Effects of anticonvulsant drugs on hippocampal neurons

Region-specific differences and areal interactions underlying transitions in epileptiform activity

Key points

Local neocortical and hippocampal territories show different and sterotypical patterns of acutely evolving, epileptiform activity.

Neocortical and entorhinal networks show tonic–clonic‐like events, but the main hippocampal territories do not, unless it is relayed from the other areas.

Transitions in the pattern of locally recorded epileptiform activity can be indicative of a shift in the source of pathological activity, and may spread through both synaptic and non‐synaptic means.

Hippocampal epileptiform activity is promoted by 4‐aminopyridine and inhibited by GABA_B receptor agonists, and appears far more sensitive to these drugs than neocortical activity.

These signature features of local epileptiform activity can provide useful insight into the primary source of ictal activity, aiding both experimental and clinical investigation.

Abstract

Understanding the nature of epileptic state transitions remains a major goal for epilepsy research. Simple in vitro models offer unique experimental opportunities that we exploit to show that such transitions can arise from shifts in the ictal source of the activity. These transitions reflect the fact that cortical territories differ both in the type of epileptiform activity they can sustain and in their susceptibility to drug manipulation. In the zero‐Mg²⁺ model, the earliest epileptiform activity is restricted to neocortical and entorhinal networks. Hippocampal bursting only starts much later, and triggers a marked transition in neo‐/entorhinal cortical activity. Thereafter, the hippocampal activity acts as a pacemaker, entraining the other territories to their discharge pattern. This entrainment persists following transection of the major axonal pathways between hippocampus and cortex, indicating that it can be mediated through a non‐synaptic route. Neuronal discharges are associated with large rises in extracellular [K⁺], but we show that these are very localized, and therefore are not the means of entraining distant cortical areas. We conclude instead that the entrainment occurs through weak field effects distant from the pacemaker, but which are highly effective at recruiting other brain territories that are already hyperexcitable. The hippocampal epileptiform activity appears unusually susceptible to drugs that impact on K⁺ conductances. These findings demonstrate that the local circuitry gives rise to stereotypical epileptic activity patterns, but these are also influenced by both synaptic and non‐synaptic long‐range effects. Our results have important implications for our understanding of epileptic propagation and anti‐epileptic drug action.

Local neocortical and hippocampal territories show different and sterotypical patterns of acutely evolving, epileptiform activity.

Neocortical and entorhinal networks show tonic–clonic‐like events, but the main hippocampal territories do not, unless it is relayed from the other areas.

Transitions in the pattern of locally recorded epileptiform activity can be indicative of a shift in the source of pathological activity, and may spread through both synaptic and non‐synaptic means.

Hippocampal epileptiform activity is promoted by 4‐aminopyridine and inhibited by GABA_B receptor agonists, and appears far more sensitive to these drugs than neocortical activity.

These signature features of local epileptiform activity can provide useful insight into the primary source of ictal activity, aiding both experimental and clinical investigation.

Multi-electrode array recordings of human epileptic postoperative cortical tissue

Epilepsy, affecting about 1% of the population, comprises a group of neurological disorders characterized by the periodic occurrence of seizures, which disrupt normal brain function. Despite treatment with currently available antiepileptic drugs targeting neuronal functions, one third of patients with epilepsy are pharmacoresistant. In this condition, surgical resection of the brain area generating seizures remains the only alternative treatment. Studying human epileptic tissues has contributed to understand new epileptogenic mechanisms during the last 10 years. Indeed, these tissues generate spontaneous interictal epileptic discharges as well as pharmacologically-induced ictal events which can be recorded with classical electrophysiology techniques. Remarkably, multi-electrode arrays (MEAs), which are microfabricated devices embedding an array of spatially arranged microelectrodes, provide the unique opportunity to simultaneously stimulate and record field potentials, as well as action potentials of multiple neurons from different areas of the tissue. Thus MEAs recordings offer an excellent approach to study the spatio-temporal patterns of spontaneous interictal and evoked seizure-like events and the mechanisms underlying seizure onset and propagation. Here we describe how to prepare human cortical slices from surgically resected tissue and to record with MEAs interictal and ictal-like events ex vivo.

Standard antiepileptic drugs fail to block epileptiform activity in rat organotypic hippocampal slice cultures

BACKGROUND AND PURPOSE

Earlier studies had demonstrated that tonic-clonic seizure-like events (SLEs) resembling electrographic correlates of limbic seizures in animals and humans can be induced in organotypic hippocampal slice cultures (OHSCs). We have explored OHSCs for their suitability to serve as in vitro models of limbic seizures for studying seizure mechanisms and screening new antiepileptic compounds.

EXPERIMENTAL APPROACH

OHSCs were cultivated according to the interface method. Neuronal activity and extracellular potassium concentration were recorded under submerged conditions. SLEs were induced by lowering magnesium concentration or by applying the potassium channel blocker 4-aminopyridine. The effects of standard antiepileptic drugs (AEDs), carbamazepine, phenytoin, valproic acid, clonazepam, diazepam and phenobarbital sodium on SLEs were analysed.

KEY RESULTS

In more than 93% of OHSCs, AEDs did not prevent the induction of SLEs or stop ongoing seizure activity even when toxic concentrations were applied. This pharmacoresistance was independent of the method of seizure provocation, postnatal age at explantation (P2-P10) and cultivation time in vitro (2 months). SLEs were reversibly blocked by glutamate antagonists or the GABA(A)-agonist muscimol.

CONCLUSIONS AND IMPLICATIONS

We present a simple to establish in vitro model of tonic-clonic SLEs that is a priori pharmacoresistant and thus has an advantage over animal models of pharmacoresistant seizures in which responders and non-responders can be sorted out only after an experiment. OHSCs could be suitable for exploring mechanisms of pharmacoresistant seizures and be used for the identification of new anticonvulsive compounds eventually effective in drug refractory epilepsy.

Methodological approaches to exploring epileptic disorders in the human brain in vitro

Brain surgery, and in particular epilepsy surgery, offers the unique opportunity to study viable human central nervous tissue in vitro. This does not only open a window to address the basic mechanisms underlying human disease, such as epilepsy, but it allows to venture into investigating neurophysiological functions per se. In the present paper, we describe the most commonly used methods in the electrophysiological (and, at least to some extent, also histochemical and molecular) analysis of human tissue in vitro. In addition, we consider the pitfalls and limitations of such studies, in particular regarding the issue of tissue sampling procedures and control experiments.

Effects of anticonvulsants on soman-induced epileptiform activity in the guinea-pig in vitro hippocampus

Seizures arising from acetylcholinesterase inhibition are a feature of organophosphate anticholinesterase intoxication. Although benzodiazepines are effective against these seizures, alternative anticonvulsant drugs may possess greater efficacy and fewer side-effects. We have investigated in the guinea-pig hippocampal slice preparation the ability of a series of anticonvulsants to suppress epileptiform bursting induced by the irreversible organophosphate anticholinesterase, soman (100 nM). Carbamazepine (300 microM), phenytoin (100 microM), topiramate (100-300 microM) and retigabine (1-30 microM) reduced the frequency of bursting but only carbamazepine and phenytoin induced a concurrent reduction in burst duration. Felbamate (100-500 microM) and clomethiazole (100-300 microM) had no effect on burst frequency but decreased burst duration. Clozapine (3-30 microM) reduced the frequency but did not influence burst duration. Levetiracetam (100-300 microM) and gabapentin (100-300 microM) were without effect. These data suggest that several compounds, in particular clomethiazole, clozapine, felbamate, topiramate and retigabine, merit further evaluation as possible treatments for organophosphate poisoning.

Effect of levetiracetam on epileptiform discharges in human neocortical slices

PURPOSE

The anticonvulsant effects of the novel antiepileptic drug (AED) levetiracetam (LEV) were tested in neocortical slice preparations from 23 patients who underwent surgery for the treatment of refractory epilepsy.

METHODS

Slices were used to evaluate the effects of LEV on two different models of epilepsy: low-Mg2+-induced untriggered and bicuculline-evoked stimulus-triggered epileptiform burst discharges and spontaneously appearing rhythmic sharp waves.

RESULTS

LEV (0.1-1 mM) did not influence spontaneously appearing rhythmic sharp waves or Mg2+-free aCSF-induced epileptiform field potentials. LEV affected neither the amplitudes or duration nor the repetition rates of burst discharges in these epilepsy models. However, LEV (100-500 microM) significantly suppressed the ictal-like discharges elicited by the gamma-aminobutyric acid subtype A (GABAA)-receptor antagonist bicuculline. A marked reduction of the amplitude and duration of bicuculline-evoked field response in the presence of LEV was observed.

CONCLUSIONS

The results indicate the potential for LEV to inhibit epileptiform burst discharges in human neocortical tissue, which is consistent with its effects in animal models of epilepsy. These results also support the seizure reduction observed in clinical trials and support that this may, in part, be related to the ability of LEV to inhibit epileptiform discharges.

Vigabatrin reduces epileptiform activity in brain slices from pharmacoresistant epilepsy patients

Human neocortical temporal lobe tissue resected for treatment of pharmacoresistant epilepsy was investigated. In slices prepared from this tissue, epileptiform field potentials (EFP) were induced by omission of magnesium from the artificial cerebrospinal fluid (ACSF). The effects of the gamma-aminobutyric acid transaminase inhibitor vigabatrin on EFP were tested. Vigabatrin exerted a dose-dependent reduction of the repetition rate of EFP: after 3 h of administration of vigabatrin in concentrations of 100 and 200 micromol/l, the repetition rate of EFP was reduced to 35% and 18% of the initial values, respectively. This effect was not reversible. In control experiments with neocortical slices from rats, vigabatrin reduced EFP in a comparable range. The results demonstrate a strong antiepileptic effect of vigabatrin on EFP in tissues from pharmacoresistant epilepsy patients.

The anticonvulsant retigabine potently suppresses epileptiform discharges in the low Ca ++ and low Mg++ model in the hippocampal slice preparation

Retigabine (N-(2-amino-4-(4-fluorobenzylamino)-phenyl) carbamic acid ethyl ester, D-23129) is a broad acting anticonvulsant currently undergoing phase II clinical trials. An opening effect on leakage conductance K+ channels, potentiation of GABA induced currents and a weak blocking effect on Na+ and Ca++ channels were previously reported. The goal of this study was to investigate whether retigabine is capable of blocking epileptiform discharges in the low Ca++ and low Mg++ model in the hippocampal slice preparations and whether the anti-burst activity can be related to the K+ channel opening effect. In the low Ca++ model, synaptic transmission is blocked and discharges evolve from ephaptically-coupled neurons. Compounds which directly interfere with the threshold for action potential induction via alteration of ion channel function (i.e. Na+ channel blocker) may alter the discharges, while compounds interfering with synaptic transmission are not active. Retigabine suppressed the discharges in a concentration-dependent manner. A significant reduction in frequency without effect on amplitude was observed after application of 1 microM, and a full block of all discharges after application of 25 microM. The opener of the ATP sensitive K+ channels cromakalim was also active. Application of 300 microM cromakalim yielded to a lower frequency with no effects on the amplitude of discharges. Treatment with phenytoin and carbamazepine resulted in a marked reduction in amplitude accompanied by a rise in frequency; only at higher concentrations was a full block observed. The effect of retigabine therefore differs from sodium channel blockers and can be related to the K+ channel opening effect. In the low Mg++ model, excitatory neurotransmission is augmented by reducing the Mg++ block of NMDA channels. This results in development of interictal-like epileptiform activity in area CA1 in isolated hippocampal slices. Treatment with retigabine 10 microM resulted in a significant reduction of the discharges, and discharges were fully blocked after application of 25 microM. Qualitatively similar effects were observed with cromakalim and valproate, albeit at higher concentrations. The data indicate that retigabine exerts potent broad spectrum activity making it an interesting candidate for treatment of drug resistant patients.

Gabapentin potentiation of the antiepileptic efficacy of vigabatrin in an in vitro model of epilepsy

1. An enhancement of promoted release of gamma-aminobutyric acid (GABA) and a change in GABA-metabolism have been suggested as mechanisms of action of gabapentin. Vigabatrin is supposed to act mainly via inhibition of GABA-transaminase but it also interferes with GABA-release and GABA-uptake. On the basis of these mechanisms of action, a pharmacodynamic interaction of the two antiepileptic drugs could be supposed which might be of relevance in the sense of a rational polypharmacy. 2. To address the aforementioned hypothesis, experiments were carried out on hippocampal slices (n=107) of guinea-pigs (n=70). Epileptiform field potentials (e.f.p.) were induced by omission of magnesium from the bath solution and recorded in the stratum pyramidale of the CA3 region. Gabapentin (30-600 microM; 5.1-102.72 microg ml(-1)), vigabatrin (50-200 microM, 6.45-25.8 microg ml(-1)) and the GABA(A)-receptor antagonist bicuculline (100 microM) were added to the bath solution for 3 h. 3. Gabapentin, in concentrations up to 600 microM, failed to decrease the repetition rate or duration of e.f.p. (n=19). However, vigabatrin, evoked a dose-dependent reduction of the repetition rate of e.f.p. For a concentration of 100 microM (12.9 microg ml(-1)) there was a reduction down to 48+/-5% (mean+/-s.e.mean) of the initial value within 3 h (n=11). With simultaneous administration of vigabatrin (100 microM) and gabapentin (60 microM) for 3 h (n=15), the repetition rate of e.f.p. decreased down to 8+/-3%, which is significantly different from the values obtained after administration of 100 microM vigabatrin alone (P<0.0001). Both, the antiepileptic effect of vigabatrin alone and the enhancement by gabapentin were blocked by the GABA(A)-receptor antagonist bicuculline (100 microM, n=16). 4. These results demonstrate that gabapentin is able to augment the antiepileptic effects of vigabatrin significantly. It is possible that a change in the GABA-release machinery is induced by vigabatrin which then can be augmented by gabapentin.

Comparison of the effects of serotonin in the hippocampus and the entorhinal cortex

Among the molecular, cellular, and systemic events that have been proposed to modulate the function of the hippocampus and the entorhinal cortex (EC), one of the most frequently cited possibilities is the activation of the serotonergic system. Neurons in the hippocampus and in the EC receive a strong serotonergic projection from the raphe nuclei and express serotonin (5-HT) receptors at high density. Here we review the various effects of 5-HT on intrinsic and synaptic properties of neurons in the hippocampus and the EC. Although similar membrane-potential changes following 5-HT application have been reported for neurons of the entorhinal cortex and the hippocampus, the effects of serotonin on synaptic transmission are contrary in both areas. Serotonin mainly depresses fast and slow inhibition of the principal output cells of the hippocampus, whereas it selectively suppresses the excitation in the entorhinal cortex. On the basis of these data, we discuss the possible role of serotonin under physiological and pathophysiological circumstances.

Serotonin blocks different patterns of low Mg2+-induced epileptiform activity in rat entorhinal cortex, but not hippocampus

Low Mg2+-induced epileptiform activity in the entorhinal cortex is characterized by an initial expression of seizure-like events followed by late recurrent discharges. Both these forms of activity as well as the transition between them were blocked by serotonin. In contrast, serotonin had little effect upon the epileptiform activity in areas CA3 and CA1 of the hippocampus. Both forms of epileptiform activity in the entorhinal cortex are sensitive to N-methyl-D-aspartate receptor antagonists and it is shown here that serotonin blocked both types of epileptiform activity through an effective concentration-dependent reduction of N-methyl-D-aspartate receptor-mediated excitatory postsynaptic potentials in deep layer entorhinal cortex cells. Serotonin also prolonged or even prevented the transition between the two types of epileptiform activity and we suggest that this may be through activation of the Na+/K+-ATPase. The resistance of epileptiform activity in CA1 and CA3 to serotonin was most likely related to the inability of serotonin to reduce Schaffer collateral-evoked excitatory postsynaptic potentials. Given the strong serotonergic inputs to both the hippocampus and entorhinal cortex, the differential sensitivity of the two regions to serotonin suggests functional differences. In addition since the late recurrent discharges in the entorhinal cortex are resistant to all clinically used anticonvulsants, serotonin may open new avenues for the development of novel anticonvulsant compounds.

Effects of gamma-aminobutyric acid (GABA) agonists and GABA uptake inhibitors on pharmacosensitive and pharmacoresistant epileptiform activity in vitro

1. Lowering of the extracellular Mg(2+)-concentration induces various patterns of epileptiform activity in combined rat entorhinal cortex-hippocampal brain slices. After a prolonged period of exposure to Mg(2+)-free medium seizure-like events in the entorhinal cortex change to a state of late recurrent discharges which cannot be blocked by clinically available antiepileptic drugs. This late epileptiform activity thus represents a useful model to test the effects of new anticonvulsant substances. 2. A mechanism possibly underlying the development of sustained seizure-like activity is the loss of synaptically released gamma-aminobutyric acid (GABA). Drugs which increase the amount of GABA available in presynaptic endings might thus be useful in the treatment of these therapeutically complicated forms of epilepsy. 3. Therefore, we studied the effects of various substances increasing GABA-mediated inhibition on early and late forms of epileptiform activity. GABA and the GABAA receptor agonist muscimol blocked both the pharmacosensitive discharges in the hippocampus and entorhinal cortex as well as the late recurrent discharges in the medial entorhinal cortex. The GABAB receptor agonist baclofen blocked the recurrent short discharges very potently, but did not consistently block seizure-like events and late recurrent discharges in the entorhinal cortex. 4. GABA uptake blockers showed a differential potency to block the various discharge patterns. Whereas nipecotic acid and beta-alanine suppressed all forms of epileptiform activity albeit at high concentrations (1-5 mM), tiagabine was much more potent in blocking the hippocampal recurrent short discharges and the seizure-like events in the medial entorhinal cortex, but could not block the late recurrent discharges. 5. Our data support the idea that prolonged neuronal overactivity might result in a loss of synaptically available GABA. Selective block of uptake into glia cells or substitution of the transmitter may therefore be an efficient strategy for the treatment of severe prolonged epileptic discharges whereas block of neuronal GABA uptake fails to counteract synchronized discharges in this situation.