~ Pergamon 0306-4522(94)00329-7

Neuroscience Vot. 64, No. 2, pp. 301 317, 1995 Elsevier Science Ltd

Copyright 4") 1994 IBRO Printed in Great Britain. All rights reserved

0306-4522/95 $9.50 + 0.00

S P I K E - A N D - W A V E EPILEPSY IN RATS: SEX D I F F E R E N C E S A N D I N H E R I T A N C E OF PHYSIOLOGICAL

TRAITS

G. J A N D O , * t D. CARPI ,* A. K A N D E L , * R. URIOSTE,* Z. H O R V A T H , * t E. P IERRE,* D. VADI,* C. VADASZ~ and G. BUZSAKI*§

*Center for Molecular and Behavioural Neuroscience, Rutgers, The State University of New Jersey, 197 University Avenue, Newark, NJ 07102, U.S.A.

~Laboratory of Neurobehavioural Genetics, Nathan Kline Institute for Psychiatric Research, Orangeburg, NY 10962 and Department of Psychiatry, New York University Medical Center, New York.

NY 10016, U.S.A.

Abstrac~Spontaneously occurring spike-and-wave patterns were examined in seven to eight-month-old rats of the inbred Fischer 344 and Brown Norway strains and their F 1 and F2 hybrids. Neocortical activity and movement were monitored for 12 night h. Spike-and-wave episodes were identified by a three-layer back-propagation neural network. The incidence, average duration and total duration of spike-and-wave episodes were significantly higher in F1 males and F2 hybrids than in the parental strains. Male rats of the Brown Norway strain had significantly more and longer episodes than females, whereas no sex differences were present in Fischer rats. The average intraepisodic frequency of spike-and-wave patterns was significantly lower in Fischer rats than in the other groups and significantly higher in males than females. Tremor (myoclonic movements) associated with spike-and-wave episodes was absent or of very small amplitude in Fischer rats but frequent and of large amplitude in Brown Norway rats and their FI and F2 descendants. Most of the interstrain differences were limited to male rats. Spike-and-wave episodes recurred at predictable short-term (10-30 s) and long-term (15 30 min) periods. The long-term oscillation corresponded to a similar fluctuation of motor activity. The maximum probability of spike-and-wave patterns occurred at a relatively narrow range of delta power (0 3.1 Hz) of the background EEG activity. Systemic administration of the adrenergic alpha-2 agonist, clonidine, increased the incidence of spike-and- wave episodes several-fold. The total duration of spike-and-wave episodes in the clonidine sessions (15 min) and night sessions (12 h test) correlated significantly. We suggest that several genes interact with maturational, environmental and endocrine factors, resulting in sex differences, and produce the variety of EEG and behavioral findings encountered. In addition, we submit that the clonidine test may be useful in genetic investigations of human absence epilepsies. The findings of this work demonstrate that genetic manipulation of rodents is a promising method for producing analogous models for the various forms of human absence epilepsies.

General ized petit mal epilepsy is characterized by loss of consciousness ("absence") and bilaterally sym- metrical spike-and-wave E E G discharges at 3-4 Hz. 2°'6° In pat ients with idiopathic seizures, bra in damage or behavioral d is turbances are not detected, but their family histories are frequently positive for epilepsy. 17'37'46'47 Fou r syndromes comprise absence

epilepsies in humans . The most c o m m o n form: (1) chi ldhood absence epilepsy, ~4 is character ized by frequent seizures with bilateral 3 Hz spike-and-wave bursts, p redominan t ly in three- to 12-year-old girls. 25'2s'61"65 (2) Juvenile absence epilepsy with onset

after 10 years of age is character ized by less frequent absences and somewhat faster (3.5-4 Hz) spike-and-

?Present address: Department of Physiology, Medical School, P6cs, Hungary.

§To whom correspondence should be addressed. Abbreviations: BN, Brown Norway strain; EEG, electro-

encephalogram; F344, Fischer 344 strain; FFT, Fast Fourier transformed; HVS, high-voltage spike-and-wave spindle.

wave patterns. Boys and girls are affected equally. Abou t 16% of these cases are complicated by my- oclonic seizures. 14'78 (3) Juvenile myoclonic epilepsy has an age specificity a round puber ty (eight to 18 years), with a clear genetic predisposit ion. 3 Occasion- aliy, it may follow juvenile absence epilepsy 79 and the clinical seizures may persist into adul thood. (4) Epi- lepsy with myoclonic absences is character ized by severe bilateral rhythmical myoclonia of the extremi- ties associated with symmetrical 3-4 Hz spike-and- wave bursts. It occurs p redominan t ly in boys at about seven years of age and these at tacks are often resist- ant to therapy. 14 Several authors have suggested that the above clinical forms are not discrete entities but form a cont inuum. 3'25'62 It is possible tha t the more serious clinical manifes ta t ions result from a combi- nat ion of gene(s) and other genetic and /o r environ- mental effects.

Genet ic animal models may play a pivotal role in elucidating the etiologic mechanisms of petit mal epilepsies. Mouse mutan t s have long been used in

301

302 G. Jand6 et al.

at tempts to elucidate genetic factors under lying the electrophysiological manifes ta t ion of the disease. 51,52 Unfor tuna te ly , a variety of neurological deficits, unrela ted to h u m a n absence epilepsy, are also associated with these mutat ions , often making in- terpreta t ions difficult. 45's4 General ized neocort ical high voltage spike-and-wave spindles (HVS) in rats, 34 is a widely accepted an imal model of h u m a n idio- pathic, generalized absence epilepsy. 12,4°,41 These

6-10 Hz discharges meet the criteria of epileptic p h e n o m e n a as being " spon taneous , episodic, recur- rent and paraxysmal" . 21 Impor tan t ly , the pha rmaco- logical profile of h u m a n absence epilepsy and HVS in rodents is remarkably similar. 4°,49,5s,59,75 Significant

genetic var ia t ions of HVS expression were found in strain compar i son studies, suggesting the presence of a robus t genetic effect. 11'29m'57 Whereas these rodent models can help in elucidating the electroclinical mani fes ta t ion of petit real epilepsies, they do not necessarily mimic the clinically more impor t an t be- haviora l d is turbances of absence seizures. Developing clinically more relevant models are of the u tmos t impor tance since most behaviora l problems associ- ated with atypical absence seizures are often resistant to drug t reatment . 63

In the present experiments we invest igated physio- logical characterist ics and inher i tance pa t te rns of spike-and-wave activity in inbred Fischer 344 (F344) and Brown Norway (BN) rats and their F1 and F2 descendents. These strains were used because our prel iminary observat ions revealed marked differences between the m o t o r correlates of HVSs in these ani- malsJ 3 In addi t ion, we asked whether shor t tests of drug- induced E E G changes can predict the incidence of HVS in long drug-free sessions in a given individ- ual.

L = _+_ 2.5 mm; relative to bregma). Two additional stainless steel screw electrodes were driven into the skull above the cerebellum and served as ground and indifferent electrodes.

Physiological testing and data analysis

At least five days after surgery animals were placed in clear plastic testing chambers (30cm diameter cylinders) where recording sessions took place. Animals were able to move freely at all times. Tremor and movement were recorded with an electrothermomechanical film accelero- meter (gift of Dr Lasse Raisanen, University of Kuopio, Finland) which served as the floor of the recording cylinder. The accelerometer was sensitive enough to detect vibrissae movements. The floor was covered with woodshavings. The cylinder was placed in a sound attenuated box with a one-way mirror. Two rats were tested simultaneously in two separate recording chambers. In preliminary experiments we found that the probability of occurrence of HVS was significantly higher during night hours than during day hours, confirming previous observations in other strains of rats. tg'4t Therefore, all rats in the present experiments were tested during 12 night h (from 8 p.m. to 8 a.m.). Neocortical EEG was amplified (2,000 x ), filtered (between 1 and 50 Hz) and sampled at 100 Hz with 12 bit precision and stored on high capacity optical disks together with the analog output of the motor activity recorder.

Data analysis was carried out off-line on a RS 6000 computer. HVSs were recognized by a three-layer backprop- agation network (Fig. 1). Details of the network have been published previously. 3° Briefly, HVSs and explicitly non- HVS epochs were manually selected from the EEG records of 16 representative animals (training set). Fast Fourier transformed (FFT) data of continuously shifting time slices of 160 ms EEG epochs (16 data points) were analysed by the network. Pilot experiments indicated that such frequency resolution was adequate for the representation of the invari- ant features of HVS. 3° Both the real and imaginary parts of FFT were presented to the network. For developing the network, several tests were carried out by varying the number of inputs, hidden layer cells and output cells. The summed squared error between the predicted and pre- evaluated output was used as a measure of the performance of the network in predicting HVS epochs. The network was trained and tested extensively and recognized HVS with better than 98% probability) ° HVSs episodes detected by

EXPERIMENTAL PROCEDURES

The experiments were performed on 195 rats. Breeding pairs of the highly inbred F344/NHsd rats and B N/SsNHsd rats were obtained from Harlan Spragu~Dawley, Inc., (Indianapolis, IN). All animals used in this experiment, the F344 and BN rats, and their F344oXBNo) FI and (F1XF1) F2 hybrids, were produced and maintained at the Animal Facility of the N. S. Kline Institute for Psychiatric Research (Orangeburg, NY) in an air conditioned room under a 12 h/12 h light/dark cycle (lights offat 6.00 p.m.), with freely available water and standard lab chow. At least two weeks before surgery, experimental animals were shipped to Rut- gers University, where physiological tests were carried out at the age of 7.6 + 0.8 months ( + S.E.). After completing the experiments, spleens of F344, BN, FI rats and individually identified F2 animals were dissected and stored at -80°C for future gene-mapping experiments.

Surgical procedures

Rats were anesthetized with a mixture of ketamine (25 mg/ml), xylazine (I.3 mg/ml) and acepromazine (0.24 mg/ml) at a dosage of 4 ml/kg of weight. Following anesthesia, the animals were placed in a Kopf stereotaxic instrument. For recording cortical field potentials (EEG), four stainless steel epidural electrodes were placed bilater- ally in the skull (AP=0 .0mm, L = _+4.5mm; A P - 2 . 5 ,

Fig. 1. Analysis of HVS by the 3-layer backpropagation network. Top: a representative HVS episode recorded from a male F344 rat. Contiguous samples of EEG (boxed area) were transformed into the frequency domain by an FFT and these data were then presented to the input cells (t"0-1"7) of the network. The output of the network indicated the presence and the duration of HVS episodes (bottom trace). The network was trained and tested extensively and recognized

HVS with better than 98% probability. 3°

Spike-and-wave epilepsy in rats 303

the network were inspected individually and the onset and offset of HVSs were corrected manually, when necessary. In addition, artifacts falsely detected by the network as HVSs were deleted. The number of HVSs in 12 h (nj) and the average duration of HVSs were calculated in every rat.

The cumulative sum of all HVSs in a session will be referred to as "total duration". The total duration of HVS in jth animal was calculated as

nj

(E,, - Bo)s, i - I

and the average duration is

~ (E,j - B,j)s, t -~ nj

where B and E are the beginning (onset) and the ending (offset) samples of a given HVS, "s'" is the real time value between two consecutive samples (e.g., 10ms at 100Hz sampling rate), " i " is the HVS count (i = 1,2,3,4...nj). Note that the number, average duration and total duration of HVS are interdependent, since total duration = number of HVS × average duration. Of these, the total duration is the most reliable measure since the determination of the offset

and onset of closely spaced HVSs was sometimes difficult. If two successively detected events by the network were separated by less than 0.5 s intervals, then the two events were arbitrarily regarded as a single HVS. Such a bias decreased the number of HVS and increased the average duration.

The temporal dynamics of HVS was examined by recurrent maps of the HVS episodes. 65" Consider x(i) as the interval between two consecutive HVSs, so X ,= (B(~+ l~j -- B•)s, where B is the beginning sample of HVS, i is the count of HVS and j is the identification of a given animal. A plot of x~+ 1 as a function of x, provides a recurrent map of the intervals between successive HVS episodes in all animals. Normally distributed patterns of points around a certain value indicate periodic reoccurence of events in such maps. Randomly distributed points indi- cate unpredictable occurrences of events.

Frequency changes within individual HVS episodes were calculated as a function of time from the onset to the offset of HVS (i.e. between B~ and El) ). The power spectrum of HVS usually consisted of four harmonic peaks (Fig. 2C). The maximum power of the first component was considered to be the dominant frequency (]At~j) in a given time window

B

_ m

1 1

movement

C

~0 10 20 30 40 , " ~ 50

: f requency (Hz)

tromor rebound

-2 -1 1 2 -2 -1 0 1 2

sec sec

Fig. 2. A. HVS and movement pattern during slow wave sleep. Note low frequency (2 Hz) rhythmic activity before HVS onset. Note also relationship between EEG and movement. The larger boxed area is shown at an expanded time scale in B. Arrows indicate relationship between HVS and movement. Inset: cross-correlation between HVS and tremor movement. (C) Power spectrum of 512 ms EEG epoch. Note dominant frequency at 8 Hz and three harmonic peaks. D. Perievent histograms of movement at HVS onset (tremor) and HVS offset (rebound). Vertical dotted lines indicate onset of offset of HVS. The timescale in D also applies to A. At, distance of the sliding window (640 ms) from the onset of HVS.

304 G. Jand6 et al.

(640 ms) at At distance from B 0 in ith HVS of the j th animal (Fig. 2A). The average frequency is expressed as

i - I H]

The average frequency was plotted as a function of At for each animal.

Whole body movement, as registered by the accelerome- ter, was continuously recorded together with EEG (Fig. 2A, B). The relationship between movement and HVS was quantified at the onset and offset of HVS. The analog trace was rectified and integrated. The summmed integral of movements occurring 2 s before and after the onset and offset of HVS was calculated for each HVS in all rats. Since rats were always motionless before the onset of HVS, this epoch was considered as the zero movement (baseline). Using the onset and offset of HVS as triggers, perievent histograms of movement were calculated (Fig. 2D). If Mk is the kth sample of movement then function

nj

Z IMkl i - I

with (B o - 200 < k < By+ 200) results in a perievent his- togram spanning from 2 s before to 2 s after the onset of HVS (Fig. 2D). Statistical significance was judged by ANOVA and t-tests.

Since movement during HVS was often rhythmic (i.e. tremor), the rhythmic movement pattern was cross-corre- lated with the peaks of HVS (Fig. 2B). The analog signal of the movement detector was differentiated and a threshold function was used to obtain digital events. The peaks of the HVS were detected by a peak searching algorithm and the resulting pulses served as reference events for the cross-cor- relating HVS with other events.

Some rats lost their electrode caps before the completion of the experiments and data from six rats were lost due to disk failure. As a result, physiological data were obtained from 183 rats (17 BN males; 14 BN females; 16 F344 males; 16 F344 females; 16 FI males; 21 F1 females; 44 F2 males; 39 F2 females).

Drug sessions were carried out two days to one week after the night recording session. Clonidine (Sigma) was given intraperitoneally at 0.05 mg/kg body weight 15 min prior to the recording session. EEG and movement were recorded for 15 min.

R E S U L T S

Previous research revealed tha t HVSs recorded from virtually the entire cerebral hemisphere were essentially synchronous . 8 The HVS episodes occurred s imultaneously and with a similar ampl i tude in mir- ror posi t ions of the two hemispheres, a l though rarely ( < 1%) were HVSs present only in one hemisphere.

For the night sessions the r ight anter ior electrode was chosen. In a few rats, however, the left an ter ior electrode was selected due to the presence of artifacts in the r ight an ter ior electrode.

HVSs occurred significantly more frequently in males of the BN and F1 and F2 strains than in their female counterpar ts . Only one of the females in the BN strain displayed HVS dur ing the 12 h of night observat ion, whereas eight of the seventeen BN males did (47%). No sex difference was present in F344 rats (100% males; 94% females had HVS). F1 males demons t ra ted 100% penet rance of HVS, whereas only hal f (56%) of the females had HVS. In the F2 generat ion 77% of the males and 69% of the females had HVS.

HVSs were character ized by two independent measures (number and total dura t ion) and the de- rived average dura t ion (total dura t ion /number ) . 2 The frequency of occurrence and the average dura t ion of HVS correlated significantly in all groups ( r s = 0 . 6 5 - 0 . 9 1 ; P < 0 . 0 1 ) except in BN females which did not exhibit HVS. Two-way A N O V A s by generat ion and sex demons t ra ted that sex was a significant source of var ia t ion (P < 0.001). With the exception of the F344 strain, males of the BN and Hybr id F1 and F2 generat ions had more and longer dura t ion HVS than their female counterpar t s (Table 1). The longest average dura t ion (8-9 s) and highest n u m b e r (120 episodes/12 h) of HVS occurred in F1 males. F1 females had significantly fewer HVS than males (P < 0.01) and were comparab le to the parent groups. In the segregating F2 generat ion, sex differ- ences of bo th the n u m b e r and average length of HVSs diminished.

Predictable recurrence of high-voltage spike-and-wave spindle episodes and arousal

Analysis of the temporal dis t r ibut ion of HVS revealed that the spike-and-wave bursts recurred at predictable intervals. In the recurrent maps of Fig. 3, we plot ted the intervals (i) between HVSs against the successive intervals (i + 1) on a logar i thmic scale. This da ta presenta t ion was chosen since it displays data f rom all animals. Al though considerable vari- a t ion was present among sexes and strains, successive

Table 1. Occurrence of high-voltage spike-and-wave spindle during 12 night hours

Total duration Number Average duration s/12 h per 12 h s

Group n (±S.D.) (±S.D.) (±S.D.)

BNf 14 0.0] 0.6 0.23 BNm 17" 33.2 (8.22) 16.2 (4.55) 1.22 (0.441) F344f 16 58.4 (8.92) 20.3 (4.66) 2.40 (0.201) F344m 16 53.2 ( l l . l ) 22.3 (5.43) 2.42 (0.173) F l f 21 68.2 (21.1) 24.2 (6.67) 1.83 (0.244) F lm 16" 1355.1 (212.41) 158.4 (23.34) 8.60 (0.465) F2f 39 311.3 (115.33) 88.4 (24.44) 2.01 (0.455) F2m 44* 712.4 (138.22) 97.4 (21.37) 3.82 (0.335)

*Significantly different from females of the same strain (P < 0.05). tOnly one rat had HVS in this group. S.D., standard deviation.

Spike-and-wave epilepsy in rats 305

NS( 64 2 B

x ( i + l )

100(}0

1000

100

10 I i

x ( i ~ - l )

1 0 0 0 0

1000

100

I0

x ( i + l )

B N m a l e

I . . . . . . : 7~ -

; , : i

• . • - . . . . , .

i . . , . .. , . - . . . . . . . • . . . . . , ,

l 0 1000

F 3 4 4 m a l e

- I r - . ]

' . : , " . . " ,. :',:..;~L-'.'.'." : ' • . . . . ~' : ~ . ':, v / !-: ~ •

i• •• Z•: •::ii:

I ~ . . . . _ L _ 4 _

10 1000

F1 m a l e

tO000 L I ~ l . . "

• • . • • . . .

I - .u.:?i":~Z-P,::f, #)-"...

100 "'.. '-"J" ' .2 ~ ",'?-~{ = " . .

.) . i : ~ . - ~ . . 2 : ~ . i . ' : : : " : ! ; •

] [ .. I . . . . . . . I 10 1000

F 2 m a l e x ( i + l )

• [ I . 1 0 0 0 0 . . . .

• ,..:-.L; . L :~.' :;-.,::~;;;~i . : . ,:...... :y,;.,..,;~,.,741~.,..~.#~:g-,,

~oo . : # ' ~ ] ~ I ~ ~ N ~ • ; ~ :•'••

. . : : . : . ~ ~.,,~:~.-.::. ~'.,'",.~ ~:-H;~":.. •

10 1000

x(i+l)

10000

BN fema le

1

1000

100 [

i

I . . . . . A x(i) - - xti)

x ( i + l )

1 0 1 0 0 0

F 3 4 4 f e m a l e

I l : [ ~, .~-~ . ' ~;4. • ' i

" .:;~ .i.:,--i;;;;:ii ~:I:I :

I O0 1 ".5, • . . .

. • . . . . . : : .

, . . . .

I0 : : ..

i

x(i) I L I x(i) I 0 1 0 0 0

F1 fenmle x(i+l)

10000 i [ - . . . . i l . i ]

100 ' . : ' . . . ~.. , *. : , . !;

I0 i " '~':" " " " i ' / ' "

x( i ) ~ i . . . . . / x( i ) 1o 1ooo

x ( i + l )

t 0 0 0 0

1000

100

I0

1

x ( i )

t i m e ( s e c )

F2 f ema le

t • I . . I • . . • : • : z . !

.'.. ,2".'~': "~'~'" ,'..;'.~,~"',:;~'7,:::" -

L :: " :" " . ( !

l 0 1000 x(i)

Fig. 3. HVS episodes recur at two predic table intervals. Return maps of the occurrence of HVS episodes. Note relat ively separa te clouds, especial ly in BN male and FI female, ind ica t ing reoccurrence of HVS at 10 30s (shor t - term) and 1,200 1,800s ( long-term) intervals• Note also tha t the faster

reoccurrence pat tern , typical in F I and F2 subjects is rare in F344 rats.

306

EEG

m o v

G. Jand6 et al.

F'*

0,5 mV

Fig. 4. Short-term oscillation of HVS in male F1 and F2 rats. Note the regular appearance of HVS episodes at approximately 30 s intervals in both rats. In the F1 rat large amplitude tremor movement (move) was present during HVSs. In the F2 rat, rebound movements occurred at the offset of HVSs.

repetitions at two different intervals, resulting in four clusters, were evident in these maps. Short intervals followed by short (SS), long intervals followed by short, short followed by long and long followed by long intervals could be distinguished. In BN males all four patterns were present at approximately 30 s and 1,200-1,800s intervals. In F344 rats the dominant cluster (long followed by long intervals), was at 1,200-1,500 s, although a less prominent short inter- val cluster at 90 lOOms was also present. In F1 descendents the patterns were similar to that of their BN parents, al though short followed by long inter- vals and long followed by long intervals were less prominent in males. In the latter group of animals a very prominent SS cluster at 30 s was present, that could often be observed in the raw records (Fig. 4). In the segregating male and female F2 rats only the SS clusters were clearly distinguishable with 30 and 10 s intervals, respectively. These analyses revealed that HVSs did not occur randomly, but recurred predictably at either short or long intervals. The pattern of short-term and long-term oscillations was characteristic of a given strain and sex.

The long-term oscillation of HVSs most likely reflected periodic changes of arousal. This not ion was supported by the periodic nature of movement which

3S

3

(,B >

O5

PM IOPM

. . . . . . . . . . movement

- - HVS

12PM 2AM 4AM

E

c0

JE}

E

~AM 8 AM

Fig. 5. Relationship between HVS and motor activity. Data are from a single F344 rat during a 12 h session. Abscissa: cumulative duration of HVS and integrated movement in 10min bins. Note that periodic peaks of HVS episodes

occur at about 10 min following movement peaks.

occurred at approximately similar intervals. A typical example of movement and HVS oscillations in a single night session is shown in Fig. 5. Movement , integrated for successive 10min bins, was plotted together with the cumulative duration of HVSs for the same bins. Based on the observation that HVSs never occurred when the animal was moving or engaged in grooming, one would expect a negative relationship between movement and the occurrence of HVSs. However, this was clearly not the case. Instead, HVS peaks regularly followed movement peaks by approximately 10rain but they did not coincide with movement minima. This relationship was true for all groups. The correlogram in Fig. 6 was constructed by cross-correlating the peaks of move- ment (reference event = 0 min) with HVS peaks from all rats of all groups. It is evident that the maximum probability of HVS occurred within 10 min after the movement peaks. The significance of the cross-correl- ogram peak was judged by comparing the derived data with the cross-correlogram of their shuffled points (Monte Carlo simulation; P < 0.01). Similar cross-correlograms were obtained when rats of the various groups were analysed separately (not shown).

After reviewing the records of individual animals and observations of several rats during the recording sessions some clues to the delay between movement peaks and the probability of HVS occurrence became evident. Exploratory bouts which included walking

-100 -80 -60 -40 -20 0 20 40 60 80 100 time (min)

Fig. 6. Cross-correlogram of HVS episodes with the move- ment peaks as zero reference, based on 12 h records of all rats. The peak of HVS activity occurred at 10 min after the movement in each rat strains (separate group data are not

shown).

Spike-and-wave epilepsy in rats 307

and frequent rearing occurred at regular intervals, similar to the long followed by long interval clusters of HVS. Cessation of exploratory behavior was typ- ically followed by grooming after which the animal sat immobile with the head in an upright position. Finally, the rat curled up with its head resting on the floor or on its forelimb with eyes closed. This sleeping posture was more typical during early morning hours but regular short sleep bouts were confirmed by the occurrence of large delta waves in the EEG through- out the night session. HVS occurred more frequently during awake immobility and were rare during sleep. Although we have not formally calculated the time that elapsed between exploratory bouts and the oc- currence of sleep, the above observations suggest that the delay between movement peaks and the maxi- mum occurrence of HVS can be explained by assum- ing that the "threshold" for the occurrence of HVS was lowest during awake immobility that preceded sleep. HVSs also occurred during sleep episodes but less frequently than in an awake and drowsy animal. This observation is also supported by our pilot observations involving 24-48h recording sessions that HVSs occurred more frequently during night hours than during daytime hours (data not shown).

In a further attempt to clarify the relationship between the occurrence of HVS and arousal, we analysed the spectral characteristics of EEG preced- ing the HVS event. The power of delta band (0 3.1 Hz) for the 2 s EEG segment preceding the HVS onset was calculated by a Fast Fournier Trans- form (FFT) program and served as an index for the level of arousal. 4 The delta power was normalized for

the entire data base in a range from 0 (zero power) to 100 (power in a given individual with the maxi- mum delta value). The incidence of HVSs was plotted as a function of delta power for the various animal groups (Fig. 7). These calculations indicated that neither deep sleep nor a highly aroused state was associated with a high probability of HVS occur- rence. Instead, HVS was associated with a preferred EEG state characterized by an optimum level of delta power.

Intraepisodic frequency changes

Intraepisodic frequency of HVS displayed signifi- cant strain and sex dependent changes (Fig. 8). Females of all strains displayed similar intraepisodic frequencies (P = 0.23. ANOVA). In contrast, the intraepisodic frequency of HVS differed significantly among the male groups (P < 0.01). F344 males had the lowest HVS frequency (6.8 Hz). These values for the other strains were BN = 7.8 Hz, FI = 7.8 Hz and F2 = 7.4 Hz. The onset frequency (first s, Table 2) was higher by 0.5-1.5 Hz than the average frequency during the entire duration of HVS (P <0.001, ANOVA), and it reached an asymptote at approxi- mately 7 Hz within 2-3 s. This tendency was ex- pressed the least in F344 rats. The onset frequency of HVS (first 1.0 s) was significantly higher in males than in females (all groups P < 0.05, except F2).

High-vohage spike-and-wave spindle-associated tremor (myoclonus) and rebound movements

Both male and female rats of the F344 strain remained motionless during the HVS episodes, with

25.

20

15

10.

O, tO c

"o -~ 140 c..- - - 120-

100-

8 0 -

60- 40- 20-

BN

~A~9 MA A 20 4'0 6'0 8'0 100

30 F344

25- ,iii

20- ~ ~

15-

10

5-

0 20 40 60 80 1 O0

F1

~AM~ 2'0 40 60 8'0 100

350 300 250. 200 150 1 O0

50 0

F2

2'0 4'0 6'0 80 100

delta power

Fig. 7. Dependence of HVS on background EEG power. Abscissa: per cent distribution of delta power of EEG recorded during 2 s before HVS onset. 100 on the abscissa corresponds to the largest power observed in a given animal. Ordinate: number of HVS episodes. Note that the highest probability of

occurrence of HVS is at a preferred delta power in all strains.

308 G. Jand6 et al.

N " l - v >., 0 c

"-i 0 "

BN 13

1 2 -

11-

1 0 -

9-

8-

7-

6-

5-

4 0 1 2 3 4

i

5 6

F344 13

11

1

8 - 2 . '2,

4° . . . . ; 0 1 2 3 4 5

F1 13

12-

11-

10-

9-

8-

7-

6-

5-

4 i i i i i

1 2 3 4 5

F2 1 3 -

11-

10-

9 -

8 -

7 -

6 -

5 -

4 i i i i i

0 1 2 3 4 5

time (sec) Fig. 8. Intraepisodic frequency shifts during HVS. Each line represents the average value from a single rat. Note that the initial 1 s epoch is similar to the rest of the HVS episode in F344 rats but substantially

higher in BN, FI and F2 subjects.

occasional movement of the vibrissae and eyelids• In contrast, most BN animals displayed rhythmic move- ments of the vibrissae, limbs, head and trunk. Recur- rent maps of the spike components of HVS and of the motor components revealed that the two events often occurred at the same frequency (Fig. 9). The HVS-associated tremor (myoclonic movement) oc- curred immediately after the onset of HVS, and often persisted throughout the duration of HVS. However, rhythmic movements during HVS were often inter- rupted by slower frequency myoclonus or irregular body jerks• These intrusive movements provide an explanation for why the cross-correlograms between HVS and movement had only one or a few peaks (e.g.

Table 2. Onset frequency of high-voltage spike-and-wave spindle

Group n Mean (Hz) S.D.

BNf 1 6.23 0.000 BNm 8 8.91 0.589 F344f 15 7.07 0.682 F344m* 15 8.29 1.034 F l f 21 7.74 0.446 F l m * 16 9.21 0.167 F2f 22 8.4l 0.943 F2m 29 8.80 1.028

*Significantly different from females of the same strain (P < 0.05). S.D., standard deviation.

• . H V S 8 0 0 -

6 0 0 -

400 ." • ".:$

2 0 0 ' "~':" . .~ ' , ,4"r , ; . . . . .. • "."

4 - 200 400 6 0 0

X 8ooJ '""' ' movement

6 0 0 - ' . L , ' , . " • . . ..

' ! , " . . i 1 , , , , , . ' . • 1:.~. . . . . .

4 0 0 - . - . , i ~ . . . l , . ' . • ' . • . •

,,~. • . , . , .

2 0 0 " ~ ' " 1 ~ , ~ ! ' " : : . " . . . " • " . ' ' " - J , ' l ' "~'; .. %: :" ' • •

~ _ _ U . "~.~,..~m~'~.~. ':"l "'" ~ I'" ' , "'

I

8 0 0 ( m s e c )

200 400 600 800 ( m s e c )

X(i) Fig. 9. Relationship between HVS and movement (tremor). Return maps of spike intervals (x(i) vs x0) + 1) within HVS (upper panel) and movement peaks (lower panel). Note dominant activity at about 110 130ms (6 9Hz) of both

HVS and movement.

Spike-and-wave epilepsy in rats 309

Fig. 2A). The HVS concurrent t remor was ex- pressed in the Fl and F2 descendents (e.g. Fig. 4). Impor tant ly , females very rarely displayed HVS-as- sociated t remor, and integrated movement during HVS was not significantly different f rom baseline (Fig. 10).

Occasionally we observed t remor in the immobile rat wi thout HVS or accompanied by low ampli tude

rhythmic EEG waves. Unfor tunate ly , the occurrence of these HVS-independent t r emor episodes could not be quantified because we could not set reliable and independent criteria for the t remor movement by using the accelerometer ou tpu t only. Nevertheless, these quali tat ive observat ions suggest tha t neuronal circuitry underlying these phenomena (HVS and t r e m o r ) can be separated. ~2

B

a i b c d

onset offset

1.5 O

C

A

2.5-

2-

0.5

F344

F344

BN

~k 2.1~

o

0.

F1 F2

BN F1 F2

| D 2.5

1.5 O

0.5.

I [ ] female

[ ] male

0" F344 BN F1 F2

Fig. 10. Relationship between HVS and motor activity at HVS onset (b), end of HVS (c) and after HVS offset (d). The ratios in B D are based on integrated values (1 s) of movement. A value of I indicates no change from baseline (a in A). Note lack of movement (tremor or rebound movement) during and after HVS in female rats as well as in F344 males. Note also the presence of tremor in BN, FI and F2 males

and rebound movements in FI and F2 males.

310 G. Jand6 et al.

A b II c d

movement ~l~J

.................................. *'< ... . . . . . g r!r T F ..................... " ..... | I 6.66 s (A) I 0.5 S (B-D) 0.5 mV

I B C D

Fig. 11. Transition from HVS to clonic seizure in an F2 male rat. A. EEG activity (upper trace) and movement. Epochs indicated by b, c and d are shown at higher time resolution in B, C and D. Note rhythmic slow delta waves before HVS onset and desynchronized EEG after HVS offset. Myoclonic jerks

(movements) were correlated by negative EEG spikes, such as shown in D.

A n o t h e r type of m o t o r pa t t e rn tha t was d o m i n a n t only in males was tha t of body movemen t which occurred after the offset of HVS (Fig. 4, F2). This " r e b o u n d " pa t te rn was not rhy thmic but corre- sponded to ei ther body ad jus tment or m o t o r arousal f rom a sleeping posture. Again, r ebound movements occurred more frequently in males than in females and these differences reached significant values in FI and F2 animals (Fig. 10).

Long duration high-voltage spike-and-wave spindle

In some rats of the F2 generat ion the ampl i tude of t r emor (myoclonic movement ) was occasionally very large, and sometimes, the rat lost its balance and fell on its side. Even when HVS was cont inuous ly present for m a n y seconds, t r emor movements disappeared and then reappeared. In many instances a reliable relat ionship between the dominance of ei ther the

spike or the wave componen t s of HVS with t remor could be established. Epochs with p rominen t spike componen t s but small wave componen t s were usually not accompanied by movement , but epochs with relatively large wave and small spike componen t s were. The longest average dura t ions of HVS occurred in F1 males (8-9 s). However, compar i son of group averages between inbred strains and their segregating F2 descendents was not meaningful , since a large variabil i ty is expected in these individuals. In some F2 males very long bouts of HVS ( > 50 s) were observed. At o ther times, long dura t ion HVSs with short in ter rupt ions were clustered. Such clusters of ten ex- ceeded 1 min.

In one F2 male rat, large ampl i tude irregular jerks built up after long HVS clusters. The body jerks cont inued after the offset of HVS with jerk-concur- rent spikes in the E E G or wi thout any EEG alter-

Table 3. Clonidine-induced frequency changes of high-voltage spike-and-wave spindle

Mean (Hz) Mean (Hz) Group n before S.D. after S.D.

BNf 1 n.t. - - - - - - BNm 8 7.71 0.546 6.18" 0.683 F344f 15 6.59 0.629 6.22* 0.544 F344m 15 6.87 0.743 6.35 0.233 F 1 f 21 6.96 0.686 6.23* 0.662 Flm 16 7.82 0.260 7.77* 0.331 F2f 22 7.08 0.808 6.26* 0.676 F2m 29 7.40 0.917 6.51 * 0.445

*Significantly different from pre-drug frequency (P < 0.05). S.D., standard deviation, n.t., not tested.

Spike-and-wave epilepsy in rats 311

100=

10-

1-

0.1 0 0 0 1

BN male

, ,

0 0 1 0 1 1

100..

1 0 .

0 .1 • OO01

all rats

• w

/ 0 0 1 0 1 1 10

¢ - , m

E If')

0

03 > -1- ¢.-

"{3 ¢-.

0 0

10o F344 male

1

0.001 0 0 1 0 1 1

F1 male

1- r . . ~ . . r , r . ,

0.001 0.01 0.1 1 10

F2 male I 0 0 -

10

0 1 ~ , 0.001 0.01 0.1 1 10

0.1

100-

10-

0 1

100~

10.

0.1 0.001

100-

10-

01 0.001

F344 female

. . .r . ' ' I 001 0 1 1 10

F1 female

.. "" ,

0.01 0 1 1 10

F2 female 100=

1

0 , 1 , ,

0001 001 0.1 1 10

overnight HVS (% of 12 hours ) Fig. 12. Relationship between spontaneously occurring (abscissa) and clonidine-induced HVS activity (ordinate) expressed as a percentage of total duration of HVS/recording time. Each dot represents a single animal. The curves in each panel represent the best fit [y = 100 - (100/eX)]. Note that clonidine increased the probability of occurrence of HVS several-fold in all groups. BN females are not shown since HVS

was expressed in only one rat.

ations. In another F2 male rat such an episode was observed visually and was classified as a clonic seizure. In the latter case, the rat was asleep and the emergence of a HVS precipitated the myoclonic seizure (Fig. 11). These examples illustrate that com- binations of genes in the segregating (F2) generation may precipitate electrobehavioral patterns that are similar to atypical absence attacks in humans.

Alpha-2 agonist-induced high-voltage spike-and-wave ,~pindle

Confirming previous research, ~°'48'64 clonidine, a noradrenergic alpha-2 agonist, powerfully increased

the incidence and duration of HVS when given intraperitoneally, 15 min before the recording ses- sion. The present experiments addressed two issues. First, whether the drug can induce HVS in rats which did not exhibit spontaneous spike-and-wave patterns during night recording sessions. Second, whether a short (15 min) drug session can reliably predict the outcome of the night recording sessions.

Clonidine failed to induce spike-and-wave patterns in rats which did not display spontaneous HVS. However, in a few rats, HVSs were not detected by the network during the night session, yet short dur- ation HVS episodes were occasionally present in the

312 G. Jand6 et al.

clonidine test. Subsequent visual tracking of the entire recording session in these rats revealed that spike-and-wave patterns of 6-9 Hz were indeed pre- sent in the absence of the drug. However, the ampli- tude of these rhythmic events was only slightly above the background EEG. At other times, unusual sleep spindles (12 16 Hz) with spike components were pre- sent. These patterns, however, did not fulfill the criteria of HVS and were not detected by the net- work.

In rats with spontaneous HVS, clonidine substan- tially increased the number and duration of HVSs. Figure 12 displays the cumulative duration of HVSs induced by clonidine as a function of the total duration of HVS during the night sessions. Overall a significant correlation (r =0.70; P <0.0001) was found between these measures, suggesting that the short drug sessions predicted the presence and inci- dence of HVS during the 12-h night sessions. Within group comparisons were significant in BN, F344 and F2 males and F1 and F2 females (r =0.624).71; P < 0.01-0.001).

The intraepisodic frequency of the clonidine- induced HVS was significantly lower than that of the spontaneously occurring ones (Table 3). The fre- quency decrease was similar in male and female rats and was different in the various strains. Tremor, when present during spontaneous HVS, also occurred in association with clonidine-induced HVS. Thus, both EEG and behavioral correlates of the drug- induced HVS were quite similar to those of the spontaneously occurring HVS.

The increased incidence of HVS under the influence of clonidine was not simply due to the sedated behavioral state of the rat. First, we have shown above that deep slow wave sleep did not promote, but in fact decreased the occurrence of HVS. Second, long periods of immobility during drug-free sessions were rarely accompanied by continuous HVS, whereas following clonidine injection HVSs were present from 10 to 100 per cent of the recording time. In fact, in several rats the virtually uninterrupted HVS episodes could be clinically classified as petit real status epilepticus. Third, intrathalamic injection of clonidine similarly enhanced the incidence of HVS without overt sedative effects. 6

DISCUSSION

The main findings of the present experiments are that (i) HVS episodes display short-term and long- term periodicity, (ii) HVS characteristics and their motor correlates are different in F344 and BN strains, (iii) several components of the EEG trait and their motor correlates are controlled by genetic factors, (iv) HVS and associated termor occur more frequently in males than in females, and (v) clonidine-induced EEG changes can predict the presence and incidence of HVS in long-term recording sessions.

Periodic' occurrence o[ high-voltage spike-and-wat~e spindle

A striking observation in the present experiment was the periodic nature of HVS occurrence during night hours. Two such oscillations were observed at 10 100s (short-term) and at 15 30min (long-term) intervals. Furthermore, the repetition frequencies were characteristic of the various strains and sex groups. The long-term periodicity was coupled to the oscillation of the arousal level as evidenced by a similar oscillation of motor activity. Such periodic changes of motor activity were also present in female BN rats without HVS. Although the physiological mechanisms of the periodic fluctuation of the arousal state has yet to be elucidated, it has been known for a long time that arousing stimuli interfere with the occurrence of spike-and-wave pattern in primates, whereas drowsy states and sleep promote its occur- r e n c e . 23"3s'5°'66~ Furthermore, within sleep states, peri- odic (< 100 min) alterations of sleep episodes have been described. 32 Importantly, periodic changes of vigilance with similar intervals have been observed during wakefulness, 5 suggesting that an endogenous oscillator, independent of sleep mechanisms, is re- sponsible for the ultradian oscillations. Based on our observations, it is tempting to conclude that the long-term periodicity in rats and the ultradian oscil- lations in humans are paced by similar mechanisms.

The short-term (10 30 s) periodicity of HVS may also be related to findings in humans. Hahisz et al., 24

and Terzano et al., 7w2 described cyclic and non-cyclic microstructural organization of slow wave sleep. The cyclic pattern consists of fluctuation of sleep spindles and delta wave bursts with 5- to 30-s-period dur- ations. Importantly in the present contexl, cyclic oscillations of spike-and-wave patterns with 5 30-s- intervals have also been reported in patients with absence seizures, w~ Although the neuronal mechan- isms of the short-term oscillation is not known, recent studies in the cat indicate that it survives extensive damage to the thalamus. 6'~

Further support for the modulatory effects on HVS by the arousal state was demonstrated by the corre- lation between the incidence of HVS and delta power. Such an analysis revealed that the maximum prob- ability of HVS occurred within a relatively narrow range of delta power. Since delta power can be taken as a reliable indicator of the arousal levelfl this observation suggests that the pattern of background EEG is an important predictor of HVS. In highly aroused states as well as in deep sleep there is a substantial reduction of HVS occurrence. On the other hand, HVSs occurring during slow wave sleep may promote the occurrence of further HVSs by increasing the arousal level of the brain. EEG activity after the offset of HVS was typically desynchronized even when slow waves dominated prior to the onset of HVS (e.g. Fig. 2A). j2 Rebound movements ob- served at the termination of HVS further support the

Spike-and-wave epilepsy in rats 313

arousing properties of HVS. As a result, a HVS occurring in the midst of a delta wave burst may aroused the brain to an optimum level for the occur- rence of subsequent HVSs.

Physiological discoveries of the past decade sub- stantially increased our understanding of the gener- ation of cyclic brain waves, including HVS. Thalamic and neocortical neurons fire rhythmic bursts in as- sociation with barbiturate spindles, sleep spindles, penicillin-induced spike-and-wave discharges in cats and HVSs in r a t s . 9"36'68 It is thought that the inter- action between GABAergic cells of the nucleus retic- ularis thalami and thalamocortical relay cells form the basis of these neuronal oscillations. 63~'4~6~ A characteristic feature of thalamocortical neurons is the presence of a low-threshold calcium conductance which is inactive at rest and de-inactivated at - 6 5 to - 7 5 inV. Release from the hyperpolarized state re- sults in low threshold bursts which will activate cells in the RT. Burst firing of RT cells, in turn, will hyperpolarize more thalamocortical cells, some of which will fire low-threshold calcium spikes upon repolarization from the hyperpolarized state. Rhyth- mic firing of RT cells is a critical event, since damage to this nucleus abolishes barbiturate spindles and H V S . 6'67 Another prerequisite for the oscillation is the polarization level of the thalamocortical neurons. Any condition that interferes with either of these requirements is regarded as anti-epileptic, l-' Arousal associated with cholinergic, nonadr, energic and sero- toninergic activation of the thalamus depolarizes relay cells and prevents the occurrence of low- threshold calcium spikes. 43 The interfering effects of deep sleep with HVSs may be explained by recent in ~'itro and in vivo observations. 15'44'55"66'6~69 The tran- sition from early sleep stages to delta wave sleep is paralleled with the progressive hyperpolarization of thalamocortical cells. In deep stages of sleep the interplay of two intrinsic currents, the hyperpolariz- ation-activated cation current (Ih) and the low- threshold calcium current in thalamocortical cells will produce a slow membrane oscillation in the delta range (1~4 H z ) 56 and prevent faster oscillations that form the basis of sleep spindles and HVS. 6,55 Our finding that HVS occurred preferentially at a rela- tively narrow range of delta power, combined with the above microphysiological observations, suggests that the polarization level of thalamocortical cells, critical for the occurrence of HVS, can be predicted from the epidural or scalp EEG.

The above short summary of the relevant in t,itro and in vivo data indicates that the cause for the large inter-strain and sex-dependent variability of HVS should be sought in the mechanisms that regulate the polarization level of thalamocortical cells. There are at least three oscillation promoting systems: (i) The nucleus reticularis thalami GABA system, 27'6~ (ii) the extrapyramidal GABAergic input to the thalamus 22 and (iii) the noradrenergic input acting on postsyn- aptic alpha-2 receptors in the thalamus. ~° Altered

sensitivity of thalamocortical cells to GABA is not very likely to play a significant role, since we failed to see differences in GABA currents in patch-clamped thalamocortical cells of rats with and without HVSs (I. Mody and Buzsaki, unpublished observations). Furthermore, no significant differences were found in the levels of radioligand binding to GABAA and GABA B receptors in selectively bred Wistar rats which exhibited high and low incidence of HVS in any of the nine areas investigated (including the thal- amus)) 5 On the other hand, GABA B agonists, given systemically or infused directly into the thalamus, promote the occurrence of HVS very effectively, z~-~'~ These findings suggest that strain-dependent hyper- polarization levels of thalamocortical cells, a prereq- uisite for the occurrence of HVS, may not be solely due to alterations of thalamic neurons and that the causes should be sought in their thalamopetal affer- ents and/or in the neocortical targets.

A main source of GABA in the thalamus is the extrapyramidal input. 3~ Alterations in the amount of GABA released onto the dendrites of thalamocortical cells (sites of postsynaptic GABA B receptors) by the extrapyramidal input may be one of the mechanisms for the differential expression of HVSs in different rat strains and for their age-dependence. Indeed, the incidence of HVS in F344, Buffalo and Sprague- Dawley rats correlated with tyrosine hydroxylase activity and D2 receptor binding values in the cau- date nucleus] t Furthermore, D2 antagonists given either systemically or directly into the caudate nu- cleus powerfully increases the incidence of HVS. j-~7~

Alpha-2 agonists given either peripherally 33-~s or directly into the thalamus l° substantially potentiate HVSs even following complete destruction of the thalamopetal noradrenergic fibers. We have sug- gested that norepinephrine may increase potassium conductance via postsynaptic alpha-2 receptors located on thalamocortical neurons. Furthermore, we have hypothesized that the overall action of nor- adrenaline in the thalamus depends on the relative density and affinity of postsynaptic alpha-1 and al- pha-2 receptors as well as on the amount of nor- adrenaline released, l° Large amounts of released noradrenaline during aroused states will prevent the occurrence of low threshold calcium spikes 43 and the consequent rhythmic discharges of thalamic cells. During lower vigilant states, such as immobility, drowsiness and light sleep noradrenaline release is decreased, and due to the higher affinity of postsyn- aptic alpha-2 receptors, the neurotransmitter will now hyperpolarize thalamocortical cells. Age-, strain- and sex-related variation of the catecholamine sys- tem, therefore, may be an important factor in the phenotypic expression of HVS.

Sex differences in the expression (~f high-t'oltage spike- and-wave spindle

An unexpected and striking finding of the present experiment was the large and consistent differences of

314 G. Jand6 et al.

various physiological and behavioral parameters in male and female rats. Male rats of the BN, FI and F2 groups had significantly more and longer duration HVSs and their intraepisodic frequency was signifi- cantly higher than in female rats. The sex differences persisted in the clonidine-induced HVS episodes. In addition, HVS-associated tremor (myoclonus) and rebound movement was present almost exclusively in male animals. With the exception of the onset fre- quency of HVS, these sex differences were not present in the F344 strain. Importantly, another study using selectively bred Wistar rats 4~ failed to reveal sex differences in the expression of HVS. At first glance it appears that sex-linked gene effects present in BN rats should be responsible for the observed inheri- tance pattern. However, for the demonstration of an X-chromosome effect the incidence of HVS in F1 and reciprocal FI males should be closer to the mother's strain and females are not expected to be different. This pattern was not seen in a previous study that involved reciprocal crosses. 73 In another genetic ex- periment including WAG and ACI strains and their descendents, F1 hybrid males had HVSs of longer duration although the mean number of HVS was not significantly different. 57 Since the offsprings of ACI mothers had longer duration HVSs than the offs- prings of WAG mothers with longer duration HVSs, then these effects probably could have arisen from other maternal effects.

One such non-chromosomal effect that is known to induce sex-dimorphic changes in the brain is the influence of sex steroids. A number of structures in the CNS, including the dopaminergic nigrostriatal system and the noradrenergic locus coeruteus, express sexual d imorph i sm . ~6'22'74 The volume and number of cells in the locus coeruleus are larger in females than in males, and in both sexes the number of neurons begins to decrease after three months of age. ~6 This age corresponds to the emergence of HVS in most rat strains. Ovarian hormones appear to be necessary for the maintenance of the normal number of locus coeruleus neurons, since ovarectomy, but not or- chidectomy, in both newborns and adult rats results in significant cell loss in this nucleusJ 6'22 Importantly, hyperinnervation of the brain by noradrenergic fibers was related to the expression of HVS in the tottering mouse. 53 Future experiments, including gonadec- tomy, are necessary to reveal whether sex hormones are responsible for the large difference in the occur- rence of HVSs in male and female rats, and whether this difference is mediated by the catecholamine system.

Inheritance o f high-voltage sp ike-and-wave spindle and its relevance to absence epilepsies

Two previous rat studies have examined the genetic architecture of HVSs in detail. In the first one 57 WAG rats with high incidence of HVSs were cross-bred with ACI rats. In that study all of the WAG but none of the ACI rats had HVSs, which was also present in

all FI rats. In the F2 generation and in B1 and B2 backcrossed rats 79, 94 and 37% of the individuals had HVSs, respectively. Using the percentage of animals with HVS, the authors concluded that one dominant gene determines whether the animal ex- presses HVS, and other genes determine their number and duration. In another experiment, however, almost all ACI rats were shown to express at least some HVSs between six to eight months of age, 29 making conclusions from these genetic analysis difficult to interpret. In another study, Wistar rats were selected for HVSs for 13 generations. 4~ In the parental groups tested, all of the HVS-high rats showed HVSs but none of the HVS-Iow animals did before 12 months of age. In the F1 cross 62% of the rats at four months and 95% at 12 months showed HVSs. In the F2 generation 55% (four months) and 86% (12 months) of the rats had HVSs. In the backcross of F2 and the parental HVS-low line 67% of the rats displayed HVSs at 12 months of age. Taking into account the duration of HVSs as well, Marescaux et al. 41 suggested that the presence of HVS depends on a single autosomal dominant gene, whereas the duration is due to another or several additional genes.

Our study differs from the above experiments in at least two important aspects. First, in the studies of Peeters et al. 57 and Marescaux et al. 4~ the average duration and incidence of HVS in the F1 hybrids fell between the parental values. In our studies, both parental groups had HVSs although the average duration and the intraepisodic frequency differed significantly between the groups. Both the duration and incidence of HVSs increased several-fold in the FI generation. Second, in our study the two parental groups differed significantly in the motor correlates to HVSs. In addition, one of the parental strains (BN) showed sex-dependent differences in the incidence, duration and intraburst frequency of HVS and in the HVS-concurrent motor correlates. In a previous quantitative genetic analysis on the total duration of HVSs, we proposed a threshold model of HVS. 73 According to the model, a dominant gene(s) and developmental-environmental factors control the level of a quantitative underlying variable. This hypo- thetical variable (e.g., hyperpolarization level of tha- lamocortical cells at rest), in turn, shows an age- dependent continuous distribution and precipitates HVSs in only those individuals in which this variable is above a certain threshold.

A prediction of the threshold model is that HVS can be found in any genetically defined population. Indeed, HVSs have been observed in at least some individuals of every inbred and outbred strains stud- ied so far. I1'12'29'41 This suggestion is in contrast to the generally held view that spike-and-wave pattern is present only in patients and relatives of patients with absence seizures. Unfortunately, the majority of gen- etic EEG studies are based on brief recordings of 25-30 rain duration. In contrast to most previous

Spike-and-wave

studies, Degen et alJ 7 recently reported the presence of spike-and-wave complexes in a very high percent- age (72%) of siblings of patients with idiopathic absence attacks. Importantly for the present study on rats, spike-and-wave patterns were detected signifi- cantly more frequently in male siblings (82%) than in female siblings (64%). In the study of Degen et al. the EEG was taken under the effect of the tranquilizer promazine. This may be of particular importance because tranquilizers have been shown to powerfully increase the incidence of HVS in rodentsJ ~,~2,v7 Never- theless, these findings suggest that if one were to test unselected individuals in the general population who have never had a seizure, a considerable percentage of individuals would manifest this EEG trait at some time in their lives, particularly between ages of five and 15 years of age. 1

If HVS and spike-and-wave complexes occur in healthy rats and humans, respectively, then is this EEG trait a good model for absence epilepsies? Rats with HVSs are considered to be an isomorphic, predictive and homologous model of human absence epilepsies. 4~ However, epilepsy is a clinical phenom- enon and not an EEG one. Rats with frequently occurring HVSs mimic the clinical picture of child- hood absence epilepsy in many important ways, but they share little with other forms of absence seizures. The present findings suggest that by crossbreeding available rat strains we can successfully develop models of other absence seizures, as well. Indeed, many EEG and behavioral parameters of some F2 hybrids resembled those of patients with myoclonic juvenile epilepsy and/or epilepsy with myoclonic ab- sences, including higher intraepisodic frequency and longer duration of HVS, severe myoclonus, rebound movements and clonic seizures. Selection for specific physiological and behavioral traits expressed by indi- viduals of the segregating F2 generation will allow us to combine various features that mimic the various subgroups of human absence seizures. Since patients with "atypical" absence seizures are considerably more difficult to treat than those with childhood absence epilepsy and are often resistant to drug treatment, developing new animal models of these absence syndromes is of considerable interest.

epilepsy in rats 315

Clonidine-induced high-voltage spike-and-wave spin-

dle

Despite slower intraepisodic frequency, HVSs in- duced by systemic administration of the noradren- ergic alpha-2 agonist, clonidine, were very similar to the spontaneously occurring ones. Clonidine never induced HVSs in rats that did not show such patterns spontaneously. Furthermore, the clinical correlates of induced and spontaneous HVS, such as HVS-associ- ated tremor (myoclonus), were also similar. The positive correlation found between drug-induced short sessions (15 min) and night recording sessions (12 h) suggests that the clonidine test is a reliable predictor for the presence and incidence of HVSs. We have recently reported that low doses of clonidine revealed the presence of spike-and-wave patterns (3 .5~Hz) in the monkey, as well. 7 Comparable results are not yet available in humans. Such reliable drug-induced tests could effectively accelerate the search for the genetic as well as other determinants of absence epilepsies.

CONCLUSIONS

The present study revealed that various aspects of the spike-and-wave patterns, such as incidence, aver- age duration, intraepisodic frequency and behavioral correlates, are genetically mediated. The significant sex differences observed in several parameters suggest further that the interaction of genetic and hormonal effects strongly influences the frequency of occurrence of the spike-and-wave phenotype. Most of the effects could be accounted for by a threshold model. A dominant gene(s) and developmental-environmental factors control the level of a quantitative underlying variable with an age-dependent continuous distri- bution. Spike-and-wave patterns precipitate in only those individuals in which these variables are above a certain threshold. The findings also suggest that gene and hormonal influences on the catecholamine system may be primarily involved in the regulation of spike-and-wave activity in rodents.

Acknowledgements We thank Drs E. L. J. M. Van Luijte- laar and D. A. McCormick for discussions. This work was supported by NIH, HFSP and the Whitehall Foundation.

REFERENCES

1. Andermann E. (1980) Multifactorial inheritance in the epilepsies. In Advance in Epileptology: Xlth Epilepsy International Symposium (eds Canger R., Angeler F. and Penry J. K.), pp. 298-309. Raven Press, New York.

2. Aporti F., Borsato R., Galderini G., Rubini R., Toffano G., Zanothi A., Valselli L. and Goldstein L. (1986) Age-dependent spontaneous EEG bursts in rats: effect of brain phosphatidylserine. Neurobiol. Aging 7, 115 120.

3. Berkovic S. F., Andermann F., Andermann E. and Gloor P. (1987) Concepts of absence epilepsies: discrete syndromes or biological continuum? Neurology 37, 993 1000.

4. Borbely A., Achermann P., Trachsel L. and Tobler I. (1988) Sleep homeostasis in humans and rats. Clin. Physiol. Sleep 13, 191-198.

5. Broughton R. J. (1975) Biorhythmic variation in consciousness and psychological functions. Can. Psy~hol. Ret,. 16, 217 239.

6. Buzsfiki G. (1991) The thalamic clock: emergent network properties. Neuroscience 41, 351-364. 7. Buzsfiki G. (1992) Network properties of the thalamic network: role of oscillatory behavior in mood disorders. In

Induced Rhythms in the Brain (eds Basar E and Bullock T. H.) pp. 235-250. Birkhauser, Boston.

316 G. Jand6 et al.

8. Buzsfi.ki G., Bickford R. G., Armstrong D. M., Ponomareff G., Chen K. S., Ruiz R., Thai L. J. and Gage F. H. (1988) Electrical activity in the neocortex of freely moving young and aged rats. Neuroscience 26, 735 744.

9. Buzsfiki G., Bickford R. G., PonomareffG., Thal L. J., Mandel R. and Gage F. H. (1988) Nucleus basalis and thalamic control of neocortical activity in the freely moving rat. Neuroscience 8, 4007-4026.

10. Buzsfiki G., Kennedy B., Solt V. B. and Ziegler M. (1991) Noradrenergic control of thalamic oscillation: the role of alpha-2 receptors. Eur. J. Neurosei. 3, 222 229.

l l. Buzsfiki G., Laszlovszky I., Lajtha A. and Vadfisz C. (1990) Spike-and-wave neocortical patterns in rats: genetic and aminergic control. Neuroseience 38, 323 333.

12. Buzsfiki G., Smith A., Berger S., Fisher L. J. and Gage F. H. (1990) Petit mal epilepsy and parkinsonian tremor: hypothesis of a common pacemaker. Neuroscience 36, 1 14.

13. Carpi D., Jando G., Pierre E., Vadi D., Fleischer A., Lajtha A., Vadasz C. and Buzsaki G. (1992) Inheritance of neocortical high voltage spike-and-wave (HVS) patterns in rats. Soc. Neurosci. Abstr. 18, 555.

14. Commission (1989) Commission on classification and terminology of the international league against epilepsy. Proposal for revised classification of epilepsies and epileptic syndromes. Epilepsia 30, 389 399.

15. Curro Dossi R., Nufiez A. and Steriade M. (1992) Electrophysiology of a slow (0 .54 Hz) intrinsic oscillation of cat thalamocortical neurones in vivo. J. Physiol., Lond. 447, 215 234.

16. de Blas M. R., Segovia S. and Guillamon A. (1990) Effect of postpuberal gonadectomy on cell population of the locus coeruleus in the rat. Med. Sci. Res. 18, 355-356.

17. Degen R., Degen H.-E. and Roth Ch. (1990) Some genetic aspects of idiopathic and symptomatic absence seizures: waking and sleep EEGs in siblings. Epilepsia 31, 784-794.

18. Doose H., Gerken H., Horstmann T. and Volzke E. (1975) Genetic factors in spike-wave absences. Epilepsia 14, 57-75.

19. Drinkenburg W. H. I. M., Coenen A. M. L., Vossen J. M. H. and Van Luijtelaar E. L. J. M. (1991) Spike-wave discharges and sleep-wake status in rats with absence epilepsy. Epilepsy Res. 9, 218 224.

20. Gibbs G. A. and Gibbs E. L. (1952) Atlas ofElectroencephalography, vol. 2: Epilepsy. Addison-Wesley, Cambridge, MA.

21. Gloor P. and Fariello R. G. (1988) Generalized epilepsy: some of its cellular mechanisms differ from those of focal epilepsy. Trends Neurosci. 11, 63~8.

22. Guillamon A., de Blas M. R. and Segovia S. (1988) Effects of sex steroids on the development of the locus coeruleus in the rat. Devl Brain Res. 40, 306 310.

23. Halasz P. and Deveny E. (1974) Petit real absences in night sleep with special reference to transitional sleep and REM periods. Acta Med. Acad. Sci. Hung. 31, 3145.

24. Halasz P., Kundra O., Rayna P., Pal I. and Vargha M. (1979) Microarousals during nocturnal sleep. Acta. Med. Acad. Sci. Hung. 54, 1 12.

25. Holmes G. L., McKeever M. and Adamson M. (1987) Absence seizures in children: clinical and electroencephalographic features. Ann. Neurol. 21, 268 273.

26. Hosford D. A., Clark S., Cao Z., Wilson W. A. Jr, Lin F., Morrisett R. A. and Huin A. (1992) The role of GABA B receptor activation in absence seizures of lethargic (lh/Ih) mice. Science 257, 398401.

27. Houser C. R., Vaughn J. E., Barber R. P. and Roberts E. (1980) GABA neurons are the major cell type of the nucleus reticularis thalami. Brain Res. 200, 431435.

28. Hughes J. R. (1985) Long-term clinical and EEG changes in patients with epilepsy. Arch. Neurol. 42, 213 223. 29. Inoue M., Peeters B. W. M. M., Van Luijtelaar E. L. J. M., Vossen J. M. H. and Coenen A. M. L. (1990) Spontaneous

occurrence of spike-wave discharges in five inbred strains of rats. Physiol. Behav. 48, 199 201. 30. Jand6 G., Siegel R. M., Horv~ith Z. and BuzsS_ki G. (1993) Pattern recognition of the electroencephalogram by artificial

neural network. Electroenceph. clin. Neurophysiol. 86, 100 109. 31. Jones E. G. (1985) The Thalamus. Plenum, New York. 32. Kellaway P. (1985) Sleep and epilepsy. Epilepsia 26, S15 $30. 33. King G. A. and Burnham W. M. (1982) Alpha-2 adrenergic antagonists suppress epileptiform EEG in a petit mal seizure

model. Life Sci. 30, 293-298. 34. Klingberg F. and Pickenhain L. (1968) Das Aufreten von "'Spindelentladungen" bei der Ratte in Bezeihung zum

Verhalten. Acta biol. med. Germ. 20, 45-54. 35. Knight A. R. and Bowery N. G. (1992) GABA receptors in rats with spontaneous generalized nonconvulsive epilepsy.

J. neural Transm. 35, 189 196. 36. Kostopoulos G., Gloor P., Pellegrini A. and Gotman J. (1981) A study of the transition from spindles to spike and

wave discharge in feline generalized penicillin epilepsy: microphysiological features. Expl Neurol. 73, 55 77. 37. Lennox W. G. and Lennox M. A. (1960) Epilepsy and Related Disorders, Little-Brown, Boston. 38. Li C. L., Jasper H. H. and Henderson L. (1952) The effect of arousal mechanisms on various forms of abnormality

in the electroencephalogram. Electroenceph. clin. Neurophysiol. 4, 513 526. 39. Liu Z., Vergnes M., Depaulis A. and Marescaux C. (1992) Involvement of intrathalamic GABA B neurotransmission

in the control of absence seizures in the rat. Neuroscience 48, 87 93. 40. Marescaux C., Micheletti G., Vergnes M., Depaulis A., Rumbach L. and Warter J. M. (1984) A model of chronic

spontaneous petit mal-like seizures in the rat: comparison with pentylenetetrazol-induced seizures. Epilepsia 25, 326 331.

41. Marescaux C., Vergnes M. and Depaulis A. (1992) Genetic absence epilepsy in rats from Strasbourg--a review. J. neural Transm. 35, Suppl. 37~9.

42. Matthes A. and Weber H. (1968) Klinische und elektroenzephalographische Familienuntersuchungen bei Pyknolepsien. Deutsche Medizinische Wissenschaft 8, 429~435.

43. McCormick D. A. (1992) Neurotransmitter actions in the thalamus and cerebral cortex and their role in neuromodu- lation of thalamocortical activity. Prog. Neurobiol. 39, 337 388.

44. McCormick D. A. and Pape H. (1990) Properties o fa hyperpolarization-activated cation current and its role in rhythmic oscillation in thalamic relay neurones. J. Physiol. Lond. 431, 291 318.

45. McKusick V. A. (1975) Mendelian Inheritance in Man. Catalogs of Autosomal Dominant, Autosomal Recessive. and X-linked Phenotypes, 4th edn. John Hopkins University Press, Baltimore.

Spike-and-wave epilepsy in rats 317

46. Metrakos J. D. and Metrakos K. (1970) Genetic factors in epilepsy. In Modern Problems o[" Pharmacopsyehiatry (ed. Niedermeyer E.), Vol. 4, Karger, Basel/New York. pp. 71 86.

47. Metrakos K. and Metrakos J. D. (1961) Genetics of convulsive disorders. II. Genetic and electroencephalographic studies in centrecephalic epilepsy. Neurology 11, 464483.

48. Micheletti G., Warter J.-M., Marescaux C., Depavlis A., Tranchart C., Rumbach L. and Vergnes M. (1987) Effects of drugs affecting noradrenergic neurotransmission in rats with spontaneous petit mal-like seizures. Eur. J. Pharmac. 135, 397 402.

49. Micheletti G., Vergnes M., Marescaux C., Reis J., Depavlis A., Rumbach L. and Warter J.-M. (1985) Antiepileptic drug evaluation in a new animal model: spontaneous petit mal epilepsy in the rat. Arzheim. Forsch./Drug Res. 35, 483 495.

50. Niedermeyer E. (1972) The Generalized Epilepsies. Charles C. Thomas, Springfield, IL. 51. Noebels J. (1986) Mutat ional analysis of inherited epilepsies. Adv. Neurol. 44, 97 113. 52. Noebels J. F. and Sidman R. L. (1979) Inherited epilepsy: spike and wave and focal seizures in the mutant mouse

tottering. Science 204, 1334 1336. 53. Noebels J. L. (1984) A single gene error of noradrenergic axon growth synchronizes central neurons. Nature 310,

409 ~11. 54. Noebels J. L. (19843 Isolating single genes of the inherited epilepsies. Ann. Neurol. 16, S18 $21. 55. Nufiez A., Curro Dossi R., Contreras D. and Steriade M. (1992) lntracellular evidence for incompatibility between

spindle and delta oscillations in thalamocortical neurons of cat. Neuroscience 1, 75 85. 56. Pape H. and McCormick D. A. (1989) Noradrenaline and serotonin selectively modulate thalamic burst firing by

enhancing a hyperpolarization-activated cation current. Nature 340, 715 718. 57. Peeters B. W. M. M., Kerbusch J. M. L., Van Luijtelaar E. L. J. M., Vossen J. M. H. and Coenen A. M. L. (19903

Genetics of absence epilepsy in rats. Behav. Genet. 20, 453460. 58. Peeters B. W. M. M.. Van Luijtelaar E. L. J. M., Coenen A. M. L. and Spooren W. P. J. M. (1988) The WAG/Rij

rat model for absence epilepsy: anticonvulsant drug evaluation. Neurosci. Res. Commun. 2, 93 97. 59. Peeters B. W. M. M.. Van Rijn C. M., Vossen J. M. H. and Coenen A. M. L. (19893 Effects of GABA-ergic agents

on spontaneous non-convulsive epilepsy, EEG and behaviour, in the WAG/Rij inbred strain of rats. L([~" Sci. 45, II71 1176.

60. Penfield W. and Jasper H. (1954) Epilepsy' and the Functional Anatom3 q[' the Human Brain; Little, Brown, Boston. 61. Penry J. K., Porter R. J. and Dreifuss F. E. (1975) Simultaneous recording of absence seizures with video tape and

electroencephalography: a study of 374 seizures with 48 patients. Brain 98, 427 440. 62. Porter R. J. (19893 Epilepsy: 100 Elementary Principles. WB Saunders, London. 63. Porter R. J. (1993) The absence epilepsies. Epilepsia 34, Suppl., $42 $48. 64. Riekkinen P. Jr, Sirvio J., Jakala P., Lammintausta R. and Riekkinen P. (1990) Effect of alpha: antagonists and an

agonist on EEG slowing induced by scopolamine and lesion of the nucleus basalis. Neuropharmacology 29, 993 999. 65. Rfitti W. (1982) Absenzen-Epilepsie im Erwachsenenalter. Sehweizer Medizinische Wissenschaft 112, 434441. 65a. Siegel R. M. (1990) Non-linear dinamical system theory and primary visual cortical processing. Physica D. 42, 385 395. 66. Sott6sz l., Lightowler S., Leresche S., Jassik-Gerschenfeld D., Pollard C. E. and Crunelli V. (1991) Two inward currents

and the transformation of low-frequency oscillations of rat and cat thalamocortical cells. J. Physiol. Lond. 441,175 197. 66a. Steriade M. (19743 lnterneuronal epileptic discharges related to spike-and-wave cortical seizures in behaving monkeys.

Eh, ctroenceph, olin. Neurophysiol. 37, 247 263. 67. Steriade M., Deschenes M., Domich L. and Mulle C. (1985) Abolition of spindle oscillation in thalamic neurons

disconnected from nucleus reticularis thalami. J. Neurophysiol. 54, 1473 1497. 68. Steriade M. and Llin~is R. R. (1988) The functional states of the thalamus and the associated neuronal interplay. Physiol.

Re~,. 68, 649 741. 68a. Steriade M., Curro Rossi R. and Nufiez A. (1991) Network modulation of a slow intrinsic oscillation of cat

thalamocortical neurons implicated in sleep delta waves: cortically induced synchronization and brainstem cholinergic suppression. J. Neurosci. I!, 3200 3217.

69. Steriade M., Nufiez A. and Amzica F. (1993) lntracellular analysis of relations between the slow (< 1 Hz) neocortical oscillation and other sleep rhythms of the electroencephalogram. J. Neurosci. 13, 3266 3283.

70. Terzano M. G., Parrino L., Anelli S. and Hal~isz P. (19893 Modulation of generalized spike-and-wave discharges during sleep by cyclic alternating pattern. Epilepsia 30, 772 781.

71. Terzano M. G., Mancia D., Salati M. R., Costani G., Decembrino A. and Parrino L. (19853 The cyclic alternating pattern as a physiologic component of normal N R E M sleep. Sleep 8, 137 145.

72. Terzano M. G., Parrino L., Fioriti G., Spaggiari M. C. and Piroli A. (1986) Morphologic and functional features of cyclic alternating pattern (CAP) sequences in normal N R E M sleep. Funct. Neurol. !, 2941 .

73. Vadasz C., Carpi D., Jand6 G., Kandel A., Urioste R., Horvfith Z., Pierre E., Vadi D., Fleischer A. and Buzsfiki G. (1994) Genetic threshold model of rodent spike-and-wave epilepsy. Am. J. Med. Genetics (in press).

74. Vadfisz C., Barber H . Fink S. and Reis D. J. (1985) Genetic effects and sexual dimorphism in tyrosine hydroxylase activity in two mouse trains and their reciprocal FI hybrids. J. Neurogenet. 2, 219 230.

75. Vergnes M., Marescaux C., Micheletti G., Depaulis A., Rumbach L. and Warter J. M. (19843 Enhancement of spike and wave discharges by GABAmimetic drugs in rats with spontaneous petit mal-like epilepsy. Neurosci. Lett. 44, 91 94.

76. Von Krosigk M., Bal T. and McCormick D. A. (1993) Cellular mechanisms of a synchronized oscillation in the thalamus. Science 261, 361 364.

77. Warter J.-M., Vergnes M., Depaulis A., Tranchant C., Rumbach L., Micheletti G. and Marescaux C. (19883 Effects of drugs affecting dopaminergic neurotransmission in rats with spontaneous petit-mal seizures. Neuropharmacology 27, 269 274.

78. Wolf P. (1985) Juvenile absence epilepsy. In Epileptic Syndromes in lnfimcy, Childhood and Adolescence (eds Roger J., Dravet C., Bureau M., Dreifuss F. E. and Wolf P.), pp. 242 246. John Libbey, London.

79. Wolf P. (1985) Juvenile myoclonic epilepsy. In Epileptic" Syndromes in lnfanc3,, Childhood and Adolescence (eds Roger J.. Dravet C., Bureau M., Dreifuss F. E. and Wolf P.), pp. 247 258. John Libbey, London.

(Accepted 25 May 19943