Brain Research 922 (2001) 299–304www.elsevier.com/ locate /bres

Short communication

Seizure susceptibility decreases with enhancement of rapid eyemovement sleep

*Pratap Kumar, T.R. RajuDepartment of Neurophysiology, National Institute of Mental Health & Neuro Sciences (NIMHANS), Bangalore 560029, India

Accepted 25 September 2001

Abstract

The study examined the effect of enhanced rapid eye movement (REM) sleep duration on the seizure threshold determined by electricalstimulation of the amygdala in rats. The duration of REM sleep was specifically increased by the microinjection of a cholinergic agonist,carbachol, into the pontine reticular formation. This was accompanied by a significant increase in the threshold current required to elicitan afterdischarge in the amygdala. The results suggest that an increase in REM sleep decreases the likelihood of cortical seizure activity,an effect that is manifest even in other stages of the sleep–wakefulness cycle and not only in the REM state, per se. 2001 Publishedby Elsevier Science B.V.

Theme: Disorders of the nervous system

Topic: Epilepsy: basic mechanisms

Keywords: Seizure; Afterdischarge; REM sleep; Carbachol

The antiepileptic nature of the rapid eye movement over long periods with a decrease in REM duration(REM) sleep state has been demonstrated in a wide array persisting at least up to 1 month after the last epilepticof experiments linking sleep and epilepsy. Seizures occur discharge [15,28]. REM sleep thus influences corticalless frequently during REM sleep, while those that do excitability over long periods and could therefore modulateoccur generalize in fewer cases [1,5,20,34]. Greater cur- seizure activity in general, and not necessarily during therents are required to elicit seizures in animal models during REM state only. The therapeutic use of the antiepilepticthis state [7,9,38]. REM sleep deprivation increases seizure property of REM sleep hinges on the persistence of thisactivity [8,10,16,32,35] and reciprocally, seizures specifi- effect into the other arousal states that are more prone tocally compromise this period of the sleep–wakefulness seizure activity.cycle [15,17,28,29,31,36,38]. Various components of REM Our aim was to explore the effects of a specific increasesleep including muscle atonia, EEG desynchronization [33] in REM sleep duration on seizure activity. This studyand the hippocampal theta rhythm [25] have also been focuses on one seizure parameter, the threshold for after-shown to suppress seizure activity, further elaborating both discharge, in the rat amygdala-kindling model. The thres-the antiepileptic nature of this sleep stage and the intimate hold for afterdischarge can be used as an index ofrelationship between sleep and cortical excitability. susceptibility to seizure activity as it reflects the minimal

Previous experiments have mostly restricted the study of current required to elicit a local spike wave discharge inseizure suppression by REM sleep to the temporal confines the animal. Evoking an afterdischarge is critical forof the state. Deprivation of REM sleep causes an increase kindling as only stimuli at or above this threshold, whenin seizure activity which is evident during wakefulness and repeated, result in the gradual establishment of a focus forslow wave sleep; conversely, seizures affect REM sleep generalized seizures [23]. Any intervention that raises the

threshold for afterdischarge could therefore inhibit thedevelopment of a seizure focus.*Corresponding author. Tel.: 191-80-6995-168; fax: 191-80-6564-

Adult male Wistar rats weighing between 200 and 250 g830.E-mail address: trraju@nimhans.kar.nic.in (T.R. Raju). were used. The animals were obtained from the Central

0006-8993/01/$ – see front matter 2001 Published by Elsevier Science B.V.PI I : S0006-8993( 01 )03174-2

300 P. Kumar, T.R. Raju / Brain Research 922 (2001) 299 –304

Animal Research Facility at the National Institute of were applied when the rats were in the active wakingMental Health and Neuro Sciences (NIMHANS), Banga- stage. After the recordings, animals were perfused trans-lore, India, and housed under a 12-h light–dark cycle with cardially with 10% formaldehyde under ether anesthesiafood and water available ad libitum. All experimental and the brains removed. Electrode and cannula placementsprocedures were carried out in accordance to the National were confirmed with cresyl violet stained sections of 30Institute of Health guidelines. The rats were anaesthetized mm thicknesses.with sodium pentabarbitone (40 mg/kg, i.p.) and im- Differences in sleep parameters before and after theplanted stereotaxically with electrodes for polysomnog- carbachol injection were compared using the paired ‘t’ testraphy (two stainless steel screw EEG (electroencephalog- while changes in threshold were analyzed using theram) electrodes over the frontal cortices, grounding screw Wilcoxan matched pairs, signed rank test for non-Gaussianover the parietal cortex, nichrome hooks into the external data. Multiple regression analysis was used to examine thecanthi for EOG (electrooculogram) activity and into the nature of the relationship between the changes in REMnuchal muscles for EMG (electromyogram) activity) and sleep duration and seizure thresholds.amygdala kindling (bipolar nichrome electrodes into the The time spent in REM sleep (REM duration) as well asright anterior amygdala; AP: 20.8 to 21.3, ML: 3.5, DV: the number of entries into REM sleep (REM frequency)8.6 mm). A 22G guide cannula was placed 1.5 mm above showed highly significant increases (P,0.0001) after thethe left pontine reticular formation (AP: 28.3 to 28.8, administration of carbachol (Figs. 1 and 2). AnimalsML: 1.5, DV: 7.0 mm). All coordinates were referenced to tended to spend more time asleep (increases in total sleepthe bregma [26]. The electrodes and cannula were perma- time, slow wave sleep and S-1 phase of SWS) and less innently fixed onto the skull using dental acrylic. The rats waking (decreased total waking and W-1 durations),were allowed to recover from the surgery over a period of though these did not reach levels of significance (Fig. 3).72 h before any further intervention was undertaken. This reflects a specific increase in the duration of the REM

The EEG, EMG, EOG and amygdala kindling electrodes component of sleep after carbachol injection.were connected to an eight-channel Grass polygraph The threshold for afterdischarge increased significantly(Model 78D) with a flexible, shielded cable. All animals (P,0.05) when tested after the sleep period influenced bywere subjected to an initial polysomnographic recording carbachol (Fig. 4). In the nine animals from which databetween 12:00 and 18:00 h. The traces were recorded on were analyzed, threshold increased in six, remained un-paper at a speed of 30 mm/s, divided into 1-min epochs changed in two and was seen to decrease in one. Theand scored manually into one of five stages of the sleep multiple regression analysis however did not reveal anywakefulness cycle [28]. statistically significant correlation between the two sleep

The threshold for afterdischarge was determined imme- parameters which changed significantly after carbacholdiately after the baseline sleep recording. Depth EEG was injection, viz. the REM duration and frequency, and the

2recorded from the amygdala for 5 min as a baseline raise in the threshold to afterdischarge (R 50.01 andreference and then again after each stimulation the 20.14, respectively).amygdala. Stimuli (1 ms square wave pulses at 75 Hz for 1 The pattern of sleep changes seen after carbachols) of increasing intensity (beginning with 10 mA, with 5 injection was similar to that observed in previous studiesmA increments) were applied with a rest pause of at least 1 of REM induction in rats [4,21]. These authors alsomin between two stimuli until a clear afterdischarge (a reported an increase in SWS duration and a decrease inspike-wave discharge with of at least three times the wakefulness apart from the enhancement of REM sleep.amplitude of the baseline depth EEG and lasting for 8 s or The increased REM duration observed in these studies andmore) was obtained. The lowest intensity of current able to the present one was a result of greater number of entriesevoke a clear afterdischarge was considered threshold. into REM sleep rather than due to longer REM episodes.This value was reconfirmed after 12 and 24 h and rats with The increases in REM sleep (over a few hours), after athresholds between 20 and 150 mA were selected for the single injection of carbachol into the pontine reticularstudy owing to instrumental constraints. formation, outlast the duration of the receptor-mediated

Once the threshold was confirmed, the rats were subject- actions of carbachol. It is likely that these increases areed to another sleep recording, this time after an injection of mediated via long-lasting intra- and inter-cellular events inthe muscarinic agent carbachol. A 0.5-ml volume of a 1.1 the PRF neurons [6,18,21].mM solution in saline [4,21] was injected into the left We observed a significant raise in the threshold forpontine reticular formation (PRF) 15 min before the afterdischarge after the increase in REM sleep duration,recording via a 26G needle inserted into the cannula placed though no correlation between the two could be deter-in-situ, using a 5 ml syringe (Hamilton). Injections were mined statistically. The absence of such a correlation,made over a 2-min period followed by a 3-min wait before however, does not preclude the existence of a cause–effectremoving the needle and replacing the stylettes. The relationship. The coexistence of an increase in the durationthreshold for afterdischarge was determined again immedi- of REM sleep and a raised threshold for afterdischargeately after this sleep period. All the stimuli to the amygdala might reflect a more complex, and not necessarily linear,

P. Kumar, T.R. Raju / Brain Research 922 (2001) 299 –304 301

Fig. 1. Change in REM sleep duration after the injection of carbachol. REM duration increased from an average baseline value of 10.7861.99 min to19.3362.45 min after the injection of carbachol (****P,0.0001).

relationship between REM sleep and the occurrence of regions thereby affecting the initiation of this sleep stateseizures. REM sleep has been studied here as an epi- [28]. Reciprocal projections from the brainstem to limbic,phenomenon of several underlying components like EEG subcortical and cortical areas may provide the basis fordesynchronization and hippocampal theta activity, both of both the effect of seizures on sleep patterns and the abilitywhich have been individually shown to suppress seizure of sleep related areas in the brainstem to modulate seizureactivity [25,33]. We propose that specific changes in these activity [14,24,37]. The selective decrease in the durationcomponents, especially EEG desynchronization and hip- of REM sleep observed in the animal epilepsy modelspocampal theta activity, and their quantification, could could therefore be the direct effect of seizure activity andreflect a more correlated raise in seizure thresholds. the compromised REM duration would in turn lead to a

Seizures directly affect REM sleep generating centers, state of increased susceptibility to seizure activity. Thiswith foci like the amygdala known to glutamatergically cycle could be arrested by the restoration of the compro-innervate brainstem reticular nuclei [22,37]. A single mised ‘anti-epileptic’ sleep stage.seizure has been proposed to cause, by excitation resulting Similar links may explain the effectiveness of vagalfrom the release of moderate amounts of glutamate, an nerve stimulation (VNS) [27,39,40] and more recently,increase in the duration of REM sleep. Repeated dis- trigeminal nerve stimulation [13] in the treatment of partialcharges result in excitotoxic death of neurons in the target seizures, both in humans and experimental animals. One

Fig. 2. Change in REM frequency after the injection of carbachol. The number of entries into REM sleep in the 6-h period increased from a baseline valueof 6.89G1.27 entries to 12.33G1.66 after carbachol injection (****P,0.0001).

302 P. Kumar, T.R. Raju / Brain Research 922 (2001) 299 –304

Fig. 3. Percentage change in sleep stages after carbachol injection. The change in percentage of the 6-h sleep period occupied by each sleep stage wascompared. The duration of REM sleep increased significantly (****P,0.0001). The time spent in slow wave sleep, especially S-1, tended to increase,while the time spent awake decreased, though not reaching levels of significance. TST: Total sleep time; TWT: total wakefulness time; REM: rapid eyemovement sleep; S1: light slow wave sleep; S2: deep slow wave sleep; W1: active wakefulness; W2: quiet wakefulness.

hypothesis to explain the antiepileptic effect of cranial both cortical excitability and seizure susceptibilitynerve stimulation is that the afferents alter the activity of [13,24,30]. Vagal signaling is cholinergic in nature andbrain stem reticular centers, which are known to affect stimulation of the reticular formation and solitary tract

Fig. 4. Change in current required to elicit an afterdischarge. In the nine animals studied, the threshold current required to elicit an afterdischarge increasedin six, decreased in one and remained unchanged in two. The increase was found to be significant (P,0.05) when compared by the Wilcoxan matchedpairs, signed rank test for non-parametric data.

P. Kumar, T.R. Raju / Brain Research 922 (2001) 299 –304 303

eye movement sleep by carbachol infusion into the pontine reticularnucleus, via the vagus, affects large areas of the cortex viaformation in the rat, Neuroreport 6 (1995) 532–536.projections to the hippocampus, amygdala, hypothalamus

[5] J.M. Calvo, R. Alverado, R. Briones, C. Paz, A. Fernandez-Guar-and the thalamus [30]. REM sleep is the result of activa- diola, Amygdaloid kindling during rapid eye movement (REM)tion of similar cholinergic centers in the brainstem as those sleep in cats, Neurosci. Lett. 29 (1982) 255–259.

[6] M.L. Capece, H.A. Baghdoyan, R. Lydic, Carbachol stimulatesmediating the action of vagal nerve stimulation [19]. The35[ S]guanylyl 59-(g-thio)-triphosphate binding in rapid eye move-effect on cortical excitability brought about by VNS could

ment sleep related brainstem nuclei of rat, J. Neurosci. 18 (1998)thus be mirrored by REM sleep induction. The enhance- 3779–3785.ment of REM sleep by non-invasive methods like the oral [7] H.P. Cohen, J. Thomas, W.C. Dement, Sleep stages, REM depriva-

tion and electroconvulsive threshold in the cats, Brain Res. 19administration of muscarinic drugs [2,3] could provide a(1970) 317–321.viable alternative to VNS in generating a seizure resistant

[8] H.P. Cohen, W.C. Dement, Electrically induced convulsions in REMstate of the cortex.deprived mice: prolongation of the tonic phase, Psychophysiology 4

Carbachol has also been studied as an epileptogenic (1968) 381.drug, capable of eliciting seizures when injected into the [9] H.P. Cohen, W.C. Dement, Sleep: changes in threshold to electro-

convulsive shock in rats after deprivation of a ‘paradoxical’ phase,pontine and mesencephalic reticular formation and also theScience 150 (1965) 1318–1319.hippocampus [12]. This effect requires the injection of

[10] W.H.I.M. Drinkenburg, A.M.L. Coenen, J.M.H. Vossen, E.L.J.M.larger doses of carbachol than those required to initiate van Luijtelaar, Sleep deprivation and spike-wave discharges inREM sleep, and is potentiated by the addition of glutamate epileptic rats, Sleep 18 (4) (1995) 252–256.

[11] Z. Elazar, A. Berchanski, Excitatory amino acids modulate epi-[11]. Such an effect is proposed to be the result of localleptogenesis in the brain stem, Neuroreport 11 (2000) 1777–1780.synchronization of reticulo-reticular connections and fur-

[12] Z. Elazar, Z. Feldman, Brainstem experimental seizures produced byther spread of this synchronization manifests as general- microinjections of carbachol, Epilepsia 28 (1987) 463–470.ized seizures. This probably reflects the dual effects of [13] E.E. Fanselow, A.P. Reid, M.A.L. Nicolelis, Reduction of

pentylenetetrazole-induced seizure activity in awake rats by seizurecarbachol in being able to elicit, in a dose-dependenttriggered trigeminal nerve stimulation, J. Neurosci. 20 (2000) 8160–manner, both REM sleep (of which EEG desynchroniza-8168.tion is an integral component) and EEG synchronization

[14] I. Ferencz, M. Kokaia, M. Keep, E. Elmer, M. Metsis, Z. Kokaia, O.(manifested as seizures); a result of varying levels of Lindvall, Effects of cholinergic denervation on seizure developmentactivation of neurons in the reticular formation. The and neurotrophin messenger RNA regulation in rapid hippocampal

kindling, Neuroscience 80 (1997) 389–399.connections mediating cortical responses to REM sleep[15] G.L. Gigli, J. Gotman, Effects of seizures, kindling, and carbamaz-initiating events in the PRF could also underlie the spread

epine on sleep organization in cats, Epilepsia 33 (1) (1992) 14–22.of seizure activity initiated in this region, highlighting the [16] S. Grahnstedt, Sleep deprivation and kindled seizures, Exp. Neurol.influence of the reticular formation on cortical excitability. 92 (1986) 248–260.

Our results indicate that an increase in REM sleep [17] T. Hiyoshi, N. Mori, J.A. Wada, Feline amygdaloid kindling andsleep, Electroencephalogr. Clin. Neurophysiol. 73 (1989) 254–259.dampens cortical excitability, a generalized effect that is

[18] L. Imeri, S. Bianchi, P. Angeli, M. Mancia, Selective blockade ofmanifest even in other stages of the sleep–wakefulnessdifferent brain stem muscarinic receptor subtypes: effects on the

cycle. This has been demonstrated as a raise in the sleep–wake cycle, Brain Res. 636 (1994) 68–72.threshold for afterdischarge. Future work in this direction [19] B.E. Jones, Paradoxical sleep and it’s chemical / structural substratesshould assess the role of individual REM sleep com- in the brain, Neuroscience 40 (1991) 637–656.

[20] B.A. Malow, X. Lin, R. Kushwaha, M.S. Aldrich, Interictal spikingponents in seizure suppression over long periods. REMincreases with sleep depth in temporal lobe epilepsy, Epilepsia 39sleep or its components could also be used as a tool to(1998) 1309–1316.

assess the state of cortical excitability, indicating the [21] G.A. Marks, C.G. Birabil, Enhancement of rapid eye movementprobability of seizure activity and therefore the necessity sleep in the rat by cholinergic and adenosinergic agonists infusedfor antiepileptic agents. The possibility of a non-invasive, into the pontine reticular formation, Neuroscience 86 (1998) 29–37.

[22] P.L. McGeer, E.G. McGeer, T. Nagai, GABAergic and cholinergicphysiological method of predicting and preventing seizuresindices in various regions of the rat brain after intracerebralis exciting and needs to be thoroughly investigated.injections of folic acid, Brain Res. 260 (1983) 107–116.

[23] J.O. McNamara, Kindling: an animal model of complex partialepilepsy, Ann. Neurol. 16 (1984) S72–S76.

[24] J.W. Miller, B.C. Gray, G.M. Turner, M.E. Bardgett, An ascendingReferences seizure-controlling pathway in the medial brainstem and thalamus,Exp. Neurol. 121 (1993) 106–112.

[1] C.W. Bazil, T.S. Walczak, Effects of sleep and sleep stage on [25] J.W. Miller, G.M. Turner, B.C. Gray, Anticonvulsant effects of theepileptic and non-epileptic seizures, Epilepsia 38 (1997) 56–62. experimental induction of hippocampal theta activity, Epilepsy Res.

[2] M. Berger, D. Riemann, J.C. Krieg, Cholinergic drugs as diagnostic 18 (1994) 195–204.and therapeutic tools in affective disorders, Acta Psychiatr. Scand. [26] G. Paxinos, C. Watson, The Rat Brain in Stereotaxic Coordinates,83 (1991) S52–S60. Academic Press, New York, 1982.

[3] A. Berkowitz, L. Sutton, D.S. Janowsky, J.C. Gillin, Pilocarpine, an [27] J.K. Penry, J.C. Dean, Prevention of intractable partial seizures byorally active muscarinic cholinergic agonist, induces REM sleep and intermittent vagal stimulation in humans: preliminary results, Epi-reduces Delta sleep in normal volunteers, Psychiatry Res. 33 (1990) lepsia 31 (1990) S40–S43.113–119. [28] Y.H. Raol, B.L. Meti, Sleep–wakefulness alterations in amygdala-

[4] P. Bourgin, P. Escourrou, C. Gaultier, J. Adrien, Induction of rapid kindled rats, Epilepsia 39 (1998) 1133–1137.

304 P. Kumar, T.R. Raju / Brain Research 922 (2001) 299 –304

[29] G. Rondouin, M. Baldy-Moulinier, P. Passouant, The influence of [35] M.N. Shouse, M.B. Sterman, P.J. Hauri, O. Belsito, Sleep disruptionhippocampal kindling on sleep organization in cats. Effects of with basal forebrain lesions decreases latency to amygdala kindlingalpha-methylparatyrosine, Brain Res. 181 (1980) 413–424. in cats, Electroencephalogr. Clin. Neurophysiol. 58 (1984) 369–377.

[30] P. Rutecki, Anatomical, physiological, and theoretical basis for the [36] W.S. Stone, P.E. Gold, Amygdala kindling effects on sleep andantiepileptic effect of vagus nerve stimulation, Epilepsia 31 (1990) memory in rats, Brain Res. 449 (1988) 135–140.S1–S6. [37] Y. Takeuchi, J.H. McLean, D.A. Hopkins, Reciprocal connections

[31] M.N. Shouse, M.B. Sterman, Sleep and kindling: II. Effects of between the amygdala and parabrachial nuclei: ultrastructure de-generalized seizure induction, Exp. Neurol. 71 (1981) 563–580. monstration by degeneration and axonal transport of horseradish

[32] M.N. Shouse, M.B. Sterman, Acute sleep deprivation reduced peroxidase in the cat, Brain Res. 239 (1982) 583–588.amygdala-kindled seizure thresholds in cats, Exp. Neurol. 78 (1982) [38] T. Tanaka, R. Naquet, Kindling effect and sleep organization in cats,716–727. Electroencephalogr. Clin. Neurophysiol. 39 (1975) 449–454.

[33] M.N. Shouse, J.M. Siegel, M.F. Wu, R. Szymusiak, A.R. Morrison, [39] D.M. Woodbury, J.W. Woodbury, Effects of vagal stimulation onMechanisms of seizure suppression during rapid-eye-movement experimentally induced seizures in rats, Epilepsia 31 (1990) S7–(REM) sleep in cats, Brain Res. 505 (1989) 271–282. S19.

[34] M.N. Shouse, J.V. Langer, P.R. Dittes, Spontaneous sleep epilepsy in [40] J. Zabara, Inhibition of experimental seizures in canines by repeti-amygdala-kindled kittens: a preliminary report, Brain Res. 535 tive vagal stimulation, Epilepsia 33 (1992) 1005–1012.(1990) 163–168.