- Email: [email protected]
Spontaneous hippocampal interictal spikes following local kindling: time-course of change and relation to behavioral seizures
Brain Research, 5 I~ ( ~t~9q))308- 314 Elsevier
308 BRES 15355
Spontaneous hippocampal interictal spikes following local kindling" time-course of change and relation to behavioral seizures Lai-Wo Stun Leung Departments of Clinical Neurological Sciences and Physiology, University Hospital, University of Western Ontario, London, Ont. (Canada) (Accepted 5 September 1989)
Key words: Kindling; Interictal spike; Long-term potentiation; Seizure
Spontaneous interictai spikes (SISs) were recorded in the hippocampus in freely behaving rats following hippocampal stimulations that resulted in afterdischarges (ADs). Hippocampal SISs were detected after an average of 5 (range 2-10) daily ADs. The rate of SISs typically increased minutes after a tetanus, and then decayed with time constants of approximately 70 min and 1.5 days. Seizure onset in the kindling paradigm was not related to a consistent change in SIS rate. Following the interruption of daily kindling, SIS rate invariably decreased to near zero by 4-8 days while seizure susceptibility, as tested by the ability to evoke generalized convulsions, remained unchanged. Despite having a low or zero SIS rate the hippocampus seemed to retain an excitability after kindling interruption, as demonstrated by the observation that an average of 1.7 rekindling stimulations resulted in a high SIS rate. In conclusion, changes in hippocampal SISs were closely time-locked to an AD, and not to evoked behavioral seizures. Hippocampal SISs probably reflect an excitability change that is more local than that necessary for evoking behavioral convulsions. The persistence of SISs in terms of hours and days suggests the involvement of long-term potentiation.
INTRODUCTION O n e main feature during epileptogenesis is the spontaneous interictal spike (SIS). In h u m a n epilepsy, interictal spiking is used as one of the signs in localizing the e p i l e p t o g e n i c site, though its relation to spontaneous or elicited seizures is not always clear 6'26. In the kindling m o d e l of epilepsy, SISs induced by amygdaloid stimulations are k n o w n to develop and s p r e a d during the course of kindling 5'12'2z. H o w e v e r , a direct relation between seizure and interictal spikes is controversial. The rate of a m y g d a l o i d SIS a p p e a r e d to decline during the stage of e v o k e d convulsions t2'22, and Engel and A c k e r m a n 4 found that a high SIS rate may correlate with an increased r a t h e r than a decreased threshold for evoking an afterdischarge ( A D ) and seizure. In amygdala-kindled s p o n t a n e o u s seizing cats, W a d a et al. 23 r e p o r t e d a m a r k e d increase of interictal discharge activity prior to the onset of s p o n t a n e o u s convulsions, while G o t m a n concluded that 'the rate of spiking a p p e a r s to have no b e a r i n g on the p r o b a b i l i t y of occurrence of spontaneous seizures' in cats 7 and humans 8. On account of its regular layered structure, the h i p p o c a m p u s m a y offer a m o d e l for the study of the mechanisms underlying interictal spikes. M a n y studies have been d o n e on the cellular correlates of convulsant-
induced interictal spikes 1"3'27, but the relation of these spikes to behavioral seizures is not known. In our previous studies, we r e p o r t e d two types o f h i p p o c a m p a l interictal spikes which differed by their polarities across the CA1 cell layer, suggesting g e n e r a t i o n by basal or apical excitation respectively 15. We also r e p o r t e d a correlation of h i p p o c a m p a l SISs with the behavioral states of the rat 14. H i p p o c a m p a l SISs were maximal during states when the local E E G showed irregular activity, such as awake immobility, chewing and slowwave sleep (SWS). In this p a p e r , we present results that indicate that an increase in the rat h i p p o c a m p a l SISs was time-locked to an A D and lasted for a duration of hours and days. H o w e v e r , the rate of SISs in kindled rats was not directly related to seizure susceptibility. MATERIALS AND METHODS Hooded (Long-Evans) rats were obtained from Charles River (St. Constant, Que.). As described in detail elsewhere13a4, bipolar electrodes were implanted bilaterally in the hippocampus, with each electrode pair straddling the CA1 cell layer. In some rats, electrodes were also implanted in the basolateral amygdala. A minimum of 10 days was allowed for recovery from surgery, followed by two sessions of baseline EEG recordings during SWS and waking immobility. All recordings were referred monopolarly to a screw in the skull over the frontal cortex or the cerebellum. Daily tetanic
Correspondence: L.S. Leung, Departments of Clinical Neurological Sciences and Physiology, University Hospital, University of Western Ontario London, Ont., Canada, N6A 5A5 0006-8993/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
309 stimulations were then delivered. The stimulation was a 1-5 s train of pulses (0.1 ms duration) at 200 Hz, and the intensity (40-300/~A) was adjusted so as to evoke an AD on every trial, typically 25-50 /~A above the AD threshold (see ref. 13 for AD threshold determination). SISs were detected visually from the polygraph, usually with the help of a spike analyzer. The latter device (Mentor 750) employed a high pass filter (3 dB at 150 Hz) and a voltage window. Standard pulses were triggered whenever the high-pass filtered EEG signal crossed an arbitrarily selected threshold. The latter threshold was adjusted for EEG from each electrode, in order that the electronically selected spikes matched the visually detected ones. In addition, EEG transients including sharp waves2 from normal rats before any tetanic stimulation were only counted infrequently (< 1/ min), since by definition, the prekindling EEG should have no interictal spike activity. At least 5 min of EEG during SWS before kindling was recorded and analyzed. The criteria for selection of SISs remained the same for each rat throughout the course of kindling. A microcomputer program was specially designed to record the full details of a SIS before and after its sharp rising phase TM. The sampling rate used for SISs was 1-2 kHz at each of the 4-5 channels of EEG, sampled almost simultaneously (5 channels sampled within 40/~s). In a few experiments, in particular those examining shortduration (< 24 h) changes in SISs, only polygraph tracings were used for SIS rate determination, with or without the aid of a window discriminator. The polygraph was set at a speed of 10 mm/s and a gain of 1 mV/em with 4 channels (2 left and 2 right) of hippocampal EEG. SISs were assessed visually from one hippocampus using both sharpness (< 50 ms duration) and amplitude (> 2 mV at deep electrode, or > 1 mV at surface electrode) criteria; electrodes straddling the CA1 cell layer had to show the same SIS with reversed polarity. SIS assessment was done without the knowledge of the time of the recording. At least 5 min of SWS recording was assessed for SISs at each time, typically on the h at 1-8 h and at 24 h after an AD. All electrode sites were confirmed by histology in 40/~m coronal sections of the brain, stained with gallocyanin.
c o r r e l a t e d with a high and relatively consistent spike rate ~4. The state of waking i m m o b i l i t y was sometimes used for recording SIS rates within 30 min following an A D since an animal rarely slept i m m e d i a t e l y after an A D . SIS rates during waking immobility a n d SWS were c o m p a r a b l e ~4. While there was some variability, the rate of SIS was typically highest at 10-20 min after an A D , declined quickly within the first h o u r (Fig. 2 A ) , increased slightly to a relatively stable level in the 2nd to 7th h (Fig. 2B), and then declined slowly over days (below). SIS rates typically increased after an A D (Fig. 2B), especially when the SIS rates b e f o r e the A D were m o d e r a t e l y high ( > 2/min). H o w e v e r , a decrease in SIS rate after an A D was occasionally found, particularly when SIS rate before an A D was low. The decay of SIS rate during the first hour was quite variable. In 38 e x p e r i m e n t s in 6 rats, a decay of SIS rate from 2 0 - 3 0 min to 60-70 min was o b s e r v e d in m o r e than 80% of the experiments, i.e., a decay in SIS rate was consistently found ( P < 0.005, X2-test). T h e decay rate was, however, quite variable and estimates of an exponential decay time constant (assuming a single time constant) varied from 13 to 533 min. T h e decay time constant in Fig. 2 A was 30 min. W h e n e x t r e m e values were excluded for each rat, the m e a n time constant was 70 + 22 min (mean + S . E . M . , 6 rats). T h e latter should be r e g a r d e d as an o r d e r of m a g n i t u d e rather than a precise estimate. h
E
prekindling
RESULTS
Changes in SIS rate following an A D T h e presence of sharp waves in the normal, unstimulated h i p p o c a m p u s m a d e the discrimination of SISs s o m e w h a t m o r e difficult. In practice, however, m a n y SISs were clearly distinguished in p o l y g r a p h tracings, with or without the aid of a slope detector. In the e x a m p l e shown in Fig. 1, SISs were distinct by being large ( > 2 m V ) , biphasic and fast (as d e t e c t e d by the window discriminator, not shown). N o r m a l sharp waves were of 40-100 ms d u r a t i o n 2'~° (Fig. 1E) and usually less than 2 m V when r e c o r d e d with the 125/~m wire-electrodes. In contrast, SISs were typically < 30 ms duration, higha m p l i t u d e and sometimes multiphasic (Fig. 1F). A n o t h e r t y p e of SIS had an initial polarity of surface-negative and apical-dendritic positive (Fig. 1G), suggesting basal dendritic excitation 25. T h e latter polarity was not found in n o r m a l sharp waves. Since the rat was free to behave and the SIS rate d e p e n d e d on b e h a v i o r 14, changes in SIS rate could only be studied when the behavioral state was fixed. T h e state of SWS was selected since it was readily o b s e r v e d and it
post 36fh AD
65 rain
B F
2 days G L] D
6 days
'
Fig. 1. A -D: EEG recording at L1 recorded during SWS (R296). A: before any tetanic stimulations (kindling). B: 65 min after the 36th AD (5th stage 5 seizure), black squares denote interictal spikes. C: 2 days after the 36th AD, and D: 6 days after the 36th AD. E: detailed digitized tracing of a normal sharp wave recorded before kindling. F: spontaneous interictal spike (type I) 2 days after the 36th AD; note the high amplitude and biphasic waveform of the interictal spike, G: type II interictal spike, note reversed initial polarity as compared to type I above; infrequent spike type for this rat. Calibration: left, 1 mV and 1 s for A-D; right, 1 mV and 20 ms for E-G. L1, deep electrode; L2, surface electrode.
310 The stage of kindling did not seem to be a major factor in the temporal changes in SIS rate following an AD, provided that SISs were observed before an AD.
Changes of SIS rate with kindling Interictal spikes were absent or very low before the onset of kindling stimulations. All animals demonstrated an increase in interictal spiking after 2-10 ADs. A SIS rate of > 2/min at > 30 min post-AD was used as the criterion for the onset of interictal spiking. The average number of ADs to evoke interictal spikes (measured at > 30 min post-AD) was 5 + 1 (mean + S.E.M., n = 10). In order to find the optimal time for recording SISs, two rats were recorded during SWS before an AD, and at about 30 min and 60 min after an AD. Each rat was left in the recording cage for most of the day and two ADs were elicited daily, separated by at least 3 h. In both rats, there was a high correlation (correlation coefficient > 0.75) between the SIS rates pre- and post-AD (Table I, Fig. 3). In 7 animals, 6 stimulated once a day (Fig. 4) and the
TABLE 1 Correlation coefficients between SIS rates
preAD, SIS rate before an AD; 30m, SIS rate at 30 min after an AD; 60m, SIS rate at 60 min after an AD. All rates measured during slow-wave sleep or waking immobility, number of observations in brackets, all correlation coefficients are significantly different from zero (P < 0.001)
R210 R2! ]
PreA D vs 30m
PreA D vs 60m
30m vs 60m
0.78 (13) 0.75 (23)
0.84 (23) 0.83 (27)
0.83 (11) 0.95 (22)
150
--t3- preK A
30m postK 60m postK
/
c 100
E (1) 4-,
/
2 O3
,/
!
50-
-# 1.4
~ o O;
i
0
A
1 0
50
100
30
#
200
150
60
90
afterdischarges
minutes poseD 150 6) u3 .E
100-
*o
E O ~.o
o 50-
a
0)
2o co
B O
a
5
10
15
20
~
o
25
hours postAD
Fig. 2. A: changes in logarithmic SIS rate following an AD in a representative experiment. SIS rate decayed quickly in the first h, as indicated by the experimental points (squares) and best-fitted line (crosses) for a single exponential decay for rates during the first h (decay time constant = 30.1 min; correlation coefficient = 0.96). B: changes in the SIS rate following an A D averaged across 6 different experiments (rats). The maximal SIS rate at 1-3 h post-AD was normalized to 100%, and the rates were then averaged. Error bars are one standard error of the mean. Note a significant increase in SIS rate after the AD; SIS rate before the A D is shown at time zero. SIS rate showed a small increase from 1 to 3 h post-AD and then remained relatively stable for several h.
o,-
lO
20
3o
40
SIS rate preK ( / r a i n )
Fig. 3. A: relation of the rates of interictal spikes immediately before (preK), 20--30 rain after (30 m postK) and 50-70 rain after (60 m postK) a stimulation train (K = kindling brain) that produced an AD each time. The last (85th) AD corresponded to a bilateral clonic seizure. Note the general positive correlations between spike rates pre- and post-kindling; spiking rates at each time interval were not recorded for every AD. B: Scatter plot of the relation between SIS rates before and at 50-70 min after a kindling train, for different days of recording. The straight line is the best linear fit through the points y = 3.31 + 1.23x, all units in min 1. Correlation coefficient is 0.83.
311
81
R99
1501
0 ................
• " ~ 30
o.
16
.
.
.
.
•
30]
0 .......................................... 0 44
R103
.
R10~
R102
..........................................................
R104
.
o I. . . . . . . . . . . . . . . . .
.
51
R196
o
58
..............................
61
38
session Fig. 4. The change in hippocampal SISs recorded in SWS during the course of kindling for 6 rats. Number of A D s is shown on the horizontal axis for each rat. The number of ADs to achieve the first generalized seizure with bilateral clonus is given. A l l generalized seizures are circled and after the onset of the first generalized seizure, non-seizure days are marked by a triangle.
seventh twice a day (Fig. 3A), the SIS rate after an AD was measured during the course of kindling, until the occurrence of at least one generalized seizure with bilateral clonus, accompanied by rearing (stage 4 of Racine17). SIS rate was measured during SWS at approximately 45-120 min after the kindling train. For practical reasons, SISs during SWS were recorded after each AD, since SWS occurred more readily after than before an AD, possibly because of the central effects of the AD and increased habituation to the recording cage. The first behavioral convulsion (bilateral clonus) required an average of 44.4 + 8.7 ADs (n = 7; mean + S.E.M.). The change in hippocampal SIS frequency during the course of kindling was rather variable (Fig. 3A, Fig. 4). In 4 of 7 rats, the peak SIS rate was clearly in the middle-late period during kindling, before the development of evoked generalized seizures. In others (R101, R103, R196), seizures were related to some of the highest spike rates. However, it may be concluded that the SIS rate did not predict seizure onset. After seizure onset, occasional days without generalized seizures (triangles in Fig. 4) were not related to any changes in spike rate. In summary, whether and when convulsions occurred or not in the course of kindling could not be predicted by the SIS rate. The relative independence of seizure susceptibility and SIS rate was more emphatically demonstrated by the following experiment. In 9 experiments on 8 rats, ~iaily kindling was interrupted after 5-8 generalized behavioral seizures were evoked. Three of the animals (295, 296, 297) were stimulated up to 5 times a day initially until one generalized seizure and then put on a daily stimulation schedule. SIS rate during SWS was measured for several sessions before and at various days following kindling interruption, for a period of 4-8 days. SIS rate in all rats
declined when recorded at > 1 day following the last stimulation train (Figs. 1 and 5). Resumption of daily kindling increased the spike rate in all rats (Fig. 5). Without exceptions, the first restimulation after 4-8 days of interruption evoked a generalized behavioral seizure. The primary AD duration on the first restimulation measured 57.2 + 5.5 s (mean _+ S.E.M., n = 9), not statistically different from the AD duration of 52.3 + 8 s (n = 9) before the interruption of kindling. Thus, a decline in SIS rate was not accompanied by a decrease in seizure susceptibility or AD duration.
250
.~
200
150
z~
.N ~ 0 E
100
50
_ 1
3
5
7
8
x T kl
k3
days postK stop Fig. 5. Normalized SIS rate in 9 experiments on 8 fully kindled (seizing) rats following the interruption of daily kindling, each experiment represented by a different symbol. SIS rate at 1-2 h post-kindling was used for normalization. Not all days were recorded for all the rats and the period of kindling interruption ranged from 4 to 8 days. Resumption of daily kindling resulted in an increase of SIS rate, in some cases, above 100%. All rats had a generalized seizure on the first day of rekindling.
312
-
no~-~eizure
D
o _2 i i 2
6
days postK stop Fig. 6. Semi-logarithmic plot of the average normalized SIS rate for the 9 experiments on fully kindled (seizing) rats shown in Fig. 5 (open square) and for 6 non-seizing rats (crosses). The straight line is the best linear regression fit for the first 6 days for the seizing rats with slope -0.28 __+0.02. The best regression line for the non-seizing rats (not shown) has a slope of -0.29 + 0.07, not significantly different from that of seizing rats.
While the SIS rate declined to almost zero during the period of kindling interruption, the hippocampus apparently retained a higher level of excitability as demonstrated by the ease in rekindling a high SIS rate. In all 8 rats, 1-3 rekindling sessions resulted in high SIS rates ( > 2/min). The mean number of ADs to rekindle SISs was 1.7 + 0.3 (n = 9), which was significantly smaller than the mean number of 5 + 1 (n = 10) ADs necessary for non-stimulated rats (P < 0.01, Mann-Whitney U-test). Thus, the hippocampus seemed to have retained a 'latent' excitability even though the rate of SISs following kindling interruption was near zero. The decline of SIS rate over days after kindling cessation varied somewhat across different experiments. However, the average decline was well represented by a single exponential decay curve. In Fig. 6, a semilogarithmic plot of the average SIS rate for each day (pooled across experiments) after the last AD was approximately linear, with a decay time constant of 1.57 days. In a different group of 6 rats that did not have behavioral signs of seizures after 4-19 ADs, the average SIS rate also declined with days of interruption of stimulations (Fig. 6). A decay time constant of 1.5 days was found (Fig. 6), which was not significantly different from that of the kindled rats with convulsions. DISCUSSION
Relation of SIS with kindling and seizures The SlSs reported in this study were clearly E E G transients that were not found in the normal unkindled animals. The duration of electrode implants for rats which underwent the complete kindling procedure was 1-4 months, all with SIS onset during the first month. In other studies, we had recorded and analyzed E E G from
non-kindled animals implanted for several months, and had not observed any SISs. In addition, the decay of SISs with time (Figs. l and 5) clearly indicated that SISs were not caused by tissue disruption by the implanted electrodes. The frequency of SISs recorded at a fixed period alter an A D appeared to peak during the middle-late course of kindling, and not usually during periods of convulsions. In fact, a decrease in SIS rate was sometimes observed after the stage of generalized convulsions was reached. Convulsions may be related to either a high or a low SIS rate, or in other words, the hippocampal SIS rate does not predict behavioral seizure. The peaking of SIS rate before convulsion onset was noted for amygdaloid SIS in the cat x2"22. Hippocampal SISs are apparently controlled by factors that are different from those governing behavioral seizures. Following interruption of daily stimulations, the decay time constant of SISs was approximately 1.5 days, and very few SISs were detected after 5 days. In the mean time, seizure susceptibility, as determined by the ability of the stimulus to evoke a generalized convulsion, remains undiminished. Thus, a change in SIS rate was not accompanied necessarily by a change in seizure susceptibility. Frequently repeated train stimulations may increase SIS rate but decrease A D or seizure susceptibility 4. Kairiss et al. 1o found few hippocampal SISs after kindling, which is at odds with the present report and with other literature 5'12'25. While Kairiss et al. recorded 24 h or more after the last AD, significant spiking was still observed in the hippocampus at this time (Figs. 1C, 2B, 5). The rate of amygdaloid SISs in amygdala-kindled animals was found to decline with a half-life of 1-2 days 5 or to disappear within 8 days 7. Waiters, in an unpublished M.A. thesis cited by Wada et al. 23, found that amygdaloid SISs were seldom detected more than 6 days after the last kindling stimulation. Waiters also concluded that seizure susceptibility was unrelated to the SIS activity. However, Wada et al. 23 reported a marked increase in SISs before spontaneous seizures in cats, though Gotman found no consistent changes in SIS rate before spontaneous seizures in cats 7 and humans s, While this paper offers no data on the issue of whether interictal spikes precede a spontaneous seizure, evidence has been presented that SISs followed an AD or seizure in the hippocampus. The literature indicate that local SISs followed an amygdaloid A D in animals 5'7 and in human spontaneous seizures s. Posterior cingulate kindling, however, only gave few SISs at any stage of kindling aS. We demonstrated a retention of local excitability without actual interictal spiking in the kindled rats. This
313 'latent' excitability is inferred from the ease of rekindling a high SIS rate; its permanence has not been studied. Other researchers have found an increased capability of in vitro slices from kindled animals to demonstrate interictal spiking when exposed to a high extracellular [K ÷ ]011. Increased Ca 2 ÷ uptake in CA1 dendrites 24 and increased N M D A receptor involvement in normal synaptic transmission in the dentate gyrus 16 were also found in slices from kindled rats. Relation of SISs to A D and physiology Many factors could have contributed to the variability of SIS rates immediately following an AD. Ongoing experiments indicate that the early component of the evoked potential in CA1 region following Schaffer's collateral stimulation was depressed for up to 3 h post-AD. This long-lasting depression may contribute to the partial suppression of SISs within the first 3 h after kindling (Fig. 2), resulting in the 'peak' in SIS rates at 2-7 h post-AD in some experiments. Gotman 7 observed peaks in amygdaloid SISs at 7-15 h post-AD. Hippocampal SISs were time-locked to the AD. There was some variability in the decay of the SISs with time, but approximate time constants of 70 min and 1.5 days were observed. The relatively long time constants of hippocampal SISs suggest a mechanism of long-term plasticity, possibly long-term potentiation (LTP). Racine et al.lS reported hippocampal LTP to decay with two time constants of 1.5 h and 5 days respectively. The similarities in the order of magnitudes between time-constants of SISs and enhanced evoked potentials (LTP) suggest a common mechanism underlying the two events. On the other hand, there are difficulties with the hypothesis that hippocampal SISs are caused by an enhancement of synaptic response (LTP) at a single synapse. First, the particular synapse has not been identified and preliminary data indicated that it was not
REFERENCES 1 Ayala, G.E, Dichter, M., Gumnit, R.J., Matsumoto, H. and Spencer, W.A., Genesis of epileptic interictal spikes, new knowledge of cortical feedback systems suggests a neurophysiological explanation of brief paroxysms, Brain Research, 52 (1973) 1-17. 2 Buzsaki, G., Leung, L.S. and Vanderwolf, C.H., Cellular bases of hippocampal EEG in the behaving rat, Brain Res. Rev., 6 (1983) 139-171. 3 Dichter, M. and Spencer, W.A., Penicillin-induced interictal discharges from the cat hippocampus. II. Mechanisms underlying origin and restriction, J. Neurophysiol., 32 (1969) 663-687. 4 Engel, J. Jr. and Ackerman, R.E, Interictal EEG spikes correlate with decreased, rather than increased epileptogenicity in amygdaloid kindled rats, Brain Research, 190 (1980) 543-548. 5 Fitz, J.G. and McNamara, J.O., Spontaneous interictal spiking in the awake kindled rat, Electroenceph. Clin. Neurophysiol., 47 (1979) 592-596.
the CA3-CA1 synapse, since LTP at the latter synapse should be blocked by N-methyl-o-aspartate antagonist 9 while SIS rate increase following an A D was not. Second, LTP could be demonstrated after a single tetanus or A D while SISs were found only after several ADs. Obviously, spontaneous interictal spiking is a different phenomenon than evoked responses commonly used to indicate the presence of LTP. The tonic and phasic inputs to the hippocampus (e.g. conditions in SWS) and multiple synapses or circuits may play a role in the generation of the SIS. Evidence so far points to plasticity within but not necessarily restricted to the hippocampal CA3 region. Most CA3 neurons possess an intrinsic capability to firing in burst 19'28. Synchronized bursts in CA3 probably generate the spontaneous sharp waves (a population EPSP) in CA1 region of normal unkindled rats 2. In vitro, when perfused with penicillin or bicuculline, CA3 drives interictal spiking recorded in the CA1 region 27. The bilaterality and polarity of CA1 interictal spikes 25 can be explained by the distribution of CA3 afferents bilaterally and in both the apical and basal dendritic layers of CA121. This and previous papers on hippocampal SISs 14'25 suggest that SISs are closely related to the normal physiology of the hippocampus. Paroxysmal interictal spiking may disrupt normal hippocampal functions and contribute to memory and cognitive deficits in epileptic patients. On account of the wealth of anatomical and physiological knowledge of the rat's hippocampus, hippocampal interictal spikes in vivo may offer a model for the study of the relation of interictal spikes to function.
Acknowledgements. This study was supported by NIH Grant NS25383 and NSERC Grant A1037. I thank Wendy Charlton, Kathy Boon and Elizabeth Szczutkowski for their patience and technical assistance, Bey Hughes for typing, and Drs. P. Cain and R. McLachlan for useful comments on the manuscript.
6 Gloor, P., Contributions of electroencephalography and electrocorticography to the neurosurgical treatment of epilepsies. In D.P. Purpura, J.K. Penry and R.D. Walter (Eds.), Neurosurgical Managementof the Epilepsies, Raven, New York, 1975, pp. 59-105. 7 Gotman, J., Relationships between triggered seizures, spontaneous seizures, and interictal spiking in the kindling model of epilepsy, Exp. Neurol., 84 (1984) 259-273. 8 Gotman, J. and Marciani, M.G., Electroencephalographic spike activity, drug levels and seizure occurrence in epileptic patients, Ann. Neurol., 17 (1985) 597-603. 9 Harris, E.W., Ganong, A.H. and Cotman, C.W., Long-term potentiation in the hippocampus involves activation of Nmethyl-D-aspartate receptors, Brain Research, 323 (1984) 132137. 10 Kairiss, E.W., Racine, R.J. and Smith, G.K., The development of the interictal spike during kindling in the rat, Brain Research, 322 (1984) 101-110. 11 King, G.L., Dingledine, R., Giacchino, J.L. and McNamara,
314
12
13 14 15 16 17 18 19
20
J.O., Abnormal neuronal excitability in hippocampal slices from kindled rats, J. Neurophysiol., 54 (1985) 1295-1304. Lange, H., Tanaka, T. and Naquet, R., Temporal-spatial pattern of subcortical spike activity in kindling epilepsy. A statistical approach, Electroenceph. Clin. Neurophysiol., 42 (1976) 564574. Leung, L.S., Hippocampal electrical activity following repetitive local stimulation. I. Afterdischarges, Brain Research, 419 (1987) 173-187. Leung, L.S., Hippocampal interictal spikes induced by kindling: relations to behavior and EEG, Behav. Brain. Res., 31 (1988) 75-84. Leung, L.S. and Boon, K.A., Kindling in the posterior cingulate cortex: electrographic and behavioral characteristics, Electroencephalogr. Clin. Neurophysiol., in press. Mody, I. and Heinemann, U., NMDA receptors of dentate gyrus granule cells participate in synaptic transmission following kindling, Nature (Lond.), 326 (1987) 701-704. Racine, R., Modification of seizure activity by electrical stimulation: II. Motor seizure, Electroenceph. Clin. Neurophysiol., 32 (1972) 281-294. Racine, R.J., Milgram, N.W. and Hafner, S., Long-term potentiation phenomena in the rat limbic forebrain, Brain Research, 260 (1983) 217-231. Ranck, J.B., Behavioral correlates and firing repertoires of neurons in the dorsal hippocampal formation and septum of unrestrained rats. In: Isaacson, R.L. and Pribram, K.H. (Eds.), The Hippocampus, Vol. 2, Plenum, New York, 1975, pp. 207-246. Suzuki, S.S. and Smith, G.K., Spontaneous EEG spikes in the
21 22
23 24 25
26
27 28
normal hippocampus. I. Behavioral correlates, laminar profiles and bilateral synchrony, Electroenceph. Clin. Neurophysiol.. 67 (1987) 348-359. Swanson, L.W., Wyss, J.M. and Cowan, W.M., An autoradiographic study of the organization of intrahippocampal association pathways in the rat, J. Comp. Neurol., 181 (1978) 681-716. Wada, J.A. and Sato, M., Generalized convulsive seizure induced by daily electrical stimulation of the amygdala in cats: correlative electrographic and behavioral features. Neurology~ 24 (1974) 565-574. Wada, J.A., Sato, M. and Corcoran, M.E., Persistent seizure susceptibility and recurrent spontaneous seizures in kindled cats, Epilepsia, 15 (1974) 465-478. Wadman, W.J., Heinemann, U., Konnerth, A. and Neuhaus, S. Hippocampal slices of kindled rats reveal calcium involvement in epileptogenesis, Exp. Brain Res., 57 (1985) 404-407. Wadman, W.J., Lopes da Silva, EH. and Leung, L.S., Two types of spontaneous interictal transients during kindling of the rat hippocampus, Electroenceph. Clin. Neurophysiol., 55 (1983) 314-319. Wieser, H.G., Bancaud, J., Talairach, J., Bonis, A. and Szikla, G., Comparative value of spontaneous and chemically and electrically induced seizures in establishing the lateralization of temporal lobe seizures, Epilepsia, 20 (1979) 47-59. Wong, R.K.S. and Traub, R.D., Synchronized burst discharge in disinhibited hippocampal slice. I. Initiation in CA2-CA3 region, J. Neurophysiol., 49 (1983) 442-458. Wong, R.K.S., Prince, D.A. and Basbaum, A.I. Intradendritic recordings from hippocampat neurons, Proc. Natl. Acad. Sci. U.S.A,, 76 (1979) 986-990.
308 BRES 15355
Spontaneous hippocampal interictal spikes following local kindling" time-course of change and relation to behavioral seizures Lai-Wo Stun Leung Departments of Clinical Neurological Sciences and Physiology, University Hospital, University of Western Ontario, London, Ont. (Canada) (Accepted 5 September 1989)
Key words: Kindling; Interictal spike; Long-term potentiation; Seizure
Spontaneous interictai spikes (SISs) were recorded in the hippocampus in freely behaving rats following hippocampal stimulations that resulted in afterdischarges (ADs). Hippocampal SISs were detected after an average of 5 (range 2-10) daily ADs. The rate of SISs typically increased minutes after a tetanus, and then decayed with time constants of approximately 70 min and 1.5 days. Seizure onset in the kindling paradigm was not related to a consistent change in SIS rate. Following the interruption of daily kindling, SIS rate invariably decreased to near zero by 4-8 days while seizure susceptibility, as tested by the ability to evoke generalized convulsions, remained unchanged. Despite having a low or zero SIS rate the hippocampus seemed to retain an excitability after kindling interruption, as demonstrated by the observation that an average of 1.7 rekindling stimulations resulted in a high SIS rate. In conclusion, changes in hippocampal SISs were closely time-locked to an AD, and not to evoked behavioral seizures. Hippocampal SISs probably reflect an excitability change that is more local than that necessary for evoking behavioral convulsions. The persistence of SISs in terms of hours and days suggests the involvement of long-term potentiation.
INTRODUCTION O n e main feature during epileptogenesis is the spontaneous interictal spike (SIS). In h u m a n epilepsy, interictal spiking is used as one of the signs in localizing the e p i l e p t o g e n i c site, though its relation to spontaneous or elicited seizures is not always clear 6'26. In the kindling m o d e l of epilepsy, SISs induced by amygdaloid stimulations are k n o w n to develop and s p r e a d during the course of kindling 5'12'2z. H o w e v e r , a direct relation between seizure and interictal spikes is controversial. The rate of a m y g d a l o i d SIS a p p e a r e d to decline during the stage of e v o k e d convulsions t2'22, and Engel and A c k e r m a n 4 found that a high SIS rate may correlate with an increased r a t h e r than a decreased threshold for evoking an afterdischarge ( A D ) and seizure. In amygdala-kindled s p o n t a n e o u s seizing cats, W a d a et al. 23 r e p o r t e d a m a r k e d increase of interictal discharge activity prior to the onset of s p o n t a n e o u s convulsions, while G o t m a n concluded that 'the rate of spiking a p p e a r s to have no b e a r i n g on the p r o b a b i l i t y of occurrence of spontaneous seizures' in cats 7 and humans 8. On account of its regular layered structure, the h i p p o c a m p u s m a y offer a m o d e l for the study of the mechanisms underlying interictal spikes. M a n y studies have been d o n e on the cellular correlates of convulsant-
induced interictal spikes 1"3'27, but the relation of these spikes to behavioral seizures is not known. In our previous studies, we r e p o r t e d two types o f h i p p o c a m p a l interictal spikes which differed by their polarities across the CA1 cell layer, suggesting g e n e r a t i o n by basal or apical excitation respectively 15. We also r e p o r t e d a correlation of h i p p o c a m p a l SISs with the behavioral states of the rat 14. H i p p o c a m p a l SISs were maximal during states when the local E E G showed irregular activity, such as awake immobility, chewing and slowwave sleep (SWS). In this p a p e r , we present results that indicate that an increase in the rat h i p p o c a m p a l SISs was time-locked to an A D and lasted for a duration of hours and days. H o w e v e r , the rate of SISs in kindled rats was not directly related to seizure susceptibility. MATERIALS AND METHODS Hooded (Long-Evans) rats were obtained from Charles River (St. Constant, Que.). As described in detail elsewhere13a4, bipolar electrodes were implanted bilaterally in the hippocampus, with each electrode pair straddling the CA1 cell layer. In some rats, electrodes were also implanted in the basolateral amygdala. A minimum of 10 days was allowed for recovery from surgery, followed by two sessions of baseline EEG recordings during SWS and waking immobility. All recordings were referred monopolarly to a screw in the skull over the frontal cortex or the cerebellum. Daily tetanic
Correspondence: L.S. Leung, Departments of Clinical Neurological Sciences and Physiology, University Hospital, University of Western Ontario London, Ont., Canada, N6A 5A5 0006-8993/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
309 stimulations were then delivered. The stimulation was a 1-5 s train of pulses (0.1 ms duration) at 200 Hz, and the intensity (40-300/~A) was adjusted so as to evoke an AD on every trial, typically 25-50 /~A above the AD threshold (see ref. 13 for AD threshold determination). SISs were detected visually from the polygraph, usually with the help of a spike analyzer. The latter device (Mentor 750) employed a high pass filter (3 dB at 150 Hz) and a voltage window. Standard pulses were triggered whenever the high-pass filtered EEG signal crossed an arbitrarily selected threshold. The latter threshold was adjusted for EEG from each electrode, in order that the electronically selected spikes matched the visually detected ones. In addition, EEG transients including sharp waves2 from normal rats before any tetanic stimulation were only counted infrequently (< 1/ min), since by definition, the prekindling EEG should have no interictal spike activity. At least 5 min of EEG during SWS before kindling was recorded and analyzed. The criteria for selection of SISs remained the same for each rat throughout the course of kindling. A microcomputer program was specially designed to record the full details of a SIS before and after its sharp rising phase TM. The sampling rate used for SISs was 1-2 kHz at each of the 4-5 channels of EEG, sampled almost simultaneously (5 channels sampled within 40/~s). In a few experiments, in particular those examining shortduration (< 24 h) changes in SISs, only polygraph tracings were used for SIS rate determination, with or without the aid of a window discriminator. The polygraph was set at a speed of 10 mm/s and a gain of 1 mV/em with 4 channels (2 left and 2 right) of hippocampal EEG. SISs were assessed visually from one hippocampus using both sharpness (< 50 ms duration) and amplitude (> 2 mV at deep electrode, or > 1 mV at surface electrode) criteria; electrodes straddling the CA1 cell layer had to show the same SIS with reversed polarity. SIS assessment was done without the knowledge of the time of the recording. At least 5 min of SWS recording was assessed for SISs at each time, typically on the h at 1-8 h and at 24 h after an AD. All electrode sites were confirmed by histology in 40/~m coronal sections of the brain, stained with gallocyanin.
c o r r e l a t e d with a high and relatively consistent spike rate ~4. The state of waking i m m o b i l i t y was sometimes used for recording SIS rates within 30 min following an A D since an animal rarely slept i m m e d i a t e l y after an A D . SIS rates during waking immobility a n d SWS were c o m p a r a b l e ~4. While there was some variability, the rate of SIS was typically highest at 10-20 min after an A D , declined quickly within the first h o u r (Fig. 2 A ) , increased slightly to a relatively stable level in the 2nd to 7th h (Fig. 2B), and then declined slowly over days (below). SIS rates typically increased after an A D (Fig. 2B), especially when the SIS rates b e f o r e the A D were m o d e r a t e l y high ( > 2/min). H o w e v e r , a decrease in SIS rate after an A D was occasionally found, particularly when SIS rate before an A D was low. The decay of SIS rate during the first hour was quite variable. In 38 e x p e r i m e n t s in 6 rats, a decay of SIS rate from 2 0 - 3 0 min to 60-70 min was o b s e r v e d in m o r e than 80% of the experiments, i.e., a decay in SIS rate was consistently found ( P < 0.005, X2-test). T h e decay rate was, however, quite variable and estimates of an exponential decay time constant (assuming a single time constant) varied from 13 to 533 min. T h e decay time constant in Fig. 2 A was 30 min. W h e n e x t r e m e values were excluded for each rat, the m e a n time constant was 70 + 22 min (mean + S . E . M . , 6 rats). T h e latter should be r e g a r d e d as an o r d e r of m a g n i t u d e rather than a precise estimate. h
E
prekindling
RESULTS
Changes in SIS rate following an A D T h e presence of sharp waves in the normal, unstimulated h i p p o c a m p u s m a d e the discrimination of SISs s o m e w h a t m o r e difficult. In practice, however, m a n y SISs were clearly distinguished in p o l y g r a p h tracings, with or without the aid of a slope detector. In the e x a m p l e shown in Fig. 1, SISs were distinct by being large ( > 2 m V ) , biphasic and fast (as d e t e c t e d by the window discriminator, not shown). N o r m a l sharp waves were of 40-100 ms d u r a t i o n 2'~° (Fig. 1E) and usually less than 2 m V when r e c o r d e d with the 125/~m wire-electrodes. In contrast, SISs were typically < 30 ms duration, higha m p l i t u d e and sometimes multiphasic (Fig. 1F). A n o t h e r t y p e of SIS had an initial polarity of surface-negative and apical-dendritic positive (Fig. 1G), suggesting basal dendritic excitation 25. T h e latter polarity was not found in n o r m a l sharp waves. Since the rat was free to behave and the SIS rate d e p e n d e d on b e h a v i o r 14, changes in SIS rate could only be studied when the behavioral state was fixed. T h e state of SWS was selected since it was readily o b s e r v e d and it
post 36fh AD
65 rain
B F
2 days G L] D
6 days
'
Fig. 1. A -D: EEG recording at L1 recorded during SWS (R296). A: before any tetanic stimulations (kindling). B: 65 min after the 36th AD (5th stage 5 seizure), black squares denote interictal spikes. C: 2 days after the 36th AD, and D: 6 days after the 36th AD. E: detailed digitized tracing of a normal sharp wave recorded before kindling. F: spontaneous interictal spike (type I) 2 days after the 36th AD; note the high amplitude and biphasic waveform of the interictal spike, G: type II interictal spike, note reversed initial polarity as compared to type I above; infrequent spike type for this rat. Calibration: left, 1 mV and 1 s for A-D; right, 1 mV and 20 ms for E-G. L1, deep electrode; L2, surface electrode.
310 The stage of kindling did not seem to be a major factor in the temporal changes in SIS rate following an AD, provided that SISs were observed before an AD.
Changes of SIS rate with kindling Interictal spikes were absent or very low before the onset of kindling stimulations. All animals demonstrated an increase in interictal spiking after 2-10 ADs. A SIS rate of > 2/min at > 30 min post-AD was used as the criterion for the onset of interictal spiking. The average number of ADs to evoke interictal spikes (measured at > 30 min post-AD) was 5 + 1 (mean + S.E.M., n = 10). In order to find the optimal time for recording SISs, two rats were recorded during SWS before an AD, and at about 30 min and 60 min after an AD. Each rat was left in the recording cage for most of the day and two ADs were elicited daily, separated by at least 3 h. In both rats, there was a high correlation (correlation coefficient > 0.75) between the SIS rates pre- and post-AD (Table I, Fig. 3). In 7 animals, 6 stimulated once a day (Fig. 4) and the
TABLE 1 Correlation coefficients between SIS rates
preAD, SIS rate before an AD; 30m, SIS rate at 30 min after an AD; 60m, SIS rate at 60 min after an AD. All rates measured during slow-wave sleep or waking immobility, number of observations in brackets, all correlation coefficients are significantly different from zero (P < 0.001)
R210 R2! ]
PreA D vs 30m
PreA D vs 60m
30m vs 60m
0.78 (13) 0.75 (23)
0.84 (23) 0.83 (27)
0.83 (11) 0.95 (22)
150
--t3- preK A
30m postK 60m postK
/
c 100
E (1) 4-,
/
2 O3
,/
!
50-
-# 1.4
~ o O;
i
0
A
1 0
50
100
30
#
200
150
60
90
afterdischarges
minutes poseD 150 6) u3 .E
100-
*o
E O ~.o
o 50-
a
0)
2o co
B O
a
5
10
15
20
~
o
25
hours postAD
Fig. 2. A: changes in logarithmic SIS rate following an AD in a representative experiment. SIS rate decayed quickly in the first h, as indicated by the experimental points (squares) and best-fitted line (crosses) for a single exponential decay for rates during the first h (decay time constant = 30.1 min; correlation coefficient = 0.96). B: changes in the SIS rate following an A D averaged across 6 different experiments (rats). The maximal SIS rate at 1-3 h post-AD was normalized to 100%, and the rates were then averaged. Error bars are one standard error of the mean. Note a significant increase in SIS rate after the AD; SIS rate before the A D is shown at time zero. SIS rate showed a small increase from 1 to 3 h post-AD and then remained relatively stable for several h.
o,-
lO
20
3o
40
SIS rate preK ( / r a i n )
Fig. 3. A: relation of the rates of interictal spikes immediately before (preK), 20--30 rain after (30 m postK) and 50-70 rain after (60 m postK) a stimulation train (K = kindling brain) that produced an AD each time. The last (85th) AD corresponded to a bilateral clonic seizure. Note the general positive correlations between spike rates pre- and post-kindling; spiking rates at each time interval were not recorded for every AD. B: Scatter plot of the relation between SIS rates before and at 50-70 min after a kindling train, for different days of recording. The straight line is the best linear fit through the points y = 3.31 + 1.23x, all units in min 1. Correlation coefficient is 0.83.
311
81
R99
1501
0 ................
• " ~ 30
o.
16
.
.
.
.
•
30]
0 .......................................... 0 44
R103
.
R10~
R102
..........................................................
R104
.
o I. . . . . . . . . . . . . . . . .
.
51
R196
o
58
..............................
61
38
session Fig. 4. The change in hippocampal SISs recorded in SWS during the course of kindling for 6 rats. Number of A D s is shown on the horizontal axis for each rat. The number of ADs to achieve the first generalized seizure with bilateral clonus is given. A l l generalized seizures are circled and after the onset of the first generalized seizure, non-seizure days are marked by a triangle.
seventh twice a day (Fig. 3A), the SIS rate after an AD was measured during the course of kindling, until the occurrence of at least one generalized seizure with bilateral clonus, accompanied by rearing (stage 4 of Racine17). SIS rate was measured during SWS at approximately 45-120 min after the kindling train. For practical reasons, SISs during SWS were recorded after each AD, since SWS occurred more readily after than before an AD, possibly because of the central effects of the AD and increased habituation to the recording cage. The first behavioral convulsion (bilateral clonus) required an average of 44.4 + 8.7 ADs (n = 7; mean + S.E.M.). The change in hippocampal SIS frequency during the course of kindling was rather variable (Fig. 3A, Fig. 4). In 4 of 7 rats, the peak SIS rate was clearly in the middle-late period during kindling, before the development of evoked generalized seizures. In others (R101, R103, R196), seizures were related to some of the highest spike rates. However, it may be concluded that the SIS rate did not predict seizure onset. After seizure onset, occasional days without generalized seizures (triangles in Fig. 4) were not related to any changes in spike rate. In summary, whether and when convulsions occurred or not in the course of kindling could not be predicted by the SIS rate. The relative independence of seizure susceptibility and SIS rate was more emphatically demonstrated by the following experiment. In 9 experiments on 8 rats, ~iaily kindling was interrupted after 5-8 generalized behavioral seizures were evoked. Three of the animals (295, 296, 297) were stimulated up to 5 times a day initially until one generalized seizure and then put on a daily stimulation schedule. SIS rate during SWS was measured for several sessions before and at various days following kindling interruption, for a period of 4-8 days. SIS rate in all rats
declined when recorded at > 1 day following the last stimulation train (Figs. 1 and 5). Resumption of daily kindling increased the spike rate in all rats (Fig. 5). Without exceptions, the first restimulation after 4-8 days of interruption evoked a generalized behavioral seizure. The primary AD duration on the first restimulation measured 57.2 + 5.5 s (mean _+ S.E.M., n = 9), not statistically different from the AD duration of 52.3 + 8 s (n = 9) before the interruption of kindling. Thus, a decline in SIS rate was not accompanied by a decrease in seizure susceptibility or AD duration.
250
.~
200
150
z~
.N ~ 0 E
100
50
_ 1
3
5
7
8
x T kl
k3
days postK stop Fig. 5. Normalized SIS rate in 9 experiments on 8 fully kindled (seizing) rats following the interruption of daily kindling, each experiment represented by a different symbol. SIS rate at 1-2 h post-kindling was used for normalization. Not all days were recorded for all the rats and the period of kindling interruption ranged from 4 to 8 days. Resumption of daily kindling resulted in an increase of SIS rate, in some cases, above 100%. All rats had a generalized seizure on the first day of rekindling.
312
-
no~-~eizure
D
o _2 i i 2
6
days postK stop Fig. 6. Semi-logarithmic plot of the average normalized SIS rate for the 9 experiments on fully kindled (seizing) rats shown in Fig. 5 (open square) and for 6 non-seizing rats (crosses). The straight line is the best linear regression fit for the first 6 days for the seizing rats with slope -0.28 __+0.02. The best regression line for the non-seizing rats (not shown) has a slope of -0.29 + 0.07, not significantly different from that of seizing rats.
While the SIS rate declined to almost zero during the period of kindling interruption, the hippocampus apparently retained a higher level of excitability as demonstrated by the ease in rekindling a high SIS rate. In all 8 rats, 1-3 rekindling sessions resulted in high SIS rates ( > 2/min). The mean number of ADs to rekindle SISs was 1.7 + 0.3 (n = 9), which was significantly smaller than the mean number of 5 + 1 (n = 10) ADs necessary for non-stimulated rats (P < 0.01, Mann-Whitney U-test). Thus, the hippocampus seemed to have retained a 'latent' excitability even though the rate of SISs following kindling interruption was near zero. The decline of SIS rate over days after kindling cessation varied somewhat across different experiments. However, the average decline was well represented by a single exponential decay curve. In Fig. 6, a semilogarithmic plot of the average SIS rate for each day (pooled across experiments) after the last AD was approximately linear, with a decay time constant of 1.57 days. In a different group of 6 rats that did not have behavioral signs of seizures after 4-19 ADs, the average SIS rate also declined with days of interruption of stimulations (Fig. 6). A decay time constant of 1.5 days was found (Fig. 6), which was not significantly different from that of the kindled rats with convulsions. DISCUSSION
Relation of SIS with kindling and seizures The SlSs reported in this study were clearly E E G transients that were not found in the normal unkindled animals. The duration of electrode implants for rats which underwent the complete kindling procedure was 1-4 months, all with SIS onset during the first month. In other studies, we had recorded and analyzed E E G from
non-kindled animals implanted for several months, and had not observed any SISs. In addition, the decay of SISs with time (Figs. l and 5) clearly indicated that SISs were not caused by tissue disruption by the implanted electrodes. The frequency of SISs recorded at a fixed period alter an A D appeared to peak during the middle-late course of kindling, and not usually during periods of convulsions. In fact, a decrease in SIS rate was sometimes observed after the stage of generalized convulsions was reached. Convulsions may be related to either a high or a low SIS rate, or in other words, the hippocampal SIS rate does not predict behavioral seizure. The peaking of SIS rate before convulsion onset was noted for amygdaloid SIS in the cat x2"22. Hippocampal SISs are apparently controlled by factors that are different from those governing behavioral seizures. Following interruption of daily stimulations, the decay time constant of SISs was approximately 1.5 days, and very few SISs were detected after 5 days. In the mean time, seizure susceptibility, as determined by the ability of the stimulus to evoke a generalized convulsion, remains undiminished. Thus, a change in SIS rate was not accompanied necessarily by a change in seizure susceptibility. Frequently repeated train stimulations may increase SIS rate but decrease A D or seizure susceptibility 4. Kairiss et al. 1o found few hippocampal SISs after kindling, which is at odds with the present report and with other literature 5'12'25. While Kairiss et al. recorded 24 h or more after the last AD, significant spiking was still observed in the hippocampus at this time (Figs. 1C, 2B, 5). The rate of amygdaloid SISs in amygdala-kindled animals was found to decline with a half-life of 1-2 days 5 or to disappear within 8 days 7. Waiters, in an unpublished M.A. thesis cited by Wada et al. 23, found that amygdaloid SISs were seldom detected more than 6 days after the last kindling stimulation. Waiters also concluded that seizure susceptibility was unrelated to the SIS activity. However, Wada et al. 23 reported a marked increase in SISs before spontaneous seizures in cats, though Gotman found no consistent changes in SIS rate before spontaneous seizures in cats 7 and humans s, While this paper offers no data on the issue of whether interictal spikes precede a spontaneous seizure, evidence has been presented that SISs followed an AD or seizure in the hippocampus. The literature indicate that local SISs followed an amygdaloid A D in animals 5'7 and in human spontaneous seizures s. Posterior cingulate kindling, however, only gave few SISs at any stage of kindling aS. We demonstrated a retention of local excitability without actual interictal spiking in the kindled rats. This
313 'latent' excitability is inferred from the ease of rekindling a high SIS rate; its permanence has not been studied. Other researchers have found an increased capability of in vitro slices from kindled animals to demonstrate interictal spiking when exposed to a high extracellular [K ÷ ]011. Increased Ca 2 ÷ uptake in CA1 dendrites 24 and increased N M D A receptor involvement in normal synaptic transmission in the dentate gyrus 16 were also found in slices from kindled rats. Relation of SISs to A D and physiology Many factors could have contributed to the variability of SIS rates immediately following an AD. Ongoing experiments indicate that the early component of the evoked potential in CA1 region following Schaffer's collateral stimulation was depressed for up to 3 h post-AD. This long-lasting depression may contribute to the partial suppression of SISs within the first 3 h after kindling (Fig. 2), resulting in the 'peak' in SIS rates at 2-7 h post-AD in some experiments. Gotman 7 observed peaks in amygdaloid SISs at 7-15 h post-AD. Hippocampal SISs were time-locked to the AD. There was some variability in the decay of the SISs with time, but approximate time constants of 70 min and 1.5 days were observed. The relatively long time constants of hippocampal SISs suggest a mechanism of long-term plasticity, possibly long-term potentiation (LTP). Racine et al.lS reported hippocampal LTP to decay with two time constants of 1.5 h and 5 days respectively. The similarities in the order of magnitudes between time-constants of SISs and enhanced evoked potentials (LTP) suggest a common mechanism underlying the two events. On the other hand, there are difficulties with the hypothesis that hippocampal SISs are caused by an enhancement of synaptic response (LTP) at a single synapse. First, the particular synapse has not been identified and preliminary data indicated that it was not
REFERENCES 1 Ayala, G.E, Dichter, M., Gumnit, R.J., Matsumoto, H. and Spencer, W.A., Genesis of epileptic interictal spikes, new knowledge of cortical feedback systems suggests a neurophysiological explanation of brief paroxysms, Brain Research, 52 (1973) 1-17. 2 Buzsaki, G., Leung, L.S. and Vanderwolf, C.H., Cellular bases of hippocampal EEG in the behaving rat, Brain Res. Rev., 6 (1983) 139-171. 3 Dichter, M. and Spencer, W.A., Penicillin-induced interictal discharges from the cat hippocampus. II. Mechanisms underlying origin and restriction, J. Neurophysiol., 32 (1969) 663-687. 4 Engel, J. Jr. and Ackerman, R.E, Interictal EEG spikes correlate with decreased, rather than increased epileptogenicity in amygdaloid kindled rats, Brain Research, 190 (1980) 543-548. 5 Fitz, J.G. and McNamara, J.O., Spontaneous interictal spiking in the awake kindled rat, Electroenceph. Clin. Neurophysiol., 47 (1979) 592-596.
the CA3-CA1 synapse, since LTP at the latter synapse should be blocked by N-methyl-o-aspartate antagonist 9 while SIS rate increase following an A D was not. Second, LTP could be demonstrated after a single tetanus or A D while SISs were found only after several ADs. Obviously, spontaneous interictal spiking is a different phenomenon than evoked responses commonly used to indicate the presence of LTP. The tonic and phasic inputs to the hippocampus (e.g. conditions in SWS) and multiple synapses or circuits may play a role in the generation of the SIS. Evidence so far points to plasticity within but not necessarily restricted to the hippocampal CA3 region. Most CA3 neurons possess an intrinsic capability to firing in burst 19'28. Synchronized bursts in CA3 probably generate the spontaneous sharp waves (a population EPSP) in CA1 region of normal unkindled rats 2. In vitro, when perfused with penicillin or bicuculline, CA3 drives interictal spiking recorded in the CA1 region 27. The bilaterality and polarity of CA1 interictal spikes 25 can be explained by the distribution of CA3 afferents bilaterally and in both the apical and basal dendritic layers of CA121. This and previous papers on hippocampal SISs 14'25 suggest that SISs are closely related to the normal physiology of the hippocampus. Paroxysmal interictal spiking may disrupt normal hippocampal functions and contribute to memory and cognitive deficits in epileptic patients. On account of the wealth of anatomical and physiological knowledge of the rat's hippocampus, hippocampal interictal spikes in vivo may offer a model for the study of the relation of interictal spikes to function.
Acknowledgements. This study was supported by NIH Grant NS25383 and NSERC Grant A1037. I thank Wendy Charlton, Kathy Boon and Elizabeth Szczutkowski for their patience and technical assistance, Bey Hughes for typing, and Drs. P. Cain and R. McLachlan for useful comments on the manuscript.
6 Gloor, P., Contributions of electroencephalography and electrocorticography to the neurosurgical treatment of epilepsies. In D.P. Purpura, J.K. Penry and R.D. Walter (Eds.), Neurosurgical Managementof the Epilepsies, Raven, New York, 1975, pp. 59-105. 7 Gotman, J., Relationships between triggered seizures, spontaneous seizures, and interictal spiking in the kindling model of epilepsy, Exp. Neurol., 84 (1984) 259-273. 8 Gotman, J. and Marciani, M.G., Electroencephalographic spike activity, drug levels and seizure occurrence in epileptic patients, Ann. Neurol., 17 (1985) 597-603. 9 Harris, E.W., Ganong, A.H. and Cotman, C.W., Long-term potentiation in the hippocampus involves activation of Nmethyl-D-aspartate receptors, Brain Research, 323 (1984) 132137. 10 Kairiss, E.W., Racine, R.J. and Smith, G.K., The development of the interictal spike during kindling in the rat, Brain Research, 322 (1984) 101-110. 11 King, G.L., Dingledine, R., Giacchino, J.L. and McNamara,
314
12
13 14 15 16 17 18 19
20
J.O., Abnormal neuronal excitability in hippocampal slices from kindled rats, J. Neurophysiol., 54 (1985) 1295-1304. Lange, H., Tanaka, T. and Naquet, R., Temporal-spatial pattern of subcortical spike activity in kindling epilepsy. A statistical approach, Electroenceph. Clin. Neurophysiol., 42 (1976) 564574. Leung, L.S., Hippocampal electrical activity following repetitive local stimulation. I. Afterdischarges, Brain Research, 419 (1987) 173-187. Leung, L.S., Hippocampal interictal spikes induced by kindling: relations to behavior and EEG, Behav. Brain. Res., 31 (1988) 75-84. Leung, L.S. and Boon, K.A., Kindling in the posterior cingulate cortex: electrographic and behavioral characteristics, Electroencephalogr. Clin. Neurophysiol., in press. Mody, I. and Heinemann, U., NMDA receptors of dentate gyrus granule cells participate in synaptic transmission following kindling, Nature (Lond.), 326 (1987) 701-704. Racine, R., Modification of seizure activity by electrical stimulation: II. Motor seizure, Electroenceph. Clin. Neurophysiol., 32 (1972) 281-294. Racine, R.J., Milgram, N.W. and Hafner, S., Long-term potentiation phenomena in the rat limbic forebrain, Brain Research, 260 (1983) 217-231. Ranck, J.B., Behavioral correlates and firing repertoires of neurons in the dorsal hippocampal formation and septum of unrestrained rats. In: Isaacson, R.L. and Pribram, K.H. (Eds.), The Hippocampus, Vol. 2, Plenum, New York, 1975, pp. 207-246. Suzuki, S.S. and Smith, G.K., Spontaneous EEG spikes in the
21 22
23 24 25
26
27 28
normal hippocampus. I. Behavioral correlates, laminar profiles and bilateral synchrony, Electroenceph. Clin. Neurophysiol.. 67 (1987) 348-359. Swanson, L.W., Wyss, J.M. and Cowan, W.M., An autoradiographic study of the organization of intrahippocampal association pathways in the rat, J. Comp. Neurol., 181 (1978) 681-716. Wada, J.A. and Sato, M., Generalized convulsive seizure induced by daily electrical stimulation of the amygdala in cats: correlative electrographic and behavioral features. Neurology~ 24 (1974) 565-574. Wada, J.A., Sato, M. and Corcoran, M.E., Persistent seizure susceptibility and recurrent spontaneous seizures in kindled cats, Epilepsia, 15 (1974) 465-478. Wadman, W.J., Heinemann, U., Konnerth, A. and Neuhaus, S. Hippocampal slices of kindled rats reveal calcium involvement in epileptogenesis, Exp. Brain Res., 57 (1985) 404-407. Wadman, W.J., Lopes da Silva, EH. and Leung, L.S., Two types of spontaneous interictal transients during kindling of the rat hippocampus, Electroenceph. Clin. Neurophysiol., 55 (1983) 314-319. Wieser, H.G., Bancaud, J., Talairach, J., Bonis, A. and Szikla, G., Comparative value of spontaneous and chemically and electrically induced seizures in establishing the lateralization of temporal lobe seizures, Epilepsia, 20 (1979) 47-59. Wong, R.K.S. and Traub, R.D., Synchronized burst discharge in disinhibited hippocampal slice. I. Initiation in CA2-CA3 region, J. Neurophysiol., 49 (1983) 442-458. Wong, R.K.S., Prince, D.A. and Basbaum, A.I. Intradendritic recordings from hippocampat neurons, Proc. Natl. Acad. Sci. U.S.A,, 76 (1979) 986-990.