- Email: [email protected]
Chronic epilepsy with damage restricted to the hippocampus: possible mechanisms
EHLEPSY P ELSEV]IER
$EARCH
Epilepsy Research 26 ,~,,~,~r.~, ~,~, _~.~qq-_.06~
Chromc epilepsy wkh damage restricted to the Nppocampus: possible mechanisms Claude Go Wasterlain
Yukiy sN Shkasaka a,b,c Andrey M. Mazarafi Igor " ' man c,d plge
a Epilepsy Research Laboratory, VA Meaica! Ce~lter (127), Sepuh'eda. CA 91343-2099, USA b Department o f Neurology, UCLA School of Medicine. Los Angeles. CA 90024. USA Brain Research b,s6mte, UCLA Sch )ol of Medicine. Los Angeles, CA 90024, USA a UCLA School of Demis~', Los Angeles, CA 90024, USA
Received 29 Novernbe:."1995; accepted 6 May 1996
Abstract
We studied the time course and possible mechanisms of the development of chronic epilepsy following unilateral stimulation of the pefforant path. After 24 h of pefforant path stimulation by a modified Sloviter method, lesions were restricted to the hippocampus, except for 2 of 24 rats with minimal eatorhinal neuronal injury in layer 3. Lesions were exclusively ipsilateral in the ?olymorph layer of the hilus and m CA4-CA 3c, predominantly ipsilateral in CA 3, in CA~ and in the granule cell layer. Feedforward and feedback inhibition were studied by paired pulse stimulation. In the week following inhibition, there was complete loss of GABAA-mediated, sbon interstimulus interval (lSI)-dependenl inhibition and frequency-dependent inhibition, and also of GABAB-mediated long ISI-dependent inhibition. Yet no spontaneous seizures were observed at that time. In the next four weeks, we saw no evidence of increasing excitatory drive such as would I~e expected from recurrent mossy fiber sprouting. On the contrary, there was progressive return of inhibition. By four weeks post-lesion, the majority of animals had developed spontaneous recurrent seizures, and/or seizures on 2 Hz stimulation (never seen in controls), in spite of complete or near-cornplete recovery of short ISI-dependent, GABAn-mediated inhibition. A small but significant loss of frequency-dependent inhibition persisted, but individual animals with complete recovery of frequency-dependent inhibition showed spontaneous seizures, suggesting that loss of GABAA-mediated inhibition was not the direct cause of chronic epilepsy. GABA B-mediated, long ISI-dependent inhibition continued to show a significant loss. The ratio of the population spike amplitude at 250 /zA to the maximal population spike amplitude, a measure of granule cell excitability, was unchanged immediately after stimulation, but increased in the next few weeks in a manner identical to that seen in kindling, suggesting the possibility that during the transient loss of inhibition, spontaneous kindling had occurred. Intracellular recordings from grar, ule cells in hippocampal slices prepared from these animals showed a significant loss of GABA B-mediated slow inhibitory postsynaptic potentials (IPSPs). These data show that the sequellae of unilateral status epilepticus with damage restricted to the hippocampus are sufficient to cause chronic recurrent seizures. There is a possibility that chronic epilepsy is not the direct result of the loss of inhibitory drive or of a sprouting-induced
* Corresponding author. 0920-1211/96/$15.00 Copyright © 1996 Published by Elsevier Science B.V. All rights reserved. PII S0920-1211 (96)00058-7
256
C.G. Wasterlain e~ al. / Epilepsy Research 26 (1996) 255-265
increase in excitatory drive, but represents plastic changes akin to spontaneous kindling, possibly facilitated by loss of GABA a-mediated inhibition. Keywords: Epilepsy; Hippocampus; Status epilepticus; Kindling; GABA; Inhibition
1. Introduction Many recent studies have shown that seizure-induced brain damage can in turn lead to chronic epilepsy [3,6,8,10,13,14,21,31]. Both the kainic acid and the pilocarpine model of status epilepticus lead to extensive brain damage and to the delayed occurrence of both amygdaloid and generalized convulsions [7,8,31]. These rats also show interictal spikes and electrographic seizures [3,13]. While most of these studies have concentrated on the hippocampus as the most easily studied epileptogenic area, their lesions are disseminated throughout the brain on both sides, and the source of the seizure activity in those animals is unknown. Stimulation of the perforant path (pfps) induces neuronal injury which has an excitotoxic appearance similar to that induced by kainic acid [18,23,27]. Lesions are far more restricted, and are associated with a measurable loss of frequency-dependent paired pulse (PP) inhibition m the dentate gyrus [25]. Therefore, this is a favorable model to unravel the physiological changes due tc seizure-induced cell loss and to understand their role in epileptogenicity. Spontaneous seizures have usually been explained either by the build-up of recurrent excitatory connections as a result of sprouting [1,8,9,11,12,15,16,26,28-30], or by the loss [17] or deafferentation [2,25] of GABAergic inhibition in the hippocampus. In this study, we examined whether the time course of development of spontaneous seizures following perforant path stimulation is compatible with these hypotheses. Our results suggest another potential explanation, witL, kindling-like changes in granule cells excitability and a long-lasting loss of GABAa-mediated IPSPs as a possible mechanism.
2. Material and methods Our method of stimulation is slightly modified from Sloviter [25]. Adult Wistar rats are stimulated
for 24 h with intermitted trains of single stimuli at 20 Hz (20 V, 0.1 ms duration and continuous 2 Hz) delivered once per minute to an electrode located in the angular bundle, under continuous intravenous urethane anesthesia. Rats which lost their population spikes during stimulation were eliminated from the study, and paired pulse stimulation was carried out before, during and at various intervals after stimulation. At key times, we varied intrastimulus interval systematically from 15 to 5,000 ms (0.1 Hz, 0.1 ms duration, 30 V monophasic stimulation). Input-output responses were studied at intensities varying from 250 to 50,000 /.,A and kindling was tested through the pefforant path 4 weeks after perforant path stimulation or sham stimulation; these methods have been described in detail previously [22]. Acute cell injury was measured 3 days after stimulation in hematoxylin-eosin stained serial sections. Fig. l C-D ~hows that eosin in the neuronal cytoplasm is easily visualized by its bright yellow fluorescence under UV light, making the identification of 'ischemic cell change' [4,5] easier. Fig. 1(1-4) provides examples of the qualitative damage scale used (see legend). Three days after stimulation, animals received a lethal injection of pentobarbital, and were perfused through the heart with a 4% solution of paraformaldehyde. The brains were embedded in paraffin and 8 /~m sections were stained with hematoxylin and eosin. Injured neurons were defined as those showing a pyknotic nucleus (viewed under regular light), and an eosinophilic cytoplasm as viewed under UV through a fluorescein filter. Estimation of damage were performed on sections located at - 2 . 3 , -2.8, -3.3, - 3 . 8 and - 4 . 3 mm posterior to bregma according to the Paxinos and Watson atlas [19] in a Minded manner for both the stimulated and nonstimulated sites. All injured neurons within the hilar region bounded by the inferior and superior blades of the dentate gyrus and by a line drawn between a lateral tips of each blade were counted. The values for each section for each animal were then summed and divided by the number of sections to obtain the
C.G. Waslertai~ el al. / t~£pitepsy Research 26 ~f996) 255-265
damage score for that animal. Since plotting of the residuals showed a normal distribution, a paired t-test was used to compare the stimulated site to the contralateral site. Chronic changes were measured at least 1 month after stimulation in sections stained with cresyl violet or with antibodies tc specific subpopulations of neurons [20]. Seven ~mplanted controls were used at each time point.
257
the hippocampus no lesions were seen in untreated co3~trc!s~ implanted controls or i~ sham stimulated controls, but all stimulated anima!s showed some neuronal loss (Fig. 1). Neuronal ~oss was exclusively ipsilaterat in hilus and in CA:~c, and in a few animals lesions were restricted to the hilus. However, the majority of rats had bilateral lesions (more severe on the side of stimulation) in CA 3, CA~, and in the dentate granule cell layer (Table 1).
2.1. Methodology of intracellular recording Whole brain was isolated from control or stimulated rats, and coronal sectio ~s were obtained from the ipsilateral hemisphere at 0°C using a vibrating blade slicer (Campden Instruments). Slices were maintained at room temperature in oxygenated (95% O~-5% CO 2) artificial cerebrospinal fluid (ASCF, in raM: NaCl, 125; KCI, 2.5; NaH2PO 4, 1.25; CaCI 2, 2; MgCl z, 2; NaHCO 3, 26 and glucose, 10; pH 7.4), and stabilized for 1 h before recording. All experiment~ were performed on slices submerged in the recording chamber at 34°C with a superfusion rate of approximately 2 ml/min. Intracellular recordings were made with 3 M potassium acetate-filled ~¢ass microelectrodes (pH = 7.25, tip resistance 120-170 Mf~) placed in the dentate granule cell layer of the hippocampus. A concentric bipolar electrode was positioned approximately 0.5 mm from tbe recording electrode in the outer molecular layer to stimulate the perforant path inputs and to minimize the variation of responses between slices. Stimuli (5 IxA-10 mA) were square wave pulses (200 ~s) delivered at 0.083 Hz. Intracellular potentials were amplified using an Axoclamp-2A amplifier. Data were digitized via a 12-bit analog-to-digital interface (TL-1) and analyzed using the pCLAMP software package (Axon Instruments).
3.2. After pfps, ihere was serere loss of iahib#iorl but no seizures Frequency-dependent and short ISI-dependent paired pulse (PP) inhibition, which appear to be GABAA-mediated [24], and long [SI-dependent PP inhibition which is probably GABA B-mediated, were completely lost during the week following pfps (Figs. 2 and 3). However, the ratio of the population spike a~,~plitude at 250 /xA to the maximal population spike amplitude, which is a measure of granule cell excitability, was not increased (Fig. 3). No spontaneous seizures were observed until the end of the second week after stimulation (n = 12) (Fig. 6). Since EEG was not monitored we cannot rule out the possibility that electrographic seizures appeared before behavioral seizures, but the latter were clearly delayed and only appeared several weeks after the lesion in a majority of animals. If the mechanism of chronic epilepsy is ti~e direct result of loss of inhibition, either because of deafferentation of basket cells [2,25] or because specific subpopulation of GABAergic cells are lost [17] it is very surprising that no spontaneous seizures occur at the time when loss o f inhibition is the greatest. 3..? The &bT,:'arance of spontaneous seizures coinci~'cd with the progressive but partial recoL,eiy of int,~ibition
3. Results and discussion 3.1. Neuronal loss was restricted to the hippocampus Of the 24 animals used for acute anatomical studies, 22 showed no lesions outside of the hippocampus. The remaining 2 animals showed a few necrotic neurons in layer 3 of entorhinal cortex. In
Over the four weeks following stimulation, short • ~I-dependent ir~:--;bition, which is thought to be GAB/~A mediated [24], appeared to recover comt,,!ete!y (Figs. 2 and 3), although one could argue that the. p,Ayspikes produced by the initial stimulus provide a stronger drive than the single spike seen in controls, and that therefore recovery is substantial
258
C.G. Wasterlain et al. / Epilepsy Research 26 (1996) 255-265
but not complete. Frequency-dependent PP inhibition, which is also thought to be G A B A a - m e d i a t e d , recovered only partially, but in some individual rats complete recovery was observed (with the same
proviso as above} while they were developing chronic recurrent spontaneous seizures (Fig. 4). Lorlg IS|-dependent inhibition and first stimulus pf~tyspike also s h o w e d a trend toward recovery, but this was far
Fig. 1. T bright fluorescence, while tlae contralateral hippocampus (B) is intact (original magnification × 10). Injured hilar neurons show a pink cytoplasm under regular light (C) and stand out under fluorescence (D) (original magnification X 40). Pictures e-h show various grades of hilar damage (e = 1; f = 2; g = 3; h = 4) (original magnification × 20).
C.G. W~¢slerlai~ e1 ~.1./ L)"#h.psy Research 26 ¢[99b~ 255-265
Table B Anatomy of neuronal injury fifllowing unilatera~ per~,~>rum path stimuiation Region
Stimulated
Col:trtHa~.eral
Dentate hilus CA 3c
2.2 +_0.5 " 2 _+0.6 ~ 2 +O.4" O.8 + 0.3 R.3 +_0.4 " 0 < O.1 0
0 O.I 4- 0.4 0.5_+_0.3 O.l + 0.2 0.4 + (1.2 0 < o, I 0
CA 3
CA) Granule cells Neocortex Entorhinal cortex Pyriform cortex
259
toting was c a m e d oat to detect seizures, both kindled-iike and generalized c o n v u l s i v e aeizures were observed in a majority o f animals (Fig. 5). Seizures star~ed at the end o f ~he second week in one animal and at the e n d o f the fourth w e e k or later in the remaining rats (Fig. 5).
3.4. Delayed seiz:ures were accompanied by a deh~ved increase in gramde celt excimbi6n,
Damage was measured 72 h after unilateral pq~s in hematoxy]meosin stained sectk)e,s at the level of habenulm nuclei (dorsal hippocampus). A qualitative scale for damage was used (I = < I(1% ir~iured neurons: 2 = 10-25%; 3 = 25-50%: 4 = > 50~). SE are provided only as a description of the population and were not used in assessing significance, p < 0.05 compared to contralateral, using a t-test for paired data. from c o m p l e t e and those m e a s u r e s r e m a i n e d quite abnormal at 4 w e e k s [22]. B e h a v i o r was o b s e r v e d only o n c e a w e e k , and while no electrographie m o n i -
in i n p u t - o u t p u t response studies, granule cell excitability was not f o u n d to increase immediately after pfps, H o w e v e r , it had increased significantly two weeks after pfps, in a way similar to the response to kindling stimulation in control rats !22]. "Ihis delayed increase in dentate grannie cell excitability might reflect a kindling-like p h e n o m e n o n precipitated by the loss o f h i p p o c a m p a l inhibition [32], or might be related to m o s s y fibers sprouting, which usually appears seve~'al w e e k s after neuronal d a m a g e [28].
Fig. 1 (continued).
260
C.G. Waswrlai, et al. / Epilepsy Reseatz'h 26 (I 996) 255-265
Tst
25
40
100
300 ,000
b
Ist
25
40
I00
300
I000
F
Fig. 2. Interstimulus interval-dependent paired pulse examination in two stimulated rats. The results of interstimulus interval-dependent paired pulse examinations in two 24 h perforant path stimulated rats (a and b) are shown. The waveforms show the average of 10 consecutive waves. The first positive deflection in each trace is the stimulation artefact, and the EPSP with/without PS discharge follows. Paired pulse stimulations (0.1 ms duration, 30 V monophasic) were given at 0.1 Hz. Abbreviation: A, before pfps in the awake state; B, hefxe pfps under urethane anesthesia; C, 30 min after pfps; D, 2 weeks after pfps in the awake state; E, 4 weeks after pfps in the awake state; F, 4 weeks after pfps under urethane anesthesia; 1st, 1st PS wave form; 25, 40, 100, 300, and 1000 indicates the interstimulus interval (ms) for that waveform. Vertical scale bar ! mV, horizontal scale bar 10 ms. Reproduced from Ref. [18], with permission,
C.G. Wasterlain el aL / Epilepsy Research 26 ¢]996t 255-265
This increase in granule cells excitability might also be responsible for the convulsive response to 2 Hz stimulation which was observed in haft of the animals three to ~imr weeks after ptps. Beth electrical and chemical kindling require activation of a critical mass o f neurons [321. W e should consider the possibility that as the result of loss o f inhibition, physiological stimuli resulting from routine activity m a y recruit e n o u g h neurons to induce a synchronous afterdischarge which exceeds the kindling threshold. If
120
26]
full kindling occurs, even a compiete recovery of inhibition later might ~,ot be able to abolish estabfished kindling, so that seizures might cormnue. We previously postulated the existence of a "filter' that limits recruitment o f bt:rsdng neurons so that their responses to physiological stimuli never exceed the kindling threshold [32]. The present data are compatible with the view thai hilar interneurons may "filter' input into the dentate gyrus, and that their loss after pfps c o m p r o m i s e s that filter function.
88
PiPS
100
61]
80
=
60
40
o
PPS
46
rnulate
20 O
~
.- z0
control
A
B !
-20
before
30 rain after
2~ after
t 20
t
before
4 wk~ after
!
30 rain after
- - 1o8
190
~
pfp$
I
2 wks a~ter
_ _
4 wt
Kindling
"=
.T~----,-~
o 80 td~
80
,U,
t
N so
~"
40
20 o -20
4o
~
rnulated
control
C
~ ~ ca
20
before
~, 30 rain
i 2 v~s
I 4wks
alter
alter
after
~ ' ~ ' ~ t ~ /
==..4 s~rnul~ ~ conlTol
o
o. - 2 0
,
t
~ 40
~
So
PfPS
D
n
I
I
i
before 30 rain 2 ~ 4 wks after alter alter
I
t --
alter Kindling 1 wk 4wI~
Fig. 3. Sequentiai changes of paired-pulse inhibition and granule ~',~11excitability. (a) Sequential changes of inhibition score [100- (PS amplitude of the 2nd puise)/PS amplitude of the 1st pulse) × 100] i~: short (25 ms) interstimulus interval (ISI)-dependent paired-pulse (PP) examinations, in stiraulated rats and control rats. (b) Those at 300 ms ISI. (c) Sequential changes of inhibition scores in frequency-dependent PP examinations, at 2 Hz in stimulated and control rats. (d) Sequential changes of PS ratio [(ps amplitude at 250 ,~A stimulus intensity)/(maximal PS amplitude in the same experiment) x lo0] in stimulated rats and control rats. All the data shown in this figure were examined in the awake state except for 30 min after 24 h perforant path stimulation (pfps) or 24 h anesthesia, which was examined under urethane anesthesia. Abbreviations: pfps or Kindling with arrows, the time point at which pfps or kindling stimulation was given; sthnulated, rats given pfps; control, non-stimulated control rats; before, before pfps in stimulated rats and before 24 h anesthesia in control rats; time after pfps or 24 h anesthesia, respectively. +, * Significantly different from before pfps (stimulated group) or before kindling (control group) in the s~maestate (awake or anesthetized), at p < 0.05 mid p < 0.01, respectively; *,* * significantly different from control at p < 0.05 and p < 0.01, respectively. Reproduced from Ref. [18], with permission.
262
C.G. Wasterlain et al. / Epilepsy Research 26 (1996) 255-265
3.5. Pj~ lesioned animals showed p~rsistent loss of GABAB-mediated slow IPSPs
Based on our results from evoked potential recordings we decided to explore the possible mechanisms which underlay chronic epileptogenicity with the use of intracellular recording techniques. We studied slices taken from animals 1 month after stimulation, a time at which they exhibited spontaneous seizures and the G A B A B receptor-mediated long ISI inhibition was still reduced. IntraceUular recordings obtained from the dentate granule (DG) cells of these animals showed that they exhibit postsynaptic membrane properties which did not differ from those seen in control (implanted; non-stimulated) rats. Their resting membrane potential, input resistance, spike amplitude, duration and
discharge pattern in response Lo depolarizing current pulses were similar to controls. In our preliminary studies comparing the respouses to synaptic stimulation of DG cells from stimulated or control rats, a trend towards reduced slow IPSPs was observed in stimulated rats. This slow IPSP results from the activation of G A B A B receptors, which in turn activates a K ÷ conductance. We stimulated synaptic inputs while changing the membrane potential of the recorded neuron with intracellular injections of hyperpolarizing or depolarizing current pulses. We used synaptic stimuli which were sufficient to evoke a large EPSP subthreshold to spike generation. As the membrane potential is changed from more hyperpolarized to more depolarized values, the slow IPSP (which peaks at 2 0 0 - 2 5 0 ms after synaptic stimulus) also changes from a
Fraquency-Dependent Paired-Pulse inhibition Nter Focal SE 0.1Hz
1Hz
Before ~ 1 ~ Awake
2Hz
.
2wksAfterI l l l | , Awake 4wksAfteJr
~
l
l
Awake 1mY
k__
lOmsec Fig. 4. Complete recovery of frequency-dependentpaired pulse inhibition in a rat exhibiting recurrent spontaneous seizures. The waveforms are average of l 0 consecutive waves. The narrow positive deflection before EPSP in each trace is a stimulatic~lartefact PP stimulations (0.1 ms duration, 30 V monophasic) were given 40 ms apart. Vertical scale bar = l mV. horizontal scale bar = 10 ms.
263
C.G. Wasterlain etal,/Epilepsy Researck 26 f ]9~6~ 2.55-265
depolariziag to a hyperpolarizing response. By extrapolating peak stow IPSP values from each recorded nemon to - 7 0 mV we were able to demonstrate significant reduction in the amplitude of the slow IPSPs in stimulated rats (3.8 + 0.7 mV, n = 7 cells fi'om 5 rats) compared to controls (7.6 + 1.5 mV, n = 6 cells, 5 rats). The reduction of slow IPSPs in stimulated rats could be due to a decrease in GABA B receptor function a n d / o r a decrease in the release of GABA from the terminals of inhibitory interneurons. While the role of this persistent los~ of inhibition is unknown, GABAa-mediated slow IPSPs are likely to be involved in chronic modulation of hippocampal excitability associated with synaptic plasticity, and represent a potential explanation of both the increase in granule cell excitability (Fig. 3) and the enormous increase in susceptibility to kindling (Fig. 5) which were observed in pefforant path-stimulated rats in this study. Further studies are in progress to test these speculative interpretations.
125
5011
~5
°L .r~.
400
"~
~
~
75
lOO
o
0
Fig. 5. Kindling in pfps rats. Stimulations were given via electrodes in the pefforant path, and responses were recorded via electrodes in the dentate gyrus and pefforant path. An afterdischarge (AD) was defined as high amplitude spike or polyspike epileptiform activity visible for 3 s or more after completion of the stimulus. AD threshold (ADT) was the lowest intensity of current necessary to elicit AD, and AD duration (ADD) as the time lapsed_ from the end of the stimulus until the organized AD terminated at either recording site. Data depicts means (SEM), ~ d (minimum-maximum). Stimulated rats were compared with nonstimulated controls by ANOVA. Abbreviations: First ADT, ADT at first trial; first ADD, ADD at first trial; first C 5, number of trials necessary to induce the first class 5 convulsion; third C S, number of trials necessary to induce three class 5 convulsions; other abbreviations as in Table I.
:C' cv, C f
c~ f
f_
x._ (D ..D
E
/v
Z PFPS
~ wk
/ 2 was
4 wks
>4
wks
Fig. 6. Time of onset of generalized convulsive seizures in pefforant path-stimulated rats (n = 10) as a function of time elapsed since stimulation (weeks). Orice seizures started they recurred in all rats in this group. ( v ) Spontaneous seizures, ( [ ] ) seizures induced by 2 Hz stimulation (which did not induce seizures in controls).
In summary, the perforant path stimulation model clearly shows that hippocampal damage resulting from unilateral status epilepticus is sufficient to cause chronic epilepsy, and therefore Sommer and Pfleger were both correct when they stated, in 188], that hippocampal lesions were the cause of epilepsy and the result of seizures, respectively [33]. This study gave no indication of an increase in excitatory drive as postulated by the recurrent mossy fiber sprouting hypothesis [l,8,15,28-30]. The relationship between loss of inhibition (be it due to death of GABAergie cells [17], to their active suppression or to their deafferentation [24]) and the occurrence of spontaneous seizures appears more complex than was previously realized. Since spontaneous behavioral seizures were not seen at the peak of inhibitory loss, an additional process must occur between the loss of inhibition and the eventual epileptogenicity. The current study ;.~ises the possibility that this might be a kindling-like phenomenon, and that the persistent loss of GABAa-mediated inhibition observed in chronic animals might play a role in that development.
Acknowledgements Supported by Research Service of the VHA and by grant NS 13515 from NWDS.
264
C.G. Wasterlain et at,/Epilepsy Research 26 (1996) 255-265
References [I] Babb, T.L., Kupfer, W.R., Pretorius, J.K., Crandall, P.H. and Levesque, M.F., Synaptic reorganization by moss? fibers in human epileptic fascia dentata, Neuroscience, 42 (1989) 351-363. [2] Bekenstein, J.W. and Lothman, E.W., D~..-mancy of inhibitory intemeurons in a model of temporal lobe epilepsy, Science, 259 (1993) 97-100. [3] Bertram, E.H. and Comett, J., The ontogeny of seizures in a rat model of limbic epilepsy: Evidence for a kindling process in the development of chronic spontaneous seizures, Brain Res., 625 (1993) 295-300. [4] Brown, A.W. and Brierley, J.B., Anoxic-ischaemic cell change in rat brain light microscopic and fine-structural observations, J. NeuroL Sci., 16(!) (1972) 59-84. [5] Brown, A.W. and Brierley, J.B., The earliest alterations in rat neurones and astrocytes after anoxia-ischaemia, Acta Neuropathol. (Berlin), 23(1) (1973) 9-22. [6] Ca,~alheiro, E.A., Riche, D.A. and Le t'al La Salle, G., Long-term effects of intrahippocampal kainic acid injection in rats: A method for inducing spontaneous recurrent seizures, Electroencephalogr. Clin. NeurophysioL, 53 (1982) 581-589. [7] Cavalheiro, E.A., Leite, J.P., Bortolotto, A.Z., Turski, W.A., Ikoaomidou, C. and Turski, L., Long-term effects of pilocarpine in rats: Structural damage of the brain triggers kindling and spontaneous recurrent seizures, Epilepsia, 32 (1991) 778-782. [8] Cronin, J. and Dudek, F.E., Chronic seizures and collateral sprouting of dentate mossy fibers after kainic acid treatment in rats, Brain Res., 474 (1988) 181-184. [9] Cronin, J., Obenhaus, A., Houser, C.R. and Dudek0 F.E., Electrophysiology of dentate granule cells after kainate-induced synaptic reorganization of the mossy fibers, Brain Res., 573 (1992) 305-310. [10] Feldblum, S. and Ackermann, R,F., Increased susceptibility to hippocampal and amygdala kindling following intrahippocampal kainic acid, Exp. Neurol,, 97 (1987) 255-269. [11] Franck, J.E. and Schwartzkroin, P.A., Do kainate-lesioned hippocampi become epileptogenic?, Brain Res., 329 (1985) 309-313. [12] Heinemann, U., Chssmann, H., Dreier, J. and Stabel, J., Changes ia synaptk: transmission in the kindled hippocampus, Adv. Exp. Med. 13iol., 268 (1990) 445-450. [13] Lothman, E.W., Ber~ am, E.H., Kapur, J. and Stringer, J.L., Recurrent spontaneous hippocampal seizures in the rat as a chronic sequelae to limbic status epilepticus, Epilepsy Res., 6 (1990) 110-118. [14] Lothman, E.W., Bertram, III, E.H. and Stringer, J.L., Functional anatomy of hippocampal seizures, Prog. NeurobioL, 37 (1991) 1-82. [15] Masukawa, L.M., Uruno, K., Sperling, M., O'Connor, M.J. and Burdett, L.J., The functional relationship between antidromically evoked field response of the dentate gyrus and mossy fiber reorganization in temporal lobe epileptic patients, Brain Res., 579 (1992) 119-127. [16] Mello, L.E., Cavalheiro, E.A., Tan, A.M., Pretorius, J.K.,
Babb, T.L. and Finch, D.M.. Granule cell dispersion in relation to mossy fiber sprouting, hippocampal cell loss, silent period and seizure frequency in the pilocarpine model of epilepsy, Epilepsy Res. Suppl., 9 (1992)51-60. [17] Obenhaus, A., Esclapez, M. and Houser, C.R., Loss of glutamate decarboxylase mRNA-containing neurons in the rat dentate gyrus following pilocarpine-induced seizures, J. Neurosci., 13 (1993) 4470-4485. [18] Oiney, J.W., de Gubareff, T. and Sloviter, R.S., 'Epileptic' brain damage in rats induced by sustained electrical stimulation of perforant path. II. Ultrastructural analysis of acute hippocampal pathology, Brain Res. Bull,, 10 (1983) 699-712. [19] Paxinos, G. and Watson, C., The rat brain in s~ereotaxic coordinates, Academic Press, New York, 1982, Figs. 21 and 30. [20] Penix, L.R. and Wasterlain, C.G., Selective protection of neuropeptide containing dentate hilar intemeurons by nonNMDA receptor blockade in an animal model of status epilepticus, Brain Res., 664 (1994) 19-24. [21] Pisa, M., Sanberg, P.R., Corcoran, M.E. and Fiber, H.C., Spontaneous recurrent seizures after intracerebral injection of kainic acid in rat: A possible model for human temporal lobe epilepsy, Brain Res.. 200 (1980) 481-487. [22] Shirasaka, Y. and Wasterlain, C.G., Chronic epileptogenicity following focal status epilepticus, Brain Res., 655 (1994) 33-44. [23] Sloviter, R.S., 'Epileptic' brain damage in rats induced by sustained electrical stimulation of the perforant path. I. Acute electrophysiological and light microscopic studies, Brain Res. Bull., 10 (1983) 675-697. [24] Sloviter, R.S., Feedforward and feedback inhibition of hippocampai principal cell activity evoked by perforant path stimulation: GABA-mediated mechanisms, that regulate excitability in vivo, Hippocampus, i (1991) 31-40. [25] Sloviter, R.S., Permanently altered hippocampal structure, excitabi!k'y and inhibition after experimental status epilepticus in the rat: The 'dormant basket cell' hypothesis and its possible relevance to temporal lobe epilepsy, Hippocampus, 1 (1991) 41-46. [26] Sloviter, R.S., Possible functional consequences of synaptic reorganization in the dentate gyms of kainate-treated rats, Neurosci. Lett., 137 (1992) 91-96. [27] Sloviter, R.S. and Dimiano, B.P., Sustained electrical stimulation of the perforant path duplicates kainate-induced electrophysiological effects and hippocampal damage in rats, Neurosci. Lett., 24 (198) 279-284. [28] Sundstrom, L.E, Mitchell, J. and Wheal, H.V., Bilateral reorganization of mossy fibers in the rat hippocampus after a unilateral intracerebroventricular kainic acid injection, Brain Res., 609 (1993) 321-326. [29] Sutula, T., Cascino, G., Cavazos, J., Parada, I. and Ramirez, L., Mossy fiber synaptic reorganization in the epileptic human temporal lobe, Ann. Neurol., 26 (1989) 321-330. [30] Tauck, D.L. and Nadler, J.V., Evidence of functional mossy fiber sprouting in hippocampal formation of kainic acid treated rats, J. Neurosci., 5 (1985) 1016-1022. [31] Tanaka, T., Kaijima, M., Yonemasu, Y. and Cepeda, C.,
C.G. W_nsterlaiJ~el a l . / Epih'p.~y Research 26 ~!996) 255-265
Spontaneous secondarily generalized seizm-es induced by a single microinjection of kainic acid into uniJateral amygdala in cats, Eleccroem'ephal(Nr. C[in. NeurophysioL, 61 (~985) 422-429. [32] WasterRain, C.G., Father, D.B.. Fairchild, MD., Synaptic mechanis.,ns in the kindled epileptic focus: A specu]alive
265
synthesis, in: A.V. Delgado-Esue~a, A.A. Ward, Jr~, D.M. Wo,~dbury and R.J. Po~ter (Eds.), Ad~ antes t~ AIeu~ol~gy, VoL 44, Raven Press, New York~ 1986, pp. 4I 1-433. [33] Wasterlain, C.G, Fujikawm D.G., Penix, L. and Sankmr, R., Pathophysiologica~ mechanisms ,~f brain damage from status epi~eplicas, Epi/eps/a, 74( t J { 1993) $37-$53.
$EARCH
Epilepsy Research 26 ,~,,~,~r.~, ~,~, _~.~qq-_.06~
Chromc epilepsy wkh damage restricted to the Nppocampus: possible mechanisms Claude Go Wasterlain
Yukiy sN Shkasaka a,b,c Andrey M. Mazarafi Igor " ' man c,d plge
a Epilepsy Research Laboratory, VA Meaica! Ce~lter (127), Sepuh'eda. CA 91343-2099, USA b Department o f Neurology, UCLA School of Medicine. Los Angeles. CA 90024. USA Brain Research b,s6mte, UCLA Sch )ol of Medicine. Los Angeles, CA 90024, USA a UCLA School of Demis~', Los Angeles, CA 90024, USA
Received 29 Novernbe:."1995; accepted 6 May 1996
Abstract
We studied the time course and possible mechanisms of the development of chronic epilepsy following unilateral stimulation of the pefforant path. After 24 h of pefforant path stimulation by a modified Sloviter method, lesions were restricted to the hippocampus, except for 2 of 24 rats with minimal eatorhinal neuronal injury in layer 3. Lesions were exclusively ipsilateral in the ?olymorph layer of the hilus and m CA4-CA 3c, predominantly ipsilateral in CA 3, in CA~ and in the granule cell layer. Feedforward and feedback inhibition were studied by paired pulse stimulation. In the week following inhibition, there was complete loss of GABAA-mediated, sbon interstimulus interval (lSI)-dependenl inhibition and frequency-dependent inhibition, and also of GABAB-mediated long ISI-dependent inhibition. Yet no spontaneous seizures were observed at that time. In the next four weeks, we saw no evidence of increasing excitatory drive such as would I~e expected from recurrent mossy fiber sprouting. On the contrary, there was progressive return of inhibition. By four weeks post-lesion, the majority of animals had developed spontaneous recurrent seizures, and/or seizures on 2 Hz stimulation (never seen in controls), in spite of complete or near-cornplete recovery of short ISI-dependent, GABAn-mediated inhibition. A small but significant loss of frequency-dependent inhibition persisted, but individual animals with complete recovery of frequency-dependent inhibition showed spontaneous seizures, suggesting that loss of GABAA-mediated inhibition was not the direct cause of chronic epilepsy. GABA B-mediated, long ISI-dependent inhibition continued to show a significant loss. The ratio of the population spike amplitude at 250 /zA to the maximal population spike amplitude, a measure of granule cell excitability, was unchanged immediately after stimulation, but increased in the next few weeks in a manner identical to that seen in kindling, suggesting the possibility that during the transient loss of inhibition, spontaneous kindling had occurred. Intracellular recordings from grar, ule cells in hippocampal slices prepared from these animals showed a significant loss of GABA B-mediated slow inhibitory postsynaptic potentials (IPSPs). These data show that the sequellae of unilateral status epilepticus with damage restricted to the hippocampus are sufficient to cause chronic recurrent seizures. There is a possibility that chronic epilepsy is not the direct result of the loss of inhibitory drive or of a sprouting-induced
* Corresponding author. 0920-1211/96/$15.00 Copyright © 1996 Published by Elsevier Science B.V. All rights reserved. PII S0920-1211 (96)00058-7
256
C.G. Wasterlain e~ al. / Epilepsy Research 26 (1996) 255-265
increase in excitatory drive, but represents plastic changes akin to spontaneous kindling, possibly facilitated by loss of GABA a-mediated inhibition. Keywords: Epilepsy; Hippocampus; Status epilepticus; Kindling; GABA; Inhibition
1. Introduction Many recent studies have shown that seizure-induced brain damage can in turn lead to chronic epilepsy [3,6,8,10,13,14,21,31]. Both the kainic acid and the pilocarpine model of status epilepticus lead to extensive brain damage and to the delayed occurrence of both amygdaloid and generalized convulsions [7,8,31]. These rats also show interictal spikes and electrographic seizures [3,13]. While most of these studies have concentrated on the hippocampus as the most easily studied epileptogenic area, their lesions are disseminated throughout the brain on both sides, and the source of the seizure activity in those animals is unknown. Stimulation of the perforant path (pfps) induces neuronal injury which has an excitotoxic appearance similar to that induced by kainic acid [18,23,27]. Lesions are far more restricted, and are associated with a measurable loss of frequency-dependent paired pulse (PP) inhibition m the dentate gyrus [25]. Therefore, this is a favorable model to unravel the physiological changes due tc seizure-induced cell loss and to understand their role in epileptogenicity. Spontaneous seizures have usually been explained either by the build-up of recurrent excitatory connections as a result of sprouting [1,8,9,11,12,15,16,26,28-30], or by the loss [17] or deafferentation [2,25] of GABAergic inhibition in the hippocampus. In this study, we examined whether the time course of development of spontaneous seizures following perforant path stimulation is compatible with these hypotheses. Our results suggest another potential explanation, witL, kindling-like changes in granule cells excitability and a long-lasting loss of GABAa-mediated IPSPs as a possible mechanism.
2. Material and methods Our method of stimulation is slightly modified from Sloviter [25]. Adult Wistar rats are stimulated
for 24 h with intermitted trains of single stimuli at 20 Hz (20 V, 0.1 ms duration and continuous 2 Hz) delivered once per minute to an electrode located in the angular bundle, under continuous intravenous urethane anesthesia. Rats which lost their population spikes during stimulation were eliminated from the study, and paired pulse stimulation was carried out before, during and at various intervals after stimulation. At key times, we varied intrastimulus interval systematically from 15 to 5,000 ms (0.1 Hz, 0.1 ms duration, 30 V monophasic stimulation). Input-output responses were studied at intensities varying from 250 to 50,000 /.,A and kindling was tested through the pefforant path 4 weeks after perforant path stimulation or sham stimulation; these methods have been described in detail previously [22]. Acute cell injury was measured 3 days after stimulation in hematoxylin-eosin stained serial sections. Fig. l C-D ~hows that eosin in the neuronal cytoplasm is easily visualized by its bright yellow fluorescence under UV light, making the identification of 'ischemic cell change' [4,5] easier. Fig. 1(1-4) provides examples of the qualitative damage scale used (see legend). Three days after stimulation, animals received a lethal injection of pentobarbital, and were perfused through the heart with a 4% solution of paraformaldehyde. The brains were embedded in paraffin and 8 /~m sections were stained with hematoxylin and eosin. Injured neurons were defined as those showing a pyknotic nucleus (viewed under regular light), and an eosinophilic cytoplasm as viewed under UV through a fluorescein filter. Estimation of damage were performed on sections located at - 2 . 3 , -2.8, -3.3, - 3 . 8 and - 4 . 3 mm posterior to bregma according to the Paxinos and Watson atlas [19] in a Minded manner for both the stimulated and nonstimulated sites. All injured neurons within the hilar region bounded by the inferior and superior blades of the dentate gyrus and by a line drawn between a lateral tips of each blade were counted. The values for each section for each animal were then summed and divided by the number of sections to obtain the
C.G. Waslertai~ el al. / t~£pitepsy Research 26 ~f996) 255-265
damage score for that animal. Since plotting of the residuals showed a normal distribution, a paired t-test was used to compare the stimulated site to the contralateral site. Chronic changes were measured at least 1 month after stimulation in sections stained with cresyl violet or with antibodies tc specific subpopulations of neurons [20]. Seven ~mplanted controls were used at each time point.
257
the hippocampus no lesions were seen in untreated co3~trc!s~ implanted controls or i~ sham stimulated controls, but all stimulated anima!s showed some neuronal loss (Fig. 1). Neuronal ~oss was exclusively ipsilaterat in hilus and in CA:~c, and in a few animals lesions were restricted to the hilus. However, the majority of rats had bilateral lesions (more severe on the side of stimulation) in CA 3, CA~, and in the dentate granule cell layer (Table 1).
2.1. Methodology of intracellular recording Whole brain was isolated from control or stimulated rats, and coronal sectio ~s were obtained from the ipsilateral hemisphere at 0°C using a vibrating blade slicer (Campden Instruments). Slices were maintained at room temperature in oxygenated (95% O~-5% CO 2) artificial cerebrospinal fluid (ASCF, in raM: NaCl, 125; KCI, 2.5; NaH2PO 4, 1.25; CaCI 2, 2; MgCl z, 2; NaHCO 3, 26 and glucose, 10; pH 7.4), and stabilized for 1 h before recording. All experiment~ were performed on slices submerged in the recording chamber at 34°C with a superfusion rate of approximately 2 ml/min. Intracellular recordings were made with 3 M potassium acetate-filled ~¢ass microelectrodes (pH = 7.25, tip resistance 120-170 Mf~) placed in the dentate granule cell layer of the hippocampus. A concentric bipolar electrode was positioned approximately 0.5 mm from tbe recording electrode in the outer molecular layer to stimulate the perforant path inputs and to minimize the variation of responses between slices. Stimuli (5 IxA-10 mA) were square wave pulses (200 ~s) delivered at 0.083 Hz. Intracellular potentials were amplified using an Axoclamp-2A amplifier. Data were digitized via a 12-bit analog-to-digital interface (TL-1) and analyzed using the pCLAMP software package (Axon Instruments).
3.2. After pfps, ihere was serere loss of iahib#iorl but no seizures Frequency-dependent and short ISI-dependent paired pulse (PP) inhibition, which appear to be GABAA-mediated [24], and long [SI-dependent PP inhibition which is probably GABA B-mediated, were completely lost during the week following pfps (Figs. 2 and 3). However, the ratio of the population spike a~,~plitude at 250 /xA to the maximal population spike amplitude, which is a measure of granule cell excitability, was not increased (Fig. 3). No spontaneous seizures were observed until the end of the second week after stimulation (n = 12) (Fig. 6). Since EEG was not monitored we cannot rule out the possibility that electrographic seizures appeared before behavioral seizures, but the latter were clearly delayed and only appeared several weeks after the lesion in a majority of animals. If the mechanism of chronic epilepsy is ti~e direct result of loss of inhibition, either because of deafferentation of basket cells [2,25] or because specific subpopulation of GABAergic cells are lost [17] it is very surprising that no spontaneous seizures occur at the time when loss o f inhibition is the greatest. 3..? The &bT,:'arance of spontaneous seizures coinci~'cd with the progressive but partial recoL,eiy of int,~ibition
3. Results and discussion 3.1. Neuronal loss was restricted to the hippocampus Of the 24 animals used for acute anatomical studies, 22 showed no lesions outside of the hippocampus. The remaining 2 animals showed a few necrotic neurons in layer 3 of entorhinal cortex. In
Over the four weeks following stimulation, short • ~I-dependent ir~:--;bition, which is thought to be GAB/~A mediated [24], appeared to recover comt,,!ete!y (Figs. 2 and 3), although one could argue that the. p,Ayspikes produced by the initial stimulus provide a stronger drive than the single spike seen in controls, and that therefore recovery is substantial
258
C.G. Wasterlain et al. / Epilepsy Research 26 (1996) 255-265
but not complete. Frequency-dependent PP inhibition, which is also thought to be G A B A a - m e d i a t e d , recovered only partially, but in some individual rats complete recovery was observed (with the same
proviso as above} while they were developing chronic recurrent spontaneous seizures (Fig. 4). Lorlg IS|-dependent inhibition and first stimulus pf~tyspike also s h o w e d a trend toward recovery, but this was far
Fig. 1. T bright fluorescence, while tlae contralateral hippocampus (B) is intact (original magnification × 10). Injured hilar neurons show a pink cytoplasm under regular light (C) and stand out under fluorescence (D) (original magnification X 40). Pictures e-h show various grades of hilar damage (e = 1; f = 2; g = 3; h = 4) (original magnification × 20).
C.G. W~¢slerlai~ e1 ~.1./ L)"#h.psy Research 26 ¢[99b~ 255-265
Table B Anatomy of neuronal injury fifllowing unilatera~ per~,~>rum path stimuiation Region
Stimulated
Col:trtHa~.eral
Dentate hilus CA 3c
2.2 +_0.5 " 2 _+0.6 ~ 2 +O.4" O.8 + 0.3 R.3 +_0.4 " 0 < O.1 0
0 O.I 4- 0.4 0.5_+_0.3 O.l + 0.2 0.4 + (1.2 0 < o, I 0
CA 3
CA) Granule cells Neocortex Entorhinal cortex Pyriform cortex
259
toting was c a m e d oat to detect seizures, both kindled-iike and generalized c o n v u l s i v e aeizures were observed in a majority o f animals (Fig. 5). Seizures star~ed at the end o f ~he second week in one animal and at the e n d o f the fourth w e e k or later in the remaining rats (Fig. 5).
3.4. Delayed seiz:ures were accompanied by a deh~ved increase in gramde celt excimbi6n,
Damage was measured 72 h after unilateral pq~s in hematoxy]meosin stained sectk)e,s at the level of habenulm nuclei (dorsal hippocampus). A qualitative scale for damage was used (I = < I(1% ir~iured neurons: 2 = 10-25%; 3 = 25-50%: 4 = > 50~). SE are provided only as a description of the population and were not used in assessing significance, p < 0.05 compared to contralateral, using a t-test for paired data. from c o m p l e t e and those m e a s u r e s r e m a i n e d quite abnormal at 4 w e e k s [22]. B e h a v i o r was o b s e r v e d only o n c e a w e e k , and while no electrographie m o n i -
in i n p u t - o u t p u t response studies, granule cell excitability was not f o u n d to increase immediately after pfps, H o w e v e r , it had increased significantly two weeks after pfps, in a way similar to the response to kindling stimulation in control rats !22]. "Ihis delayed increase in dentate grannie cell excitability might reflect a kindling-like p h e n o m e n o n precipitated by the loss o f h i p p o c a m p a l inhibition [32], or might be related to m o s s y fibers sprouting, which usually appears seve~'al w e e k s after neuronal d a m a g e [28].
Fig. 1 (continued).
260
C.G. Waswrlai, et al. / Epilepsy Reseatz'h 26 (I 996) 255-265
Tst
25
40
100
300 ,000
b
Ist
25
40
I00
300
I000
F
Fig. 2. Interstimulus interval-dependent paired pulse examination in two stimulated rats. The results of interstimulus interval-dependent paired pulse examinations in two 24 h perforant path stimulated rats (a and b) are shown. The waveforms show the average of 10 consecutive waves. The first positive deflection in each trace is the stimulation artefact, and the EPSP with/without PS discharge follows. Paired pulse stimulations (0.1 ms duration, 30 V monophasic) were given at 0.1 Hz. Abbreviation: A, before pfps in the awake state; B, hefxe pfps under urethane anesthesia; C, 30 min after pfps; D, 2 weeks after pfps in the awake state; E, 4 weeks after pfps in the awake state; F, 4 weeks after pfps under urethane anesthesia; 1st, 1st PS wave form; 25, 40, 100, 300, and 1000 indicates the interstimulus interval (ms) for that waveform. Vertical scale bar ! mV, horizontal scale bar 10 ms. Reproduced from Ref. [18], with permission,
C.G. Wasterlain el aL / Epilepsy Research 26 ¢]996t 255-265
This increase in granule cells excitability might also be responsible for the convulsive response to 2 Hz stimulation which was observed in haft of the animals three to ~imr weeks after ptps. Beth electrical and chemical kindling require activation of a critical mass o f neurons [321. W e should consider the possibility that as the result of loss o f inhibition, physiological stimuli resulting from routine activity m a y recruit e n o u g h neurons to induce a synchronous afterdischarge which exceeds the kindling threshold. If
120
26]
full kindling occurs, even a compiete recovery of inhibition later might ~,ot be able to abolish estabfished kindling, so that seizures might cormnue. We previously postulated the existence of a "filter' that limits recruitment o f bt:rsdng neurons so that their responses to physiological stimuli never exceed the kindling threshold [32]. The present data are compatible with the view thai hilar interneurons may "filter' input into the dentate gyrus, and that their loss after pfps c o m p r o m i s e s that filter function.
88
PiPS
100
61]
80
=
60
40
o
PPS
46
rnulate
20 O
~
.- z0
control
A
B !
-20
before
30 rain after
2~ after
t 20
t
before
4 wk~ after
!
30 rain after
- - 1o8
190
~
pfp$
I
2 wks a~ter
_ _
4 wt
Kindling
"=
.T~----,-~
o 80 td~
80
,U,
t
N so
~"
40
20 o -20
4o
~
rnulated
control
C
~ ~ ca
20
before
~, 30 rain
i 2 v~s
I 4wks
alter
alter
after
~ ' ~ ' ~ t ~ /
==..4 s~rnul~ ~ conlTol
o
o. - 2 0
,
t
~ 40
~
So
PfPS
D
n
I
I
i
before 30 rain 2 ~ 4 wks after alter alter
I
t --
alter Kindling 1 wk 4wI~
Fig. 3. Sequentiai changes of paired-pulse inhibition and granule ~',~11excitability. (a) Sequential changes of inhibition score [100- (PS amplitude of the 2nd puise)/PS amplitude of the 1st pulse) × 100] i~: short (25 ms) interstimulus interval (ISI)-dependent paired-pulse (PP) examinations, in stiraulated rats and control rats. (b) Those at 300 ms ISI. (c) Sequential changes of inhibition scores in frequency-dependent PP examinations, at 2 Hz in stimulated and control rats. (d) Sequential changes of PS ratio [(ps amplitude at 250 ,~A stimulus intensity)/(maximal PS amplitude in the same experiment) x lo0] in stimulated rats and control rats. All the data shown in this figure were examined in the awake state except for 30 min after 24 h perforant path stimulation (pfps) or 24 h anesthesia, which was examined under urethane anesthesia. Abbreviations: pfps or Kindling with arrows, the time point at which pfps or kindling stimulation was given; sthnulated, rats given pfps; control, non-stimulated control rats; before, before pfps in stimulated rats and before 24 h anesthesia in control rats; time after pfps or 24 h anesthesia, respectively. +, * Significantly different from before pfps (stimulated group) or before kindling (control group) in the s~maestate (awake or anesthetized), at p < 0.05 mid p < 0.01, respectively; *,* * significantly different from control at p < 0.05 and p < 0.01, respectively. Reproduced from Ref. [18], with permission.
262
C.G. Wasterlain et al. / Epilepsy Research 26 (1996) 255-265
3.5. Pj~ lesioned animals showed p~rsistent loss of GABAB-mediated slow IPSPs
Based on our results from evoked potential recordings we decided to explore the possible mechanisms which underlay chronic epileptogenicity with the use of intracellular recording techniques. We studied slices taken from animals 1 month after stimulation, a time at which they exhibited spontaneous seizures and the G A B A B receptor-mediated long ISI inhibition was still reduced. IntraceUular recordings obtained from the dentate granule (DG) cells of these animals showed that they exhibit postsynaptic membrane properties which did not differ from those seen in control (implanted; non-stimulated) rats. Their resting membrane potential, input resistance, spike amplitude, duration and
discharge pattern in response Lo depolarizing current pulses were similar to controls. In our preliminary studies comparing the respouses to synaptic stimulation of DG cells from stimulated or control rats, a trend towards reduced slow IPSPs was observed in stimulated rats. This slow IPSP results from the activation of G A B A B receptors, which in turn activates a K ÷ conductance. We stimulated synaptic inputs while changing the membrane potential of the recorded neuron with intracellular injections of hyperpolarizing or depolarizing current pulses. We used synaptic stimuli which were sufficient to evoke a large EPSP subthreshold to spike generation. As the membrane potential is changed from more hyperpolarized to more depolarized values, the slow IPSP (which peaks at 2 0 0 - 2 5 0 ms after synaptic stimulus) also changes from a
Fraquency-Dependent Paired-Pulse inhibition Nter Focal SE 0.1Hz
1Hz
Before ~ 1 ~ Awake
2Hz
.
2wksAfterI l l l | , Awake 4wksAfteJr
~
l
l
Awake 1mY
k__
lOmsec Fig. 4. Complete recovery of frequency-dependentpaired pulse inhibition in a rat exhibiting recurrent spontaneous seizures. The waveforms are average of l 0 consecutive waves. The narrow positive deflection before EPSP in each trace is a stimulatic~lartefact PP stimulations (0.1 ms duration, 30 V monophasic) were given 40 ms apart. Vertical scale bar = l mV. horizontal scale bar = 10 ms.
263
C.G. Wasterlain etal,/Epilepsy Researck 26 f ]9~6~ 2.55-265
depolariziag to a hyperpolarizing response. By extrapolating peak stow IPSP values from each recorded nemon to - 7 0 mV we were able to demonstrate significant reduction in the amplitude of the slow IPSPs in stimulated rats (3.8 + 0.7 mV, n = 7 cells fi'om 5 rats) compared to controls (7.6 + 1.5 mV, n = 6 cells, 5 rats). The reduction of slow IPSPs in stimulated rats could be due to a decrease in GABA B receptor function a n d / o r a decrease in the release of GABA from the terminals of inhibitory interneurons. While the role of this persistent los~ of inhibition is unknown, GABAa-mediated slow IPSPs are likely to be involved in chronic modulation of hippocampal excitability associated with synaptic plasticity, and represent a potential explanation of both the increase in granule cell excitability (Fig. 3) and the enormous increase in susceptibility to kindling (Fig. 5) which were observed in pefforant path-stimulated rats in this study. Further studies are in progress to test these speculative interpretations.
125
5011
~5
°L .r~.
400
"~
~
~
75
lOO
o
0
Fig. 5. Kindling in pfps rats. Stimulations were given via electrodes in the pefforant path, and responses were recorded via electrodes in the dentate gyrus and pefforant path. An afterdischarge (AD) was defined as high amplitude spike or polyspike epileptiform activity visible for 3 s or more after completion of the stimulus. AD threshold (ADT) was the lowest intensity of current necessary to elicit AD, and AD duration (ADD) as the time lapsed_ from the end of the stimulus until the organized AD terminated at either recording site. Data depicts means (SEM), ~ d (minimum-maximum). Stimulated rats were compared with nonstimulated controls by ANOVA. Abbreviations: First ADT, ADT at first trial; first ADD, ADD at first trial; first C 5, number of trials necessary to induce the first class 5 convulsion; third C S, number of trials necessary to induce three class 5 convulsions; other abbreviations as in Table I.
:C' cv, C f
c~ f
f_
x._ (D ..D
E
/v
Z PFPS
~ wk
/ 2 was
4 wks
>4
wks
Fig. 6. Time of onset of generalized convulsive seizures in pefforant path-stimulated rats (n = 10) as a function of time elapsed since stimulation (weeks). Orice seizures started they recurred in all rats in this group. ( v ) Spontaneous seizures, ( [ ] ) seizures induced by 2 Hz stimulation (which did not induce seizures in controls).
In summary, the perforant path stimulation model clearly shows that hippocampal damage resulting from unilateral status epilepticus is sufficient to cause chronic epilepsy, and therefore Sommer and Pfleger were both correct when they stated, in 188], that hippocampal lesions were the cause of epilepsy and the result of seizures, respectively [33]. This study gave no indication of an increase in excitatory drive as postulated by the recurrent mossy fiber sprouting hypothesis [l,8,15,28-30]. The relationship between loss of inhibition (be it due to death of GABAergie cells [17], to their active suppression or to their deafferentation [24]) and the occurrence of spontaneous seizures appears more complex than was previously realized. Since spontaneous behavioral seizures were not seen at the peak of inhibitory loss, an additional process must occur between the loss of inhibition and the eventual epileptogenicity. The current study ;.~ises the possibility that this might be a kindling-like phenomenon, and that the persistent loss of GABAa-mediated inhibition observed in chronic animals might play a role in that development.
Acknowledgements Supported by Research Service of the VHA and by grant NS 13515 from NWDS.
264
C.G. Wasterlain et at,/Epilepsy Research 26 (1996) 255-265
References [I] Babb, T.L., Kupfer, W.R., Pretorius, J.K., Crandall, P.H. and Levesque, M.F., Synaptic reorganization by moss? fibers in human epileptic fascia dentata, Neuroscience, 42 (1989) 351-363. [2] Bekenstein, J.W. and Lothman, E.W., D~..-mancy of inhibitory intemeurons in a model of temporal lobe epilepsy, Science, 259 (1993) 97-100. [3] Bertram, E.H. and Comett, J., The ontogeny of seizures in a rat model of limbic epilepsy: Evidence for a kindling process in the development of chronic spontaneous seizures, Brain Res., 625 (1993) 295-300. [4] Brown, A.W. and Brierley, J.B., Anoxic-ischaemic cell change in rat brain light microscopic and fine-structural observations, J. NeuroL Sci., 16(!) (1972) 59-84. [5] Brown, A.W. and Brierley, J.B., The earliest alterations in rat neurones and astrocytes after anoxia-ischaemia, Acta Neuropathol. (Berlin), 23(1) (1973) 9-22. [6] Ca,~alheiro, E.A., Riche, D.A. and Le t'al La Salle, G., Long-term effects of intrahippocampal kainic acid injection in rats: A method for inducing spontaneous recurrent seizures, Electroencephalogr. Clin. NeurophysioL, 53 (1982) 581-589. [7] Cavalheiro, E.A., Leite, J.P., Bortolotto, A.Z., Turski, W.A., Ikoaomidou, C. and Turski, L., Long-term effects of pilocarpine in rats: Structural damage of the brain triggers kindling and spontaneous recurrent seizures, Epilepsia, 32 (1991) 778-782. [8] Cronin, J. and Dudek, F.E., Chronic seizures and collateral sprouting of dentate mossy fibers after kainic acid treatment in rats, Brain Res., 474 (1988) 181-184. [9] Cronin, J., Obenhaus, A., Houser, C.R. and Dudek0 F.E., Electrophysiology of dentate granule cells after kainate-induced synaptic reorganization of the mossy fibers, Brain Res., 573 (1992) 305-310. [10] Feldblum, S. and Ackermann, R,F., Increased susceptibility to hippocampal and amygdala kindling following intrahippocampal kainic acid, Exp. Neurol,, 97 (1987) 255-269. [11] Franck, J.E. and Schwartzkroin, P.A., Do kainate-lesioned hippocampi become epileptogenic?, Brain Res., 329 (1985) 309-313. [12] Heinemann, U., Chssmann, H., Dreier, J. and Stabel, J., Changes ia synaptk: transmission in the kindled hippocampus, Adv. Exp. Med. 13iol., 268 (1990) 445-450. [13] Lothman, E.W., Ber~ am, E.H., Kapur, J. and Stringer, J.L., Recurrent spontaneous hippocampal seizures in the rat as a chronic sequelae to limbic status epilepticus, Epilepsy Res., 6 (1990) 110-118. [14] Lothman, E.W., Bertram, III, E.H. and Stringer, J.L., Functional anatomy of hippocampal seizures, Prog. NeurobioL, 37 (1991) 1-82. [15] Masukawa, L.M., Uruno, K., Sperling, M., O'Connor, M.J. and Burdett, L.J., The functional relationship between antidromically evoked field response of the dentate gyrus and mossy fiber reorganization in temporal lobe epileptic patients, Brain Res., 579 (1992) 119-127. [16] Mello, L.E., Cavalheiro, E.A., Tan, A.M., Pretorius, J.K.,
Babb, T.L. and Finch, D.M.. Granule cell dispersion in relation to mossy fiber sprouting, hippocampal cell loss, silent period and seizure frequency in the pilocarpine model of epilepsy, Epilepsy Res. Suppl., 9 (1992)51-60. [17] Obenhaus, A., Esclapez, M. and Houser, C.R., Loss of glutamate decarboxylase mRNA-containing neurons in the rat dentate gyrus following pilocarpine-induced seizures, J. Neurosci., 13 (1993) 4470-4485. [18] Oiney, J.W., de Gubareff, T. and Sloviter, R.S., 'Epileptic' brain damage in rats induced by sustained electrical stimulation of perforant path. II. Ultrastructural analysis of acute hippocampal pathology, Brain Res. Bull,, 10 (1983) 699-712. [19] Paxinos, G. and Watson, C., The rat brain in s~ereotaxic coordinates, Academic Press, New York, 1982, Figs. 21 and 30. [20] Penix, L.R. and Wasterlain, C.G., Selective protection of neuropeptide containing dentate hilar intemeurons by nonNMDA receptor blockade in an animal model of status epilepticus, Brain Res., 664 (1994) 19-24. [21] Pisa, M., Sanberg, P.R., Corcoran, M.E. and Fiber, H.C., Spontaneous recurrent seizures after intracerebral injection of kainic acid in rat: A possible model for human temporal lobe epilepsy, Brain Res.. 200 (1980) 481-487. [22] Shirasaka, Y. and Wasterlain, C.G., Chronic epileptogenicity following focal status epilepticus, Brain Res., 655 (1994) 33-44. [23] Sloviter, R.S., 'Epileptic' brain damage in rats induced by sustained electrical stimulation of the perforant path. I. Acute electrophysiological and light microscopic studies, Brain Res. Bull., 10 (1983) 675-697. [24] Sloviter, R.S., Feedforward and feedback inhibition of hippocampai principal cell activity evoked by perforant path stimulation: GABA-mediated mechanisms, that regulate excitability in vivo, Hippocampus, i (1991) 31-40. [25] Sloviter, R.S., Permanently altered hippocampal structure, excitabi!k'y and inhibition after experimental status epilepticus in the rat: The 'dormant basket cell' hypothesis and its possible relevance to temporal lobe epilepsy, Hippocampus, 1 (1991) 41-46. [26] Sloviter, R.S., Possible functional consequences of synaptic reorganization in the dentate gyms of kainate-treated rats, Neurosci. Lett., 137 (1992) 91-96. [27] Sloviter, R.S. and Dimiano, B.P., Sustained electrical stimulation of the perforant path duplicates kainate-induced electrophysiological effects and hippocampal damage in rats, Neurosci. Lett., 24 (198) 279-284. [28] Sundstrom, L.E, Mitchell, J. and Wheal, H.V., Bilateral reorganization of mossy fibers in the rat hippocampus after a unilateral intracerebroventricular kainic acid injection, Brain Res., 609 (1993) 321-326. [29] Sutula, T., Cascino, G., Cavazos, J., Parada, I. and Ramirez, L., Mossy fiber synaptic reorganization in the epileptic human temporal lobe, Ann. Neurol., 26 (1989) 321-330. [30] Tauck, D.L. and Nadler, J.V., Evidence of functional mossy fiber sprouting in hippocampal formation of kainic acid treated rats, J. Neurosci., 5 (1985) 1016-1022. [31] Tanaka, T., Kaijima, M., Yonemasu, Y. and Cepeda, C.,
C.G. W_nsterlaiJ~el a l . / Epih'p.~y Research 26 ~!996) 255-265
Spontaneous secondarily generalized seizm-es induced by a single microinjection of kainic acid into uniJateral amygdala in cats, Eleccroem'ephal(Nr. C[in. NeurophysioL, 61 (~985) 422-429. [32] WasterRain, C.G., Father, D.B.. Fairchild, MD., Synaptic mechanis.,ns in the kindled epileptic focus: A specu]alive
265
synthesis, in: A.V. Delgado-Esue~a, A.A. Ward, Jr~, D.M. Wo,~dbury and R.J. Po~ter (Eds.), Ad~ antes t~ AIeu~ol~gy, VoL 44, Raven Press, New York~ 1986, pp. 4I 1-433. [33] Wasterlain, C.G, Fujikawm D.G., Penix, L. and Sankmr, R., Pathophysiologica~ mechanisms ,~f brain damage from status epi~eplicas, Epi/eps/a, 74( t J { 1993) $37-$53.