- Email: [email protected]
Amygdala kindling effects on sleep and memory in rats
Brain Research, 449 (1988) 135-140 Elsevier
135
BRE 13530
Amygdala kindling effects on sleep and memory in rats William S. Stone and Paul E. Gold Department of Psychology, University of Virginia, Gilmer Hall, Charlottesville, VA 22903-2477(U.S.A_) (Accepted 3 November 1987) Key words: Kindling, Sleep; Memory; Aging; Rat; Seizure
Sleep d~sturbances accompany the development of amygdaloLd-kindledseLzuresin cats. Some of these sleep defiots resemble those seen m aged rats; these latter changes m sleep patterns are correlated with memory impairments in the aged animals. In the present study, we examined the hypothesis that sleep deficits after kindling may be related to memory tmpaitments Rats were kindled for 4 weeks (2-2.5 weeks after stage 5 smzures) and were then allowed a one week recovery period. Sleep patterns were assessed throughout the kindling and recovery periods The ammals were then trained on an inhibitory avoidance apparatus and tested for retention 24 h later. Only transient sleep changes occurred during the development of kindling ( to stage 5 seizures). However, continued kindling resulted in significant reductions m several sleep measures which remained depressed for at least one week after the termination of the kindling trials. As a group, kindled rats were impaired in retention of the inhibitory avoidance learned response. In kindled animals, retention performance was slgmflcantly correlated with total paradoxical sleep, the ratio of paradoxical/total sleep, and paradoxical sleep bout duration. These correlations are consistent with the view that deficits m paradoxical sleep may be related to deficits in some forms of memory. INTRODUCTION Daily administration of electrical stimulation of any of several brain structures, in a variety of species, can lead ~0 the progressive development of electrographic and behavioral seizures 7'14. Due to the reliability of the phenomenon, and the distinct behavioral progression that accompanies seizure development, this procedure, termed kindling, has been utilized as a model of epileptogenesis l°. Epileptic conditions are often accompanied by modifications in sleep 6. A number of studies have identified relationships between the development of kindled seizures and changes in sleep patterns. In cats, amygdaloid kindling reduced paradoxical and non-paradoxical sleep levels 1s'23, hippocampal kindling also resulted in similar reductions in paradoxical sleep 15. Changes in sleep states also influence kindled seizures, as demonstrated by the inverse correlation between the extent of sleep pathology in kindled cats and the levels of subsequent seizure thresholds 19. Additionally, kindling during sleep
states differentially alters the rate of seizure development. Stimulation during paradoxical sleep 4 or non-Faradoxicai ~ieep 16 produce kindling at rates slower and faster, respectively, than those seen when animals are kindled during wake.Culness. Kindling is also facilitated when sleep levels are reduced with either basal forebrain lesions, which reduces the number of stimulations required for generalized convulsions 2°, or by sleep deprivation, which lowers seizure thresholds s. Other behavioral changes, including deficits in inhibitory avoidance retention in rats I, also accompany kindled seizures. The sleep reductions seen after amygdaloid kindling in cats, along wtth memory deficits seen in rats, appear similar to sleep and memory changes we recently observed in old rats ~l. These sleep changes include, among others, significant reductions m total and paradoxical sleep. In addition, the extent of impairment found in several paradoxical sleep parameters was highly correlated with deficits in retention of a passive avoidance behavior in old, but not young, rats. The present study addresses
Correspondence: W S. Stone, Department of Psychology, Gilmer Hall, Charlottesville, VA 22903-2477, U S A. 0006-8993/88/$03 50 © 1988 Elsevier Science Publishers B.V (Biomedical Division)
136 the hypothesis that sleep deficits accompany kindled seizures in rats and, in addition, that the sleep deficits are related to memory impairments after kindling. MATERIALS AND METHODS Male Sprague-Dawley rats were obtained from Dominion laboratories (Dublin, VA). The animals were housed individually in plastic cages, maintained on a 12:12 light:dark cycle, and allowed free access to food and water throughout the experiment. For surgery, the animals were anesthetized with Nembutal (45 mg/kg, i.p.) and also received atropine sulfate (0.3 mg/kg, i.p.). Two jeweler's screws were threaded through the skull to obtain cortical EEGs from all animals. One of these was placed at the level of bregma, about 1 mm from the midline, while the other was positioned in the contralateral hemisphere at the level of lambda, at about 4 mm lateral from the midline. Each electrode was connected to a male Amphenol pin placed in an Amphenol microminiature connector strip. A few animals additionally received E M G electrodes in the posterior neck muscles. In addition to the cortical electrodes, other rats were bilaterally implanted with bipolar electrodes in the basolateral amygdala. These electrodes consisted of twisted stainless steel wires that were 0.25 mm in diameter and insulated with enamel except at the tips. They were stereotaxically implanted at the following coordinates13:1.0 mm posterior from bregma, 4.6 mm lateral from the midline, and 8.5 mm below the level of dura. These electrodes were also connected to the Amphenol strip and the entire assembly was cemented to the calvarium. Beginning one week after surgery, the animals were habituated in a recording room for 3 h/day for 4 days before baseline sleep measures were obtained. In the initial recordings, evaluations of awake, nonparadoxical sleep and paradoxical sleep states were determined in 30 s scoring epochs with animals who received both EEG and E M G electrodes (n = 3). Three-hour sleep records were obtained on an 8channel Grass model 7D polygraph with a chart speed of 10 mm/s. For these animals, movement was simultaneously recorded on separate channels from the E M G electrodes and from a combination of an E E G and open lead. The epochs were sco-ed using first the E E G and EMG channels, and then by view-
ing only the E E G and the movement channels. After 100 epochs which included all 3 states were obtained, a high level of agreement (>95%) was found between the two methods of movement evaluation for each animal. Since using an open lead to identify movement both simplified the surgical procedure and demanded fewer electrode leads to the polygraph (allowing simultaneous recordings from more animals), the remainder of the experiment was performed using this method. Following baseline sleep recordings, a group of amygdala-implanted animals ( , = 7) was kindled daily (unilaterally, biphasic square waves, 1 ms, 60 Hz, 250 #A, 1 s duration) for 4 weeks. The stimulation was administered from a Grass S-4 stimulator through a circuit with a 500,000 ~ resistor to assure constant current, and a 1000 Q resistor across which the voltage drop was monitored on a Tektronix Type 502 storage oscilloscope. Seizures were assessed by measuring the duration of the electrographic afterdischarge and by rating the behavioral seizures with the 5-point scale developed by Racine ~4. As in previous studies 24,25, seizure development was found to be complete after 2 consecutive stage 5 kindled convulsions. At that time (usually after 1-2 weeks of daily stimulation), both the behavioral severity and the afterdischarge of the seizure stabilize. Sleep recordings were obtained every 1-2 days throughout the kindling period, and also during a 7 day recovery period after kindling was terminated. The records were always taken in the 3 b period before the stimulation in order to minimize acute effects of the stimulation itself on sleep. There were also two nonkindled control groups. One of these was implanted with amygdala electrodes (n = 3), and the other only received cortical electrodes (n = 5). Sleep was recorded from control animals at the same time period as the kindled rats. One week after the last kindling trial, all animals were trained in an inhibitory (passive) avoidance task. The training apparatus was a 2-compartment acrylic box, in which a white start compartment (10 c m x 14 cm x 23 cm) was separated from a larger dark compartment (10 cm x 14 c m x 37 cm) by a sliding white door. On the training trial, animals were placed in the start compartment and the sliding door was opened. When they had completely entered the dark compartment (all 4 paws), a brief footshock (1
137 mA, I s) was delivered through a grid floor, and the animals were then removed from the chamber. Twenty-four hours later they were returned to the start box and the sliding door was again opened. Retention performance was assessed by measuring the latencies to re-enter the dark compartment up to a maximum score of 600 s. At the conclusion of the inhibitory avoidance test, the animals were sacrificed for histological examinations (40-pm sections, Cresyl violet stains) of electrode placements.
The effects of kindfing on sleep were analyzed for changes from baseline values with dependent t-tests, while comparisons with non-kindled animals were assessed with independent t-tests. Inhibitory avoidance
comparisons were made with Mann-Whitney Utests. RESULTS
Histological examination verified amygdala place-
1A 15o 125I 140 0J
E
p-
13~
{11
0) 120 O3 tlO
CO +, 100
8O
Basehne
4Weeks +1Week Stage5 Recovery (X :13 Days)
2-Year
Basehne
Olds
li251
Stage5
4Weeks
(x-13Days)
÷1Week
2-Year
Recovery OIds
1D8
Cono.
I
Kindled .~
e.12o
g~
Basehne
Stage 5 4 Weeks (X=13Days)
÷1 Week
Recovery
2-Year
Olds
Base-
Stage 5
line
(~=13Days)
4 Weeks
+1Week
2-Year
Recovery Olds
Fig. 1 Mean (+ S.E.M.)valueso(sleepparameters after 4w,'eksofkindhng. The time pointsin the graphs include baseline values, 13 days (stage 5), 4 weeks, and 1 week of recovery. Sleep values from 2-year-old rats are also included for comparison. The parameters showing significant reductions after 4 weeks of kindling included total sleep time (A), non-paradoxical sleep (B), and paradoxical sleep (C); these changes persisted after 1 week recovery period. The number of paradoxical sleep bouts (D) was also slgmficandy reduced after the one week recovery period *P < 0 05 vs baseline; **P < 0.01 vs baseline: e p < 0 05 vs non-st~mubted controls' e e p < 0.01 vs non.stimulated controls; e e e p < 0 001 vs non-stimulated controls.
138 ment of electrodes in all implanted animals. Implanted and the non-implanted control groups exhibited no significant differences in sleep or memory and these animals were combined for statistical analyses. Of the sleep changes seen in kindled rats, some were transient and others lasted beyond the termination of kindling trials. At the time of the second stage 5 seizure, a significant decrease in non-paradoxical sleep bout numbers (t = 3.03, P < 0.05) was found, which coincided with an increase in non-paradoxical sleep bout duration that approached statistical significance (t = 2.38, P < 0,06). Paradoxical sleep bout duration also exhibited a non-significant increase similar to that seen for non-paradoxical sleep (t -- 1.91, P < 0.1). No other kindling effects on sleep reached or approached significance at the time of the second stage 5 seizure. When the animals were kindled for an additional 2-2.5 weeks, both non-paradoxical bout number and duration returned to their baseline values. A second, more durable set of sleep changes devel-
600
i v
oped after the appearance of 2 consecutive stage 5 seizures. This later group of alterations persisted during the 1 week recovery period. Fig. 1 displays sleep parameters which changed after extensive kindling. Data on sleep parameters in old rats from our earlier study 2t are also included here for comparison. At the end of 4 weeks, total sleep time (Fig. 1A), and both of its major components, non-paradoxical (Fig. 1B) and paradoxical sleep (Fig. 1C), were significantly reduced relative both to their own baseline values, and to those of the non-kindled control group. These parameters were still significantly depressed after the one week recovery period. Kindled rats also exhibited other sleep alterations at this time point. The ratio of paradoxical sleep/total sleep time and the number of paradoxical sleep bouts (Fig. 1D) were both significantly lower than their initial values and were also lower than those of the non-kindled control group. Results from the inhibitory avoidance tests, obtained one week after the last kindling trial (minus one kindled animal whose electrode assembly fell off just prior to training, leaving an n of 6) are displayed in Fig. 2. Inhibitory avoidance data from old rats 21 are also presented for comparison. Relative to the non-kindled control animals, the kindled rats were significantly impaired in their retention of the avoidance response (P < 0.02). Relationships between sleep and retention were
,oo TABLE I Correlations between sleep parameters and inhtbttory avmdance latency scores in kindled and non-kindled rats (n = 14) Sleep parameter
200
E *l 100
Young
Young
Control
Kindled
2 - Year Olds
Fig 2. Median (+ mterquartlle range) inhibitory avoidance re-
tentton test latencies in kindled and non-stimulated controls Note that kindled animals had retention scores significantly lower than those of controls Also included for comparison are retention scores from 2-year-old animals, which are s~mflarto those seen in the kindled group. Op < 0.02 (Mann-Whitney Utest)
Group (n) All (14)
Kindled (6)
r Total sleep time 0.46 Non-paradoxical sleep time 0 31 Non-paradoxical sleep boutnumber -0.21 Non-paradoxical sleep bout duration 0.35 Paradoxical sleep time 0.69 Paradoxical sleep bout number 0.37 Paradoxical sleep bout duration 0.58 Paradoxical sleep/total sleeptime 0.68
P(12df)
r
e(4df)
ns ns
0 32 0 22
ns ns
ns
0.22
ns
ns < 0 01
-0.16 0.52
ns ns
ns
-0.65
ns
<0.05
0.87
<0.05
<0.01
0.45
ns
139 further examined by correlating sleep parameters with retention scores in individual rats (kindled and controls). As depicted in Table I, total paradoxical sleep, the ratio of paradoxical sleep/total sleep time, and paradoxical sleep bout duratio.~ were each significantly correlated with retention performance. One of these parameters, paradoxical sleep bout duration, was also significantly correlated with retention performance within the kindled group. DISCUSSION The present results demonstrate several sleep alterations during kindling in rats. These changes include two distinct sets of findings. During development of kindling, sleep exhibited decreased bout numbers of longer duration, findings which are in the direction of greater sleep continuity and fewer state changes. However, these responses were transient and returned to baseline values by the end of the 4week kindling period. A second pattern of sleep changes, seen after stage 5 kindling was attained, was primarily characterized by reductions in the amount of sleep. A primary difference between the early and later elterations was that while the earlier changes appeared to enhance aspects of sleep function (i.e. greater continuity and less fragmentation), the later changes more clearly demonstrated sleep deterioration (Le. reduced sleep). The reductions m non-paradoxical and paradoxical sleep found after 4 weeks of kindling persisted through the one week recovery period. In addition, other indices of paradoxical sleep were significantly impaired at that time, including the ratio of paradoxical sleep/total sleep time and the number of paradoxical sleep bout numbers. The reductions in rats of both the major sleep components, and their persistence beyond the kindling period, are in agreement with previous sleep and kindling relationships that have been reported to occur m cats 1s'19"23. However, unlike these previous reports, the appearance of the major sleep disorders in rats did not accompany the development of kindled seizures, but only appeared after the generalized convulsio~,~ had become established. This finding suggests that c,.,.nges in the brain continue with kindling trials administerd after the development of stage 5 seizures. It is not clear from these results whether paradoxi-
cal sleep selectively deteriorated after kindling was terminated, or whether the phenomenon was secondary to the recovery of non-paradoxical sleep. Previous studies of recovery from sleep deprivation in rats demonstrated an earlier recovery of non-paradoxical sleep possibly occurring at the expense of paradoxical sleep time2. The findings of fewer bout numbers and a lower paradoxical sleep/total sleep percentage would be consistent with such a recovery process. However, other studies that have examined reduced paradoxical sleep in association with kindling ~7"I9"23or electroconvulsive shock6 in cats, did not observe paradoxical sleep rebound during prolonged recovery periods. The continued impairment in paradoxical sleep, then, could represent a longer lasting concomitant of the kindling process. The presence of sleep changes is one of several functional alterations which accompany kindled seizures. Learning and memory impairments have been reported for conditioned emotional responding9, interference with conditioned taste and odor aversions 1~, and inhibitory avoidance retention I which is also seen in the present experiment. The present study further found significant correlations between individual retention scores and paradoxical sleep measures. Although correlations cannot specify causal relationships, the results are consistent with the notion that deficits in paradoxical sleep may be related to deficits in cognitwe processrag. This view is supported by convergent findings of similar sleep and memory relationships in 2-year-old rats, and in young rats with lesions of the nucleus basalis magnocellularis (NBM21). These analogies between kindled and old animals, as well as between kindled and NBM-lesioned young animals, suggest that the kindling procedure, in addition to its utility in other contexts, may also model some of the characteristics of aging. In this context, a comparison of different neuroblological markers of aging may be usefully compared with changes accompanying extensive kindling to determine further the significance of the analogies described here. Central cholinergic functions, for example, exhibit similar age and kindling-related alterations3"12"21'22. In summary, several significant effects of extensive kindling were seen in Sprague-Dawley rats. First, kindled animals displayed significant reductions in total sleep, paradoxical sleep, and non-para-
140 doxical sleep times. These changes occurred after the development of stable stage 5 convulsions, indicating that neural modifications continue when additional kindling trials are administered. The sleep alterations induced by kindling also persisted at least 1 week after the last kindling trial. Finally, significant correlations were found between the extent of deterioration in several paradoxical sleep parameters and the extent of deterioration in inhibitory avoidance behavior. These sleep and memory relationships resemble those found in intact old and NBM-lesioned young rats, supporting the view that deterioration of
REFERENCES 1 Boast, C A and Mclntyre, D C., Bilateral kindled loci and inhibitory behavtor in rats: a functional lesion effect, Physml. Behav., 18 (1977) 25-28. 2 Borbely, A.A. and Ne'ahaus, H.U., Sleep deprivation- effects on sleep and EEG in the rat, J. Comp Phystol, 133 (1979) 71-87. 3 Byrne, M.C., Gottlieb, R. and McNamara, J.O., Amygdala kindling induces muscarinic cholinergic receptor declines in a highly specific distril"¢tion within the hmbic system, Exp. Neurol., 69 (1980) 85-98. 4 Calvo, J.M., Alvarado, R., Brioaes, R., Paz, C. and Fernandez-Guardiola, A., Amygdalold kindling during rapid eye movement (REM) sleep in cats, Neurosci. Leu., 29 (1980) 255-259. 5 Cohen, H.B. and Dement, W.C, Suppression of rapid eye movement phase m the cat after electroconvulstve shock, Science, 154 (1966) 396-398. 6 Dahl, M. and Dam, M., Sleep and epdepsy, Ann. Clin. Res., 17 (1985) 235-242. 7 Goddard, G.V., Mclntyre, D.C and Leech, C.K., A permanent change in brain function resulting from daily electrical stimulatmn, Exp. Neurol., 25 (1969) 295-330. 8 Grahnstedt, S., Sleep deprivation and kindled seizures, Exp. Neurol., 92 (1986) 248-260. 9 McIntyre, D.C and Molino, A., Amygdala lestons and CER learmng: long term effect of kmdhng, Physiol. Behay., 8 (1972) 1055-1058. 10 McNamara, J.O, Byrne, M.C., Dasheiff, R.M. and Fitz, J.G., The kindling model of epilepsy: a review, Prog. Neuroblol., 15 (1980) 139-159. 11 Mikulka, PJ. and Freeman, F G., The effect of amygdalakindled seizures on the acqms~t~onof taste and odor averslons, Physiol. Behav., 32 (1984) 967-972 12 Noda, Y , Wada, J.A and McGeer, E.G., Lasting influence of amygdaloid kindling on cholinergic neurotransmisslon, Exp. Neurol, 78 (1982) 91-98. 13 Pellegrino, L J., Pellegrmo, A S and Cushman, A.J., A Stereotaxw Atlas of the Rat Brain, Plenum Press, New
paradoxical sleep and deterioration of selective cognitive functions may often be related.
ACKNOWLEDGEMENTS This research was supported by research grants from the National Institute of Mental Health (MH 31141), the Office of Naval Research (N00014-85K0472), the A m e r i c a n Diabetes Association, and by a National Institute of Aging Postdoctoral Fellowship (AG05408) to W.S.S.
York, 1979 14 Racine, R.J., Kindling: the fu~t decade, Neurosurgery, 3 (1978) 234-252. 15 Rondoum, G., Baldy-Mouhmer, M. and Passouant, P., The influence of hippocampal kindling on sleep organization in cats. Effects of alpha-methylparatyrosme, Brain Research, 181 (1980) 413-424 16 Sato, M. and Nakeshima, T., Kindling: secondary epileptogenesis and catecholamines, Can. J. Neurol., 2 (1975) 439-446. 17 Shouse, M.N, State disorders and state-dependent seizures in amygdala kindled cats, Exp. Neurol., 92 (1986) 601-608. 18 Shouse, M.N and Sterman, M.B., Sleep and kmdhng: II. Effects of generalized seizure induction, Exp Neurol., 71 (1981) 563-580. 19 Shouse, M.N. and Sterman, M.B., 'Klndhng' a seizure disorder' degree of sleep pathology predicts kindled seizure susceptibilitym cats, Brain Research, 271 (1983) 196-200. 20 Shouse, M.N., Sterman, M.B, Haurt, P. and Belstto, O., Sleep disruption with basal forebrain les~ons decreases latency to amygdala kindling m cats, Electroencephalogr Cl,n. Neurophystol., 58 (1984) 369-377. 21 Stone, W.S., Altman, H J., Berman, R.F., Caldwell, D.F. and Kilbey, M.M., Association of sleep parameters and memory in intact old and nucleus basalis lesioned young rats, in prep.. 22 Strong, R., Hicks, L., Hsu, R., Bartus, R T. and Enna, S.J., Age-related alterations in the rodent brain chohnergtc system and behavior, Neuroblol. Aging, 1 (1980) 59-63 23 Tanaka, T. and Nacquet, R , Kindling effect and sleep orgamzatmn m cats, Electroencephalogr Clin. Neurophyszol, 39 (1975) 449-454. 24 Welsh, K.A and Gold, P.E., Attenuation of epileptogenesJs Proacttve effect of a single epinephrine inlect~or, on amygdalold'kindling, Behavtor Neur. Btol, 40 (1984) 179185. 25 Welsh, K A and Gold, P.E., Epinephrine proactwe retardation of amygdala-kmdled eplleptogenesls, Behavior Neuroscl., 100 (1986) 236-245.
135
BRE 13530
Amygdala kindling effects on sleep and memory in rats William S. Stone and Paul E. Gold Department of Psychology, University of Virginia, Gilmer Hall, Charlottesville, VA 22903-2477(U.S.A_) (Accepted 3 November 1987) Key words: Kindling, Sleep; Memory; Aging; Rat; Seizure
Sleep d~sturbances accompany the development of amygdaloLd-kindledseLzuresin cats. Some of these sleep defiots resemble those seen m aged rats; these latter changes m sleep patterns are correlated with memory impairments in the aged animals. In the present study, we examined the hypothesis that sleep deficits after kindling may be related to memory tmpaitments Rats were kindled for 4 weeks (2-2.5 weeks after stage 5 smzures) and were then allowed a one week recovery period. Sleep patterns were assessed throughout the kindling and recovery periods The ammals were then trained on an inhibitory avoidance apparatus and tested for retention 24 h later. Only transient sleep changes occurred during the development of kindling ( to stage 5 seizures). However, continued kindling resulted in significant reductions m several sleep measures which remained depressed for at least one week after the termination of the kindling trials. As a group, kindled rats were impaired in retention of the inhibitory avoidance learned response. In kindled animals, retention performance was slgmflcantly correlated with total paradoxical sleep, the ratio of paradoxical/total sleep, and paradoxical sleep bout duration. These correlations are consistent with the view that deficits m paradoxical sleep may be related to deficits in some forms of memory. INTRODUCTION Daily administration of electrical stimulation of any of several brain structures, in a variety of species, can lead ~0 the progressive development of electrographic and behavioral seizures 7'14. Due to the reliability of the phenomenon, and the distinct behavioral progression that accompanies seizure development, this procedure, termed kindling, has been utilized as a model of epileptogenesis l°. Epileptic conditions are often accompanied by modifications in sleep 6. A number of studies have identified relationships between the development of kindled seizures and changes in sleep patterns. In cats, amygdaloid kindling reduced paradoxical and non-paradoxical sleep levels 1s'23, hippocampal kindling also resulted in similar reductions in paradoxical sleep 15. Changes in sleep states also influence kindled seizures, as demonstrated by the inverse correlation between the extent of sleep pathology in kindled cats and the levels of subsequent seizure thresholds 19. Additionally, kindling during sleep
states differentially alters the rate of seizure development. Stimulation during paradoxical sleep 4 or non-Faradoxicai ~ieep 16 produce kindling at rates slower and faster, respectively, than those seen when animals are kindled during wake.Culness. Kindling is also facilitated when sleep levels are reduced with either basal forebrain lesions, which reduces the number of stimulations required for generalized convulsions 2°, or by sleep deprivation, which lowers seizure thresholds s. Other behavioral changes, including deficits in inhibitory avoidance retention in rats I, also accompany kindled seizures. The sleep reductions seen after amygdaloid kindling in cats, along wtth memory deficits seen in rats, appear similar to sleep and memory changes we recently observed in old rats ~l. These sleep changes include, among others, significant reductions m total and paradoxical sleep. In addition, the extent of impairment found in several paradoxical sleep parameters was highly correlated with deficits in retention of a passive avoidance behavior in old, but not young, rats. The present study addresses
Correspondence: W S. Stone, Department of Psychology, Gilmer Hall, Charlottesville, VA 22903-2477, U S A. 0006-8993/88/$03 50 © 1988 Elsevier Science Publishers B.V (Biomedical Division)
136 the hypothesis that sleep deficits accompany kindled seizures in rats and, in addition, that the sleep deficits are related to memory impairments after kindling. MATERIALS AND METHODS Male Sprague-Dawley rats were obtained from Dominion laboratories (Dublin, VA). The animals were housed individually in plastic cages, maintained on a 12:12 light:dark cycle, and allowed free access to food and water throughout the experiment. For surgery, the animals were anesthetized with Nembutal (45 mg/kg, i.p.) and also received atropine sulfate (0.3 mg/kg, i.p.). Two jeweler's screws were threaded through the skull to obtain cortical EEGs from all animals. One of these was placed at the level of bregma, about 1 mm from the midline, while the other was positioned in the contralateral hemisphere at the level of lambda, at about 4 mm lateral from the midline. Each electrode was connected to a male Amphenol pin placed in an Amphenol microminiature connector strip. A few animals additionally received E M G electrodes in the posterior neck muscles. In addition to the cortical electrodes, other rats were bilaterally implanted with bipolar electrodes in the basolateral amygdala. These electrodes consisted of twisted stainless steel wires that were 0.25 mm in diameter and insulated with enamel except at the tips. They were stereotaxically implanted at the following coordinates13:1.0 mm posterior from bregma, 4.6 mm lateral from the midline, and 8.5 mm below the level of dura. These electrodes were also connected to the Amphenol strip and the entire assembly was cemented to the calvarium. Beginning one week after surgery, the animals were habituated in a recording room for 3 h/day for 4 days before baseline sleep measures were obtained. In the initial recordings, evaluations of awake, nonparadoxical sleep and paradoxical sleep states were determined in 30 s scoring epochs with animals who received both EEG and E M G electrodes (n = 3). Three-hour sleep records were obtained on an 8channel Grass model 7D polygraph with a chart speed of 10 mm/s. For these animals, movement was simultaneously recorded on separate channels from the E M G electrodes and from a combination of an E E G and open lead. The epochs were sco-ed using first the E E G and EMG channels, and then by view-
ing only the E E G and the movement channels. After 100 epochs which included all 3 states were obtained, a high level of agreement (>95%) was found between the two methods of movement evaluation for each animal. Since using an open lead to identify movement both simplified the surgical procedure and demanded fewer electrode leads to the polygraph (allowing simultaneous recordings from more animals), the remainder of the experiment was performed using this method. Following baseline sleep recordings, a group of amygdala-implanted animals ( , = 7) was kindled daily (unilaterally, biphasic square waves, 1 ms, 60 Hz, 250 #A, 1 s duration) for 4 weeks. The stimulation was administered from a Grass S-4 stimulator through a circuit with a 500,000 ~ resistor to assure constant current, and a 1000 Q resistor across which the voltage drop was monitored on a Tektronix Type 502 storage oscilloscope. Seizures were assessed by measuring the duration of the electrographic afterdischarge and by rating the behavioral seizures with the 5-point scale developed by Racine ~4. As in previous studies 24,25, seizure development was found to be complete after 2 consecutive stage 5 kindled convulsions. At that time (usually after 1-2 weeks of daily stimulation), both the behavioral severity and the afterdischarge of the seizure stabilize. Sleep recordings were obtained every 1-2 days throughout the kindling period, and also during a 7 day recovery period after kindling was terminated. The records were always taken in the 3 b period before the stimulation in order to minimize acute effects of the stimulation itself on sleep. There were also two nonkindled control groups. One of these was implanted with amygdala electrodes (n = 3), and the other only received cortical electrodes (n = 5). Sleep was recorded from control animals at the same time period as the kindled rats. One week after the last kindling trial, all animals were trained in an inhibitory (passive) avoidance task. The training apparatus was a 2-compartment acrylic box, in which a white start compartment (10 c m x 14 cm x 23 cm) was separated from a larger dark compartment (10 cm x 14 c m x 37 cm) by a sliding white door. On the training trial, animals were placed in the start compartment and the sliding door was opened. When they had completely entered the dark compartment (all 4 paws), a brief footshock (1
137 mA, I s) was delivered through a grid floor, and the animals were then removed from the chamber. Twenty-four hours later they were returned to the start box and the sliding door was again opened. Retention performance was assessed by measuring the latencies to re-enter the dark compartment up to a maximum score of 600 s. At the conclusion of the inhibitory avoidance test, the animals were sacrificed for histological examinations (40-pm sections, Cresyl violet stains) of electrode placements.
The effects of kindfing on sleep were analyzed for changes from baseline values with dependent t-tests, while comparisons with non-kindled animals were assessed with independent t-tests. Inhibitory avoidance
comparisons were made with Mann-Whitney Utests. RESULTS
Histological examination verified amygdala place-
1A 15o 125I 140 0J
E
p-
13~
{11
0) 120 O3 tlO
CO +, 100
8O
Basehne
4Weeks +1Week Stage5 Recovery (X :13 Days)
2-Year
Basehne
Olds
li251
Stage5
4Weeks
(x-13Days)
÷1Week
2-Year
Recovery OIds
1D8
Cono.
I
Kindled .~
e.12o
g~
Basehne
Stage 5 4 Weeks (X=13Days)
÷1 Week
Recovery
2-Year
Olds
Base-
Stage 5
line
(~=13Days)
4 Weeks
+1Week
2-Year
Recovery Olds
Fig. 1 Mean (+ S.E.M.)valueso(sleepparameters after 4w,'eksofkindhng. The time pointsin the graphs include baseline values, 13 days (stage 5), 4 weeks, and 1 week of recovery. Sleep values from 2-year-old rats are also included for comparison. The parameters showing significant reductions after 4 weeks of kindling included total sleep time (A), non-paradoxical sleep (B), and paradoxical sleep (C); these changes persisted after 1 week recovery period. The number of paradoxical sleep bouts (D) was also slgmficandy reduced after the one week recovery period *P < 0 05 vs baseline; **P < 0.01 vs baseline: e p < 0 05 vs non-st~mubted controls' e e p < 0.01 vs non.stimulated controls; e e e p < 0 001 vs non-stimulated controls.
138 ment of electrodes in all implanted animals. Implanted and the non-implanted control groups exhibited no significant differences in sleep or memory and these animals were combined for statistical analyses. Of the sleep changes seen in kindled rats, some were transient and others lasted beyond the termination of kindling trials. At the time of the second stage 5 seizure, a significant decrease in non-paradoxical sleep bout numbers (t = 3.03, P < 0.05) was found, which coincided with an increase in non-paradoxical sleep bout duration that approached statistical significance (t = 2.38, P < 0,06). Paradoxical sleep bout duration also exhibited a non-significant increase similar to that seen for non-paradoxical sleep (t -- 1.91, P < 0.1). No other kindling effects on sleep reached or approached significance at the time of the second stage 5 seizure. When the animals were kindled for an additional 2-2.5 weeks, both non-paradoxical bout number and duration returned to their baseline values. A second, more durable set of sleep changes devel-
600
i v
oped after the appearance of 2 consecutive stage 5 seizures. This later group of alterations persisted during the 1 week recovery period. Fig. 1 displays sleep parameters which changed after extensive kindling. Data on sleep parameters in old rats from our earlier study 2t are also included here for comparison. At the end of 4 weeks, total sleep time (Fig. 1A), and both of its major components, non-paradoxical (Fig. 1B) and paradoxical sleep (Fig. 1C), were significantly reduced relative both to their own baseline values, and to those of the non-kindled control group. These parameters were still significantly depressed after the one week recovery period. Kindled rats also exhibited other sleep alterations at this time point. The ratio of paradoxical sleep/total sleep time and the number of paradoxical sleep bouts (Fig. 1D) were both significantly lower than their initial values and were also lower than those of the non-kindled control group. Results from the inhibitory avoidance tests, obtained one week after the last kindling trial (minus one kindled animal whose electrode assembly fell off just prior to training, leaving an n of 6) are displayed in Fig. 2. Inhibitory avoidance data from old rats 21 are also presented for comparison. Relative to the non-kindled control animals, the kindled rats were significantly impaired in their retention of the avoidance response (P < 0.02). Relationships between sleep and retention were
,oo TABLE I Correlations between sleep parameters and inhtbttory avmdance latency scores in kindled and non-kindled rats (n = 14) Sleep parameter
200
E *l 100
Young
Young
Control
Kindled
2 - Year Olds
Fig 2. Median (+ mterquartlle range) inhibitory avoidance re-
tentton test latencies in kindled and non-stimulated controls Note that kindled animals had retention scores significantly lower than those of controls Also included for comparison are retention scores from 2-year-old animals, which are s~mflarto those seen in the kindled group. Op < 0.02 (Mann-Whitney Utest)
Group (n) All (14)
Kindled (6)
r Total sleep time 0.46 Non-paradoxical sleep time 0 31 Non-paradoxical sleep boutnumber -0.21 Non-paradoxical sleep bout duration 0.35 Paradoxical sleep time 0.69 Paradoxical sleep bout number 0.37 Paradoxical sleep bout duration 0.58 Paradoxical sleep/total sleeptime 0.68
P(12df)
r
e(4df)
ns ns
0 32 0 22
ns ns
ns
0.22
ns
ns < 0 01
-0.16 0.52
ns ns
ns
-0.65
ns
<0.05
0.87
<0.05
<0.01
0.45
ns
139 further examined by correlating sleep parameters with retention scores in individual rats (kindled and controls). As depicted in Table I, total paradoxical sleep, the ratio of paradoxical sleep/total sleep time, and paradoxical sleep bout duratio.~ were each significantly correlated with retention performance. One of these parameters, paradoxical sleep bout duration, was also significantly correlated with retention performance within the kindled group. DISCUSSION The present results demonstrate several sleep alterations during kindling in rats. These changes include two distinct sets of findings. During development of kindling, sleep exhibited decreased bout numbers of longer duration, findings which are in the direction of greater sleep continuity and fewer state changes. However, these responses were transient and returned to baseline values by the end of the 4week kindling period. A second pattern of sleep changes, seen after stage 5 kindling was attained, was primarily characterized by reductions in the amount of sleep. A primary difference between the early and later elterations was that while the earlier changes appeared to enhance aspects of sleep function (i.e. greater continuity and less fragmentation), the later changes more clearly demonstrated sleep deterioration (Le. reduced sleep). The reductions m non-paradoxical and paradoxical sleep found after 4 weeks of kindling persisted through the one week recovery period. In addition, other indices of paradoxical sleep were significantly impaired at that time, including the ratio of paradoxical sleep/total sleep time and the number of paradoxical sleep bout numbers. The reductions in rats of both the major sleep components, and their persistence beyond the kindling period, are in agreement with previous sleep and kindling relationships that have been reported to occur m cats 1s'19"23. However, unlike these previous reports, the appearance of the major sleep disorders in rats did not accompany the development of kindled seizures, but only appeared after the generalized convulsio~,~ had become established. This finding suggests that c,.,.nges in the brain continue with kindling trials administerd after the development of stage 5 seizures. It is not clear from these results whether paradoxi-
cal sleep selectively deteriorated after kindling was terminated, or whether the phenomenon was secondary to the recovery of non-paradoxical sleep. Previous studies of recovery from sleep deprivation in rats demonstrated an earlier recovery of non-paradoxical sleep possibly occurring at the expense of paradoxical sleep time2. The findings of fewer bout numbers and a lower paradoxical sleep/total sleep percentage would be consistent with such a recovery process. However, other studies that have examined reduced paradoxical sleep in association with kindling ~7"I9"23or electroconvulsive shock6 in cats, did not observe paradoxical sleep rebound during prolonged recovery periods. The continued impairment in paradoxical sleep, then, could represent a longer lasting concomitant of the kindling process. The presence of sleep changes is one of several functional alterations which accompany kindled seizures. Learning and memory impairments have been reported for conditioned emotional responding9, interference with conditioned taste and odor aversions 1~, and inhibitory avoidance retention I which is also seen in the present experiment. The present study further found significant correlations between individual retention scores and paradoxical sleep measures. Although correlations cannot specify causal relationships, the results are consistent with the notion that deficits in paradoxical sleep may be related to deficits in cognitwe processrag. This view is supported by convergent findings of similar sleep and memory relationships in 2-year-old rats, and in young rats with lesions of the nucleus basalis magnocellularis (NBM21). These analogies between kindled and old animals, as well as between kindled and NBM-lesioned young animals, suggest that the kindling procedure, in addition to its utility in other contexts, may also model some of the characteristics of aging. In this context, a comparison of different neuroblological markers of aging may be usefully compared with changes accompanying extensive kindling to determine further the significance of the analogies described here. Central cholinergic functions, for example, exhibit similar age and kindling-related alterations3"12"21'22. In summary, several significant effects of extensive kindling were seen in Sprague-Dawley rats. First, kindled animals displayed significant reductions in total sleep, paradoxical sleep, and non-para-
140 doxical sleep times. These changes occurred after the development of stable stage 5 convulsions, indicating that neural modifications continue when additional kindling trials are administered. The sleep alterations induced by kindling also persisted at least 1 week after the last kindling trial. Finally, significant correlations were found between the extent of deterioration in several paradoxical sleep parameters and the extent of deterioration in inhibitory avoidance behavior. These sleep and memory relationships resemble those found in intact old and NBM-lesioned young rats, supporting the view that deterioration of
REFERENCES 1 Boast, C A and Mclntyre, D C., Bilateral kindled loci and inhibitory behavtor in rats: a functional lesion effect, Physml. Behav., 18 (1977) 25-28. 2 Borbely, A.A. and Ne'ahaus, H.U., Sleep deprivation- effects on sleep and EEG in the rat, J. Comp Phystol, 133 (1979) 71-87. 3 Byrne, M.C., Gottlieb, R. and McNamara, J.O., Amygdala kindling induces muscarinic cholinergic receptor declines in a highly specific distril"¢tion within the hmbic system, Exp. Neurol., 69 (1980) 85-98. 4 Calvo, J.M., Alvarado, R., Brioaes, R., Paz, C. and Fernandez-Guardiola, A., Amygdalold kindling during rapid eye movement (REM) sleep in cats, Neurosci. Leu., 29 (1980) 255-259. 5 Cohen, H.B. and Dement, W.C, Suppression of rapid eye movement phase m the cat after electroconvulstve shock, Science, 154 (1966) 396-398. 6 Dahl, M. and Dam, M., Sleep and epdepsy, Ann. Clin. Res., 17 (1985) 235-242. 7 Goddard, G.V., Mclntyre, D.C and Leech, C.K., A permanent change in brain function resulting from daily electrical stimulatmn, Exp. Neurol., 25 (1969) 295-330. 8 Grahnstedt, S., Sleep deprivation and kindled seizures, Exp. Neurol., 92 (1986) 248-260. 9 McIntyre, D.C and Molino, A., Amygdala lestons and CER learmng: long term effect of kmdhng, Physiol. Behay., 8 (1972) 1055-1058. 10 McNamara, J.O, Byrne, M.C., Dasheiff, R.M. and Fitz, J.G., The kindling model of epilepsy: a review, Prog. Neuroblol., 15 (1980) 139-159. 11 Mikulka, PJ. and Freeman, F G., The effect of amygdalakindled seizures on the acqms~t~onof taste and odor averslons, Physiol. Behav., 32 (1984) 967-972 12 Noda, Y , Wada, J.A and McGeer, E.G., Lasting influence of amygdaloid kindling on cholinergic neurotransmisslon, Exp. Neurol, 78 (1982) 91-98. 13 Pellegrino, L J., Pellegrmo, A S and Cushman, A.J., A Stereotaxw Atlas of the Rat Brain, Plenum Press, New
paradoxical sleep and deterioration of selective cognitive functions may often be related.
ACKNOWLEDGEMENTS This research was supported by research grants from the National Institute of Mental Health (MH 31141), the Office of Naval Research (N00014-85K0472), the A m e r i c a n Diabetes Association, and by a National Institute of Aging Postdoctoral Fellowship (AG05408) to W.S.S.
York, 1979 14 Racine, R.J., Kindling: the fu~t decade, Neurosurgery, 3 (1978) 234-252. 15 Rondoum, G., Baldy-Mouhmer, M. and Passouant, P., The influence of hippocampal kindling on sleep organization in cats. Effects of alpha-methylparatyrosme, Brain Research, 181 (1980) 413-424 16 Sato, M. and Nakeshima, T., Kindling: secondary epileptogenesis and catecholamines, Can. J. Neurol., 2 (1975) 439-446. 17 Shouse, M.N, State disorders and state-dependent seizures in amygdala kindled cats, Exp. Neurol., 92 (1986) 601-608. 18 Shouse, M.N and Sterman, M.B., Sleep and kmdhng: II. Effects of generalized seizure induction, Exp Neurol., 71 (1981) 563-580. 19 Shouse, M.N. and Sterman, M.B., 'Klndhng' a seizure disorder' degree of sleep pathology predicts kindled seizure susceptibilitym cats, Brain Research, 271 (1983) 196-200. 20 Shouse, M.N., Sterman, M.B, Haurt, P. and Belstto, O., Sleep disruption with basal forebrain les~ons decreases latency to amygdala kindling m cats, Electroencephalogr Cl,n. Neurophystol., 58 (1984) 369-377. 21 Stone, W.S., Altman, H J., Berman, R.F., Caldwell, D.F. and Kilbey, M.M., Association of sleep parameters and memory in intact old and nucleus basalis lesioned young rats, in prep.. 22 Strong, R., Hicks, L., Hsu, R., Bartus, R T. and Enna, S.J., Age-related alterations in the rodent brain chohnergtc system and behavior, Neuroblol. Aging, 1 (1980) 59-63 23 Tanaka, T. and Nacquet, R , Kindling effect and sleep orgamzatmn m cats, Electroencephalogr Clin. Neurophyszol, 39 (1975) 449-454. 24 Welsh, K.A and Gold, P.E., Attenuation of epileptogenesJs Proacttve effect of a single epinephrine inlect~or, on amygdalold'kindling, Behavtor Neur. Btol, 40 (1984) 179185. 25 Welsh, K A and Gold, P.E., Epinephrine proactwe retardation of amygdala-kmdled eplleptogenesls, Behavior Neuroscl., 100 (1986) 236-245.