Address correspondence and reprint requests to Dr. M. Avoli at 3801 University Street, Montreal, QC, H3A 2B4 Canada. E-mail:firstname.lastname@example.org
Summary: Purpose: We determined how CA3-driven interictal discharges block ictal activity generated in the entorhinal cortex during bath application of 4-aminopyridine (4AP, 50 μM).
Methods: Field potential and [K+]o recordings were obtained from mouse combined hippocampus–entorhinal cortex slices maintained in vitro.
Results: 4AP induced N-methyl-d-aspartate (NMDA) receptor–dependent ictal discharges that originated in the entorhinal cortex, disappeared over time, but were reestablished by cutting the Schaffer collateral (n = 20) or by depressing CA3 network excitability with local application of glutamatergic receptor antagonists (n = 5). In addition, two types of interictal activity occurred throughout the experiment. The first type was CA3 driven and was abolished by a non-NMDA glutamatergic receptor antagonist. The second type was largely contributed by γ-aminobutyric acid type A (GABAA) receptor–mediated conductances and persisted during blockade of glutamatergic transmission. The absence of CA3-driven interictal discharges in the entorhinal cortex after Schaffer collateral cut facilitated the GABA-mediated interictal potentials that corresponded to large [K+]o elevations and played a role in ictal discharge initiation. Accordingly, ictal discharges along with GABA-mediated interictal potentials disappeared during GABAA-receptor blockade (n = 7) or activation of μ-opioid receptors that inhibit GABA release (n = 4).
Conclusions: Our findings suggest that CA3-driven interictal events restrain ictal discharge generation in the entorhinal cortex by modulating the size of interictal GABA-mediated potentials that lead to large [K+]o elevations capable of initiating ictal discharges in this structure.
Interictal activity originating in the CA3 area may prevent the occurrence of sustained epileptiform discharges, resembling electrographic seizures, in the entorhinal cortex (1–4). However, the mechanisms by which such a control is achieved are as yet unknown.
4-Aminopyridine (4AP)-induced epileptiform activity is accompanied by enhancement of both excitation and inhibition (5–7). Indeed, 4AP-induced ictal discharges in the rat entorhinal cortex are initiated by γ-aminobutyric acid (GABA)-mediated interictal potentials that result from the depolarizing action of GABA acting on type A receptors (8) and are associated with transient increases in [K+]o(9,10). Accordingly, both ictal discharges and GABA-mediated interictal potentials are abolished by inhibiting GABA release through μ-opioid receptor activation or by GABAA-receptor antagonists (8,11).
This evidence predicts that 4AP-induced interictal activity may block ictal activity by depressing the occurrence of GABA-mediated interictal potentials in the entorhinal cortex. To test this hypothesis, we used combined mouse hippocampus–entorhinal cortex slices treated with 4AP to obtain epileptiform activity that, over time, consisted only of interictal discharges. After cutting of the Schaffer collaterals, CA3-driven interictal events did not propagate to the entorhinal cortex where ictal activity was released (2,11). Here we show that after Schaffer collateral cut, GABA-mediated interictal potentials in the entorhinal cortex are less frequent and are associated with larger elevations in [K+]o. Our findings indicate that 4AP-induced CA3 interictal discharges may block the generation of ictal activity in the entorhinal cortex by altering the occurrence of GABA-mediated interictal potentials and their associated increases in [K+]o.
MATERIALS AND METHODS
Adult, male, CD-1 or Balb-C mice (25–35 g) were decapitated under halothane anesthesia (according with the guidelines established by the Canadian Council of Animal Care). The patterns of synchronous activity induced by 4AP in these two mouse strains were remarkably similar, and thus data were pooled together. Horizontal slices (550–600 μm) were prepared and maintained at 33.5 ± 0.5°C in an interface chamber, as previously reported (2). Slices included the entorhinal cortex and the hippocampus proper, comprising subiculum and dentate gyrus. The composition of the artificial cerebrospinal fluid (aCSF) was (in mM): NaCl 124, KCl 2, KH2PO4 1.25, MgSO4 2, CaCl2 2, NaHCO3 26, and glucose 10 (pH 7.4). 4AP, bicuculline methiodide (BMI), picrotoxin, P3-amino-propyl, P-diethoxymethyl phosphonic acid (CGP 35348), 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), 3,3-(2-carboxy-piperazin-4-yl)-propyl-1-phosphonate (CPP), and (d-Ala2-N-Me-Phe, Gly-ol)enkephalin (DAGO) were bath applied. CGP 35348 was a kind gift from Novartis (Basel, Switzerland). All other chemicals were acquired from Sigma (St. Louis, MO, U.S.A.) or Tocris Cookson (Langford, U.K.).
Field potential recordings were performed with aCSF-filled electrodes (tip diameter, ∼8 μm; resistance, 2–10 MΩ), whereas K+ electrodes were prepared according to the technique of Heinemann et al. (12). Details concerning the measurements of [K+]o can be found in our previous publications (9,10). Field-potential and ion-selective electrodes were usually positioned in layers IV–V of the medial entorhinal cortex, the granule cell layer of the dentate gyrus, and the CA3 or CA1 stratum radiatum. Cut of the Schaffer collateral pathway was accomplished under visual control with a razor blade mounted on a micromanipulator. In some experiments, kynurenic acid or CNQX was diluted in aCSF (5 or 0.5 mM, respectively) and were pressure-applied to the CA3 area through a broken pipette (tip diameter, ∼12 μm). To avoid diffusion of the drug from the perfused area to distant regions, hippocampus–entorhinal cortex slices were positioned in such a way that the treated area was downstream with respect to the aCSF flow. The spread of the drop-applied solution was visualized by adding phenol red to the perfused medium. Measurements in the text are expressed as mean ± SD and n represents the number of slices studied. Data were compared with Student's t test or the analysis of variance test and were considered significantly different if p < 0.05.
Characteristics of the 4AP-induced activity
Application of 4AP (50 μM) to mouse hippocampus–entorhinal cortex slices induced brief interictal events (duration, 200–350 ms; rate of occurrence, 0.7–1.2 Hz) originating in CA3 (arrows in Fig. 1A, 1h panel) and prolonged ictal discharges (duration, 30–160 s; continuous line in Fig. 1A, 1h panel) that were initiated in the entorhinal cortex (2,11). Both types of epileptiform discharges occurred in all areas of the slice. Ictal discharges were suppressed within ∼2 h of continuous 4AP application (Fig. 1A and C), whereas the CA3-driven interictal activity occurred throughout the experiment (n > 20 slices); however, ictal discharges could be revealed again by cutting the Schaffer collateral (n = 20), which blocked propagation of CA3 output activity to the entorhinal cortex (Fig. 1A).
These data confirm the evidence obtained in previous studies in which we reported that removing CA3 output activity makes ictal discharges reappear in the entorhinal cortex (2,11). To rule out the possibility that this phenomenon was due to nonspecific changes in excitability caused by the surgical cut, we used a different protocol. In five slices, we lesioned the Schaffer collaterals at the beginning of the experiment, and after having allowed a recovery of ∼1 h, we applied medium that contained 4AP. The spontaneous activity recorded in these experiments consisted of interictal discharges that were confined to the CA3 area, along with ictal discharges that occurred throughout the experiment (≤4 h) both in the entorhinal cortex and in the hippocampus proper (Fig. 1B). In addition, we observed a further type of slow interictal activity (duration, 0.8–2 s; interval of occurrence, 5–20 s) that appeared in all areas of the slice but did not have a fixed site of origin (9,10). These field potentials, which were presumably contributed by GABA receptor–mediated conductances, displayed negative polarity in the entorhinal cortex (Fig. 1B, asterisks in the 1h panel), and appeared to be associated with ictal-discharge onset (Fig. 1B arrow head; 9,10). The histogram in Fig. 1C shows the time course of the ictal discharge occurrence in intact slices (n = 5) and in those experiments (n = 4) in which the Schaffer collaterals were cut before starting the 4AP treatment.
Pharmacologic characteristics of the 4AP-induced activity
4AP-induced ictal discharges recorded initially in intact hippocampus–entorhinal cortex slices were readily abolished by the N-methyl-d-aspartate (NMDA)-receptor antagonist CPP (10 μM; n = 4; Fig. 2A, +CPP). Further addition of the non-NMDA glutamatergic receptor antagonist CNQX (10 μM; n = 4) blocked the CA3-driven interictal activity in all areas of the slice. However, slow, presumably GABA-mediated, field potentials continued to occur under these conditions (Fig. 2A, +CNQX, asterisk). These events were most often of negative polarity and displayed duration, rate of occurrence, and shape that were similar to the slow interictal activity recorded at times under control conditions (Fig. 1B) or during blockade of NMDA receptors (Fig. 2A, CPP panel; 9,10).
In five intact slices we also analyzed the effects induced by the NMDA-receptor antagonist CPP (10 μM) on the CA3-driven interictal activity recorded when ictal discharges had disappeared (Fig. 2B, +CPP). NMDA-receptor blockade in these experiments did not induce any significant change in the rate of occurrence of these interictal discharges, but it caused a 15–30% decrease of their amplitude (see inserts in Fig. 2B).
We also investigated the pharmacologic properties of the 4AP-induced epileptiform discharges after having uncovered ictal activity in the entorhinal cortex through Schaffer collateral cut (n = 4). As shown in Fig. 3 (control), ictal discharges were accompanied by [K+]o elevations ≤10–16 mM in the entorhinal cortex and 8–15 mM in the dentate gyrus from resting values of 3.2 mM(11). As observed in the intact slice preparation, the NMDA-receptor antagonist CPP (10 μM) blocked, in all areas, the ictal discharges along with the associated increases in [K+]o (Fig. 3, +CPP). By contrast, interictal discharges in CA3 continued to occur in the presence of CPP. These discharges were at times recorded in the dentate gyrus where they were presumably volume conducted (13). Moreover, CPP could uncover slow interictal discharges that were not noticeable under control conditions and were associated with elevations in [K+]o (≤5.5 mM) in dentate gyrus, in entorhinal cortex, and in CA3.
Additional application of the non-NMDA glutamatergic receptor antagonist CNQX (10 μM) abolished the CA3-driven interictal activity, whereas GABA-mediated interictal potentials and the associated elevations in [K+]o continued to occur (Fig. 3, +CNQX). Further activation of μ-opioid receptors by DAGO [10 μM; n = 4; a procedure that selectively decreases GABA release (14)] abolished these slow field potentials (Fig. 3, +DAGO). They also were blocked by bath application of the GABAA-receptor antagonists BMI (10 μM; n = 3) or picrotoxin [50 μM; n = 4; not illustrated, but see (9)]. This pharmacologic evidence indicates that the slow field potentials recorded during application of 4AP and glutamatergic receptor antagonists reflected the activation of GABA (mainly type A) receptors after the release of GABA from interneurons that fire synchronously (8–10,15,16).
The generation of ictal discharges, and presumably their initiation, depended on the occurrence of the GABA-mediated interictal potentials. To test this hypothesis we first applied DAGO (10 μM) to intact slices (n = 3) in which 4AP-induced ictal activity was still present. In these experiments, GABA-mediated interictal potentials were abolished, and ictal discharges were replaced by a pattern of recurrent, robust interictal discharges that occurred in all areas of the slice (Fig. 4A). Second, we obtained similar findings by testing the effects induced by the GABAA-receptor antagonists picrotoxin (50 μM; n = 3) or BMI (10 μM, n = 3) on the ictal discharges generated by slices in which the Schaffer collaterals had been cut (Fig. 4B). As illustrated in the inserts of Fig. 4B, picrotoxin (or BMI) abolished the initial, negative-going field potential that preceded the ictal discharge onset under control conditions. It also is possible to appreciate in this experiment that the ictal discharges recorded under control conditions as well as the recurrent interictal event induced by further addition of picrotoxin initiated in the entorhinal cortex.
CA3-driven interictal discharges modulate GABA-mediated interictal events and ictal activity
First, we studied the properties of the GABA-mediated interictal potentials recorded with field potential and [K+]o measurements in intact slices (after ictal discharges stopped occurring) and compared them with those in which the Schaffer collaterals had been cut (and thus ictal discharges were released). We found that in intact slices, after ∼2 h of 4AP incubation, CA3-driven interictal discharges were associated with peak elevations in [K+]o that attained values ≤4.0 ± 0.2 mM (n = 8) in entorhinal cortex and 4.1 ± 0.2 mM (n = 8) in the dentate gyrus from a baseline level of 3.2 mM(Fig. 5A). By using [K+]o recordings, we also could distinguish the GABA-mediated interictal potentials that were often best identified in the dentate gyrus (Fig. 5A, asterisks in the Control panel). With field-potential recordings, these events were not easily distinguished from the CA3-driven interictal discharges in the entorhinal cortex. Hence we measured the [K+]o elevations recorded in the entorhinal cortex at the same time as the dentate gyrus GABA-mediated interictal events. These discharges (a) occurred every 13.2 ± 6.2 s (n = 8); (b) lasted 0.5 ± 0.3 s in dentate gyrus and 0.7 ± 0.3 s in entorhinal cortex (n = 4); and (c) were associated with [K+]o increases to 4.4 ± 0.6 mM (n = 4) in the dentate gyrus and to 4.0 ± 0.2 mM (n = 8) in the entorhinal cortex from baseline levels of 3.2 mM (Fig. 5B–D).
We then examined these slices after Schaffer collateral cut, which uncovered: (a) prolonged ictal discharges and (b) slow, GABA-mediated interictal potentials that were seen either in isolation or before ictal events [Fig. 5A, asterisks in Schaffer collateral cut panel; (9)]. These GABA-mediated interictal potentials occurred every 28.5 ± 14.4 s (n = 8) and lasted 1.5 ± 1 s in dentate gyrus and 1.5 ± 0.5 s in entorhinal cortex. Under these conditions, the GABA-mediated potentials had corresponding increases in [K+]o≤4.6 ± 0.7 mM (n = 4) in the dentate gyrus and 4.3 ± 0.3 mM (n = 7) in the entorhinal cortex. However, when they preceded the ictal discharges, the elevations in [K+]o associated with the GABA-mediated potentials reached values of 4.8 ± 0.6 mM (n = 5) in the dentate gyrus and of 5.3 ± 0.6 mM (n = 8) in the entorhinal cortex (Fig. 5D).
We obtained similar results in five intact hippocampus–entorhinal cortex slices in which the CA3-driven interictal activity was depressed by local application of the glutamatergic antagonists kynurenic acid or CNQX to the CA3 subfield (Fig. 6A). In two of these experiments, we also measured the [K+]o in the entorhinal cortex and hippocampus. As illustrated in Fig. 6B (Control panel), 2 h of 4AP application induced a pattern of continuous CA3-driven interictal discharges occurring at a rate of 0.6 Hz; these events were associated with increases in [K+]o that had peak values of 4.2–4.9 mM in entorhinal cortex and 3.7–4.4 mM in the CA3 area from baseline levels of ∼3.2 mM. Drop application of CNQX depressed the CA3-driven interictal activity and uncovered slow, GABA-mediated interictal events along with ictal discharges (Fig. 6B, CNQX panel). It also is possible to appreciate in this experiment that the GABA-mediated interictal events recorded after the CNQX drop application were characterized by increases in [K+]o to peak values ranging from 5.0 to 6.9 mM in the entorhinal cortex and 4.1–5.0 mM in the CA3 area from a baseline level of 3.2 mM.
GABAB-receptor activation does not suppress ictal activity
In Schaffer collateral cut hippocampus–entorhinal cortex slices, we further analyzed the effects induced by repetitive electrical stimuli that were delivered at 0.5–1 Hz in the CA1-subicular area. We previously reported that this procedure mimics the CA3-driven interictal activation of entorhinal cortex networks that prevent the generation of ictal discharges in intact slices during prolonged (>1 h) application of 4AP-containing medium (2). As shown in Fig. 7A, we found that ictal discharges were abolished by this pattern of electrical stimulation in a reversible manner (n = 15). Moreover, the slow GABA-mediated interictal events decreased in rate of occurrence and in amplitude during the stimulation (Fig. 7A and B).
Activation of GABAB receptors leads to presynaptic inhibition of GABA and excitatory transmitter release (17,18) and causes a postsynaptic hyperpolarization that decreases the excitability of principal cells (19) and interneurons (20,21). In addition, application of baclofen decreases the rate of occurrence of 4AP-induced synchronous GABA-mediated potentials (22). Thus we tested the hypothesis that GABAB receptor–mediated mechanisms may participate in the block of 4AP-induced ictal discharges induced by repetitive electrical stimuli. Application of the GABAB-receptor antagonist CGP 35348 (1 mM) to Schaffer collateral cut slices did not modify the ictal activity induced by 4AP (n = 3). Under these conditions, rhythmic electrical stimulation of hippocampal output fibers at 0.5–1 Hz could still block ictal discharges, suggesting that CA3-driven interictal discharges did not control the occurrence of ictal activity in the entorhinal cortex through the activation of GABAB receptors.
The main findings of this study are that (a) slow, interictal discharges that are largely contributed by GABA-mediated conductances occur in mouse hippocampus–entorhinal cortex slices treated with 4AP; (b) the onset of ictal discharges, which are dependent on the function of NMDA receptors, is associated in entorhinal cortex with these GABA-mediated interictal events, presumably through the concomitant increases in [K+]o; (c) CA3-driven, non-NMDA glutamatergic receptor–dependent interictal activity modulates the interictal GABA-mediated potentials by increasing their rate of occurrence and by diminishing their strength, as measured in terms of associated increase in [K+]o; and (d) as a result, the [K+]o increases that are associated in entorhinal cortex with the GABA-mediated potentials in a “closed” hippocampus–entorhinal cortex circuit do not reach a level sufficient to induce ictal discharges (9).
Glutamatergic receptor subtypes and epileptiform activity in combined mouse hippocampus–entorhinal cortex slices
We have shown here that NMDA receptor–dependent conductances are instrumental for the generation of ictal discharges induced by bath application of 4AP in mouse hippocampus–entorhinal cortex slices. These findings are in agreement with those obtained in previous studies performed in similar slices from rat brain, in which ictal discharges were evoked by bath application of 4AP (8,9) or pilocarpine (13). Remarkably, these ictal discharges originate in the entorhinal cortex, a limbic structure in which prolonged NMDA receptor–mediated depolarizations occur (23,24). It has been shown that entorhinal cortex networks play a prominent role in the generation of seizure activity in patients with temporal lobe epilepsy (25,26).
In line with evidence obtained from rat hippocampus–entorhinal cortex slices (9) and isolated hippocampal slices (5,7,10), mouse CA3 networks respond to bath application of 4AP by generating a pattern of interictal discharge that (a) propagates to the entorhinal cortex via the Schaffer collateral–CA1–subiculum path (2), and (b) is readily abolished by the non-NMDA glutamatergic antagonist CNQX. We also noted here that this CA3-driven interictal activity continues to occur at similar rates during blockade of NMDA receptors (7,9,10,13), even though this pharmacologic procedure causes a decrease in the amplitude of the interictal events.
Evidence for the occurrence of interictal GABA-mediated potentials
In addition to ictal and CA3-driven interictal discharges, 4AP induced the appearance of slow, interictal potentials that occurred synchronously in all areas of the slice, but did not have a fixed site of origin, and were largely contributed by GABA receptor–mediated conductances. A similar synchronous activity has been reported to occur during 4AP application in combined entorhinal cortex–hippocampus slices from rats (9,27) as well as in isolated hippocampal slices from different rodent species (7,10,15,16). In line with these previous studies, we have confirmed here that these events continue to occur during application of glutamatergic antagonists. Indeed, by abolishing ionotropic excitatory transmission, we could clearly show that these events were caused by the activation of GABA (mainly of type A) receptors. Accordingly, they were abolished by pharmacologic procedures that activated μ-opioid receptors or antagonized GABAA receptors. These pharmacologic manipulations also blocked the slow GABA-mediated interictal events under control conditions (i.e., during application of 4AP only) and caused the appearance of recurrent epileptiform activity. A similar pattern of recurrent interictal discharge was recently reported by Bruckner et al. (28) in rat hippocampus–entorhinal cortex slices that were treated with 4AP and bicuculline.
Previous studies demonstrated that the interictal GABA-mediated events induced by 4AP are associated with transient [K+]o increases even when glutamatergic transmission is blocked (9,10,16,29) . It is believed that these GABA receptor–dependent elevations in [K+]o reflect depolarization of neurons and glia caused by the activation of GABAA receptors that may lead to outward counter/cotransport of K+ with Cl–/ HCO3+ anion shift (9,10,30). Moreover, it has been shown that during application of Mg2+-free medium, GABAergic conductances lead to the generation of ictal discharge (31). We have here confirmed these findings by recording field potential and [K+]o in entorhinal and hippocampal regions of the mouse slice. Interestingly, the GABA-mediated interictal potentials recorded in intact slices were not distinguished as clearly as in rat slices because of the ongoing CA3-driven interictal activity. However, even with frequent CA3-driven interictal activity, we could identify the GABA-mediated interictal potentials by their pronounced increases in [K+]o. Thus measuring [K+]o was critical in identifying this type of activity.
Dual control of ictal discharges by CA3-driven and GABA-mediated interictal activity
Schaffer collateral cut prevents interictal discharges from reaching the entorhinal cortex and uncovers the occurrence of robust GABA-mediated potentials associated with large elevations in [K+]o, thus contributing to ictal discharge initiation. Schaffer collateral cut modifies some properties of the GABA-mediated potentials. In particular, they occur at a lower frequency after Schaffer collateral cut, whereas their duration is longer and the associated [K+]o increases become larger in the entorhinal cortex, where ictal activity is initiated. The most pronounced increases in [K+]o are indeed observed for the GABA-mediated potentials preceding ictal discharges. Therefore uncovering GABA-mediated potentials and their [K+]o increases is required to initiate the ictal discharges in entorhinal cortex (9). In line with this view, preliminary experiments have demonstrated that during 4AP application, ictal discharges can be triggered by local microinjection of K+ in the entorhinal cortex (M. D'Antuono and M. Avoli, unpublished data).
A causative relation between GABA-mediated potentials and ictal discharges was originally identified in the rat entorhinal cortex (8,9). Moreover, the role played by GABA-mediated potentials in triggering ictal discharges has been associated with developmental changes in [K+]o homeostasis in the rat CA3 field (10). Here we have further shown that depression of GABA release as well as blockade of GABAA receptors leads to disappearance of ictal activity in the mouse combined slice preparation.
Mechanism of interictal-induced block of ictal discharge
Hyperpolarizing GABAA and GABAB conductances reduce some forms of NMDA receptor– dependent epileptiform activities (32) or excitatory postsynaptic potentials (EPSPs) (24). We have shown here that in 4AP-treated hippocampus–entorhinal cortex slices, the interictal-induced block of ictal activity does not involve activation GABAB conductances. Because GABAA receptor blockade depresses ictal discharges in slices where such events are recorded (8), we could not verify whether GABAA-receptor activation by interictal events affects ictal activity. However, because GABA-mediated potentials (mostly contributed by GABAA-receptor activation) are more frequent in intact hippocampus–entorhinal cortex slices (in which ictal discharges are depressed by the interictal activity), it is unlikely that CA3-driven interictal discharges control ictal discharge generation in entorhinal cortex through a GABAA-receptor–mediated mechanism.
4AP-induced GABA-mediated potentials occurring at frequencies resembling CA3-driven interictal discharges can occur in the entorhinal cortex (26). These reverberant potentials consist of short-lasting, Cl–-dependent depolarizations and depend on the recruitment of local, non-NMDA glutamatergic receptor–dependent circuits. Staley and Mody (33) suggested that GABAA depolarizing conductances may exert an inhibitory effect on NMDA-mediated responses by shunting this type of excitation input. It is tempting to speculate that 4AP-induced GABA-mediated potentials that initiate NMDA-dependent ictal discharges are masked through a shunting effect produced by reverberating GABA potentials (26), whose occurrence is entrained by hippocampal-driven interictal activity. Thus the interictal control of ictal discharges may involve activation of non–NMDA-dependent circuits in the entorhinal cortex, whereby non-NMDA glutamatergic receptors become functional at the postsynaptic membrane of previously silent (only NMDA-containing) synapses (34).
Acknowledgment: This study was supported by the Canadian Institute of Health Research (grant MT-8109) and the Savoy Foundation. M.B. was the recipient of an FRSQ studentship; M.D. is a Fellow of the Fragile X Research Foundation of Canada. We thank Toula Papadopoulos for secretarial assistance.