sleep and cessation of activity of those neurons induced REM sleep possibly through withdrawal of inhibition. However, the mecha-nism of cessation of activities of the LC neurons was not known. The presence of adrenergic receptors in the brain was shown long ago[49,50]. Since the
LC neurons are noradrenergic, agonists and antagonists of NE were used to study the role and mechanism of action of NE released by the LC-neurons in REM sleep regulation. REM sleep
Acta Physiologica Sinica, August 25, 2005, 57 (4): 401-413 404
Fig. 1.  Percent changes in REM sleep under various experimental conditions. The original reference is shown beneath each pie diagram. The numbers in parenthesis show respective references in the reference list. The pie chart shows sleep-wakefulness recordings (8 h except where mentioned otherwise) under the following conditions: A: Without any treatment (control study).  B: Single bilateral microinjection (250 nl) of normal saline into LC (control study).  C: Low frequency, low amplitude electrical stimulation of bilateral LC for 8 h.  D: Single bilateral microinjection (250 nl) of picrotoxin into LC.  E: Single bilateral microinjection (250 nl) of GABA into LC.  F: Continous low frequency, low amplitude electrical stimulation of bilateral PrH.  G: Continuous low frequency, low amplitude electrical stimulation of bilateral PrH in presence of single injection of picrotoxin into LC.  H: Repeated intermittent microinjections (250 nl) of picrotoxin into bilateral LC for 48 h at an interval of 6 h. The sleep-wakefulness recording was also done for 48 h continuously.
was facilitated by systemic injection of drugs that stimu-
lated β-adrenoceptors[51,52] and by blocking α
1-
adrenoceptors[53-55]. On the other hand, REM sleep was inhibited by blocking β-adrenoceptors[51,52] and stimulation
405 Dinesh Pal et al: Neural Mechanism of Rapid Eye Movement Sleep Generation: Cessation of Locus Coeruleus Neuronsmodulate
of α
1
-adrenoceptors[56]. Oral administration of prazosin in rats was found to shorten quiet waking and REM sleep while it increased active waking and slow wave sleep[55].α-2 agonist, clonidine, when injected intraperitoneally, re-duced REM sleep in rats and cats[57,58]. A similar decrease in REM sleep was observed in man with a dose roughly five times smaller than that used in the rat[59]. Yohimbine,α-2-antagonist, increased active wakefulness immediately after administration but did not affect REM sleep. Though systemic injections advanced our understanding of the LC mediated regul
ation of REM sleep, localized injections of adrenergic agonist and antagonist provided a more robust evidence for the role of LC in the regulation of REM sleep.    The REM sleep was decreased when methoxamine, α-1-agonist, was injected into the dorsal pontine tegmentum of cats. The decrease in total REM sleep was found to be due to both, an increased REM sleep latency and a reduced number as well as duration of REM sleep episodes[60]. Bi-lateral injection of α-2-agonist, clonidine, in the dorsal pon-tine tegmentum of cat produced an almost complete sup-pression of REM sleep[61]. β-agonist isoproterenol almost suppressed REM sleep, while β-antagonist propranolol consistently enhanced it, mainly through an increase in the number of REM sleep episodes[62]. Microinjection of β-agonist isoproterenol into medial septal region of basal fore-brain significantly increased the time spent awake and a near complete suppression of REM sleep[63]. Norepineph-rine in peri-LC-α caused a dose-dependent inhibition of REM sleep and induction of REM sleep without atonia. These effects were also produced by clonidine, an alpha-2-agonist, whereas alpha-2-antagonists were found to block the effect of norepinephrine. When co-applied with carba-chol into the caudal peri-LC-α, clonidine completely blocked the marked REM sleep inducing effect of carbachol[64]. Thus, the interaction of various networks of neurons hav-ing different adrenoceptors plays a crucial role for the gen-eration and regulation of REM sleep. Although systemic and local injection studies advanced the knowledge about the role of NE in REM sleep regulation, they were u
nable to elucidate the role of NE released from the LC in such regulation. Based on our earlier studies[65] and the results obtained in a recent study where LC stimulation was car-ried out in presence of adrenergic agonists and antagonists, we have proposed a model showing the possible mecha-nism of action of NE release from the LC-neurons and its role in REM sleep regulation[66].
The studies mentioned above suggest that although some NE may be needed, excess NE is inhibitory for the genera-tion of REM sleep. The brain receives most of NE from the LC-neurons, which cease firing during REM sleep and they continue firing during REM sleep deprivation. Thus, cessation of firing of the LC-neurons is likely to be at least one of the key factors for the generation of REM sleep. Further, increased NE in the brain during REM sleep dep-rivation due to non-cessation of firing of the LC-neurons is likely to be the primary factor for REM sleep deprivation induced effects.
How are the LC neurons kept active through wake-fulness
Normally REM sleep does not appear during wakefulness or immediately after going to sleep. It appears after certain period of non-rapid eye movement (NREM) sleep. At least in humans, the duration and number of REM sleep epi-sodes increase with progress and depth of sleep through the night. Although it was known that the REM sleep can-not be initiated as long as the LC noradrenergic
REM-OFF neurons continue firing[9,47], the cellular mechanism(s) of sleep-wake state dependent changes in the LC neuronal firing from highly active state during wakefulness to slow-ing down during NREM sleep and finally cessation of firing during REM sleep was not known. Mallick’s group proposed that the wakefulness inducing area, the midbrain reticular formation (MRF), possibly exerted opposite influence on REM-OFF and REM-ON neurons and hypothesized that the MRF would excite REM-OFF and inhibit REM-ON neurons during wakefulness. In a com-bined single unit recording and MRF stimulation study carried out in freely moving normally behaving cats it was observed that a majority of the neurons whose firing rate increased during spontaneous wakefulness, including the REM-OFF neurons, were excited, while the REM-ON neurons were inhibited[67] by the MRF wakefulness inducing area. These results supported our hypothesis and suggested that the wake active neurons in MRF continu-ously excite the NE-ergic REM-OFF neurons in the LC and inhibit the REM-ON neurons throughout the waking period[67,68]. This view may also be supported by the fact that activation of the REM-OFF neurons is reported to prevent REM sleep[9] and is likely to increase the level of NE in the brain causing cortical activation and desynchronization of the EEG[5,69-71]. Therefore, it is likely that continuous activation of the noradrenergic REM-OFF neurons contributes to EEG desynchronization associated with wakefulness, but not with that of REM sleep when the effect of cholinergic REM-ON neurons is pronounced.
Acta Physiologica Sinica, August 25, 2005, 57 (4): 401-413 406
This view may be supported by the power spectrum analysis study in the freely moving cats that the adrenergic and cholinergic antagonists affected different higher frequency bands of the desynchronized EEG. Additionally, as men-tioned above, MRF wakefulness-inducing area inhibited the REM-ON neurons and this may be the cause for non-acti-vation of the REM-ON neurons during waking period. It may be supported by the fact that activation of the area containing the REM-ON neurons increases REM sleep[72]. It has also been reported that activation of the REM-ON neurons in the peri-LC during REM sleep is responsible for muscle atonia during REM sleep[73,74]. All these results con-sidered together provide possible explanation for neural mechanism as to why does muscle atonia, associated with REM sleep, not appear during wakefulness although the EEG is desynchronized during both those stages. Moreover, logical extrapolation of these observations is that in case of narcolepsy possibly there occurs an error in this neural pathway resulting in appearance of muscle atonia during wakefulness.
Thus, the following is likely to be the working model of neuronal mechanism of REM sleep generation. The wake active neurons in the MRF are active during wakefulness. Activity of these neurons keeps activating the REM-OFF neurons and inhibiting the REM-ON neurons, which do not a
llow REM sleep signs to appear during wakefulness. Experimentally, we found that the REM-OFF neurons are normally active during all the stages except during REM sleep, while the REM-ON neurons behave in an opposite manner. As a mechanism of action one or more of the following possibilities may exist. One, that MRF neurons exert independent excitatory and inhibitory effects on the REM-OFF and the REM-ON neurons, respectively; two, that the MRF neurons exert an excitatory effect on the REM-OFF neurons that in turn (may be through GABAergic neurons) inhibit the REM-ON neurons[68]; and three, that the MRF neurons exert an inhibitory effect on the REM-ON neurons and that in turn (may be through withdrawal of GABAergic inhibition) exert an excitatory effect on the REM-OFF neurons[48]. It was known that neurons in the wake (MRF) and sleep (caudal brainstem) areas are mutu-ally inhibitory to each other[75,76] and at the onset of sleep the activity of the wake active neurons in the MRF is sig-nificantly reduced[77]. This reduction in the activity of wake active neurons in MRF gradually withdraws the excitatory and the inhibitory effects from the REM-OFF and the REM-ON neurons, respectively[67]. Gradually NREM sleep sets in when the sleep inducing neurons in the caudal brain-stem[75,78] and basal forebrain further increase firing[79-81]. At some point when certain (yet unknown) conditions are satisfied, the sleep active neurons stimulate the REM-ON neurons which in turn actively inhibit and cease firing of the REM-OFF neurons (directly or indirectly) and initiate REM sleep[67,68]. Also, recently it was reported from our l
aboratory that picrotoxin, a GABA-A antagonist, in PPT, site of REM-ON neurons decreases REM sleep[82]. We expect that GABA from interneurons within the PPT[83] or from sleep area (caudal brainstem) dis-facilitates the in-hibitory influence of NE-ergic inputs from LC onto PPT neurons resulting in an increase in REM sleep. These find-ings have been summarized in Fig.2.
How do the LC neurons cease firing during REM sleep
Reciprocal interaction model
The reciprocal interaction model was proposed by Hobson et al[37]. The model hypothesized that the REM-OFF neu-rons are inhibitory to REM-ON neurons and to themselves, but the latter are excitatory to the former and to themselves. Electrophysiologically it has been shown that presumably noradrenergic LC neurons and serotonergic raphe neurons are REM-OFF while the cholinergic FTG neurons are REM-ON. Thus, according to the hypothesis, inactivation of putative monoaminergic REM-OFF neurons plays a criti-cal role in the generation and maintenance of REM sleep. Mutual inhibitory model
The mutual inhibitory model between the NE-ergic REM-OFF and cholinergic REM-ON neurons was proposed by Sakai[84]. It was based on the hypothesis that cessation of firing of the REM-OFF neur
ons excites the REM-ON neu-rons by disinhibition while the excitation of the REM-ON neurons inhibits the REM-OFF neurons. Therefore, REM sleep can appear either by excitation of the REM-ON neu-rons or by inhibition of the REM-OFF neurons. This hy-pothesis seems to imply that for the generation of REM sleep, cholinergic neurons directly inhibit NE-ergic REM-OFF neurons.
Lacuna in the above mentioned models
The two hypotheses mentioned above did not consider the type and role of neurotransmitters involved in mediating such actions. As mentioned earlier, during REM sleep the cholinergic REM-ON neurons increased firing, the REM-OFF neurons in the LC ceased firing and continuous acti-vation of the LC neurons by electrical or by chemical means prevented generation of REM sleep. Since NE inhibits cho-linergic tegmental neurons[85], it is reasonable to understand

版权声明：本站内容均来自互联网，仅供演示用，请勿用于商业和其他非法用途。如果侵犯了您的权益请与我们联系QQ:729038198，我们将在24小时内删除。