467

Acetylcholine±GABA interaction in LC and REM sleep Neuroscience Vol. 104, No. 2, pp. 467±485, 2001
q 2001 IBRO. Published by Elsevier Science Ltd
Pergamon Printed in Great Britain. All rights reserved
PII: S0306-4522(01)00062-8 0306-4522/01 $20.00+0.00
www.elsevier.com/locate/neuroscience

INTERACTIONS BETWEEN CHOLINERGIC AND GABAERGIC

NEUROTRANSMITTERS IN AND AROUND THE LOCUS COERULEUS FOR
THE INDUCTION AND MAINTENANCE OF RAPID EYE MOVEMENT SLEEP
IN RATS

B. N. MALLICK,* S. KAUR and R. N. SAXENA²

School of Life Sciences, Jawaharlal Nehru University, New Delhi 110 067, India

AbstractÐThe noradrenergic ªREM-offº neurons in the locus coeruleus cease ®ring, whereas some cholinergic and non-
cholinergic ªREM-onº neurons increase ®ring during rapid eye movement sleep. A reciprocal interaction between these
neurons was proposed. However, acetylcholine did not inhibit neurons in the locus coeruleus. Nevertheless, since GABA
levels increase during rapid eye movement sleep and picrotoxin injections into the locus coeruleus reduced rapid eye
movement sleep, it was hypothesized that GABA in the locus coeruleus might play an intermediary inhibitory role for
rapid eye movement sleep regulation. Therefore, the effects of GABA or carbachol (a mixed cholinergic agonist receptor)
alone, as well as an agonist of one in presence of an antagonist of the other, in the locus coeruleus were investigated on
sleep±wakefulness and rapid eye movement sleep.
The cholinergic agonist carbachol increased, while the muscarinic antagonist receptor scopolamine decreased, the
frequency of induction of rapid eye movement sleep per hour. In contrast, GABA and picrotoxin increased and decreased,
respectively, the duration of rapid eye movement sleep per episode. However, when carbachol was injected in the presence
of picrotoxin or GABA was injected in the presence of scopolamine, the effect of GABA or picrotoxin was dominant.
Microinjection of both scopolamine and picrotoxin in combination reduced both the frequency of initiation as well as the
duration per episode of rapid eye movement sleep.
From these results we suggest that in the locus coeruleus cholinergic input modulates the frequency of induction of rapid
eye movement sleep and this action is mediated through GABA interneurons, whereas the length of rapid eye movement
sleep per episode is maintained by the presence of an optimum level of GABA. A model of neural connections for initiation
and maintenance of rapid eye movement sleep is proposed and discussed. q 2001 IBRO. Published by Elsevier Science Ltd.
All rights reserved.

Key words: carbachol, microinjection, noradrenergic neurons, picrotoxin, REM sleep duration per episode, REM sleep frequency.

Rapid eye movement (REM) sleep is an essential physio- may be supported by the facts that: (i) these neurons
logical phenomenon, which is identi®ed by characteristic cease ®ring during REM sleep; 8 (ii) cooling of the LC
electrophysiological signals in the electroencephalogram increased REM sleep; 16 (iii) cholinergic stimulation of
(EEG), electro-oculogram (EOG) and electromyogram dorsolateral pons, including the LC and adjoining
(EMG). The pontine brainstem is reported to regulate areas, increased REM sleep; 82,83 (iv) ®ring rates of LC
REM sleep. 66,72 Although the detail mechanisms of neurons decreased, possibly a compensatory phenom-
initiation and maintenance of REM sleep are not enon, during REM sleep deprivation; 49 and (v) continu-
known, several populations of neurons including those ous activation (non-cessation) of LC neurons, by mild
in the locus coeruleus (LC), have been reported to play a electrical stimulation, reduced REM sleep. 73 Besides
signi®cant role in its regulation. The involvement of the noradrenergic neurons (ªREM-offº), presumably
noradrenergic neurons in LC for REM sleep regulation cholinergic neurons (ªREM-onº) are reported to increase
®ring during REM sleep. 22,41,51 Also, mild electrical
*Corresponding author. Tel.: 111-610-7676, ext. 2522; fax: 111- stimulation of cholinergic neurons increased REM
616-5886. sleep. 76 Based on the existing knowledge, a reciprocal
E-mail address: [email protected] (B. N. Mallick). interaction between the noradrenergic ªREM-offº and
²Present address: Department of Zoology, University of Delhi, New cholinergic ªREM-onº neurons for the genesis of REM
Delhi 110 007, India
Abbreviations: ACh, acetylcholine; ANOVA, analysis of variance;
sleep has been suggested. 29,67 This hypothesis may be
AW, active awake; DAB, 3,3 0 -diaminobenzidine; EEG, electroen- supported by the fact that the ªREM-onº and the
cephalogram; EOG, electro-oculogram; EMG, electromyogram; ªREM-offº neuronal ®ring rates altered in a reciprocal
LC, locus coeruleus; LDT/PPT, laterodorsal/pedunclopontine manner both during spontaneous 29 and induced REM
tegmental nuclei; NE, norepinephrine; PBS, phosphate-buffered sleep, 77 as well as during REM sleep deprivation. 49 The
saline; PGi, paragigantocellularis nucleus; PrH, prepositus hypo-
glossal nucleus; QA, quite awake; REM, rapid eye movement; LC receives cholinergic inputs, 35,36 and cholinergic
SWS, slow wave sleep; TH, tyrosine hydroxylase. receptors have been identi®ed on the LC neurons. 12,57
467
468 B. N. Mallick et al.

Table 1. Experimental protocol

Group Number of rats Chemical injected Volume injected (molarity) Type of recording

1 5 ± ± Baseline
2 5 Saline 2.25 mg/250 nl (17.1 mM) Control
3 5 Carbachol 2.5 mg/250 nl (54.8 mM) Experimental
4 5 Scopolamine 2.5 mg/250 nl (29.4 mM) Experimental
5 5 Picrotoxin 250 ng/250 nl (1.6 mM) Experimental
6 5 GABA 250 ng/250 nl (9.7 mM) Experimental
7 5 Picrotoxin followed 250 ng/250 nl (1.6 mM) Experimental
by carbachol
2.5 mg/250 nl (54.8 mM)
8 5 Scopolamine followed 2.5 mg/250 nl (29.4 mM) Experimental
by GABA
250 ng/250 nl (9.7 mM)
9 5 Scopolamine followed 2.5 mg/250 nl (29.4 mM) Experimental
by picrotoxin
250 ng/250 nl (1.6 mM)

Microdialysis sample showed that the levels of acetyl- episode studied. Preliminary ®ndings of this study were
choline (ACh) increased in dorsomedial tegmental ®eld reported earlier. 40
(near the LC) during REM sleep. 43 However, the
mechanism(s) of action of ACh in the LC and inhibition EXPERIMENTAL PROCEDURES
of the LC noradrenergic neurons during REM sleep were
not known. It was also not known if an identical mechan- Animals and recording apparatus
ism in the LC would be responsible for both the initiation The experiments were performed on male Wistar rats (250±
and the maintenance of REM sleep. 300 g) bred at the University Animal Facility were maintained in
As mentioned above, ACh levels increased in areas a 12-h light/dark cycle with food and water available ad libitum.
Under surgical anesthesia (Nembutal, pentobarbital sodium,
around the LC and also LC neurons cease ®ring during 35 mg/kg i.p.), rats were positioned in a stereotaxic apparatus
REM sleep. Hence, it may be suggested that increased and surgically implanted with bilateral EEG, EOG and EMG
ACh might be responsible for cessation of LC neurons electrodes to monitor sleep±waking states. For recording bipolar
for the initiation of REM sleep. However, ACh is EEG, two stainless steel screw electrodes were ®xed in the skull
(2 mm anterior and 4 mm lateral to bregma). Flexible insulated
reported to depolarize and increase LC neuronal ®ring (except at the tip) wires were connected bilaterally on either side
rate, 20,21 i.e., LC neurons were not inhibited by ACh. This of dorsal neck muscles and to the muscles near external canthus
led us to propose that an inhibitory intermediary neuro- of eyes for recording bipolar EMG and EOG, respectively.
transmitter GABA could mediate the cholinergic action Another screw electrode was ®xed on the midline over frontal
on LC neurons to induce REM sleep possibly by inhibit- sinus to serve as animal ground. A pair of stainless steel guide
cannulae with indwelling stylet was stereotaxically implanted
ing the LC neurons. 4,39,40,53 This proposal may be bilaterally so that their tips reached 1.5 mm above the LC
supported by the presence of GABAergic interneurons (9.3 mm posterior to bregma, 1 mm lateral and 5.5 mm deep) 63
and terminals in LC, 31,35,36 the presence of GABA- to minimize cellular damage in LC. The free ends of EEG, EMG
ergic 46,61 and cholinergic receptors 12,57 on the neurons and EOG electrodes were soldered to 9-pin female plug that was
then ®xed to the skull with dental acrylic. After surgery, a mini-
in LC, the presence of acetylcholinesterase in LC, 6 mum of four days was allowed for recovery from surgical trauma
increased GABA levels in LC during REM sleep, 60 before recording was started. During recovery period the rats
reduction of REM sleep by microinjection of picrotoxin, were acclimatized to the recording chamber and connecting
a GABA-A receptor antagonist, in the LC, 39 activation of cables. All efforts were made to minimize both the suffering
GABAergic neurons during REM sleep, 55 and inhibition and number of animal used, and the experimental protocols
were approved by the Institutional Animal Ethics Committee.
of noradrenergic neurons in the LC by GABA. 26
However, in our earlier studies it was observed that
Recording
stimulation of LC reduced the initiation of REM sleep
only, and the duration of REM sleep per episode On the day of recording each rat was placed in a semi-sound
remained unaffected; 73 while microinjection of picro- proof shielded recording chamber, at least 1 h before the start of
recording. Bipolar EEG, EOG and EMG were continuously and
toxin into the LC did not affect the initiation but simultaneously recorded with a Grass Model 7 polygraph. All the
reduced the duration of REM sleep signi®cantly. 39 recordings were done between 09.00 and 19.00 h in nine groups
Hence, it was hypothesized that in the LC the cholinergic of rats (Table 1). No injection was made in the baseline group of
action is possibly mediated through GABA and both rats. In all other groups pre-injection recording was done for 1 h
and thereafter, 250 nl of saline, agonist or antagonist (individu-
GABA and ACh interact for the induction and main- ally or in combination) was locally microinjected between 10.00
tenance of REM sleep. Therefore in this study, either and 10.30 h as described earlier. 3 The injected chemicals and
the agonist or the antagonist of these neurotransmitters their molarity are given in Table 1. The microinjection was
alone or in selected combinations was locally micro- done with a 30G-injector connected to a 2 ml Hamilton syringe
injected into the LC of chronically prepared freely through a thin 10-cm polyethylene tubing tightly ®t to the needle.
The injector cannula had a pre®xed stopper at a height so that
moving rats and the effects on total REM sleep, rate of when introduced through the guide cannula, its tip projected
induction of REM sleep and duration of REM sleep per 1.5 mm beyond the latter to reach the LC. The injection was
 Acetylcholine±GABA interaction in LC and REM sleep 469

made bilaterally, one side at a time, at the rate of 100 nl/min. TH (1:1000; Eugene Tech., USA), in 0.1 M PBS, containing
Thereafter, the injector was left in place for 1 min for diffusion 0.3% Triton X-100 and 2% normal goat serum, for four days at
and to prevent back¯ow of the injected chemical. After injection, 48C. The sections were washed in PBS, and then incubated in a
the injector cannula was replaced with the blockers. The injection biotinylated secondary antibody (goat anti-rabbit anti IgG; Vector
procedure took 7±8 min for bilateral injection of one chemical Laboratories, USA) for 3 h at 48C. The sections were again
and double the time for serial injection of two chemicals. When washed in PBS and incubated in PBS containing 1% avidin±bio-
injecting a combination of an agonist and an antagonist, the tinylated horseradish peroxidase complex (Vector Laboratories)
antagonist was injected ®rst followed by injection of the agonist. for 1 h at room temperature. Finally, sections were treated in
After injection the recording was continued for 8 h. Each animal 50 mM Tris buffer (pH 7.4) containing 0.025% 3,3 0 -diamino-
was only used for one study, involving either a single chemical benzidine (DAB) and 0.0075% H2O2 for 8±10 min, and then
injection or a combination of chemical injections. However, to the reaction was stopped by washing in PBS. The sections were
make sure that the injection did not damage the area, in randomly then mounted on gelatin subbed slides, dehydrated and prepared
selected rats, a second injection was also made after an interval of for examination under light microscope (Fig. 2).
one day after the ®rst injection to study if the effects could be
reproduced. Similarly, in randomly selected rats in which the
RESULTS
microinjection induced a signi®cant effect, normal recordings
were made after an interval of one day to make sure that the
injection did not have a lasting or permanent effect and also to
The percent time spent in sleep and wakefulness states
show that the effect of the chemical injection was reversible. after microinjection into the LC of saline or cholinergic
and GABAergic receptor agonist or antagonists (indivi-
Analysis dually or in combination) are shown in Table 2. Saline
did not have any signi®cant effect on the total sleep, or
The polygraphic records were ®rst scanned visually to mark total waking and REM sleep. However, an hourwise
obvious artifacts, if any, which were often marked during record-
ing session. The records were then scored in bins of 10 s epoch analysis showed that REM sleep was reduced during
and subdivided into active wake (AW), quiet awake (QA), slow the ®rst hour. This is likely to be a non-speci®c effect
wave sleep (SWS1), deep sleep (SWS2) and REM sleep as due to handling because the effects were similar after
reported earlier. 73,78 The AW was identi®ed by the presence of all the injections. Microinjection of cholinergic and
desynchronized EEG accompanied by high EMG tone and/or
muscle movement and eye movements in the EOG. The QA GABAergic receptor agonist or antagonists into the LC
was characterized by the presence of occasional spindling, no signi®cantly affected sleep±waking (Table 2). Represen-
active muscle movement and reduced eye movements. The tative recording samples of REM sleep during baseline
SWS1 was characterized by the presence of EEG synchronization and after saline injection into the LC (control) are shown
up to 50% of recording time, reduced EMG tone and reduced eye in Fig. 3; those after microinjection of carbachol and
movements. The SWS2 was characterized by the presence of
synchronized EEG (.75% time), signi®cantly reduced muscle scopolamine are shown in Fig. 4 and those after GABA
tone and no eye movements. The REM sleep was scored when and picrotoxin in Fig. 5. The microinjection did not
EEG desynchronization was accompanied by muscle atonia and appear to have damaged the brain area because the
frequent eye movements, usually following a stage of SWS2. The sleep±waking periods returned to baseline when the
mean of each of total sleep, wakefulness, REM sleep, REM sleep
duration per episode and REM sleep frequency values for the recording was done after a gap of one day after the injec-
entire recording period as well as for every hour in all the experi- tion, and a second injection in randomly selected rats
mental groups were statistically compared separately with that of produced reproducible results.
respective baseline and saline injection values by using ANOVA
(and the statistical package, BIOSTAT). The levels of signi®-
cance of differences at a minimum of P , 0.05 were calculated Effects of cholinergic agonist and antagonist
applying the Newman±Keuls multiple range post hoc test. Effect of carbachol. Injection of carbachol into the LC
induced a signi®cant increase in total time spent in REM
Reagents injected sleep (Table 2) compared to both baseline (F1,8 ˆ 12.7;
Carbachol, a mixed cholinergic agonist; scopolamine, a P , 0.01) as well as saline injection (F1,8 ˆ 19.1;
muscarinic receptor antagonist; picrotoxin, a GABA-A receptor P , 0.01). This increase was due to a signi®cant increase
antagonist; and GABA. These drugs were obtained from Sigma (Fig. 6A) in the frequency of initiation of REM sleep
(USA) and dissolved in sterile isotonic saline.
episodes, compared to baseline (F1,8 ˆ 5.2; P , 0.01)
and saline (F1,8 ˆ 4.4; P , 0.01), whereas the mean dura-
Histological identi®cation of the site and spread of injection tion of REM sleep per episode was not signi®cantly
At the end of each experiment and under deep anesthesia affected (Fig. 7A). Carbachol did not affect other
(pentobarbital sodium 45 mg/kg) the rats were injected with sleep±wakefulness stages signi®cantly (Table 2). Carba-
250 nl (same volume as that of the injected chemical) of 2% chol injections in areas away from the LC (Fig. 1) were
pontamine Sky Blue in the same manner as that of the agonist/
antagonist. After 20±30 min the animals were intracardially ineffective.
perfused with 80±100 ml of 10% formalin in saline. The brains An hourwise analysis revealed that after carbachol
were then stored in formal for one to four weeks. The brainstem microinjection the frequency of REM sleep initiations
was trimmed and kept in 30% sucrose solution for 24 h, before increased signi®cantly (Fig. 6B) up to the seventh post-
making 40-mm thick sections. The sections were stained with
eosin, and the site of injection was identi®ed by the presence of injection hour, while the mean duration per episode
blue coloration and plotted in reconstruction diagrams (Fig. 1). remained unaffected except for the ®rst post-injection
To con®rm if the injection was made in the noradrenergic LC hour (Fig. 7B).
region, some of the randomly selected experimental rat brains
were perfused intracardially with 4% paraformaldehyde in
0.1 M phosphate-buffered saline (PBS) for tyrosine hydroxylase
Effect of scopolamine. The percent time spent in REM
(TH) immunostaining using a standard avidin±biotin method. 30 sleep, compared to both baseline (F1,8 ˆ 21.9; P , 0.01)
The free-¯oating sections were incubated in primary antisera, anti and after saline injection (F1,8 ˆ 30.1; P , 0.01), was
470 B. N. Mallick et al.

Fig. 1. Schematic reconstruction diagrams of histological sections through LC showing the anatomical locations of the microinjec-
tion sites. Coronal sections of the rat brainstem through A8.8 mm, A9.3 mm, A9.8 mm and A10.3 mm according to the atlas of
Paxinos and Watson 63 are represented. The sites where the drug injection signi®cantly affected REM sleep are marked as effective
(hatched). Those sites where the injections did not signi®cantly affect REM sleep are marked as ineffective (®lled). The shaded areas
show the centres of the site of injection of saline and drugs in all the experimental animals. Although the injections were made
bilaterally, they are presented unilaterally for convenience. 4V, fourth ventricle, Me5, mesencephalic nucleus of the trigeminal
nerve, Pnc, Pontine reticular nucleus (caudal), Rpn, raphe pontis nucleus, SubC, subcoeruleus nucleus.

signi®cantly reduced after scopolamine microinjection in signi®cantly reduced the frequency of REM sleep ini-
the LC (Table 2). This decrease in REM sleep was due to tiation for the majority of the post-injection record-
a reduction (Fig. 6A) in the frequency of initiation of ing hours (Fig. 6B); however, the mean duration per
REM sleep episodes per hour, compared to baseline episode remained unaffected (Fig. 7B) except for the
(F1,8 ˆ 13.0; P , 0.01 and saline (F1,8 ˆ 18.5; P , 0.01), ®rst 2 h.
and was not due to a reduction in mean duration per
episode (Fig. 7A). Although there was an increase in
AW, compared to baseline (F1,8 ˆ 3.2; P , 0.05) and Effects of GABA and its antagonist
saline (F1,8 ˆ 3.7; P , 0.05), other stages remained un- Effect of GABA. Microinjection of GABA into the
affected (Table 2). LC signi®cantly increased AW compared to both
An hourwise analysis showed that scopolamine baseline (F1,8 ˆ 1.9; P , 0.05) and saline injection
 Acetylcholine±GABA interaction in LC and REM sleep
Fig. 2. A representative histological section through TH-positive neurons of the LC at different magni®cations. The site of injection is shown by the Pontamine Sky Blue (Arrow). (a) A
photomicrograph of TH-stained histological section showing the site and spread of Pontamine Sky Blue (shown by arrow) in a drug-injected rat. Bilateral cannula tracts (CT) are seen. (b) A
photomicrograph at a higher magni®cation of the right side of the same TH-stained section as in panel a. The arrow points to the prescence of Pontamine Sky Blue. The tip of the cannula tract and
TH-positive neurons are also seen. (c) Magni®ed photomicrograph of the section as shown in panel b showing the injection site (arrow) in the vicinity of the TH-positive LC neurons. (d) Higher
magni®ed view of the TH-positive neurons of the same section shows that the neurons were not damaged.

471
472 B. N. Mallick et al.

Fig. 3. Polygraphic records of EEG, EOG and EMG showing signs of REM sleep (marked by arrows). Polygraphic tracings from
baseline (A) and after saline microinjection (B) are shown here. The duration of REM sleep episodes were comparable.

Table 2. Mean ^ S.E.M. percent duration of sleep±wakefulness REM sleep stages

Group Chemical injected AW QA SWS1 SWS2 REM sleep

1 Baseline 20.65 ^ 2.08 23.05 ^ 5.16 24.82 ^ 1.35 25.67 ^ 1.35 5.26 ^ 0.29
2 Saline 19.50 ^ 3.3 20.91 ^ 5.17 22.48 ^ 1.78 29.51 ^ 7.89 5.39 ^ 0.14
3 Carbachol 17.89 ^ 2.59 23.86 ^ 3.65 19.99 ^ 1.51 29.63 ^ 4.39 9.46 ^ 0.89**²²
4 Scopolamine 27.91 ^ 2.26*² 18.17 ^ 1.04 19.66 ^ 1.80 28.58 ^ 0.89 2.39 ^ 0.60**²²
5 GABA 26.65 ^ 2.94*² 26.99 ^ 3.46 18.02 ^ 1.88**² 22.32 ^ 2.35 7.04 ^ 0.80*²
6 Picrotoxin 11.57 ^ 2.4**²² 22.22 ^ 2.92 29.48 ^ 1.42**²² 32.48 ^ 2.47 3.48 ^ 0.42**²²
7 Picrotoxin 1 Carbachol 10.96 ^ 1.26**²² 21.18 ^ 3.53 24.22 ^ 2.57 38.20 ^ 6.41*² 3.50 ^ 0.95**²²
8 Scopolamine 1 Carbachol 28.15 ^ 4.26*² 17.17 ^ 2.9 20.35 ^ 3.54 24.36 ^ 1.81 6.87 ^ 1.44*²
9 Scopolamine 1 Picrotoxin 40.03 ^ 5.83**²² 17.10 ^ 1.86 17.38 ^ 2.79**² 22.72 ^ 3.43 0.83 ^ 0.09**²²

The data are for 8 h recording after injection as described in the Experimental procedures.
*P , 0.05 as compared to baseline and ²P , 0.05 as compared to saline; **P , 0.01 as compared to baseline and ²²P , 0.01 as compared to saline.

(F1,8 ˆ 2.2; P , 0.05) and decreased SWS1, compared to An hourwise analysis showed that after GABA injec-
baseline (F1,8 ˆ 7.2; P , 0.01) and saline injection tion in the LC the frequency of REM sleep initiation
(F1,8 ˆ 2.9; P , 0.05) (Table 2). It also resulted in a signi- remained unaffected, except for the ®rst post-injection
®cant increase in percent time spent in REM sleep, hour (Fig. 8B) which was likely to be due to the handling
compared to baseline (F1,8 ˆ 3.5; P , 0.05) and saline and non-speci®c factors. However, the mean duration
injection (F1,8 ˆ 3.2; P , 0.05) (Table 2). Although the per episode increased signi®cantly until 6 h (Fig. 9B)
frequency of initiation of REM sleep remained unaffected post-injection.
(Fig. 8A), this increase in REM sleep after GABA injection
was due to a signi®cant increase (Fig. 9A) in the mean Effect of picrotoxin. Microinjection of picrotoxin, a
duration of REM sleep per episode compared to both base- GABA-A receptor antagonist, into the LC signi®cantly
line (F1,8 ˆ 6.2; P , 0.01) and saline injection (F1,8 ˆ 10.0; reduced (Table 2) the percent total time spent in REM
P , 0.01). The injection of GABA in areas outside the LC sleep compared to both baseline (F1,8 ˆ 15.0; P , 0.01)
(Fig. 1) did not affect REM sleep. and saline injection (F1,8 ˆ 9.7; P , 0.01). This decreased
 Acetylcholine±GABA interaction in LC and REM sleep 473

Fig. 4. Simultaneous polygraphic traces of EEG, EOG and EMG showing a period of REM sleep, after microinjection of scopo-
lamine (A) and carbachol (B) into the LC of two rats. REM sleep durations are marked by arrows. The REM sleep duration per
episode in these groups was comparable to those of the baseline and the saline groups (as seen in Fig. 3).

REM sleep was not due to a reduction in the frequency of understand the possible sequence of events and accord-
initiation of REM sleep (Fig. 8A), but was due to a ingly the possible functional connections between
signi®cant reduction (Fig. 9A) in the mean duration neurons. Representative polygraphic traces for identi®ca-
of REM sleep per episode, compared to baseline tion of REM sleep after combined injection of the chemi-
(F1,8 ˆ 12.7; P , 0.01) and saline injection (F1,8 ˆ 9.9; cals into LC are shown in Fig. 10 (cf. Fig. 3 for controls).
P , 0.01). An hourwise analysis showed that although
the frequency of initiation of REM sleep remained un-
affected throughout (Fig. 8B), the mean duration per Effect of picrotoxin followed by carbachol. This injec-
episode was considerably reduced (Fig. 9B) at least for tion combination signi®cantly decreased REM sleep
6 h after picrotoxin injection. (Table 2). Sample polygraphic traces of REM sleep are
After picrotoxin microinjection, although there was a shown in Fig. 10A. Although the frequency of initiation
signi®cant reduction in the percent time spent in AW, of REM sleep was unaffected (Fig. 11A), the mean dura-
compared to baseline (F1,8 ˆ 5.9; P , 0.01) and saline tion of REM sleep per episode (Fig. 12A) was reduced
injection (F1,8 ˆ 5.6; P , 0.01), there was an increase in signi®cantly, compared to baseline (F1,8 ˆ 5.1, P , 0.05)
the time spent in SWS1, compared to baseline and saline (F1,8 ˆ 2.2; P , 0.05). Total time spent in
(F1,8 ˆ 10.9; P ,0.01) and saline injection (F1,8 ˆ 17.9; REM sleep was signi®cantly reduced, compared to base-
P , 0.01) (Table 2). line (F1,8 ˆ 31.8; P , 0.01) and saline (F1,8 ˆ 30.7;
P , 0.01). This reduction was comparable to that after
picrotoxin alone. Also, this combination of injection
Effects of combined agonist/antagonist injections signi®cantly decreased AW compared to both baseline
A combination of cholinergic and GABAergic agonist (F1,8 ˆ 9.9; P , 0.01) and saline (F1,8 ˆ 8.9; P , 0.01),
and antagonist were microinjected bilaterally in the LC, but increased SWS2, compared to baseline (F1,8 ˆ 7.3;
and their effects on sleep±waking states analysed. As P , 0.05) and saline (F1,8 ˆ 3.7; P , 0.05) (Table 2).
mentioned in the Experimental procedures section, the An hourwise analysis showed that the combination did
injection of agonists followed that of antagonists in all not affect the frequency of initiation of REM sleep (Fig.
the combined injections. This approach was followed to 11B); however, the mean duration per episode was
474 B. N. Mallick et al.

Fig. 5. This ®gure shows (simultaneous) polygraphic tracings of EEG, EOG and EMG during a REM sleep episode (marked by
arrows) after microinjection of picrotoxin (A) or GABA (B) into the LC. The duration of REM sleep per episode signi®cantly
decreased after picrotoxin (A) and signi®cantly increased after GABA (B) microinjection into LC. In B the REM sleep lasted for
120 s.

signi®cantly reduced during the second, ®fth and seventh (F1,8 ˆ 208.9; P , 0.01) (Table 2). In this case the reduc-
hours of the recording period after injection (Fig. 12B). tion was due to a signi®cant decrease in both the
frequency of initiation of REM sleep compared to both
Effect of scopolamine followed by GABA. A represen- baseline (F1,8 ˆ 8.27; P , 0.05) and saline (F1,8 ˆ 4.39;
tative polygraphic trace of an episode of REM sleep after P , 0.05) (Fig. 11A), as well as the duration per episode
microinjecting scopolamine followed by GABA is compared to both baseline (F1,8 ˆ 50.12; P , 0.01) and
shown in Fig. 10B. The percent time spent in REM saline (F1,8 ˆ 43.03; P , 0.01) (Fig. 12A). The REM
sleep during the entire recording period after such injec- sleep episodes in these rats, whenever present, mostly
tion was signi®cantly increased compared to both base- terminated shortly after its onset. The classical electro-
line (F1,8 ˆ 2.1; P , 0.05) and saline injection (F1,8 ˆ 1.9; graphic signals of REM sleep (EEG desynchrony, REM
P , 0.05) (Table 2). This was due to a signi®cant and EMG atonia) were present for a very brief period, as
increase in mean duration of REM sleep per episode, if the REM sleep could not continue (Fig. 10C). This also
compared to baseline (F1,8 ˆ 11.4; P , 0.01) and saline led to a signi®cant increase in AW, compared to base-
(F1,8 ˆ 2.4; P , 0.05) (Fig. 12A). This combination also line (F1,8 ˆ 8.811; P , 0.01) and saline (F1,8 ˆ 9.25;
led to a signi®cant increase in AW compared to both P , 0.01); however, the SWS1 duration was reduced,
baseline (F1,8 ˆ 1.9; P , 0.05) and saline (F1,8 ˆ 2.2; compared to baseline (F1,8 ˆ 5.18; P , 0.01) and saline
P , 0.05) (Table 2). (F1,8 ˆ 2.37; P , 0.05) (Table 2).
An hourwise analysis revealed that the frequency of An hourly analysis showed that both the frequency of
REM sleep initiation remained unaffected (Fig. 11B), REM sleep initiation (Fig. 11B) and the mean duration
whereas the mean duration of REM sleep per episode per episode of REM sleep (Fig. 12B) were signi®cantly
was signi®cantly increased for the entire recording reduced throughout post-injection recording time.
period (Fig. 12B).
DISCUSSION
Effect of scopolamine followed by picrotoxin. This
combination signi®cantly reduced REM sleep compared The noradrenergic ªREM-offº neurons, 8 cholinergic 35,36
to both baseline (F1,8 ˆ 735.5; P , 0.01) and saline and GABAergic inputs, 64,74 as well as GABAergic
 Acetylcholine±GABA interaction in LC and REM sleep 475

Fig. 6. The mean ^ S.E.M. frequency of initiation of REM sleep during the entire recording time (A) and every hour (B) in baseline
(BSLN) and after microinjection of saline, carbachol (CARB) and scopolamine (SCOP) into the LC. * Signi®cant as compared to
baseline, # signi®cant as compared to saline; *# P , 0.05; **## P , 0.01.

neurons 23,31,36 are present in and immediately around the the animal was disturbed during injections and therefore,
LC. The independent in¯uence of cholinergic 82,83 and the response up to about 30 min of post-injection was
GABAergic 39,40 inputs in the LC in REM sleep has confounded with non-speci®c effects. It may be argued
been studied in isolation in different species. A reciprocal that at least part of this limitation could have been
interaction between ªREM-onº and ªREM-offº neurons avoided by the use of remote control injection by
was proposed earlier. 29,67 However, the neurochemical pump. However, in the remote control injection proce-
nature of these neurons could not explain the mechanism dure sequential injection of two chemicals as has been
of inhibition of noradrenergic ªREM-offº neurons by the done in this study, was not possible without disturbing
cholinergic ªREM-onº neurons. Hence, in this study an the animal. Hence, as a compromise, local injection was
interaction between the cholinergic and the GABAergic done manually as before 50,52 and to rule out the non-
systems in the LC in REM sleep regulation was investi- speci®c effects appropriate control studies were carried
gated. Local microinjection of cholinergic and GABA- out.
ergic agonists and antagonists alone or in a speci®c The results of this study showed that in the LC both the
sequence was used in this study. This method has been cholinergic and the GABAergic agonists increased while
used earlier by us 50,52 and also recently by another their antagonists decreased REM sleep; however, the
group. 14 The primary limitation of this method is that former affected the frequency of initiation while the latter
476 B. N. Mallick et al.

Fig. 7. The effects of scopolamine (SCOP) and carbachol (CARB) on mean ^ S.E.M. REM sleep duration per episode as compared
to that of baseline (BSLN) and saline for the entire 8-h post-injection recording period (A) and at each post-injection recording hour
(B). Signi®cance levels are as in Fig. 6.

affected the duration per episode of REM sleep. These in LC is spontaneously active in regulating REM sleep.
suggest that in the LC, ACh is possibly involved in the An increase in REM sleep by cholinergic stimulation of
initiation, while GABA in the maintenance, of REM dorsolateral pontine tegmentum 10,11,15,25 and an area
sleep. Although the cholinergic agonist, carbachol, did around LC has been reported earlier. 82,83 The effective
not signi®cantly affect other sleep±waking stages, the dose (the dose that signi®cantly increased REM sleep)
antagonist, scopolamine, increased AW. However, of carbachol used in the present study was comparable to
GABA signi®cantly increased AW and decreased that used in earlier studies in rats where the effect lasted
SWS1, while its blocker had opposite effects. for 4±6 h and the number of REM sleep episodes
increased. 13,18,27,32,58 In cats, enhanced REM sleep was
evoked by varying doses of carbachol (8.76±
Effects of cholinergic agonist and antagonist 87.68 mM, 20±500 nl) and the effects lasted from 4 h
Microinjection of a cholinergic agonist in the LC to four days. 11,15,25,82,83 In cats there have been reports
increased, whereas its antagonist decreased, the percent of an increase in the frequency as well as the duration
time spent in REM sleep. Since the cholinergic antagon- of REM sleep after carbachol injection; 10,11,15,25,71
ist reduced REM sleep, it is likely that cholinergic input however, in rats only the frequency of REM sleep
 Acetylcholine±GABA interaction in LC and REM sleep 477

Fig. 8. The effects of picrotoxin (PICRO) and GABA on mean ^ S.E.M. frequency of initiation of REM sleep during the entire
recording period (A) and during each post-injection recording hour (B). The baseline (BSLN) and the effect of saline are also shown.
Signi®cance levels are as in Fig. 6.

initiation was signi®cantly affected. 13,18,27,32,58 Differ- extent of diffusion of the drug, which depends on several
ences in the doses used and spread of injection, differ- factors including the solvent, solute, its concentration,
ences in the chemical nature and the total number of binding coef®cient, dissociation constant, degradation,
affected neurons between injection sites in cats and tissue metabolism and blood ¯ow, could not be esti-
rats, differential distributions of cholinergic and amin- mated. To overcome this limitation, injections were
ergic neurons within pontine tegmentum and their also made outside the LC, which did not produce an
projections are possible contributing factors for the varia- effect similar to that observed when the injections were
tions in the effects. The results of this study are in agree- made in the LC. The sites where carbachol or GABA
ment with the ®ndings in the Flinder sensitive line of rats, injections signi®cantly increased REM sleep and their
which showed increased muscarinic receptors and hyper- antagonists reduced REM sleep were considered to be
sensitivity to cholinergic agents. 70 Thus, the increased effective sites, while the sites where those injections
REM sleep in the Flinder sensitive line of rats was due did not affect REM sleep signi®cantly were termed as
to increased cholinergic activity. ineffective sites.
The site of injection was identi®ed histologically by A contiguity of the cholinergic innervation on LC
the presence of a Pontamine Sky Blue spot. The centre of neurons and a close apposition of their axonal and dendritic
the injection site and the spread of the drug to the processes have been shown. 35,36 These suggest a possible
immediate surrounding of the injection site could be interaction between cholinergic inputs and aminergic
localized with reasonable accuracy. However, the precise neurons in the LC for the initiation of REM sleep. This
478 B. N. Mallick et al.

Fig. 9. The effect of saline, picrotoxin (PICRO) and GABA on mean ^ S.E.M. REM sleep duration per episode for the entire 8-h
post-injection recording period (A) and for each post-injection recording hour (B). The baseline values are also shown. Signi®cance
levels are as in Fig. 6.

view may also be supported by the fact that presumably inhibitory neurotransmitter, such as GABA. 4,39,40,53 To
cholinergic ªREM-onº neurons increased ®ring, 22,41,51 con®rm this, the rat LC was microinfused with a
while the aminergic ªREM-offº neurons decreased ®ring, GABA-ergic agonist and an antagonist alone, and also
during spontaneous 8 and induced REM sleep77 as well as in the presence of cholinergic agonist or antagonist, and
during REM sleep deprivation. 49 However, in vitro intra- the effect on the REM sleep was investigated. The
cellular studies failed to show inhibition of rat LC neurons hypothesis was supported by the fact that GABA is
by ACh.20,21 In contrast, continuous activation of the LC an inhibitory neurotransmitter; LC receives a dense
signi®cantly reduced the frequency of initiation of REM GABAergic innervation from local as well as other
sleep while the mean duration of REM sleep per episode GABAergic neurons including those from the preoptic
remained unaffected. 73 Since LC noradrenergic neurons area; 64,74 GABA levels increase in the LC; 60 and
normally cease ®ring during REM sleep and continue ®ring GABAergic neurons are active during REM sleep. 55
during REM sleep deprivation, it was interpreted that
stimulation of LC did not allow neurons to cease ®ring
and therefore, REM sleep was not initiated. The above Effects of GABA and its antagonist
results led us to propose that in the LC the cholinergic GABA increased, while its antagonist picrotoxin
initiation of REM sleep is possibly mediated through an decreased, REM sleep. Unlike the cholinergic effects
 Acetylcholine±GABA interaction in LC and REM sleep 479

Fig. 10. Simultaneously recorded polygraphic traces of EEG, EOG and EMG representing an REM sleep episode in the rats
microinjected with picrotoxin followed by carbachol (A), scopolamine followed by GABA (B), or scopolamine followed by
picrotoxin (C) into the LC. As compared to baseline and saline (see Fig. 3A, B) the REM sleep duration per episode decreased
in A and increased in B. The REM sleep, although initiated in C (as marked by an arrow), could not continue further.

where the frequency of initiation of REM sleep was was much smaller than those used by others. 37,59,65 This
affected, microinjection of GABA or picrotoxin into dose of GABA was used because the same amount of
the LC altered REM sleep by affecting primarily the picrotoxin was also used. Although the molarity of carba-
duration of REM sleep per episode. It is likely that the chol was approximately double that of the scopolamine,
cholinergic and the GABAergic inputs in the LC play a for GABA it was six times that of picrotoxin. This is
mutually supportive role where they interact for initiation justi®ed because GABA is present naturally as a neuro-
and maintenance of REM sleep, respectively. The persis- transmitter in the brain where it can be degraded. The
tence of the effect of GABAergic agonist and antagonist reason for an increase in wakefulness after GABA micro-
for 6 h may be due to local reverberatory connections of injection into the LC cannot be explained from this study.
GABAergic neurons to and other neurons within and Since in the LC, the cholinergic input modulated the
outside the LC causing inhibition and disinhibition of frequency and GABAergic input the duration of REM
the neurons. The GABAergic effect in the preoptic area sleep, it may be said that GABA acted after the cholin-
also lasted for a comparable period. 7 The picrotoxin dose ergic system had initiated its action, because the process
was similar to that used earlier and the effects were of maintenance would be required only after the phenom-
comparable. 39 The effect of GABA was opposite to that enon has been initiated. Thus, the individual injection
of picrotoxin, suggesting that the effects were not non- studies brought forward the following possibilities: (i)
speci®c. Since the GABA agonist muscimol has a strong the cholinergic input is likely to act on the GABAergic
depressing effect in the CNS, 2 GABA itself was used in neurons in the LC for the initiation of REM sleep; and
this study. The dose of GABA used in this study (250 ng) (ii) GABA acts on the cholinergically primed system to
480 B. N. Mallick et al.

Fig. 11. The effects of the combined injections of picrotoxin followed by carbachol (PICRO 1 CARB), scopolamine followed by
GABA (SCOP 1 GABA) and scopolamine followed by picrotoxin (SCOP 1 PICRO) on the frequency of initiation of REM sleep
(mean ^ S.E.M.) during the entire 8-h post-injection recording period (A) and during every hour of the post-injection recording
period (B). The baseline (BSLN) and saline values are also shown. Signi®cance levels are as in Fig. 6.

maintain the duration of REM sleep. To test these possi- LC signi®cantly decreased REM sleep due to a reduction
bilities, the LC was microinfused with combination of in REM sleep duration per episode, while the frequency
either of the following: (i) a GABAergic antagonist of initiation of REM sleep remained unaffected. This
followed by a cholinergic agonist; (ii) a cholinergic antag- effect on REM sleep was similar to that of the micro-
onist followed by GABA; or (iii) a cholinergic antagonist injection of picrotoxin only. In contrast, when scopol-
followed by a GABAergic antagonist. These were done amine was followed by GABA microinjection, it
with the assumption that if a cholinergic input acted on resulted in a signi®cant increase in the percent of REM
the GABAergic neurons, the carbachol in the presence of sleep due to an increase in mean duration per episode of
picrotoxin would induce an effect similar to that of picro- REM sleep, whereas the frequency of initiation of REM
toxin alone, while GABA in the presence of scopolamine sleep remained unaffected. Thus, when cholinergic and
would show an effect similar to that of GABA alone. GABAergic agonists and antagonists were injected in
any combination, the effect of the agonist or antagonist
of GABA injected alone was observed. One possibility
Combined injection studies to explain such a result in simple terms is that one of
Injections of picrotoxin followed by carbachol into the the neurotransmitters is acting on the neuron containing
 Acetylcholine±GABA interaction in LC and REM sleep 481

Fig. 12. The effects of saline and combined injections of picrotoxin followed by carbachol (PICRO 1 CARB), scopolamine followed
by GABA (SCOP 1 GABA) and scopolamine followed by picrotoxin (SCOP 1 PICRO) into the LC on the REM sleep duration per
episode (mean ^ S.E.M.). The baseline value is also shown for comparison. The mean ^ S.E.M. for the entire 8-h recording period is
shown in A while values for each post-injection hour are shown in B. Signi®cance levels are as in Fig. 6.

the other neurotransmitter, i.e. in this particular case neurons, and their inhibition by GABA leading to the
cholinergic input is acting on that of GABAergic initiation of REM sleep needs to be con®rmed. A total
neurons. Thus, the observation supported our contention loss of REM sleep was not observed after blocking
that the cholinergic in¯uence in the LC was mediated either cholinergic receptors or GABAergic receptors.
through GABA transmission. The GABAergic neurons However, there was a complete suppression of REM
are present in and immediately around the LC. The latter sleep when scopolamine and picrotoxin were injected
may be supported by a recent report that there are some together. This supports mutually permissive and co-
carbachol-sensitive, non-cholinergic ªREM-onº neurons operative roles between the cholinergic and GABAergic
in the peri-LC / . 69 Although these neurons are systems.
suspected to be glutamatergic, 69 we propose that at Besides GABAergic terminals, glycinergic and gluta-
least some of them are GABAergic because glutamate, matergic neurons and terminals also have been reported
unlike GABA, has not been shown to increase in con- in the LC. 19,24 However, since glycine and glutamate
centration in the LC during REM sleep. 60 Nevertheless, concentrations did not change during REM sleep 60 and
the presence of GABA receptors on LC ªREM-offº glycine tonically inhibits noradrenergic LC neurons
482 B. N. Mallick et al.

represented as a model (Fig. 13) for initiation and main-

Medial Preoptic area
sleep active neurons tenance of REM sleep. The cholinergic ªREM-onº
neurons act simultaneously on the GABAergic as well
(GABAergic)

(-) as the noradrenergic ªREM-offº neurons in the LC.

(+) Noradrenergic During non-REM sleep-state the noradrenergic ªREM-
REM-OFF (LC) offº neurons continue ®ring possibly due to inputs from
(+)
(-)
(-)
neurons other than the ªREM-onº type. The activity of
(+)
GABAergic the ªREM-offº neurons would cause norepinephrine
Noncholinergic REM-ON
(NE) release through their collaterals onto themselves,
(+) (+) and this hyperpolarizes the same neurons. 1,80 This
? mechanism possibly causes the ªREM-offº neurons to
(-) GABAergic ®re rhythmically. Since it is reported that there are
Cholinergic
REM ON projections of the noradrenergic LC neurons near the
(-) ªREM-onº neurons, 34 there would be an increase in
Fig. 13. Based on the ®ndings of this study, the possible neuronal NE around the cholinergic ªREM-onº neurons also.
network for REM sleep initiation and maintenance by the LC is During non-REM sleep, this increased NE release is
schematically represented in this model. It is a modi®cation of the
earlier models proposed by Hobson and McCarley 29 and Sakai, 67 that
likely to maintain inhibition on presumably cholinergic
overcomes the drawbacks of those models. Continuous ®ring of the neurons. 28,81
noradrenergic (ªREM-offº) neurons in the LC keeps the ®ring of the During the gradual onset of sleep, there would be an
cholinergic ªREM-onº neurons (which may be non-cholinergic as increase in ®ring rate of the sleep-related neurons includ-
well) at a very low level (or may be inhibited). Increased release of ing those in the preoptic area. 38,48,75 Some of these sleep-
NE by the LC neurons due to their continuous activity tends to hyper-
polarize and autoinhibit LC neurons. This in turn would disinhibit related neurons in the preoptic area are proposed to be
cholinergic ªREM-onº neurons, which would start ®ring. The latter GABAergic. 5,7 Also, there are projections from the
would activate the GABAergic neurons as well as the adrenergic preoptic area to the LC and some of which are possibly
neurons in the LC. The inhibition by GABA together with the auto- GABAergic. 64,74 Thus, during sleep there will be gradual
inhibition would cause the LC noradrenergic neurons to cease ®ring
and initiate REM sleep. The continuation of REM sleep would depend
increase in GABA levels in the LC. This may be
on the duration for which the GABA will be available. The GABA- supported by the facts that a gradual increase in GABA
ergic inputs may come from the local interneurons present within and levels is reported in the LC during the transition from
around the LC or from neurons located in distant areas, which could wake to sleep, 60 and GABAergic neurons are likely to be
include the PrH 9,68 and non-cholinergic ªREM-onº neurons. 69 (1) active during REM sleep. 55 This increased GABA release
Excitation; (2) inhibition; continuous lines, known neural connec-
tions; broken lines, proposed neural connections. would consequently slow down the noradrenergic
ªREM-offº neuronal activity in the LC. 26 The slowing
of the noradrenergic LC neurons is likely to withdraw the
throughout the sleep±waking cycle and not only dur- inhibition (directly or through GABA) from the cholin-
ing REM sleep, 17 they may not be involved in the primary ergic ªREM-onº neurons, which would start ®ring,
regulation of REM sleep. Optimum doses of the antagonist resulting in a pre-REM period. The cholinergic input to
or the kinetics of dissociation of receptor-antagonist the noradrenergic ªREM-offº neurons as well as to the
complex also are likely to play a signi®cant role. GABAergic neurons (which may be non-cholinergic
Apart from the cholinergic inputs from laterodorsal/ ªREM-onº neurons) 69 in and around the LC would
pedunculopontine tegmental nuclei (LDT/PPT), 34 cholin- cause further release of NE and GABA, respectively, in
ergic 68 and GABAergic 9 inputs may be reaching LC from the LC. Both these inputs would hyperpolarize the LC
the paragigantocellularis (PGi) and prepositus hypoglossi noradrenergic ªREM-offº neurons and shut them off
(PrH) nuclei, respectively. Neurons in the PrH were found resulting in the initiation of REM sleep. 1,26 Once the
to be immunoreactive for c-Fos, a marker of neural activ- ªREM-offº neurons cease ®ring, there would be no NE
ity, following initiation of REM sleep. 84 Also presumptive at the projection sites; however, due to continuous acti-
cholinergic neurons in and around PGi have been shown to vation of the cholinergic ªREM-onº neurons, GABA
increase ®ring during REM sleep. 68 These suggest a possi- release will continue to maintain REM sleep. Thus,
bility of an increased activity of GABAergic neurons in the from this model it is likely that the sleep areas in the
PrH and cholinergic neurons in PGi during REM sleep, brain modulate the activities of the REM sleep related
which might also be responsible, in addition to the neurons, which plays a crucial role in REM sleep regula-
GABAergic neurons in the LC, for inhibiting the LC nor- tion. This is currently under investigation and prelimin-
adrenergic neurons for the initiation and maintenance of ary reports support the hypothesis and the model. 54
REM sleep. Also, the preoptic area GABAergic inputs to The trigger for activation of the cholinergic neurons is
LC 64,74 may inhibit LC neurons. This is supported by the likely to be the withdrawal of inhibition from the LC
fact that sleep 38,45,48,75 and REM-active neurons 45,62 are ªREM-offº neurons. This may be supported by the
present in the preoptic area. fact that LC neurons have projections to the LDT/PPT
cholinergic area, 34 and those neurons, cholinergic 69,81
as well as non-cholinergic, 69 are reported to be inhibited
Possible interactions between different groups of by NE either directly 81 or indirectly. 44 This may explain
neurons and their mechanisms of action the continuation of REM sleep after the lesion of
The neuronal connections in the LC may be noradrenergic-LC neurons 33 and after pharmacological
 Acetylcholine±GABA interaction in LC and REM sleep 483

depletion of NE. 33,42,47,56 Nevertheless, the possibility of initiation; 73 and picrotoxin administration in the LC
an inhibition of those neurons by GABA cannot be ruled reduced REM sleep by decreasing the mean duration of
out because GABA neurons are also present there. 23,35,36 REM sleep. 39 In addition, there are inputs to the LC
This model also ®ts in with previous reports that carba- ªREM-offº neurons from other brain areas as well,
chol injections in the dorsolateral pons and around the which may in¯uence REM sleep, and this needs further
LC increased REM sleep by increasing the frequency of investigation.
REM sleep initiation, 13,18,27,32,58 while clonidine (a / 2
noradrenergic receptor agonist) injections around the
LC decreased REM sleep affecting both the frequency AcknowledgementsÐWe thank Prof. J. Siegel, UCLA for sending
of REM sleep initiation and its mean duration; 79 mild us a batch of TH antibody. Financial support from Council of
continuous electrical stimulation of LC resulted in Scienti®c and Industrial Research, India, to B.N.M. is duly
reductions of REM sleep by decreasing its frequency of acknowledged.

REFERENCES

1. Aghajanian G. K., Cedarbaum J. M. and Wang R. Y. (1977) Evidence for norepinephrine-mediated collateral inhibition of locus coeruleus
neurons. Brain Res. 136, 570±577.
2. Akhondzadeh S. and Stone T. W. (1998) Potentiation of muscimol-induced long-term depression by benzodiazepines and prevention or reversal
by pregnenolone sulfate. Pharmac. Res. 38, 441±448.
3. Alam M. N. and Mallick B. N. (1990) Differential acute in¯uence of medial and lateral preoptic areas on sleep wakefulness in freely moving
animals. Brain Res. 525, 242±248.
4. Alam M. N., Kumari S. and Mallick B. N. (1993) Role of GABA in acetylcholine induced locus coeruleus mediated increase in REM sleep.
Sleep Res. 22, 541.
5. Alam M. N., McGinty D. and Szymusiak R. (1995) Neuronal discharge of preoptic/anterior hypothalamic thermosensitive neurons: relation to
non-rapid eye movement sleep. Am. J. Physiol. 269, R1240±R1249.
6. Albanese A. and Butcher L. L. (1980) Acetylcholinesterase and catecholamine distribution in the locus coeruleus of the rat. Brain Res. Bull. 5,
127±134.
7. Ali M., Jha S., Kaur S. and Mallick B. N. (1999) Role of GABA-A receptor in the preoptic area in the regulation of sleep-wakefulness and rapid
eye movement sleep. Neurosci. Res. 33, 245±250.
8. Aston-Jones G. and Bloom F. E. (1981) Activity of norepinephrine-containing locus coeruleus neurons in behaving rats anticipates ¯uctuations
in the sleep-waking cycle. J. Neurosci. 1, 876±886.
9. Aston-Jones G., Shipley M. T., Ennis M., Williams J. T. and Pieribone V. A. (1990) Restricted afferent control of locus coeruleus neurons
revealed by anatomical, physiological and pharmacological studies. In The Pharmacology of Noradrenaline in the Central Nervous System (eds
Marsden C. A. and Heal D. J.), pp. 187±247. Oxford University Press, Oxford.
10. Baghdoyan H. A., Rodrigo-Angulo M. L., McCarley R. W. and Hobson J. A. (1984) Site-speci®c enhancement and suppression of desynchro-
nized sleep signs following cholinergic stimulation of three brain stem regions. Brain Res. 306, 39±52.
11. Baghdoyan H. A., Rodrigo-Angulo M. L., McCarley R. W. and Hobson J. A. (1987) A neuroanatomical gradient in the pontine tegmentum for
the cholinoceptive induction of desynchronized sleep signs. Brain Res. 414, 245±261.
12. Baghdoyan H. A., Mallios V. J., Duckrow R. B. and Mash D. C. (1994) Localization of muscarinic receptor subtypes in brain stem areas
regulating sleep. NeuroReport 5, 1631±1634.
13. Bourgin P., Escourrou P., Gaultier C. and Adrien J. (1995) Induction of rapid eye movement sleep by carbachol infusion into the pontine
reticular formation in the rat. NeuroReport 6, 532±536.
14. Bourgin P., Ahnaou A., Laporte A. M., Hamon M. and Adrien J. (1999) Rapid eye movement sleep induction by vasoactive intestinal peptide
infused into the oral pontine tegmentum of the rat may involve muscarinic receptors. Neuroscience 89, 291±302.
15. Calvo J. M., Datta S., Quattrochi J. and Hobson J. A. (1992) Cholinergic microstimulation of the peribrachial nucleus in the catÐII. Delayed
and prolonged increases in REM sleep. Arch. ital. Biol. 130, 285±301.
16. Cespuglio R., Gomez H., Faradji H. and Jouvet M. (1982) Alterations in the sleep-waking cycle induced by cooling of the locus coeruleus area.
Electroenceph. clin. Neurophysiol. 54, 570±578.
17. Darracq L., Gervasoni D., Souliere F., Lin J. S., Fort P., Chouvet G. and Luppi P. H. (1996) Effect of strychnine on rat locus coeruleus neurons
during sleep and wakefulness. NeuroReport 8, 351±355.
18. Deurveilher S., Hars B. and Hennevin E. (1997) Pontine microinjection of carbachol does not reliably enhance paradoxical sleep in rats. Sleep
20, 593±607.
19. Dube G. R. and Marshall K. C. (1997) Modulation of excitatory synaptic transmission in locus coeruleus by multiple presynaptic metabotrobic
glutamate receptors. Neuroscience 80, 511±521.
20. Egan T. M. and North R. A. (1985) Acetylcholine acts on m2-muscarinic receptors to excite rat locus coeruleus neurons. Eur. J. Pharmac. 85,
733±735.
21. Egan T. M. and North R. A. (1986) Actions of acetylcholine and nicotine on rat locus coeruleus neurons in vitro. Neuroscience 19, 565±571.
22. El Mansari M., Sakai K. and Jouvet M. (1989) Unitary characteristics of presumptive cholinergic tegmental neurons during the sleep-wake
cycle in freely moving cats. Expl Brain Res. 76, 519±529.
23. Ford B., Holmes C. J., Mainville L. and Jones B. E. (1995) GABA-ergic neurons in the rat pontomesencephalic tegmentum: codistribution with
cholinergic and other tegmental neurons projecting to the posterior lateral hypothalamus. J. comp. Neurol. 363, 177±196.
24. Fort P., Luppi P. and Jouvet M. (1993) Glycine-immunoreactive neurons in the cat brain stem reticular formation. NeuroReport 4, 1123±1126.
25. Garzon M., De Andres L. and Reinoso-Suarez E. (1998) Sleep patterns after carbachol delivery in the ventral oral pontine tegmentum of the cat.
Neuroscience 83, 1137±1144.
26. Gervasoni D., Darracq L., Fort P., SoulieÁre F., Chouvet G. and Luppi P. H. (1998) Electrophysiological evidence that noradrenergic neurons of
the rat locus coeruleus are tonically inhibited by GABA during sleep. Eur. J. Neurosci. 10, 964±970.
27. Gnadt J. M. and Pegram G. V. (1986) Cholinergic brain stem mechanisms of REM sleep in the rat. Brain Res. 384, 29±41.
28. Greene R. W. and Carpenter D. O. (1985) Actions of neurotransmitters on pontine medial reticular formation neurons of the cat. J. Neuro-
physiol. 54, 520±531.
29. Hobson J. A., McCarley R. W. and Wyzinski P. W. (1975) Sleep cycle oscillation: reciprocal discharge by two brain stem neuronal groups.
Science 189, 55±58.
484 B. N. Mallick et al.

30. Hsu S. M., Raine L. and Fanger H. (1981) Use of avidin±biotin±peroxidase complex (ABC) in immunoperoxidase techniques: a comparison
between ABC and unlabelled antibody (PAP) procedures. J. Histochem. Cytochem. 29, 577±580.
31. Iijima K. and Ohtomo K. (1988) Immunocytochemical study using GABA antiserum for demonstration of inhibitory neurons in the rat locus
coeruleus. Am. J. Anat. 181, 43±52.
32. Imeri L., Bianchi S., Angeli P. and Mancia M. (1994) Selective blockade of different brain stem muscarinic receptor subtypes: effects on the
sleep±wake cycle. Brain Res. 63, 68±72.
33. Jones B. E., Harper S. T. and Halaris A. E. (1977) Effects of locus coeruleus lesions upon cerebral monoamine content, sleep wakefulness states
and the response to amphetamine in the cat. Brain Res. 124, 473±496.
34. Jones B. E. and Yang T. (1985) The efferent projections from the reticular formation and the locus coeruleus studied by anterograde and
retrograde axonal transport in the rat. J. comp. Neurol. 212, 56±92.
35. Jones B. E. (1990) Immunohistochemical study of choline acetyltransferase-immunoreactive processes and cells innervating the ponto-
medullary reticular formation in the rat. J. comp. Neurol. 295, 485±514.
36. Jones B. E. (1991) Paradoxical sleep and its chemical/structural substrates in the brain. Neuroscience 40, 637±656.
37. JuhaÁsz G., Emri Z., Kekesi K. and Pungor K. (1989) Local perfusion of the thalamus with GABA increases sleep and induces long-lasting
inhibition of somatosensory event-related potentials in cat. Neurosci. Lett. 103, 229±233.
38. Kaitin K. I. (1984) Preoptic unit activity during sleep and wakefulness in the cat. Expl Neurol. 83, 347±357.
39. Kaur S., Saxena R. N. and Mallick B. N. (1997) GABA in locus coeruleus regulates spontaneous rapid eye movement sleep by acting on GABA-
A receptors in freely moving rats. Neurosci. Lett. 223, 105±108.
40. Kaur S., Saxena R. N. and Mallick B. N. (1997) Role of GABA in locus coeruleus mediated cholinergic induction of rapid eye movement sleep.
Second Cong. Asian Sleep Res. Soc., p. 63. (Jerusalem, 1997), Israel.
41. Kayama Y., Ohta M. and Jodo E. (1992) Firing of `possibly' cholinergic neurons in the rat laterodorsal tegmental nucleus during sleep and
wakefulness. Brain Res. 569, 210±220.
42. King C. D. and Jewett R. E. (1971) The effects of alpha-methyltyrosine on sleep and brain norepinephrine in cats. J. Pharmac. exp. Ther. 177,
188±195.
43. Kodama T., Takahashi Y. and Honda Y. (1990) Enhancement of acetylcholine release during paradoxical sleep in the dorsal tegmental ®eld of
the cat brain stem. Neurosci. Lett. 114, 277±282.
44. Kohlmeier K. A. and Reiner P. B. (1999) Noradrenaline excites non-cholinergic laterodorsal tegmental neurons via two distinct mechanisms.
Neuroscience 93, 619±630.
45. Koyama Y. and Hayaishi O. (1994) Firing of neurons in the preoptic/anterior hypothlamic areas in rat: its possible involvement in slow wave
sleep and paradoxical sleep. Neurosci. Res. 19, 31±38.
46. Luque J. M., Malherbe P. and Richards J. G. (1994) Localization of GABA-A receptor subunit mRNAs in the rat locus coeruleus. Molec. Brain
Res. 24, 219±226.
47. Lidbrink P., Corrodi H., Fuxe K. and Olson L. (1973) On the role of central monoamine neurons in sleep mechanisms. In Pharmacology and the
Future of Man, pp. 90±102. Proc. 5th Int. Congr. Pharmacology (San Francisco, 1972). Vol. 4. Basel, Karger.
48. Mallick B. N., Chhina G. S., Sundaram K. R., Singh B. and Mohan Kumar V. (1983) Activity of preoptic neurons during synchronization and
desynchronization. Expl Neurol. 81, 586±597.
49. Mallick B. N., Siegel J. M. and Fahringer H. (1989) Changes in pontine unit activity during REM sleep deprivation. Brain Res. 515,
94±98.
50. Mallick B. N. and Alam M. N. (1992) Different types of norepinephrinergic receptors are involved in preoptic area mediated independent
modulation of sleep±wakefulness and body temperature. Brain Res. 591, 8±19.
51. Mallick B. N., Thakkar M., Shiromani P. J. and McCarley R. W. (1994) REM sleep-selective unit activity in the cholinergic pedunculopontine
tegmental nucleus of the freely moving cats. Sleep Res. 23, 24.
52. Mallick B. N. and Joseph M. M. (1997) Role of cholinergic inputs to the medial preoptic area in regulation of sleep-wakefulness and body
temperature in freely moving rats. Brain Res. 750, 311±317.
53. Mallick B. N., Kaur S., Jha S. K. and Siegel J. M. (1999) Possible role of GABA in the regulation of REM sleep with special reference to REM-
OFF neurons. In The Rapid Eye Movement Sleep (eds Mallick B. N. and Inoue S.), pp. 153±166. Marcel and Dekker, New York.
54. Mallick B. N. (1999) Cellular mechanism of rapid eye movement sleep generation: role of sleep and waking areas in the brain. 11th Annual
Conference of the Physiological Society of India. (New Delhi) India.
55. Maloney K. J., Mainville L. and Jones B. E. (1999) Differential c-fos expression in cholinergic, monoaminergic and GABA-ergic cell groups of
the pontomesencephalic tegmentum after paradoxical sleep deprivation and recovery. J. Neurosci. 19, 3057±3072.
56. Marantz R. and Rechtschaffen A. (1967) Effect of alpha-methyltyrosine on sleep in the rat. Percept. Mot. Skills 25, 805±808.
57. Mash D. C. and Potter L. T. (1986) Autoradiographic localization of M1 and M2 muscarinic receptors in the rat brain. Neuroscience 19,
551±564.
58. Mastrangelo D., De Saint-Hilaire-Ka® Z. and Gaillard J. M. (1994) Effects of clonidine and alpha-methyl-p-tyrosine on the carbachol
stimulation of paradoxical sleep. Pharmac. Biochem. Behav. 48, 93±100.
59. Nishikawa T. and Scatton B. (1985) Inhibitory in¯uence of GABA on central serotonergic transmission raphe nuclei as the neuroanatomical site
of the GABA-ergic inhibition of cerebral serotonergic neurons. Brain Res. 331, 91±103.
60. Nitz D. and Siegel J. M. (1997) GABA release in the locus coeruleus as a function of the sleep/wake state. Neuroscience 78, 795±801.
61. Olpe H. R., Steinmann M. W., Hall R. G., Brugger F. and Pozza M. F. (1988) GABA-A and GABA-B receptors in locus coeruleus: effects of
blockers. Eur. J. Pharmac. 149, 183±185.
62. Osaka T. and Matsumura H. (1994) Noradrenergic inputs to sleep-related neurons in the preoptic area from the locus coeruleus and the
ventrolateral medulla in the rat. Neurosci. Res. 19, 39±50.
63. Paxinos G. and Watson C. (1982) The Rat Brain in Stereotaxic Coordinates. Academic, Australia.
64. Peyron C., Luppi P. H., Rampon C. and Jouvet M. (1995) Location of the GABA-ergic neurons projecting to the dorsal raphe nucleus and the
locus coeruleus of the rat. Soc. Neurosci. Abstr. 21, 373.
65. Rada P. V., Mark G. P. and Hoebel B. G. (1993) In vivo modulation of acetylcholine in the nucleus accumbens of freely moving rats: II.
Inhibition by g-aminobutyric acid. Brain Res. 619, 105±110.
66. Sakai K. (1985) Anatomical and physiological basis of paradoxical sleep. In Brain Mechanism of Sleep (eds McGinty D. J., Drucker-Colin R.,
Morrison A. and Parmeggiani P. L.), pp. 111±137. Raven, New York.
67. Sakai K. (1988) Executive mechanisms of paradoxical sleep. Arch. ital. Biol. 126, 239±257.
68. Sakai K. (1991) Physiological properties and afferent connections of the locus coeruleus and adjacent tegmental neurons involved in the
generation of paradoxical sleep in the cat. Prog. Brain Res. 88, 31±45.
69. Sakai K. and Koyama Y. (1996) Are there cholinergic and non-cholinergic paradoxical sleep-on neurones in the pons? NeuroReport 7,
2449±2453.
 Acetylcholine±GABA interaction in LC and REM sleep 485

70. Shiromani P. J., Overstreet D. H., Levy D., Goodrich C. A., Campbell S. S. and Gillin J. C. (1988) Increased REM sleep in rats selectively bred
for cholinergic hyperactivity. Neuropsychopharmacology 1, 127±133.
71. Shuman S. L., Capece M. L., Baghdoyan H. A. and Lydic R. (1985) Pertussis toxin-sensitive G proteins mediate carbachol induced REM sleep
and respiratory depression. Am. J. Physiol. 269, R308±317.
72. Siegel J. M. (1989) Brain stem mechanisms generating REM sleep. In Principles and Practice of Sleep Medicine (eds Kryger M. H., Roth T. and
Dement W. C.), pp. 104±120. W.B. Saunders, Philadelphia.
73. Singh S. and Mallick B. N. (1996) Mild electrical stimulation of pontine tegmentum around locus coeruleus reduces rapid eye movement sleep.
Neurosci. Res. 24, 227±235.
74. Singewald N. and Philippu A. (1998) Release of neurotransmitters in the locus coeruleus. Prog. Neurobiol. 56, 237±267.
75. Szymusiak R. (1995) Magnocellular nuclei of the basal forebrain: substrates of sleep and arousal regulation. Sleep 18, 478±500.
76. Thakkar M, Chiara P and McCarley R. W. (1996) Chronic low-amplitude stimulation of the laterodorsal tegmental nucleus of freely moving cats
increases REM sleep. Brain Res. 723, 223±227.
77. Thankachan S., Islam F. and Mallick B. N. (1997) Behavior of brain stem neurons during spontaneous and induced EEG desynchronization
during wakefulness and rapid eye movement sleep in free moving cats. Sleep Res. 26, 54.
78. Timo-Iaria C., Negrao N., Schmidak W. R., Hoshino K., Lobato D., Menezes D. and Rocha T. L. (1970) Phases and states of sleep in the rat.
Physiol. Behav. 5, 1057±1062.
79. Tononi G., Pompeiano M. and Cirelli C. (1991) Effects of local pontine injection of noradrenergic agents on desynchronized sleep of the cat.
Prog. Brain Res. 88, 545±553.
80. Watabe K. (1980) Mode of neuronal interaction in rat locus coeruleus. Arch. ital. Biol. 118, 303±329.
81. Williams J. A. and Reiner P. B. (1993) Noradrenaline hyperpolarizes identi®ed rat mesopontine cholinergic neurons in vitro. J. Neurosci. 13,
3878±3883.
82. Vanni-Mercier G., Sakai J. S. and Jouvet M. (1989) Mapping of cholinoceptive brain stem structures resposible for the generation of paradoxical
sleep in the cat. Arch. ital. Biol. 127, 133±164.
83. Yamamoto K., Mamelak A. N., Quattrochi J. J. and Hobson J. A. (1990) A cholinoceptive desynchronized sleep induction zone in the
anterodorsal pontine tegmentum: locus of the sensitive region. Neuroscience 39, 279±293.
84. Yamuy J., Mancillas J. R., Morales F. R. and Chase M. H. (1993) C-fos expression in the pons and medulla of the cat during carbachol-induced
active sleep. J. Neurol. 136, 2703±2718.

(Accepted 7 February 2001)