107

Epilepsy Res., 8 (1991) 107-116 Elsevier EPIRES 00390

Transition from normal to epileptiform activity in kindled rat hippocampal slices

S.M. BawinaYb*c , W.M. SatmaryC, M.D. Mahoney” and W.R. AdeyC Departments

of ‘Physiology and bNeurosurgery, Loma Linda University, and ‘J. L. Pettis Memorial Veterans Hospital, Loma Linda CA (U.S.A.)

(Received 7 August 1990; revision received November 1990; accepted 8 November 1990) Key words: Kindling; Hippocampal slices; Late EPSPs; Interictal bursts; N-methyl-D-aspartate

We previously demonstrated kindling of synchronized bursts (1%) by repeated sine-wave stimulation (SW: 2-5 set, 60 Hz, 20-50 PA, every 5 min) in the CAY3 area of rat hippocampal slices. Here we report the behavior of individual CA2/3 neurons during the kindling procedure. Intra- and extracellular recordings were obtained concurrently before, during and following SW. Test pulses and SWs were applied in CA2/3 or CA1 stratum radiatum. Neuronal response to intracellular stimulation was tested by 100 msec depolarizing dc pulses or by 2-20 set sinusoidal currents. The role of the N-methyl-D-aspartate (NMDA) receptor in the transition from normal responses to ISs was assessed by perfusing the slices with a specific antogonist (DL-2-amino-5-phosphono-vale& acid, APV, 50-2OOpM). Our results show that kindling of ISs occurred in two steps: (1) via NMDA-dependent depolarizations during SW, or during SW-induced afterdischarges, and (2) through the recruitment of secondary, late EPSPs (IEPSPs), between consecutive SWs. ISs developed from the IEPSPs, while the early responses (action potentials, EPSPs, and population spikes) remained unchanged. Kindling of ISs occurred with no changes in resting membrane potential, membrane resistance, or threshold of action potentials. APV did not block kindled ISs, but considerably reduced their amplitude and duration, and increased their frequency. These latter findings suggest that APV-insensitive mechanisms, activated through NMDA-dependent processes, were responsible for the triggering of ISs, and that NMDA receptor systems participated in the control of their rate of occurrence.

INTRODUCTION The kindling phenomenon, first reported by Goddard et al. in 19679, has become one of the most studied animal models of epilepsy. The procedure, which involves repeated stimulation of a brain structure with a subconvulsive stimulus, pro-

Correspondence too: Suzanne M. Bawin, J.L. Pettis Memorial Veterans Hospital, Research Service (151) 11201, Benton Street, Loma Linda, CA 92357, U.S.A. Note: Part of the results described here have been reported in Neurosci. Abstr., 15 (1989) 454.

0920-1211/91/$03.50

0

1991 Elsevier Science Publishers B.V.

gressively intensifies afterdischarges (ADS) and culminates in generalized seizures. This model has widely contributed to the present knowledge of electrographic activity in epileptic foci, to anatomical studies of epileptogenesis, and to pharmacological studies involving both anticonvulsants and epileptogenic agents (for reviews, see refs 18 and 23). However, the cellular mechanisms involved in the progressive transformation of normal neurons into bursting cells and in the synchronization of neuronal firing cannot easily be studied in this in vivo model. Since the development of brain-slice techniques

108

more than two decades ago, manipulation of excitatory and inhibitory mechanisms by convulsant agents has considerably increased the understanding of the cellular processes involved in epileptic bursts (for example, see refs. 22 and 33). Still, the transition from normal to bursting activity is difficult to study in these biochemical models, because of the rapid onset of fully developed epileptiform activity. Recently, the kindling model of epilepsy was extended to an in vitro preparation. Stasheff et al. first showed that repeated electrical stimulation with trains (2 set, 60 Hz) of pulses (50psec) at high intensity (400-1200 ,uA) induced prolonged ADS as well as spontaneous and test-pulse evoked synchronized bursts in rat hippocampal slices*‘. We later demonstrated that repeated weak 60 Hz sine-wave stimulation (SW, 2-5 set, 40-100 ,uA, pp) of the Schaffer collaterals or the mossy fibers also kindled synchronized bursts in CA2/3*. Hippocampal synchronized bursts in disinhibited slices2v35,in slices from previously kindled rats13, or in kindled slices (see above), are believed to be equivalent to the interictal spikes (ISs) seen in the epileptic brain in vivo. ISs were shown to increase in frequency, and spread from the site of stimulation to other parts of the brain, during kindling of seizures in rats’. In view of these findings, it appears paradoxical that ISs decrease the frequency of fits in vivo7,15,as well as in hippocampal slices treated with low magnesium solution3’. Whether or not ISs are refractory to seizures, they do appear to play an important role in epileptogenicity. The experiments reported here represent an attempt to study the dynamics of the development of ISs kindled in the CA2/3 area of rat hippocampal slices. The kindling stimulus (SW) was the same as in our previous studies*. Since the artifact induced by SW in intracellular recordings is easily reduced by digital filtering of the raw data (see below), we could investigate the behavior of individual CA2/3 neurons during SW, as well as during the transition from normal to bursting activity, between kindling stimulations. Concurrent extracellular recordings in CA2/3 provided information about the development of synchronicity in the region. The possibility that SW induced changes in intrinsic (non-synaptic) neuronal properties was tested by intracellular

stimulation with depolarizing DC pulses or sinusoidal currents. Finally, we investigated the role of the N-methyl-D-aspartate (NMDA) receptor in the induction and maintenance of ISs. Our results show that ISs resulted from the NMDA-dependent depolarizations during SW or ADS, and activation of secondary, late EPSPs in CA2/3. Once initiated, ISs were reduced but not completely abolished by the NMDA antagonist DL-2-amino-phosphono-valeric acid (APV, 50-200 ,uM), suggesting co-activation of other receptor systems. Neuronal responses to intracellular stimuli were unchanged by kindling. We suggest that the procedure mainly affected recurrent synaptic excitatory mechanisms in CA2/3, although the possibility that SW decreased local inhibition’9’31 cannot be ruled out. These results have been reported in part elsewhere3. MATERIALS

AND METHODS

Adult male rats (Sprague-Dawley, 60-100 days old, 250-375 g, n = 74) were used in this study. Transverse hippocampal slices were obtained as previously described4 and transferred to an interface-type recording chamber perfused with warmed (35 “C), oxygenated (95% 0,/5% CO,) artificial cerebrospinal fluid (ACSF). Warmed, moistened oxygen flowed over the slices. ACSF contained (in mM): 2 CaC1,/3.75 KCY1.2 MgCl,/ 1.2 NaH,POJ124 NaCV26 NaHCO,/lO glucose. Treatment with oL-2-amino-5-phosphono-valeric acid (APV, 50-200 ,uM) was by addition of the drug to ACSF. Standard electrophysiological techniques were employed. Intracellular recordings were made with a capacity-compensated amplifier with bridge circuit to transmit constant currents. Extracellular recordings used either AC- or DC-coupled amplifier. Intra- and extracellular signals were recorded concurrently in the CA2/3 cell layer, with glass micropipettes filled with 3 M K+ acetate (60-120 MQ) and 1 M NaCl (5-15 MQ), respectively. Monopolar test pulses (50 ,usec, 25-150 ,uA) were delivered at 0.1 or 0.2 Hz via a bipolar electrode in the CA2/3 stratum radiatum (sr). In some experiments, intracellular depolarizing pulses (lo-100 msec, 0.5-2 nA, 0.1 Hz) were delivered 5 set after

109

test pulse stimulation (also at 0.1 Hz). Hyperpolarizing pulses of 100 msec (0.2-2 nA) were used to measure the membrane input resistance (R,). Slices with extracellular responses (single population spikes, PSs) exhibiting robust paired-pulse potentiation (pulse intervals of 40-100 msec) and inhibition (pulse intervals of lo-40 msec) were chosen for the study (n = 74). CA2/3 neurons with stable resting membrane potential (RMP) larger than 55 mV and action potentials with overshoots, were selected among these slices (n = 57). R, ranged from 15 to 50 MQ (28.6 f 11.1, mean + S.D., n = 16). All studies described below were of 1 neuron per slice per animal, unless otherwise stated. All neurons were monitored for a minimum of 15 min following stabilization of the resting membrane potential with injection of negative current (-0.2 to -0.7 nA). During that period, neuronal responses to extracellular test pulses, as well as to depolarizing and hyperpolarizing intracellular DC steps were evaluated. PSs and membrane potentials were monitored on a digitizing oscilloscope and a chart recorder, and stored on magnetic tapes for further analysis. The kindling stimulus (SW) was a 60 Hz, 2-4 set sinusoidal constant current delivered every 5 min in CA2/3 sr, or in CA1 sr, via a second bipolar electrode. SWs ranged from 25 to 100 ,uA (50-200 ,uA p-p). The first SW was delivered at half the intensity of the test pulse which produced a l-5 mV PS in the CA2/3 cell layer. If this SW failed to induce short- or long-term increase in the intra- and extracellular responses to the test pulses, the next SWs were increased in steps of 12.5 ,uA. In the following, the first SW that induced changes in excitability will be referred to as SWl. Intracellular recordings during SWs were later digitized at 1 kHz and processed with a 50-70 Hz band-stop digital filter to reduce the artifacts and allow the study of depolarizations during the kindling stimulations.

A

I

B& 4

SW*= -+-sw* -YW

1”

-J----i

4

ccG

t

Fig. 1. Progressive kindling of ISs by 3 successive SWs. SW (50 ,uA, 2 s) and recordings were in CA2/3. (A) Depolarizations during and following SW1 and SW3. The intracellular recordings were processed with a 50-70 Hz band-stop digital filter to reduce the stimulus artifact. The hatched bars under the traces indicate the SWs. The apparent irregular height of the action potentials is due to the digital processing. (B) Intracellular EPSPs evoked by test pulses (arrows) before (top) and immediately after SW1 and SW3 (middle and bottom traces, see asterisks in A). A late EPSP can be seen after SW3. (C) The progressive increase of synchronized 1EPSPs led to the development of ISs. Extracellular and intracellular responses (top and bottom traces, respectively) to test pulses (arrows) and spontaneous bursts are shown 1, 4, 10 and 20 min after SW3. RMP was -69 mV. Calibration: horizontal bar, 2 set (A) and 40 msec (B,C); vertical bars, 20 and 1 mV for intra- and extracellular recordings, respectively.

RESULTS Kindling of interictal bursts by SW

Typically, the kindling of ISs by SW involved 2 steps: (1) potentiation of depolarizations and associated firing during successive SWs; and (2) postSW potentiation of secondary, late EPSPs

(1EPSPs). Fig. 1. illustrates these steps in a slice where ISs were kindled by 3 successive SWs. All intracellular recordings are from the same CA3 neuron. The neuronal responses during SW1 and SW3 (50 PA) are shown in Fig. 1A. As in most neurons that responded to the kindling procedure,

110

depolarization increased during SW, and was prolonged by an after-depolarization terminated by hyperpolarization of the membrane potential (AHP). Fig. 1B shows intracellular EPSPs evoked before (top trace), and immediately after SW1 and SW3 (see asterisks in Fig. 1A). A 1EPSP is clearly seen following SW3 (bottom trace). The progressive development of 1% is illustrated in Fig. 1C. Field potentials and intracellular recordings (top and lower traces, respectively) are shown 1,4, 10 and 20 min after SW3. Comparison between the responses at 1 and 4 min shows that bursts originally occurred during the IEPSPs while the ‘early’ responses (eEPSP and population spike) remained practically unchanged. The latency of the 1EPSPs and bursts decreased progressively until early and late responses were fused together (see responses at 10 min), and spontaneous bursts of action potentials riding on large depolarization shifts (see traces at 20 min) started to appear between test pulse stimulations. Similar intra- and extracellular kindling patterns were observed concurrently in 15 other experiments. An additional series of 11 slices where only field potentials were recorded showed the same transition from single population spikes to either evoked or spontaneous synchronized bursts. The number of SWs necessary to kindled ISs varied from 1 to 16 (4.7 f 3.4, n = 27) with a mean intensity of 47.3 rt 22.8 PA. Kindling of ISs occurred without long-term changes in RMP or R,. Measurements in 7 neurons showed values of 30.8 + 11.1, and 31.4 * 10.9 Ma, before and after kindling of ISs. respectively. Bursts could be evoked by test pulses delivered at 0.1 or 0.2 Hz, but not faster. When spontaneous bursts developed, test pulse stimulation was discontinued and ISs were seen to occur at 0.2-0.05 Hz. Fig. 2Aa shows chart recordings of kindled spontaneous extracellular (top) and intracellular bursts (bottom) in the same slice and neuron as in Fig. 1. Hyperpolarization (start of record to first deflection of marker bar) and depolarization (second deflection of marker bar to right-hand end of trace) of the membrane potential with injection of current (-1.0 and + 0.5 nA, respectively) did not change the rate of occurrence of the intracellular bursts (0.1 Hz), which remained synchronous with

Fig. 2. Examples of kindled activities in CA213. (A) Same slice and neuron as in Fig. 1. (Aa) Chart recordings of extra- and intracellular spontaneous bursts (top and bottom charts, respectively) after test pulse stimulation was discontinued. Bursts occurred regularly at a frequency of 0.1 Hz. The bars under the charts indicate currents injected in the neuron. From left to right: -1.0 nA, 0 nA, and +0.5 nA. (Ab) Examples of intracellular bursts during current injections. From left to right: (1) during injection of-l.0 nA; (2) 0 nA; (3 and 4) f 0.5 nA. The fourth trace contrasts the slow and fast rising slopes of intrinsic and kindled depolarizing shifts. The dotted line across the figure indicates the RMP (-69 mV). (B) Intracellular data from a different experiment. (Ba) Chart recording showing spontaneous baseline spikes between responses evoked every 5 set (arrows). Spike occurred as single APs or as doublets. (Bb) details of evoked (left) and spontaneous (right) APs, showing the differences between the rising time and duration of the depolarizing potentials. RMP was -60 mV. Calibration: horizontal bars 100 set (Aa), 80 msec (Ab) and 10 msec (Bb); vertical bars 40 mV (Ab) and 20 mV (Bb). There are no vertical calibrations for the chart recordings showed at low speed in this (Aa) and following figures, due to the clipping of fast events by the lowfrequency response of the recorder.

the field potentials. Examples of intracellular bursts at the three different membrane potentials are shown in Fig. 2Ab. As expected, hyperpolarization increased the amplitude of the depolarization shift, while depolarization decreased it. The third and fourth traces from the left show activity induced by the depolarizing current, in addition to the kindled bursts. Spontaneous APs, rising abruptly from the baseline with no evidence of preceding depolarization, also occurred in 5 of the kindled slices. These baseline spikes were either single APs, doublets, or

111

triplets. Fig. 2Ba shows a chart recording of intracellular activity in a kindled slice where baseline spikes occurred between the evoked responses (arrows). The baseline bursts exhibited longer depolarizations than the evoked APs, and were not followed by afterhyperpolarizations. Examples of evoked (arrow) and spontaneous doublets are shown in Fig. 2Bb. In 14 of the kindling experiments mentioned above, test pulses were alternated with intracellular depolarizing pulses to study the intrinsic neuronal responses (data not shown). For a constant RMP, there were no changes in the depolarization and the number of APs induced by the intracellular pulses. Neurons that exhibited AHPs following the depolarizing steps continued to do so after kindling of bursts by SW. SWs failed to kindle ISs in 6 other neurons/slices, although bursts of action potentials and multiplepeak field potentials occurred immediately following the SWs. Late EPSPs also failed to develop in these slices. Intracellular stimulation with sinusoidal currents (2-10 s, 0.6-8 nA, p-p) induced depolarizations with bursts of 2-4 APs occurring at approximately 1 Hz through the stimulation (13 slices, 15 neurons, data not shown). There were no increases in depolarization and no sign of increased firing during successive sine-wave stimulations. There were no changes in subsequent neuronal responses to either extracellular test pulses or intracellular depolarizing steps of 100 msec. There were no signs of concurrent extracellular activity, whether or not the slices had been previously kindled by SW (n = 3). Role of NMDA in the kindling of interictal bursts

The effect of APV on the kindling of ISs was first studied in 7 neurons (6 slices). The kindling procedure was started in ACSF. As soon as a SW induced transient bursts or IEPSPs, kindling was interrupted. The slices were then perfused for at least 30 min with 50 or 100 ,uM APV before SW was restarted. APV reduced depolarization both during and after SW, by comparison with the responses observed previously in ACSF. Furthermore, repeated SWs failed to facilitate depolarization during and following stimulation. Fig. 3A and

Fig. 3. Effects of APVon the kindling of 1%. (A,B) SW (50yA, 2 set) was in CA2/3 ST.Intracellular recordings in CA3 during SWs before (A) and after (B) treatment with 50,~M APV show that the NMDA antagonist curtailed SW-induced depolarizations. The SW artifacts were reduced as in Fig. 1. The hatched bars indicate SWs. The APs were clipped in order to clearly show the depolarizations on an expanded scale. The inserts show responses to test pulses immediately after the SWs (asterisks). RMP was -66 mV. (C) The signals in A and B were superimposed after being processed with a low-pass, 10 Hz digital filter. (D) Examples of kindled extracellular spontaneous bursts before (arrow) and after perfusion with 50 PM APV in a different experiment, showing that APV did not always block kindled activity, although it considerably decreased burst duration. Calibration bars in C (2 s, 20 mV) also apply to A and B; inserts: 40 msec, 40 mV; D 20 msec, 1 mV.

B show ‘intracellular responses to identical SWs before and after 60 min of perfusion with APV. The inserts show responses to test pulses immediately following SW. The recordings were processed by a low-pass 10 Hz digital filter and superimposed in Fig. 3C to facilitate comparison. Evoked or spontaneous bursts could not be kindled in APV treated slices. In one of these experiments, the blockade of kindling by APV was reversed by subsequent perfusion with ACSF. When applied to previously kindled slices (n = 9), APV (50-200 PM) considerably reduced the amplitude and duration of the synchronized bursts, but did not block their occurrence. On the contrary, the frequency of spontaneous ISs increased during APV treatment. For example, the rate of occurrence of I%, measured in 3 slices during 15 min epochs before APV, and after 30 min in APV, increased from 0.09 + 0.001 to 0.15 + 0.01 Hz (P < 0.05). Examples of spontaneous kindled ISs in one of these slices, before (arrow) and after 60 min of perfusion with 100pM APV are shown in Fig. 3D.

112 A a.

B a. t-L1--

Fig. 4. Potentiation of SW effects by after-discharges. SW (50 PA, 2 set) was in CA1 stratum radiatum. The chart recordings show extra- and intracellular activities in CA3 (top and bottom traces, respectively) during and after SW1 (Aa) and SW2 (Ba). The extracellular recording was AC-coupled. The spikes on the charts in Ba are responses to test pulse stimulation. Examples of post-SW1 synchronized bursts (see arrows above top chart recording) are shown at faster speed in Ab. Details of neuronal activity at the beginning and end of SW2 are shown in Bb. See text for further discussion. RMP was -72 mV. Calibrations: the bar under the top trace represents 5 set in Aa and Bb, 50 msec in Bb; the bars in Ab represent 400 msec and 20 mV.

Kindling of interictal bursts by after-discharges

After-discharges (ADS) appeared to potentiate the effects of SW in 7 other neurons/slices. Fig. 4 illustrates an experiment where the kindling electrode was positioned in CA1 sr. Chart recordings of extracellular (top) and intracellular (bottom) activities are shown in Fig. 4Aa and Ba. SW1 induced only a small, decreasing depolarization of the CA3 neuron. High-frequency discharges, probably due to an electrographic seizure at the site of stimulation, were first registered by the extracellular electrode which was closer to CAl. The

discharges were followed by a series of large waves which coincided with a depolarization shift and associated bursting in the impaled CA2/3 neuron. Details of the extra- and intracellular activities occurring between the two arrows at the top of the chart recording are shown in Fig. 4Ab. The field potentials were processed with a 50 Hz lowpass digital filter to eliminate the artifactual pickup of APs due to coupling between the 2 recording electrodes. The next SW induced a much larger neuronal depolarization, but produced no after-discharges (Fig. 4Ba). Fig. 4Bb shows the beginning and end of SW2 at a faster chart speed. The initial rapid firing of single APs was replaced by bursts riding on small depolarizing waves during the later part of the kindling stimulation. Thereafter, single test pulses delivered either in CA1 or in CA2/3 apical dendrites elicited both extra- and intracellular ISs. In 3 slices of this series, ADS induced by SW1 were immediately followed by ISs, which persisted until the end of the experiment, without further kindling stimulation. In slices where SWs were continued after induction of ISs, ADS could only by induced after 2-5 successive SWs, thus after 10 to 25 min in our protocol. Spreading depression

Epochs of spreading depression (SD) occurred in 6 slices during the kindling procedure. In those slices, successive SWs of progressively higher intensity typically failed to induce epileptiform activity until SD was induced. As with ADS, SD was induced only by every second to fifth SW. Successive SWs between SD episodes induced only moderate depolarizations without after-effects. Comparable SDS, induced in hippocampal slices by repeated trains of pulses had been previously reported by others25. However, the maximum refractory period for SD was shorter (l-2 min) in those studies, which were mainly conducted in a static bath. SD and ADS could also be induced by SW in slices where spontaneous activity had previously been kindled. Fig. 5A shows extracellular (top) and intracellular (bottom) chart recordings of an episode of SD in CA2/3, following SW in the CA1 region. The intracellular recording shows high-fre-

113 DISCUSSION

B

Fig. 5. APV did not prevent SW-induced spreading depression. SWs (50 yA, 2 set) were in CAU3 sr, and are indicated by the arrows under the traces. (A) Chart recordings of extracellular (top) and intracellular (bottom) activities in CA3 before perfusion with APV. The gain of the extracellular recording was decreased in order to show the large negative potential during SD, and PSs appears very small on the chart. The neuron exhibited baseline spikes between the APs evoked by test pulses (see hyperpolarizations and concurrent extracellular responses before SW). (B) After 30 min perfusion with 50pM APV. The intracellular recording is from a different CA3 neuron, which exhibited fewer baseline spikes than the neuron shown in A, prior to perfusion with APV. RMPs were -64 mV in A and -66 mV in B. Horizontal calibration bar: 60 set (A,B). Peak extracellular negative shifts were -8 mV (A) and -7.8 mV (B). Membrane potentials at the peak of depolarization were -24 mV (A) and -23 mV (B).

quency spontaneous baseline spikes, together with responses evoked by test pulses at 0.2 Hz. Fig. 5B shows that perfusion of the slice with APV (50,&i, 30 min) did not prevent the triggering of SD by SW. At that time APV had already blocked IEPSPs and bursts. The intracellular recording is from a different neuron which exhibited much less frequent baseline spikes than the cell in Fig. 5A before APV treatment. Similar failure of APV to block SW-induced SDS was observed in the other 5 slices. In complementary studies3, we showed that APV also failed to block SW-induced SD in slices perfused with low-M2+ solutions. These results do not agree with those of Mody et al., who reported that APV blocked SD induced by tetani in similarly perfused hippocampal slice?. Whether or not SWs are more potent than trains of pulses in inducing SD remains to be studied.

Our results showed that the kindling of synchronized bursts in CA2/3 by sinusoidal electrical stimulation involved neuronal depolarizations sustained beyond SWs (after-depolarizations), potentiation of these depolarizations by repeated stimulation, and post-SW development of IEPSPs, leading to bursting. SW-induced ADS were also shown either to kindle ISs directly, or to potentiate responses to the next SW. In addition to bursts of APs riding on depolarizing shifts, spontaneous APs (singles, doublets and triplets) rising from the baseline with no preceding depolarization, also occurred in kindled slices (Fig. 2B). Similar ‘baseline spikes’ were previously reported by others, during the kindling of electrographic seizures by trains of pulses, and were assumed to be generated at ectopic sites, away from the soma *’. These spikes could trigger other events and contribute to the overall epileptiform activity in the tissue (see regulation of [K+], below). At this point, their precise role in epileptogenicity remains unknown. The kindling of ISs and baseline spikes occurred with no permanent changes in RMP or R,, as reported by others26. Furthermore, neuronal responses to intracellular depolarizing pulses or sinusoidal currents remained unchanged, suggesting that the kindling procedure did not lower the thresholds for action potentials. Our studies further showed that APV curtailed both the depolarization occurring during SW and the after-depolarization. The contribution of NMDA receptors to depolarization during highfrequency stimulation with trains of pulses had been previously demonstrated in CA1 neurons”. It is likely that in those and our experiments, additional entry of Na+ and Ca*+ via the NMDA channel was responsible for the larger depolarizations seen in the absence of APV (Fig. 3). Voltage-dependent inward Ca*+ currents were shown to underlie depolarizing afterpotentials in bursting hippocampal neurons3*. Similar Ca2+ currents, through NMDA and/or other channels, were probably responsible for the after-SW depolarizations. APV also inhibited the potentiation of depolarizations induced by successive SWs. This sug-

114

gests that NMDA-sensitive mechanisms activated during one SW were sustained for at least the length of the inter-stimulus interval (5 min), and were cumulative. The progressive increase in 1EPSPs and resulting late bursts (Fig. 1) suggests that recruitment of excitatory synaptic pathways continued after SW. Latent excitatory pathways between CA3 neurons, revealed after tetanization of the mossy fibers, have been described previously”. As in our experiments, the new connections developed slowly following tetanization, and were reinforced by successive tetani. These events were accompanied by a decrease in recurrent inhibition which could have facilitated synchronization of excitatory events in the CA2/3 area. Other studies31 have also demonstrated that long-lasting (30-60 set) repetitive stimulation at low frequency (3-10 Hz) reduced IPSP driving force and conductance in CA3. It is thus possible that SW and/or associated after-discharges also decreased inhibition in our experiments. Further studies involving local application of GABA, use of GABA agonists/antagonists, and intracellular studies of IPSPs will clarify the issue. It is of interest that the role of GABA inhibition in the kindling of epilepsy in vivo is still not understood23. We found that APV prevented the kindling of 1EPSPs and associated bursts. These findings concur with those of Anderson et al., who demonstrated that APV blocked the induction of synchronized bursts by trains of electrical stimuli in rat hippocampal slices’. Potentiation of APV-sensitive 1EPSPs was also demonstrated following high frequency stimulation in slices from rat frontal cortex28,29. Furthermore, all or none 1EPSPs were observed in slices from rat piriform cortex following bursting induced by perfusion with a low-Mg2+ solution l2 . Thus, the NMDA-dependent recruitment of 1EPSPs appears to be a general cortical phenomenon which probably plays an important role in epileptogenicity. SW-induced ADS and epochs of spreading depression were separated by similar lo-30 min intervals (2-5 SWs). Recently, the regulation of [K’10 was shown to be a highly critical factor for controlling both seizures and SD in the CA3 area”. It was prop osed that local rise in [K+lo

could trigger ‘positive feedback loops’ leading to epileptiform activity. This model could explain how SW-induced release of K+ led to the development of isolated synchronized bursts, ADS, or SD, depending: (1) on [K+], at the time of stimulation (ISs and/or persistent baseline spike activity in kindled slices could have contributed to rising [K+],); and (2) on the activity of the Na-K pump and other local K+ regulating mechanisms in CA213 at the time of SW. How long these mechanisms remain activated following a surge of potassium ions in the extracellular space could determine the interval between SW-induced ADS or SDS, and perhaps the minimum inter-stimulus interval for successful kindling in viva”. The results presented here differ from our previous findings2 in 2 ways. First, SWs were shown here to readily induce ADS and occasionally SD. These activities had not been observed before in SW-kindled slices. Second, perfusion with APV (50-200 PM) did not completely block kindled ISs, although their amplitude and duration were considerably reduced, as shown in Fig. 3D. The kindling procedures were identical in both series of experiments. The only obvious difference was that the slices were submerged in our earlier work, while they were kept in an interface chamber for this series of experiments. Electrographic seizures, as well as APV-resistant bursts, have been kindled in submerged slices by trains of pulses at much higher intensity than the SW used in our studies’,27. It could be that SW was less effective in our early experiments than in the present studies. It is possible that less current was applied to the submerged slices, due to leakage through the surrounding solution. In addition, slower washout of ions and neurotransmitters released by SW in nonsubmerged slices could have increased and prolonged the responses to SW. It is not known how larger and longer SW-induced depolarizations and ADS could have recruited APV-insensitive receptor systems. Ongoing work in our laboratory shows that APV-resistant, kindled ISs, and bursts induced by perfusion with NMDA, are blocked by 5 PM DNQX (6,7-dinitroquinoxaline-2,3-dione). These results suggest a role for kainate and/or quisqualate receptors (K/Q) in the initiation of ISS.

115

In summary, the results presented here suggest that the NMDA-sensitive level of cell membrane depolarization achieved during SW or ADS was critical in triggering secondary mechanisms responsible for the transition from normal to bursting activity in CA2l3. We speculate that Ca*’ influx through NMDA-controlled channels24 triggered a cascade of intracell~ar events, perhaps involving protein kinases and calmodulin, as seen in the induction of LTP in CA1’4,‘6*‘7.These intracellular events would activate a K/Q-dependent network of recurrent excitatory synapses” responsible for the triggering of ISs. The maximum rate of occurrence of ISs in CA3 was recently shown to be controlled by hippocampal cell ~terhype~olarization (AHP)‘. We propose a model where NMDA receptors would: (1) be required for the induction of 1% via the activation of a latent WQ

receptor system; and (2) participate in the control of the frequency of ISs by enhancing the K/Q induced depolarizations, associated Ca*+ influxes, and Ca2+-dependent AHPs. We plan to test and expand this model with further studies of in vitro kindling involving the regulation of GABAergic processes, [K’lo and AHP, K/Q and NMDA receptors, and second messenger systems.

REFERENCES

9 Goddard, G.V., McIntyre, D.C. and Leech, C.K., A permanent change in brain function resulting from daily electrical stimulation, Exp. Neural., 25 (1969) 295-330. 10 Haglund, M.M. and Schwartzkroin, P.A., Role of Na-K pump potassium regulation and IPSPs in seizures and spreading depression in immature rabbit hippocampal slices, J. Neurophysiol., 63 (1990) 225-239. 11 Herron, C.E., Lester, R.A.J., Coan, E.J. and Collingridge, G.L., Frequency-dependent involvement of NMDA receptors in the hippocampus: a novel synaptic mechanism,

1 Anderson, W.W., Swartzwelder, H.S. and Wilson, W.A., The NMDA receptor antagonist 2-amino&phosphonovalerate blocks stimulus train-induced epileptogenesis but not epileptiform bursting in the rat hippocampal slice, J. Neurophysiol., 57 (1987) 1-21. 2 Bawin, SM., Shahhal, I., Mahoney, M.D. and Adey, W.R., Induction of delayed synchronized bursts in hippocampal slices by weak sine-wave stimulation; role of the NMDA receptor, Epilepsy Res., 3 (1989) 41-48. 3 Bawin, S.M., Mahoney, M.D. and Adey, W.R., Intracellular correlates of the kindling of synchronized bursts in rat hippocampal slices, Neurosci. A&r., 1.5 (1989) 454. 4 Bawin, SM., Sheppard, A.R., Mahoney, M.D. and Adey, W.R., Influence of sinusoidal electrical fields on excitability in the rat hippocampal slice, Brain Res., 323 (1989) 227-237. 5 Chamberlin, N.L. and Dingledine, R., Control of epileptiform burst rate by CA3 hippocampal cell afterhyperpolarizations in vitro, Bruin Res., 492 (1989) 337-346. 6 Connors, B.C. and Gutnick, M.J., Cellular mechanisms of neocortical epileptogenesis in an acute experimental model. Tn P.A. Schwartzkroin and H. Wheal (Eds.), Electrophysiology of Epilepsy, Academic Press, 1984, pp. 79-105. 7 Engel, J., Jr. and Ackermann, R.F., Interictal EEG spikes eorretate with decreased, rather than increased, epileptogenicity in amygdaloid kindled rats, Bruin Res., 190 (1980) 543-548. 8 Fitz, J.G. and McNamara, J.O., Spontaneous interictal spiking in the awake kindled rat, Elec~oe~ce~halog. C&n. ~euro~hysio~., 47 (1979) 592-596.

ACKNOWLEDGEMENTS We thank Drs. J.W. Ashe, C.L. Cox and F.E. Dudek for their advice, support and helpful discussions during these studies. Mr. R. Jones provided skilled assistance in the design and development of the intracellular recording apparatus and computerized management of the data. This work was supported by the Veterans Administration.

Nature, 322 (1986) 265-268. 12 Hoffman, W.H. and Haberly, L.B., Bursting induces per-

sistent all-or none EPSPs by an NMDA-dependent process in piriform cortex, .I. Neurosci., 9 (1989) 206-215. 13 King, G.L., Dingledine, R., Giacchino, J.L. and McNamara, J.O., Abnormai neuronal excitability in hippocampal slices from kindled rats, J. Neurophysiol., 54 (1985) 129.5-1304. 14 Linden, D.J. and Routtenberg,

A., The role of protein kinase C in long-term potentiation: a testable model, Brain Res. Rev., 14 (1989) 279-296. 15 Lockard, J.S., Ojemann, G.A., Congdon, W.C. and DuCharme, L.L., Cerebellar stimulation in aiumina-gel monkey model: inverse relationship between clinical seizures and EEG interictal bursts, Epilepsia, 20 (1979) 223-234. 16 Malenka, R.C., Kauer, J.A., Perkel, D.J., Mauk, M.D., Kelly, P.T., Nicoil, R.A. and W~ham, M.N., An essential role for postsynaptic calmoduhn and protein kinase activity in long-term potentiation, Nature, 340 (1989) 554-557. 17 Malinow, R., Schulman, H. and Tsien, R.W., Inhibition of postsynaptic PKC or CaMKII blocks induction but not expression of LTP, Science, 245 (1989) 862-866.

116 18 McNamara, J.O., Bonhaus, D.W. and Shin, C., Mechanisms of kindling: a speculative hypothesis. In M.A. Dichter (Ed.), ~ec~un~s~ of Epileptogenesis. The Transition to Seizure, Plenum Press, 1988, pp~ 85-99. 19 Miles, R. and Wong, R.K.S., Latent synaptic pathways revealed after tetanic stimulation in the hippocampus, Nature, 329 (1987) 724-726. 20 Miles, R. and Wong, R.K.S., Excitatory synaptic interactions between CA3 neurones in the guinea-pig hippocampus. 1. ~kys~o~., 373 (1986) 397-418. 21 Mody. I., Lambert, J.D.C. and Heinemann, U., Low extracellular magnesium induces epileptiform activity and spreading depression in rat hippocampal slices, 1. Neuropkysiol., 57 (1987) 869-888. 22 Prince, D.A., Mechanisms of epileptogenesis in brain-slice model systems. In A.A. Ward, Jr., J.K. Penry and D. Purpura (Eds.), Epilepsy, Raven Press, New York, 1983, pp. 29-52. 23 Racine, J.R. and Burnham, W., The kindling model. In P.A. Schwartzkroin and H.V. Wheal (Eds.), Electropkysiology of Epilepsy, Academic Press, New York, 1984, pp. 153-171. 24 Regehr. W.G., Connor. J.A. and Tank, D.W., Optical imaging of calcium accumulation in hippocampal pyramidal cells during synaptic activation, Nature, 341 (1989) 533-536. 2.5 Snow, R. W., Taylor, C.P. and Dudek, F.E., Electrophysiological and optical changes in slices of rat hippocampus during spreading depression, J. Neuropkys~o~., 50 (1983) 561-572. 26 Stasheff, S.F, and Wilson, W.A., Increased ectopic action

potential generation accompanies epileptogenesis

in vitro,

Neurosci. Letf., ll(l990) 144-150. 27 Stasheff, S.F., Bragdon, A.C. and Wilson, W.A., Induction of epileptiform activity in hippocampal slices by trains of electrical stimuli, Bruin Res., 344 (1985) 2%-302. 28 Sutor, B. and Hablitz, J.J., EPSPs in rat neocortical neurons in vitro. I. Electrophysiological evidence for two distinct EPSPs, J. Neurophysiol. 61 (1989) 607-620. 29 Sutor, B. and Hablitz, J.J., EPSPs in rat neocortical neurons in vitro. II. Involvement of N-methyl+aspartate receptors in the generation of EPSPs, J. Neuropkys~oZ., 61 (1989) 621-634. 30 Swartzwelder,

H.S., Lewis, D.V., Anderson, W.W. and Wilson, W.A., Seizure-like events in brain slices: suppression by interictal activity, Brain Res., 410 (1987) 362-366. 31 Thompson, SM. and Gahwiler, B.H., Activity-dependent disinhibition. I. Repetitive stimulation reduces IPSP driving force and conductance in the hip~ampus in vitro, I. Neurophysiol.,

61(1989)

501-511.

32 Wong, R.K.S. and Prince, D.A., Afterpotential

generation in hippocampal pyramidal cells, J. Neurophysiol., 45 (1981)

86-97. 33 Wong, R.K.S., Traub, R.D. and Miles, R., Epileptogenic

mechanisms as revealed by studies of the hippocampal slice. In P.A. Schwartzkroin and H. Wheal (Eds.), Electrophysiology of Epilepsy, Academic Press, 1984, pp. 253-275. 34 Wong, R.K.S., Traub, R.D. and Miles, R., Cellular basis of neural synchrony. In A.V. Delgado-Escueda, A.A. Ward, Jr., D.M. Woodbury and R.J. Porter (Eds.). Advances in Neurology, Vol. 44, Raven Press, New York, 1986, pp. 583-592.