Epilepsia, 33(5):789-798, 1992 Raven Press, Ltd., New York 0 International League Against Epilepsy

Ontogeny of Feline Temporal Lobe Epilepsy, 11: Stability of Spontaneous Sleep Epilepsy in Amygdala-Kindled Kittens Margaret N. Shouse, Paul Dittes, James Langer, and Robert Nienhuis Department of Anatomy and Cell Biology, WCLA School of Medicine, Los Angeles, and V A Medical Center, Sepulveda, California, U . S . A .

Summary: We previously described a model of spontaneous “sleep epilepsy” in kindled kittens with temporal lobe epilepsy (TLE). We now describe the postkindling course of this model from preadolescence to maturity and suggest pathophysiologic mechanisms. Spontaneous epilepsy, particularly generalized tonic-clonic convulsions (GTCs), developed l h to 4 months after amygdala kindling and persisted to adulthood. At first, GTCs were detected only in sleep; later, convulsions also occurred during wakefulness. Two factors were consistently associated with the sequential onset of sleep and waking GTCs: seizure clusters and anatomic seizure localization. (1) Seizure clusters. Cats with infrequent or unclustered GTCs continued to exhibit “sleep epilepsy,” defined by convulsions occurring exclusively during sleep. In contrast, cats with frequent seizure clusters developed recurrent or terminal convulsive status in conjunction with GTCs during waking and sleep. Severe seizure manifestations therefore appeared to contribute to the dissociation of convulsions from the sleepwake cycle. (2) Anatomical seizure localization. Focal seizure origin ap-

peared to differentiate sleep from waking GTCs. Onset during sleep was first recorded in the kindled amygdala, whereas onset during waking was initially detected outside the temporal lobe. Findings thus suggest secondary “kindling” of multifocal epilepsy. Secondary epileptogenesis is consistent with “transsynaptic” kindling effects. This phenomenon is defined in mature animals by rapid secondary site kindling (transfer) and subtle morphologic changes distal to the stimulating electrode. Transfer may be accentuated by youth, because kittens developed spontaneous seizure foci in previously unstimulated tissue. Moreover, multifocal interactions and diffuse cell loss were implicated as possible mechanisms. Collectively, the findings indicate complications with early onset TLE in kindled cats. Onset during youth can have an unfavorable prognosis, reflected by recurrent status epilepticus and multifocal epilepsy with convulsions distributed throughout the sleep-wake cycle. Key Words: Sleep-Epilepsy-Developmental biology-AmygdalaNeurologic models-Kindling-Kittens-West syndrome-Spontaneous seizures.

We previously described the ontogeny of feline temporal lobe epilepsy (TLE) after amygdala kindling in 24 preadolescent kittens or adult cats (Shouse et al., 1990 a , b , c ) .Kindled kittens were far more likely than adults to develop epilepsy, reflected by spontaneous seizures. The finding is consistent with clinical literature indicating that epilepsy is a disorder with onset in youth (Janz, 1962, 1974; Hauser and Kurland, 1975; Woodbury, 1977). Even human TLE is prone to early onset, although it can develop at any age. We subsequently established a second parallel to the clinical literature by documenting convulsive

“sleep epilepsy” in this model. Amygdala-kindled kittens displayed 380% of spontaneous convulsions during sleep, all during slow wave sleep (SWS) and particularly during the transition from SWS into rapid-eye-movement (REM) sleep. Findings resemble evoked seizure patterns in amygdala-kindled adult cats (Sato and Nakeshima, 1975; Calvo et al., 1982; Shouse, 1986, 1987) as well as convulsive seizure patterns in human TLE (Janz, 1962, 1974; Stevens et al., 1972; Billiard, 1982; Cadhillac, 1982; Shouse, 1989). TLE is the most common “convulsive sleep epilepsy” in humans, because nearly 60% of the patients have generalized tonic-clonic seizures (GTCs) only during sleep (Janz, 1962; Billiard, 1982). We now describe the postkindling course of this model from preadolescence to maturity and suggest

Received April 1991; revision accepted February 1992. Address correspondence and reprint requests to Dr. M. N. Shouse at Sleep Disturbance Research (151A3), VA Medical Center, Sepulveda, CA 91343, U.S.A.

789

M . N . SHOUSE ET AL.

790

pathophysiologic mechanisms for stable versus unstable temporal seizure patterns.

METHODS Subjects We studied 15 preadolescent kittens weighing 0.8-2.2 kg and 13 adult cats weighing 3.3-5.2 kg. At procurement, kittens were aged 2-6 months (n = 14 female, 1 male), and adult cats were aged at least 1 year (n = 8 female, 5 male). Female kittens reach puberty at age 7-9 months, and males reach puberty at 1 year (Inglis, 1980). Kittens were from five separate litters. One was obtained from a random source mother, who delivered in the Sepulveda VA Animal Research Facility. The other 14 were procured after weaning from the specific pathogen-free (SPF) colony at The University of California, Davis (UC Davis). The four litters from UC Davis yielded 5, 4, 3, and 2 kittens each. The 13 adult cats were from an unknown number of litters, with 11 from UC Davis; 2 from random source. We cannot specify genetic background further except to state that inbreeding occurs in the UC Davis colony. Some kittens in each of the five litters developed spontaneous epilepsy after kindling. The epileptic adult cat was from UC Davis. Stereotaxic surgery Aseptic neurosurgery was performed under intraperitoneal (i.p.) sodium pentobarbital anesthesia at 35 mg/kg. Electrodes were implanted for basolateral amygdala kindling and sleepwaking state evaluation. Stereotaxic coordinates were extrapolated from brain atlases for kittens aged 3.5 kg (Snider and Niemer, 1961), as described previously (Shouse et al., 1990~). All cats had tripolar strut electrodes aimed at the basolateral amygdala bilaterally and indifferent electrodes (skull screws over the intact frontal sinus) for kindling. Implants for sleep-waking state assessment included screw electrodes threaded into the bone over motor cortex and above the orbit to register cortical EEG and eye movements electrooculogram (EOG), stainless-steel wires inserted in the nuchal musculature to measure muscle tone (electromyography , EMG) and bilateral, tripolar leads in the lateral geniculate nucleus (LGN) to register pontogeniculooccipital (PGO) spikes. LGN leads were implanted in cats aged 2 4 months. Initial kindling or surgical control procedures After the animals were allowed 1-2 weeks of postoperative recovery, kindling or postoperative Epilepsia, Vol. 33, No. 5 , 1992

observation was initiated. Preadolescent kittens (n = 13) had initial afterdischarge (AD) evoked at age 2.5-6.5 months. Two kittens were not kindled but had comparable implants and postoperative observation beginning at age 4 months (surgical controls). All 13 adults were kindled between the ages of 1 and 3 years. The protocol for kindling kittens and adult cats was described previously (Shouse et al., 1990a,b). Focal AD threshold was determined by a methodof-limits procedure adapted from Wada and Sat0 (1974). This involved daily stimulation beginning with a 100-pA stimulus (1-s train of I-ms biphasic square waves at 60 Hz) applied to either an indifferent lead paired with an amygdala lead or through two leads in the ipsilateral amygdala. Intensity was increased by 100- to 200-FA increments until the first AD appeared and then decreased by 100-pA decrements until AD disappeared. One stimulus was applied daily at initial AD threshold intensity until a stage 6 seizure was provoked (GTC). We then reestablished AD and GTC thresholds, using the same procedures as for initial AD determination. Four modifications of our standard kindling protocol in adult cats compensated for high focal AD thresholds in young kittens (n = 4 at 15-20 mA). Changes involved (a) raising the ceiling on stimulus intensity from 2 to 40 mA, (b) reducing stimulus duration from 1 to 0.1 s at intensities >2 mA, (c) multiple daily stimulation to determine initial focal AD thresholds, and (d) bilateral amygdala stimulation of one kitten in which focal AD could not be evoked ipsilaterally at 40 mA. Spontaneous epilepsy was detected in kittens undergoing modified as well as standard kindling protocols, as previously reported (Shouse et al., 1990a,b). Accordingly, procedural adjustments did not appear to be critical to onset of spontaneous TLE. The first evoked stage 6 seizure (GTC) marks the end of initial kindling. None of the kittens had reached puberty by that time. Postkindling follow-up Kittens were followed from 3 days to > I year after initial kindling to document the postkindling course of spontaneous versus stimulus-evoked seizures and their timing in the sleepwake cycle. Surgical controls and adult cats had 9-24 month postoperative or postkindling observation. Spontaneous epilepsy was defined by seizures occurring at least 1 h after amygdala stimulation (Shouse et al., 1990a,b,c), as detected by routine behavioral, polygraphic, or split-screen video monitoring. Seizures first observed behaviorally were

791

DEVELOPMENTAL SLEEP EPILEPSY followed by polygraphic or split-screen video recordings and vice versa. Behavioral seizure detection was based on neardaily observation by laboratory personnel and by twice-daily clinical rounds at the Animal Research Facility (ARF). Six- to 36-h polygraphic or splitscreen video recordings were also performed in conjunction with stimulus-evoked or spontaneous seizures during the first 9 months after kindling (a mean of one each week) and were repeated at 1-5 months after the last known evoked or spontaneous seizure. Incidence and morbidity of convulsive status epilepticus were documented. Status, defined as two or more GTCs in 5 min or without intervening consciousness, was terminal (undetected at outset and intractable) or not terminal (detected soon enough for effective treatment with 35 mg/kg nembutal administered i.p. during the tonic phase of the GTC). Data and statistical analysis Population data were summarized for 24 of 26 kindled kittens and adults (Shouse et al., 1990b), but longitudinal factors have not been described before. This report focuses mainly on spontaneous GTCs because convulsions are more entrained to the sleepwake cycle than any other seizure manifestation (Janz, 1962, 1974; Shouse, 1989) and are also the most frequent spontaneous seizure type we observed in kindled kittens (Shouse et al., 1 9 9 0 ~ ) . Several factors were evaluated with respect to onset and postkindling course of spontaneous sleep or waking GTCs. Simple or repeated-measures ANOVA was used to evaluate distribution of spontaneous seizures in the sleepwake cycle as well as the contribution of various factors to the postkindling course. Pearson product-moment correlations was used to assess interactions between individual variables. Verification of electrode sites Electrode placements were verified histologically in the 13 kindled adults, visually or histologically in 7 cats kindled as kittens, and radiographically in the 6 kittens that were still under observation (n = 3 epileptic, 3 nonepileptic). Histologic confirmation of electrode sites was based on light microscopic examination of frozen coronal sections (30 km) stained with thionine. RESULTS Onset, frequency, and timing of spontaneous convulsions after kindling Onset of spontaneous GTCs after kindling Spontaneous convulsions developed 1 h to 4 months after initial kindling in 8 of 13 kittens (62%)

and persisted to adulthood in the surviving animals. Neither of the neurosurgical control animals manifested seizures during follow-up, and only 1 of 13 kindled adults had spontaneous GTCs (8%). Number of spontaneous GTCs after kindling One hundred spontaneous GTCs were detected during behavioral (n = 56), polygraphic (n = 24), or split-screen video monitoring (n = 20). Of 44 convulsions recorded on polygraph or split-screen video, five were obscured by electrical interference, leaving 39 convulsions that could be reliably classified for sleep or waking state dependency. Moreover, the epileptic adult cat and 2 kittens with terminal status epilepticus died before GTCs could be recorded, leaving 6 animals whose seizures could be verified in relation to sleep-wake cycle. Timing of spontaneous GTCs Table 1 shows the sleep-waking state distribution of 39 spontaneous GTCs in the 6 epileptic kittens with recorded seizures. Thirty-two GTCs were recorded during sleep (82%); only seven were detected during wakefulness (18%). Sleep GTCs occurred during SWS, mostly in the REM sleep transition, and none occurred during stable REM sleep. Time of seizure occurrence is further specified in Table 1 (e.g., SWS onset, transitions into vs. out of REM). Postkindling course of sleep versus waking convulsions Latency from evoked to spontaneous GTCs Most spontaneous convulsions (93%) were observed within 72 h after stimulus-evoked seizures, which were distributed throughout the first 9 TABLE 1. Sleep-waking state distribution of 39 spontaneous convulsions recorded on polygraph or split-screen video in 6 amygdala-kindled kittensa Sleepwaking state distribution Waking

SWS

REM transition

REM sleep

39

7

14

18

0

100

18’

35

47

0*

Parameter Total no. Percentage of total

GTCs, generalized tonic-clonic convulsions; SWS, slow wave sleep; REM, rapid eye movement. a GTCs were more likely to occur during sleep (82%) than waking (18%). Of 14 GTCs classified as SWS proper, two occurred 30-60 s after SWS onset, and 12 began during stable SWS (>1 rnin after SWS onset). The remaining 18 convulsions occurred during REM transition; 16 began during transition into REM, and 2 occurred in the transition out of REM. In the latter case, GTCs began 30-45 s after REM ended. No GTCs occurred during stable REM sleep ( 3 1 min after REM onset). p < 0.05 from SWS (including REM transition). Epilepsia, Vol. 33, N o . 5 , I992

M . N . SHOUSE ET AL.

792

months of the follow-up period. The other 7% were observed 5-9 months after the last evoked or detected seizure, including one GTC in the epileptic adult cat and six GTCs in the 3 epileptic kittens surviving past the initial 9-month postkindling follow-up. Kittens were mature (aged 2-3 years) at the time they exhibited the first spontaneous GTCs with prolonged interictal interval ( 2 5 months after last known seizure); nevertheless, 80% of these GTCs were detected during sleep. The timing of convulsions thus appeared comparable regardless of interictal interval. Seizure clusters Spontaneous convulsions typically occurred in

Severity of seizure disorder Table 2 summarizes data suggesting severity of seizure disorder as a prominent factor in unstable longitudinal seizure patterns. Cats that continued to exhibit convulsions exclusively during sleep (n = 3) had a history of infrequent or unclustered GTCs. In contrast, cats with GTCs during waking and sleep (n = 3) also had a history of frequent seizure clusters, with recurrent or terminal convulsive status and/or multiple seizure types. At least four indexes of severity were implicated in the onset and/or persistence of sleep or waking GTCs, including seizure incidence, density of seizure clusters, status epilepticus, and multiple seizure tvnes. These variahles are shnwn in Tahle 3

H . OGUNl ET AL.

808

rollowea by polygraphlc recordings.

Or

spllt-screen

‘ldeo

Sequential onset of sleep and waking GTCs Longitudinal data were consistent even though the first sleep or waking GTC was not recorded in 1 of the 6 kittens (17%) and in 5 of the 16 seizure clusters captured on polygraph or split-screen video. At first, convulsions were recorded exclusively during sleep; later, GTCs also occurred during wakefulness. Factors implicated in sequential onset of sleep and waking GTCs Several variables differentiated cats that retained “convulsive sleep epilepsy” from those that developed random seizure patterns, defined by GTCs in waking and sleep (Janz, 1962, 1974; Billiard, 1982; Shouse, 1989).

spontaneous GTCs per cat (Table 2), was higher in cats with sleep and waking GTCs than in those with chronic sleep epilepsy. The number of GTCs recorded during wakefulness also increased in relation to the total number of recorded convulsions per cat (r = 0.99, p < O.Ol), but the correlation was much lower after behaviorally detected seizures were included (total GTCs, r = 0.78, p < 0.05, Table 2); e.g., 1 kitten had 36 isolated convulsions during behavioral observation, but only two GTCs recurred during follow-up polygraphic or splitscreen video recordings, both during sleep. This result indicated seizure density rather than seizure incidence as a preponderant influence in onset of waking GTCs. Seizure density per cluster Seizure density, defined as the average number of GTCs per detection day (Table 2), was more closely

TABLE 2. Severe seizure manifestations (seizure incidence, density of seizure clusters, or number of convulsions per detection day, status epifepticus, and multiple seizure types“) differentiated chronic distribution of GTCs in the sleep-wake cycle in kittens with sleep convulsions only versus kittens with sleep and waking convulsions ( n = 3 eachIh Seizure incidence (mean f SD) GTCs Sleep (n = 3) Sleep and waking (n = 3)

Recorded GTCs

Seizure density (mean f SD) .-

Total GTCs

Recorded GTCs

Total GTCs

Convulsive status epilepticus (no. of kittens)

Multiple seizure tYPe.s (no. of kittens)

1.3

f

0.5

12.7

10

1.0 t 0.0

1.3

f

0.5

1

1

11.7

f

5.7b

23.0 t 12’

2.8 t 3.8‘

3.0

f

0.7‘

3

2

f

Abbreviations as in Table 1. Spontaneous GTCs coexisted with one of the following subconvulsive seizure types in 3 kittens (one subconvulsive seizure type each): Focal EEG seizures-shown in Fig. 1, (n = 10); complex-partial seizures-generalized EEG discharge with chewing, lipsmacking, limb clonus (n = 5 ) ; “catnip” seizures-diffuse spikes and multispike-wave complexes accompanied by sustained absencelike staring with hyperventilation, periodic head jerks (dorsoflexion), and occasional “jackknife” seizures (n = 2). Catnip seizures may be a variant of complex-partial epilepsy or of West syndrome (West, 1840-1841). Of these, increased density of recorded and total GTCs most consistently preceded dissociation of convulsions from sleep. p < 0.05 from kittens with sleep convulsions only. a

Epilepsia, Vol. 33, N o . 5 , 1992

DEVELOPMENTAL SLEEP EPILEPSY

associated with onset of waking convulsions than any other factor. The number of waking GTCs increased as a function of per capita seizure density, including recorded convulsions (r = 0.91, p < 0.01) and total GTCs (recorded plus behavioral seizure detection, Y = 0.94, p < 0.01,Table 2). Indeed, waking GTCs were documented only in animals with dense seizure clusters and were first recorded after three or more sleep GTCs earlier that day. After onset, however, waking GTCs sometimes recurred without previous sleep GTCs. Status epilepticus Convulsive status epilepticus (two or more GTCs in 5 min or without intervening consciousness) did not appear to be critical to sequential onset of GTCs during sleep and waking (Table 2). None of the kittens manifested convulsive status before the first recorded sleep GTC. Moreover, status preceded the first waking GTC in only 1 . Recurrent status was nevertheless common in cats with waking GTCs and may have influenced the persistence of waking and sleep convulsions after onset. Multiple seizure types Multiple seizure types did not appear to be essential to distribution of convulsions in the sleepwake cycle, even though cats with sleep and waking convulsions were more likely to manifest subconvulsive seizures (n = 2 of 3) than were cats with GTCs only in sleep (n = 1 of 3, Table 2). Two subconvulsive seizure types have been described previously (Shouse et al., 1990a,b,c) and are briefly characterized in the footnote to Table 2. A third electrographic seizure type is shown in Fig. 1. The timing of recorded GTCs was comparable in

793

cats with or without subconvulsive seizures (n = 3 each). The majority of GTCs occurred in sleep (>SO%), and a minority occurred during waking in both groups; in contrast, subconvulsive seizures manifested randomly in the sleepwake cycle (41% in waking, 47% in SWS, and 12% in REM sleep). On the other hand, the present findings may underestimate the role of undetected or detected seizure types as well as other seizure foci in sequential onset of sleep and waking GTCs. Subconvulsive seizures were detected before GTCs in 2 of 3 kittens, 1 of whom later manifested waking GTCs. Moreover, local geniculate seizures were recorded after onset of spontaneous convulsions in the cat with the most numerous LGN-onset GTCs (n = 4 of 5) and may have escaped earlier detection because of minor clinical accompaniment. Geniculate seizure discharge could occur at any time but rarely propagated during sleep; instead, focal LGN seizures during SWS often ended in REM sleep (Fig. 1). Anatomic seizure origin Table 3 summarizes anatomic localization of convulsions, defined by focal versus generalized onset at the three recording sites. GTCs with apparent generalized onset were recorded only during SWS, whereas GTCs with focal onset could occur during sleep or waking. Anatomic localization appeared to differentiate sleep from waking GTCs throughout the postkindling course. Sleep convulsions with focal onset almost always started in the amygdala (20 of 21 = 95%, Fig. 2). Conversely, waking convulsions with focal onset rarely began in the amygdala (n = 1 of

FIG. 1. Focal geniculate EEG seizures in the cat with the most frequent lateral geniculate nucleus (LGN)-onset generalized t o n i c - c l o n i c seizures (GTCs). A 7-s polyspike train occurred in the transition from slow-wave sleep to REM and ended at REM onset. Clinical accompaniment, if any, is underscored beneath the tracing and consisted of eye movements and/or muscle twitches that normally accompany clustered PGO spikes during active REM sleep. Paper speed was 10 mmis. EOG, electrooculogram; EMG, electromyogram.

Epilepsia, Vol. 33, N o . 5 , 1992

794

M . N . SHOUSE ET AL. TABLE 3 . Anatomic origin of sleep versus waking convulsions* Sleep-waking state Waking n(%)

SWS n(%)

REM transition n (%I

Generalized onset (n = 11) Focal onset (n = 28) Motor cortex or LGN (n = 7) Amygdala (n = 21)

0 (0)

6 (38)

5 (28)

6 (83) l(17)

0 (0) 8 (62)

1 (6) 12 (66)

Total

7 (100)

14 (100)

18 (100)

Anatomic origin

n = 39 polygraphic or split-screen video monitored GTCs LGN, lateral geniculate nucleus; other abbreviations as in Table 1. a Simple analysis of variance showed that sleep GTCs were more likely to originate from the kindled temporal lobe, whereas waking GTCs were recorded first from extrafocal structures.

7); instead, initial EEG seizure manifestations were initially recorded from structures outside the temporal lobe, including either the LGN or motor cortex (6 of 7 = 86%, Fig. 3 ) . The number of waking GTCs with recorded onset

FIG. 2. A continuous 4-min tracing showing first spontaneous generalized tonic-clonic convulsion (GTC) in an arnygdala-kindled, preadolescent kitten. The GTC occurred during stable slow wave sleep, evidenced by frequent sleep spindles over motor cortex for nearly 2 min beforehand. The GTC originated from the kindled amygdala (AMY, channel 3); onset is identified by the solid arrow near the end of the second tracing. lctal discharge propagated rapidly to the contralateral AMY and motor cortex and was detectable in the ipsilateral lateral geniculate nucleus (LGN; channel 5) 7 s later. The contralateral LGN and electromyogram were not recorded until later in the day. Paper speed was 10 mm/s.

Epilepsia, Vol. 33, No. 5, 1992

outside the amygdala increased as a function of frequent seizure clusters. LGN was the first detected and most common extrafocal site that appeared to initiate waking GTCs (n = 5 ; at least one each in the 3 kittens with waking GTCs); the other two waking GTCs were recorded first from motor cortex or amygdala (n = 1 each) in the animal with the highest seizure density (mean = 4.5). Ontogenetic factors The severity and timing of spontaneous epilepsy were differently affected by immaturity, at least when compared as a function of age at initial AD (2.5-4 months vs. 5-6.5 months at initial AD; n = 3 each). Youth appeared to facilitate onset of epilepsy with severe complications because the 3 younger kittens were more likely than the 3 older preadolescents to develop multiple seizure types (67 vs. 33%), terminal status epilepticus (67 vs. 33%), and multifocal epilepsy with sleep and waking GTCs (100 vs. 0%). In contrast, seizure density and recorded seizure origin were associated with sleep-waking state dis-

795

DEVELOPMENTAL SLEEP EPILEPS Y

EOG AMY AMY LG N

LGN

EMG

a

tribution of GTCs throughout preadolescence. Kittens with low seizure densities retained “pure sleep epilepsy” (mean seizure density = 1.0 5 0 in both age groups), and kittens with high seizure densities progressed to random timing patterns regardless of age at initial AD (2.7 1 0 in younger kittens vs. 2.3 & 1 in older kittens). Similarly, most sleep GTCs were first recorded from the amygdala at both ages (100 vs. 91%), whereas most waking GTCs appeared to begin in structures outside the temporal lobe (100 vs. 86%). Morphologic changes Morphologic alterations were estimated by light microscopic examination of histologic sections from cats kindled as preadolescents (n = 5 ) versus adults (n = 13) or by radiograph. Electrode placements and size of lesions at electrode tips were similar in nonepileptic kittens and adults. Sites adjacent to electrode sites also seemed intact, as long as the animals were nonepileptic and healthy at time of death (Fig. 4). Extensive cell loss was more common after kindling of kittens (n = 4 of 5 ) than of adults (n = 2 of 8). The 4 kittens with marked structural pathology also had a history of spontaneous epilepsy, whereas only 1 of the 2 adults with cell loss was epileptic, but the 4 kittens with cell loss also died of terminal

FIG. 3. A continuous 3-min tracing showing the first spontaneous generalized tonic-clonic seizure (GTC) during waking in the same preadolescent kitten described in the legend to Fig. 2. The contralateral lateral geniculate nucleus (LGN) and electromyogram (EMG) were added (channels 5 and 7 in this tracing) and gains adjusted after three sleep GTCs earlier in the day. The tracing begins with 9 s of slow wave sleep (dense cortical sleep spindles are apparent) and -1.5 min of sustained waking, characterized by EEG desynchronization with muscle tone. The arrow beneath the second tracing denotes the beginning of the GTC, which appeared to originate with a high-voltage discharge in the LGN ipsilateral to kindled amygdala (AMY) (channel 6 in this tracing). lctal discharge spread rapidly to motor cortex; sustained ictal discharge was delayed at other subcortical recording sites, including contralateral LGN and AMY bilaterally. Paper speed was 10 mmls.

status or infectious disease before brain perfusion. Similar attrition was apparent in sections from the two kindled adults evaluated under comparable postmortem conditions, regardless of spontaneous epilepsy (Fig. 4). Roentgenograms taken under ketamine anesthesia verified stereotaxic coordinates and stability of depth electrode placements in kittens still under observation. Figure 5 is a lateral radiograph of a 1.5year-old, epileptic cat kindled at age 4 months. Immobility of chronic implants suggested comparable brain size in preadolescent and adult cats. DISCUSSION Kittens with amygdala-kindled TLE manifested a substantial majority of convulsions during sleep (82%), regardless of coexisting seizure types. Similar findings have been reported in association with secondarily generalized TLE in the clinical literature (Janz, 1962, 1974; Billiard, 1982; Shouse, 1989), yet convulsions were documented during wakefulness in 50% of epileptic kittens with recorded seizures and in 42% of patients with TLE (Janz, 1962). The basis for unstable longitudinal seizure patterns is unknown, but our findings lend credence to clinical data suggesting frequent seizures, Epileppsia, Vol. 33, No. 5 , 1992

796

M . N . SHOUSE ET AL.

multiple foci, and early onset as potential contributing factors (Janz, 1962, 1974; Shouse, 1989). Conclusions are based on data from only 6 kindled kittens and are further limited by the number of seizures and anatomic sites recorded. Sequential analysis nevertheless showed some systematic trends. At first, spontaneous GTCs appeared to be entrained to amygdala discharge propagation during sleep. Later, however, convulsions became disentrained from the kindled amygdala and from sleep onset. Severe seizure manifestations, such as frequent seizures, high-density seizure clusters, status epilepticus, and multiple seizure types, accompanied shifts in seizure timing. However, high seizure density was the most reliable precursor of “waking and sleep epilepsy,” defined by the random occurrence of convulsions during waking and sleep (Janz, 1962, 1974; Billiard, 1982). Afterward, spontaneous waking convulsions could occur without prior sleep GTCs and were still observed at maturity, even 5-9 months after the last known seizure. This result is compatible with clinical data indicating that once GTCs dissociate from the sleep-wake cycle random convulsive seizure patterns can be permanent (Janz, 1962, 1974; Shouse, 1989). Frequent or clustered seizures could influence longitudinal seizure patterns through various mechanisms. Of these, anatomic seizure origin appeared to be the most persistent correlate. The first spontaneous waking GTCs were recorded after multiple sleep GTCs on the same day and did not originate in the amygdala; based on the limited electrode montage, ictal onset appeared to start in the thalamic LGN (72%) or motor cortex (14%). Moreover, subsequent waking convulsions still appeared to begin at sites other than the amygdala, whereas sleep con4A

4B

FIG. 4. Coronal sections through the amygdala show electrode placements, lesions at electrode tips and cell loss as a function of age at kindling. Histology is shown for 3 cats kindled as kittens (A) or as adults (8): healthy, left; with status epilepticus, middle; and with infectious disease, right. Kindled cats were either healthy at time of death (nonepileptic), had terminal status epilepticus, or died of infectious disease. Epilepsia, Vol. 33, No. 5 , 1992

FIG. 5. Lateral roentgenogram (performed under ketamine anesthesia) showing stability of depth electrodes in amygdala and lateral geniculate nucleus in an adult cat kindled as a preadolescent kitten.

vulsions appeared to originate from the amygdala throughout the postkindling course. Undetected seizure foci cannot be excluded as a factor, but results resemble normative clinical data suggesting temporal lobe origins for most sleep GTCs with focal onset; in contrast, extratemporal foci generate most secondarily generalized convulsions in humans (Janz, 1962, 1974). Localization of waking convulsions to unstimulated structures suggests the secondary kindling of multifocal epilepsy. Secondary site epileptogenesis is consistent with the concept of transsynaptic kindling effects. Kindling produces permanent pathophysiologic changes at sites distal to the stimulated electrode, evident by rapid secondary site kindling or transfer (Cain, 1986) and by subtle morphologic changes at extrafocal sites in mature animals (Geinisman et al., 1988; Sutula et al., 1988; Sutula and Steward, 1987, 1989). Transfer may be accelerated by youth because kindled kittens developed active spontaneous seizure foci in previously unstimulated tissue. The mechanism may involve increased morphologic change at these sites, potentially related to developing inhibitory mechanisms and/or increased neural plasticity in the young (MoshC et al., 1992). Multifocal interactions may have influenced chronic seizure patterns, because focal seizure discharge can potentiate or interfere with propagation at other sites and is believed to influence generalization (Burchfiel et al., 1986). Accordingly, detected or undetected seizure foci and subconvulsive seizures could have facilitated or suppressed recurrent GTCs at different times.

797

DEVELOPMENTAL SLEEP EPILEPSY

Pronounced structural pathology may have contributed to primary and/or secondary epileptogenesis in epileptic kittens, but the histologic consequences of frequent or terminal bouts of status epilepticus cannot be overlooked. Status induces cell loss in kindled rats (McIntyre et al., 1982) and may influence clinical course independently of the kindling process. Collectively, the longitudinal data show that complications occur with early onset TLE in cats, potentially mediated by diverse neural mechanisms. Multifocal epilepsy tends to have an unfavorable prognosis and to be medically refractory in humans, particularly when seizures are disentrained from a specific sleep or waking state (Janz, 1962; Niedermeyer, 1987; Shouse, 1989). Kindled kittens also displayed terminal status epilepticus, multifocal seizure discharge, and random temporal seizure patterns. Findings thus suggest that TLE in immature organisms need not have a benign course; indeed, onset during youth may even increase the risk of developing a progressive, intractable seizure disorder. Acknowledgment: This w o r k was supported b y t h e Veterans Administration a n d PHS Grant No. NS 25629. We thank Diane Mahadeen, Paul Farber and Dr. Yuji W a d a for editorial assistance, Alison King a n d Kenneth King for technical laboratory assistance, Drs. John Young and Melvyn Richkind f o r veterinary supervision and advice, and Evelyn Beninati and Susan Nichols for daily animal care and monitoring. Joe Holston, AAALAS certified technologist, was a special consultant o n this project.

REFERENCES Billiard M. Epilepsies and the sleep-wake cycle. In: Sterman MB, Shouse MN, Passouant P, eds. Sleep and epilepsy. New York: Academic Press, 1982:269-86. Burchfiel JL, Appelgate CD, Konkol RJ. Kindling antagonism: a role for norepinephrine in seizure suppression. In: Wada JA, ed. Kinding 3. New York: Raven Press, 1986213-30. Cadhillac J. Complex partial seizures and REM sleep. In: Sterman MB, Shouse MN, Passouant P, eds. Sleep and epilepsy. New York: Academic Press, 1982:315-24. Cain DP. The transfer phenomenon in kindling. In: Wada JA, ed. Kindling 3. New York: Raven Press, 1986:21345. Calvo JM, Alvarado R, Briones R, Paz C, Fernandez-Guardiola A. Amygdaloid kindling during rapid eye movement (REM) sleep in cats. Neurosci Lett 1982;29:255-9. Geinisman Y, Morrell F, de Toledo-Morel1 L. Remodeling of synaptic architecture during hippocampal “kindling.” Proc Natl Acad Sci U S A 1988;85:32604. Hauser WA, Kurland LT. The epidemiology of epilepsy in Rochester, Minnesota, 1935 through 1967. Epilepsia 1975;16:146. lnglis JK. Introduction to laboratory animal science and technology. New York: Pergamon Press, 1980. Janz D. The grand ma1 epilepsies and the sleeping waking cycle. Epilepsia 1962;3:69-109. Janz D. Epilepsy and the sleeping-waking cycle. In: Vincken PJ, Bruyn GW, eds. Handbook of clinical neurology. Amsterdam: Elsevier, 1974:457-90. (The epilepsies, vol. 15.)

McIntyre D, Nathanson D, Edson N,A new model of partial status epilepticus based on kindling. Bruin Res 1982;250:5363. MoshC SL, Sperber EF, Albala BJ. Kindling as a model of epilepsy in developing animals. In: Morrell F, ed. Kindling and synaptic plasticity. Boston: Dirkauser 1992 (in press). Niedermeyer E. Epileptic seizure disorders. In: Niedermeyer E, Lopes da Silva F, eds. Electroencephalography, 2nd ed. Baltimore: Urban and Schwartzenberg, 1987:405-5 10. Rose GH, Goodfellow EF. A stereotaxic atlas of the kitten brain: coordinates of 104 selected structures. Los Angeles: Brain Information Service/Brain Research Institute, University of California, Los Angeles, 1973. Sato M, Nakeshima T. Kindling: secondary epileptogenesis, sleep and catecholamines. Can J Neurol Sci 1975;3:43946. Shouse MN. State disorders and state dependent seizures in amygdala-kindled cats. Exp Neurol 1986;91:601-9. Shouse MN. Thalamocortical mechanisms of state dependent seizures during amygdala kindling and systemic penicillin epilepsy in cats. Epilepsia 1987;28:399408. Shouse MN. Seizures and epilepsy during sleep. In: Dryger MH, Roth T, Dement WC, eds. Principles and practice of sleep medicine. Philadelphia: W.B. Saunders, 1989:104-20. Shouse MN, King A, Langer J, Vreeken T, King K , Richkind M. The ontogeny of feline temporal lobe epilepsy 1. Kindling a spontaneous seizure disorder in kittens. Brain Res 1990a;525: 215-24. Shouse MN, Langer JV, Dittes PR. Spontaneous sleep epilepsy in amygdala kindled kittens: a preliminary report. Brain Res 1990h;535:163-8. Shouse MN, King A, Langer J, et al. Basic mechanisms underlying seizure prone and seizure resistant sleep and awakening states in feline kindled and penicillin epilepsy. In: Wada JA, ed. Kindling 4 . New York: Plenum, 1990c:313-27. Snider RS, Niemer WT. A stereotaxic atlas of rhe cat brain. Chicago: University of Chicago Press, 1961. Stevens JR, Lonsbury BL, Goel SL. Seizure occurrence and interspike interval. Arch Neurol 1972;26:409-19. Sutula T, Steward 0. Quantitative analysis of synaptic potentiation during kindling of the perforant path. J Neurophysiol 1986;56:73246. Sutula T, Steward 0 . Facilitation of kindling by prior induction of long-term potentiation in the perforant path. Brain Res 1987;420:109-17. Sutula T, Xiao-Xian H, Cavozos J, Scott G. Synaptic reorganization in the hippocampus induced by abnormal functional activity. Science 1988;239:1147-50. Wada JA, Sato M. Generalized convulsive seizures induced by daily stimulation of the amygdala in kindled cats. Neurology 1974;24:565-74. West WJ. On a peculiar form of infantile convulsion. Lancet 1840-184 1;1 :724-725. Woodbury LA. Incidence and prevalence of seizure disorders including the epilepsies in the U.S.A. A review and analysis of the literature. In: Plan for the Nationwide Action of Epilepsy, vol. IV, DHEW Publication No. (NIH) 78-276, 1977: 24-77.

&SUME Les auteurs ont rCcemment dtcrit un modtle d’kpilepsie spontanCe du sommeil chez des chatons ayant benCficiC d’un kindling, et presentant une Cpilepsie du lobe temporal (ELT). Dans ce travail ils dkcrivent I’evolution aprks kindling de ce modele, depuis l’adolescence jusqu’a la maturitC et sugg6rent les mtcanismes physiopathologiques impliguks. Une Cpilepsie spontanCe, comportant en particulier des crises gCnCralisCes tonicocloniques (GTC) s’est installCe une heure 4 mois aprks kindling amygdalien, et a persist6 jusqu’a l’fige adulte. Au debut, GTC n’Ctaient dCtectCes que pendant le sommeil; plus tard, les convulsions sont survenues Cgalement a 1’Ctat de veille. Deux facteurs Ctaient nettement associts a I’installation sequentielle des Epilepsia, Vol. 33, N o . 5 , 1992

M . N . SHOUSE ET AL.

798

GTC du sommeil et de 1’6tat de veille. 1”) Les sCries de crises. Les chats ne presentant que des crises rares et non organisees en series ont continue B avoir une “epilepsie du sommeil”, dCfinie par l’existence de convulsions survenant exclusivement pendant le sommeil. A I’opposC, les chats presentant de frkquentes sCries de crises ont d6veloppC des Etats de Ma1 recurrents ou terminaux, associCs B des GTC survenant a la veille et pendant le sommeil. Des manifestations critiques sCveres semblent donc avoir contribue B la dissociation entre les convulsions et le cycle veille-sommeil. 2”) La localisation anatomique des crises. L’origine focale des crises semble avoir diffkrencit les crises du sommeil et celles de la veille. Le debut pendant le sommeil Btait d’abord enregistre au niveau de I’amygdale qui avait subi le kindling, alors que I’installation pendant la veille Ctait initialement dttectke en dehors du lobe temporal. Ces constatations sugggerent la survenue d’un kindling secondaire dans le cadre d’une Cpilepsie multi-focale. Une CpileptogCnese secondaire est cohCrente avec des effets de kindling trans-synaptiques. Ce phCnomt n e est defini chez I’animal adulte par un kindling rapide ainsi que secondaire (transfert) et par des modifications morphologiques discrbtes a distance de 1’6lectrode de stimulation. Ce transfert peut &re accentue par l’bge jeune, car les chatons developpent spontankment des foyers critiques au niveau de zones non stimulees auparavant. De plus, des interactions multifocales et des pertes neuronales diffuses ont CtC impliquees comme mtcanismes possibles. Globalement, ces donnCes mettent en evidence des complications dans l’kvolution d’une ELT prCcoce chez le chat apres kindling. Le debut a un bge jeune peut comporter un pronostic dtfavorable, qui se traduit par des Etats de Ma1 recurrents et par une Cpilepsie multifocale avec convulsions survenant tout au long du cycle veille-sommeil.

(P. Genton, Marseille)

RESUMEN Los autores describen un modelo de “epilepsia del suefio” espontanea en gatos condicionados (kindled) con epilepsia en 16bulo temporal (TLE). Describen la posibilidad de una evoluci6n post-kindling en este modelo desde la preadolescencia a la madurez y se proponen 10s mecanismos patofisiologicos. La ep-

Epilepsia, Vol. 33, No. 5 , 1992

ilepsia esponthnea, particularmente generalizada con convulsiones tonico-clonicas (GTCs), se produjo desde una hora a cuatro rneses despuCs del “kindling” de la amigdala y persistio hasta la edad adulta. A1 principio 10s GTCs se detectaron solamente en el suefio; mhs tarde las convulsiones podian tambiCn ocurrir durante la vigilia. Dos factores se asociaron repetidamente con el comienzo secuencial de las GTCs durante el suefio y el despertar. (1) Ataques acumulados. Gatos con GTCs y frecuentes o dispersos, es decir, no acumulados, continuaron exhibiendo “epilepsia del suefio” definida por convulsiones que ocurrian exclusivamente durante el suefio. En contraste 10s gatos con frecuentes ataques acumulados desarrolbaron un status convulsivo recurrente o terminal en union a 10s GTCs durante el despertar y el sueno. Las severas manifestaciones de 10s ataques, parecen haber contribuido a la disociacion entre convulsiones apareciendo durante el ciclo suefio-vigilia. (2) Localizacidn anatdmica de 10s ataques: Parece ser que el origen focal de 10s ataques ayuda a diferenciar 10s GTCs durante el suefio de 10s que ocurren en la vigilia. El comienzo durante el sueno fue inicialmente registrado en la amigdala condicionada mientras que el comienzo durante el despertar fue inicialmente detectado en zonas alejadas del16bulo temporal. Estos halkazgos sugieren que existe una epilepsia multifocal secundaria a1 “kindling”. La epileptogenesis secundaria es consistente con efectos “transsinapticos” del condicionamiento. Este fenomeno se define en animales maduros por la rapida aparicion de una zona secundaria de “kindling” (transferencia) y cambios morfol6gicos sutiles alejados del electrodo de estimulacion. Esta transferencia puede ser acentuada en edades jovenes de 10s animales puesto que 10s gatos de corta edad desarrollaron ataques espontkneos focales en zonas previamente no estimuladas. Ademas las interacciones multifocales y la pCrdida difusa de cClulas se interpretaron como posibles mecanismos implicados en este proceso. El conjunto de estos hallazgos anticipan posibles complicaciones con el comienzo precoz de TLA en 10s gatos condicionados. El comienzo durante edades jdvenes puede proporcionar pronosticos desfavorables manifestados por status epilepticus recurrente y epilepsia multifocal con convulsiones distribuidas a lo largo del ciclo suefio-vigilia. (A. Portera-Sanchez, Madrid)