Anticonvulsant drug effects on spontaneous thalamocortical rhythms in vitro: phenytoin, carbamazepine, and phenobarbital

Validated animal models for antiseizure drug (ASD) discovery: Advantages and potential pitfalls in ASD screening

Since 1993, over 20 new anti-seizure drugs (ASDs) have been identified in well-established animal seizure and epilepsy models and subsequently demonstrated to be clinically effective in double-blinded, placebo-controlled clinical trials in patients with focal onset seizures. All clinically-available ASDs on the market today are effective in at least one of only three preclinical seizure and epilepsy models: the acute maximal electroshock (MES), the acute subcutaneous pentylenetetrazol (scPTZ) test, or the kindled rodent with chronic evoked seizures. Thus, it reasons that preclinical ASD discovery does not need significant revision to successfully identify ASDs for the symptomatic treatment of epilepsy. Unfortunately, a significant need still persists for more efficacious and better tolerated ASDs. This is particularly true for those patients whose seizures remain drug resistant. This review will focus on the continued utility of the acute MES and scPTZ tests, as well as the kindled rodent for current and future ASD discovery. These are the only "clinically validated" rodent models to date and been heavily used in the search for novel and more efficacious ASDs. This is to say that promising ASDs have been brought to the clinic on the basis of efficacy in these particular seizure and epilepsy models alone. This review also discusses some of the inherent advantages and limitations of these models relative to existing and emerging preclinical models. It then offers insight into future efforts to develop a preclinical model that will advance a truly transformative therapy for the symptomatic treatment of difficult to treat focal onset epilepsy. This article is part of the special issue entitled 'New Epilepsy Therapies for the 21st Century - From Antiseizure Drugs to Prevention, Modification and Cure of Epilepsy'.

Epidural Electrotherapy for Epilepsy

Penetrating electronics have been used for treating epilepsy, yet their therapeutic effects are debated largely due to the lack of a large-scale, real-time, and safe recording/stimulation. Here, the proposed technology integrates ultrathin epidural electronics into an electrocorticography array, therein simultaneously sampling brain signals in a large area for diagnostic purposes and delivering electrical pulses for treatment. The system is empirically tested to record the ictal-like activities of the thalamocortical network in vitro and in vivo using the epidural electronics. Also, it is newly demonstrated that the electronics selectively diminish epileptiform activities, but not normal signal transduction, in live animals. It is proposed that this technology heralds a new generation of diagnostic and therapeutic brain-machine interfaces. Such an electronic system can be applicable for several brain diseases such as tinnitus, Parkinson's disease, Huntington's disease, depression, and schizophrenia.

Differences in cortical versus subcortical GABAergic signaling: a candidate mechanism of electroclinical uncoupling of neonatal seizures

Electroclinical uncoupling of neonatal seizures refers to electrographic seizure activity that is not clinically manifest. Uncoupling increases after treatment with Phenobarbital, which enhances the GABA(A) receptor (GABA(A)R) conductance. The effects of GABA(A)R activation depend on the intracellular Cl(-) concentration ([Cl(-)](i)) that is determined by the inward Cl(-) transporter NKCC1 and the outward Cl(-) transporter KCC2. Differential maturation of Cl(-) transport observed in cortical versus subcortical regions should alter the efficacy of GABA-mediated inhibition. In perinatal rat pups, most thalamic neurons maintained low [Cl(-)](i) and were inhibited by GABA. Phenobarbital suppressed thalamic seizure activity. Most neocortical neurons maintained higher [Cl(-)](i), and were excited by GABA(A)R activation. Phenobarbital had insignificant anticonvulsant responses in the neocortex until NKCC1 was blocked. Regional differences in the ontogeny of Cl(-) transport may thus explain why seizure activity in the cortex is not suppressed by anticonvulsants that block the transmission of seizure activity through subcortical networks.

Valproate decreases EEG synchronization in a use-dependent manner in idiopathic generalized epilepsy

INTRODUCTION

In order to explore the mechanism of action of valproate (VPA) in idiopathic generalized epilepsy (IGE), the effect of VPA on cortical EEG activity was investigated.

HYPOTHESIS

VPA decreases EEG synchronization in the delta and theta frequency bands in a use-dependent manner in IGE patients.

METHODS

First setting: EEG records of 17 untreated IGE patients (NAE group) were analyzed and compared to those of 15 healthy controls (NC group). Second setting: EEG recorded in the untreated condition (NAE) was compared to the EEG recorded in the treated condition (VPA) of the patient group. Technique and analysis: 2 min of eyes-closed, waking EEG background activity (without epileptiform potentials and artifacts) were analyzed. Absolute power (AP) and mean frequency (MF) were computed for 19 electrodes and four frequency bands (delta=1.5-3.5 Hz, theta=3.5-7.5 Hz, alpha=7.5-12.5 Hz, beta=12.5-25.0 Hz). Log-transformed data entered further analysis. Group differences were computed by means of parametric statistics including correction for multiple comparisons. The VPA-related changes (APvpa-APnae) were correlated with the degree of the baseline abnormality (APnae) and the daily dose/serum levels of VPA.

MAIN RESULTS

Statistically significant (p<0.05, corrected) changes in the first setting: diffuse delta, theta, alpha AP increase, mainly right hemispheric beta AP increase was found in the NAE group, as compared to the NC group. Second setting: VPA decreased delta and theta AP. Strong correlation was demonstrated between the degree of the initial AP abnormality and the VPA-related AP decrease. AP decrease did not correlate with the daily dose and the serum level of the drug.

CONCLUSION

The hypothesis that VPA decreased EEG synchronization in the delta and theta frequency bands in a use-dependent manner was supported. The findings contribute to the understanding of the action of VPA at the network level.

Lamotrigine decreases EEG synchronization in a use-dependent manner in patients with idiopathic generalized epilepsy

OBJECTIVE

To investigate the quantitative EEG effects of lamotrigine (LTG) monotherapy.

HYPOTHESIS

LTG was predicted to decrease thalamo-cortical neuronal synchronization in idiopathic generalized epilepsy (IGE).

METHODS

Waking EEG background activity of 19 IGE patients was investigated before treatment and in the course of LTG monotherapy. Raw absolute power (RAP), raw percent power (RRP), and raw mean frequency (RMF) were computed for 19 electrodes and four frequency bands (delta=1.5-3.5Hz, theta=3.5-7.5Hz, alpha=7.5-12.5Hz, and beta=12.5-25.0Hz). Inter- and intrahemispheric coherence was computed for eight electrode pairs and the four frequency bands. In addition, scalp-averages were calculated for each variable. Group differences were computed by means of nonparametric statistics including correction for multiple comparisons.

RESULTS

Main results were decreased delta and theta RAP (p<0.05 for scalp-averages). LTG compressed the delta, theta, and alpha RAP datasets, reducing the upper limit of the scatter in particular. Spearman r-values indicated marked correlation between the starting values (RAPuntreated) and the LTG-related decrease (RAPtreated-RAPuntreated) in three bands: delta (r=-0.72; p=0.0005), theta (r=-0.59; p=0.007), and alpha (r=-0.61; p=0.006). Thus, the greater the baseline neuronal synchronization, the marked the dampening effect of LTG on it. The remaining findings were decreased theta RRP, theta RMF, and increased alpha RMF (p<0.05 for scalp-averages). The electrode-related changes were small but topographically consistent across the 19 electrode sites. LTG did not affect coherence.

CONCLUSIONS

1. LTG partially normalized the spectral composition of EEG background activity. LTG decreased pathological thalamo-cortical synchronization in use-dependent manner. 2. LTG did not cause quantitative EEG alterations suggesting worsening of the physiological brain functions. Instead, its profile suggested a mild psychostimulant effect.

SIGNIFICANCE

The results contribute to the understanding of the effect of LTG at the network level.

Brain malformations, epilepsy, and infantile spasms

The models of cortical dysplasia discussed earlier--the Lis1 knockout, the MAM-induced cobblestone LIS, the spontaneous tish mutant, and focal freeze injury-induced PMG--illustrate several important insights into epileptogenesis in malformed brain. First, the appearance of epilepsy varies according to the pathogenesis of the dysplasia and may well depend more on the intrinsic properties of the neurons in these models rather than on the disturbed position of the cells. This is supported by models such as the reeler mouse, in which the dysfunctional extracellular matrix molecule leads to a form of lissencephaly in mouse and human, but there is a far less impressive association with seizures than for LIS1 mutations. However, Lis1 and Dex mutations that appear to affect the cytoskeleton and perhaps intracellular protein trafficking are frequently associated with infantile spasms and epilepsy. Second, the possible mechanisms of epileptogenesis in these models include (a) a loss of subsets of neurons, (b) altered neurotransmitter release, (c) differences in neurotransmitter receptor levels and changes in receptor subunit composition, (d) altered neurite density and/or synaptogenesis, (e) changed membrane properties (e.g., altered voltage-gated channels), (f) altered cell morphology (neuronal differentiation), and (g) effects on cytoskeletal function. Finally, it is important to note that the "generator" of excitability in affected brain may be within the heterotopia or in the normotopic cortex. As additional genetic models come to light and the ability to distinguish their clinical counterparts improves, more individually tailored therapies, including standards for surgical interventions, will surely evolve.

Anticonvulsant actions of lamotrigine on spontaneous thalamocortical rhythms

PURPOSE

This study examined the actions of lamotrigine (LTG) on epileptiform discharges resembling generalized absence (GA) and primary generalized tonic-clonic (GTC) seizures in rat thalamocortical (TC) brain slices and attempted to characterize further the cellular mechanisms of action of LTG on neuronal ionic conductances.

METHODS

Rat TC slices generated spontaneous generalized epileptiform discharges after perfusion with a medium containing no added Mg(2+). Using multiple channel extracellular field-potential recordings in thalamus and cortex, the effects of LTG were characterized on two principal variants of activity that are similar to spike-wave discharges (SWDs) of GA epilepsy and GTC seizure discharges. These were termed simple TC burst complexes (sTBCs) and complex TC burst complexes (cTBCs), respectively. With whole-cell patch-clamp recording techniques in acutely dissociated TC neurons, the effects of LTG on GABA (gamma-aminobutyric acid)(A)-receptor-mediated currents and the low-threshold calcium current (I(T)) were examined.

RESULTS

In field-potential recording studies in TC slices, both sTBCs and cTBCs were blocked by clinically relevant concentrations of LTG. In patch-clamp recording studies, LTG was found to be ineffective in the modulation of both GABA(A) receptors (GABARs) and I(T) in TC neurons.

CONCLUSIONS

The efficacy of LTG on both variants of epileptiform discharges in TC slices clearly parallels its broad human clinical spectrum of action. This demonstrates that neurons within the TC system constitute one probable therapeutic target of LTG. However, LTG did not block either GABAR-mediated responses or I(T) in TC neurons. Modulation of these conductances represent likely cellular mechanisms of action of other antiepileptic drugs effective in the control of GA epilepsy. This suggests that LTG may have as yet uncharacterized effects that could combine with its previously defined sodium channel-blocking actions to explain its clinical utility in the control GA seizures.

Burst firing of action potentials in central snail neurons elicited by d-amphetamine: effect of anticonvulsants

The effect of anticonvulsants on the burst firing of action potentials in snail central neuron elicited by d-amphetamine was studied in the identified RP4 neuron of the African snail, Achatina fulica Ferussac. Oscillation of membrane potential and burst firing of action potentials were elicited by d-amphetamine in a concentration-dependent manner. Voltage clamped studies revealed that d-amphetamine elicited a negative slope resistance (NSR) in steady-state I-V curve between - 40 and - 10 mV. The burst firing of action potentials was alleviated following extracellular application of phenytoin, but was not affected after ethosuximide, carbamazepine, and valproic acid. The NSR elicited by d-amphetamine was blocked by phenytoin. However, the NSR was not altered if carbamazepine was added. These results suggest that of the four anticonvulsants tested, only phenytoin could alleviate the burst firing of action potentials elicited by d-amphetamine in snail neuron.

Regionally selective blockade of GABAergic inhibition by zinc in the thalamocortical system: functional significance

The thalamocortical (TC) system is a tightly coupled synaptic circuit in which GABAergic inhibition originating from the nucleus reticularis thalami (NRT) serves to synchronize oscillatory TC rhythmic behavior. Zinc is colocalized within nerve terminals throughout the TC system with dense staining for zinc observed in NRT, neocortex, and thalamus. Whole cell voltage-clamp recordings of GABA-evoked responses were conducted in neurons isolated from ventrobasal thalamus, NRT, and somatosensory cortex to investigate modulation of the GABA-mediated chloride conductance by zinc. Zinc blocked GABA responses in a regionally specific, noncompetitive manner within the TC system. The regional levels of GABA blockade efficacy by zinc were: thalamus > NRT > cortex. The relationship between clonazepam and zinc sensitivity of GABA(A)-mediated responses was examined to investigate possible presence or absence of specific GABA(A) receptor (GABAR) subunits. These properties of GABARs have been hypothesized previously to be dependent on presence or absence of the gamma2 subunit and seem to display an inverse relationship. In cross-correlation plots, thalamic and NRT neurons did not show a statistically significant relationship between clonazepam and zinc sensitivity; however, a statistically significant correlation was observed in cortical neurons. Spontaneous epileptic TC oscillations can be induced in vitro by perfusion of TC slices with an extracellular medium containing no added Mg(2+). Multiple varieties of oscillations are generated, including simple TC burst complexes (sTBCs), which resemble spike-wave discharge activity. A second variant was termed a complex TC burst complex (cTBC), which resembled generalized tonic clonic seizure activity. sTBCs were exacerbated by zinc, whereas cTBCs were blocked completely by zinc. This supported the concept that zinc release may modulate TC rhythms in vivo. Zinc interacts with a variety of ionic conductances, including GABAR currents, N-methyl-D-aspartate (NMDA) receptor currents, and transient potassium (A) currents. D-2-amino-5-phosphonovaleric acid and 4-aminopyridine blocked both s- and cTBCs in TC slices. Therefore NMDA and A current-blocking effects of zinc are insufficient to explain differential zinc sensitivity of these rhythms. This supports a significant role of zinc-induced GABAR modulation in differential TC rhythm effects. Zinc is localized in high levels within the TC system and appears to be released during TC activity. Furthermore application of exogenous zinc modulates TC rhythms and differentially blocks GABARs within the TC system. These data are consistent with the hypothesis that endogenously released zinc may have important neuromodulatory actions impacting generation of TC rhythms, mediated at least in part by effects on GABARs.

Single gene defects in mice: the role of voltage-dependent calcium channels in absence models

Nineteen genes encoding alpha1, beta, gamma, or alpha2delta voltage-dependent calcium channel subunits have been identified to date. Recent studies have found that three of these genes are mutated in mice with generalised cortical spike-wave discharges (models of human absence epilepsy), emphasising the importance of calcium channels in regulating the expression of this inherited seizure phenotype. The tottering (tg) locus encodes the calcium channel alpha1 subunit gene Cacna1a, lethargic (lh) encodes the beta subunit gene Cacnb4, and stargazer (stg) encodes the gamma subunit gene Cacng2. These calcium channel mutants should provide important insights into the basic mechanisms of neuronal synchronisation, and the genes may be considered candidates for involvement in similar human disorders. The mutant models offer an important opportunity to elucidate the molecular, developmental, and physiological mechanisms underlying one subtype of absence epilepsy. Since calcium channels are involved in numerous cellular functions, including proliferation and differentiation, membrane excitability, neurite outgrowth and synaptogenesis, signal transduction, and gene expression, their role in generating the absence epilepsy phenotype may be complex. A comparative analysis of channel function and neural excitability patterns in tottering, lethargic, and stargazer brain should be useful in identifying the common elements of calcium channel involvement in these absence models.

Carbamazepine withdrawal in children with previous symptomatic partial epilepsy: effects on neuropsychologic function

Neurocognitive performance was evaluated in seven children with symptomatic partial epilepsy prior to, and at least 12 months after, discontinuation of carbamazepine. The patients treated with carbamazepine monotherapy were seizure-free for at least 2 years and without electroencephalographic anomalies for at least 1 year. Results indicated that carbamazepine at therapeutic levels does not affect intellectual, memory, or attentional functions, or more complex frontal functions. Nevertheless, after therapy withdrawal scores on frontal function tests used in this study improved significantly. This suggests that these functions could have been better without carbamazepine therapy. The fact that carbamazepine decreases neuron membrane excitability and could reduce the information circuity, particularly in the frontal areas, is offered as a possible explanation. Further studies on larger samples using the same design are required to validate these results.

Antiepileptic drug cellular mechanisms of action: where does lamotrigine fit in?

Current frontline antiepileptic drugs tend to fall into several cellular mechanistic categories, and these categories often correlate with the clinical spectrum of action of the various antiepileptic drugs. Many antiepileptic drugs effective in control of partial and generalized tonic-clonic seizures are use- and voltage-dependent blockers of sodium channels. This mechanism selectively dampens pathologic activation of sodium channels, without interacting with normal sodium channel function. Examples include phenytoin, carbamazepine, valproic acid, and lamotrigine. Many antiepileptic drugs effective in control of generalized absence seizures block low threshold calcium currents. Low threshold calcium channels are present in high densities in thalamic neurons, and these channels trigger regenerative bursts that drive normal and pathologic thalamocortical rhythms, including the spike wave discharges of absence seizures. Examples include ethosuximide, trimethadione, and methsuximide. Several antiepileptic drugs that have varying clinical actions interact with the gamma-amino-butyric acid (GABA)ergic system. Diazepam and clonazepam selectively augment function of a subset of GABAA receptors, and these drugs are broad-spectrum antiepileptic drugs. In contrast, barbiturates augment function of all types of GABAA receptors, and are ineffective in control of generalized absence seizures, but effective in control of many other seizure types. Tiagabine and vigabatrin enhance cerebrospinal levels of GABA by interfering with reuptake and degradation of GABA, respectively. These antiepileptic drugs are effective in partial seizures. Lamotrigine is effective against both partial and generalized seizures, including generalized absence seizures. Its sole documented cellular mechanism of action is sodium channel block, a mechanism shared by phenytoin and carbamazepine. These drugs are ineffective against absence seizures. Consequently, unless there are unique aspects to the sodium channel block by lamotrigine, it seems unlikely that this mechanism alone could explain its broad clinical efficacy. Therefore, lamotrigine may have as yet uncharacterized cellular actions, which could combine with its sodium channel blocking actions, to account for its broad clinical efficacy.

Physiological and pharmacological alterations in postsynaptic GABA(A) receptor function in a hippocampal culture model of chronic spontaneous seizures

Cultured rat hippocampal neurons previously exposed to a media containing no added Mg2+ for 3 h begin to spontaneously trigger recurrent epileptiform discharges following return to normal medium, and this altered population epileptiform activity persisted for the life of the neurons in culture (> 2 wk). Neurons in "epileptic" cultures appeared similar in somatic and dendritic morphology and cellular density to control, untreated cultures. In patch-clamp recordings from hippocampal pyramidal cells from "epileptic," low Mg2+ pretreated hippocampal cultures, a rapid (within 2 h of treatment), permanent (lasting > or = 8 days) and statistically significant 50-65% reduction in the current density of functional gamma-aminobutyric acid-A (GABA(A)) receptors was evident when the GABA responses of these cells were compared with control neurons. Functional GABA receptor current density was calculated by determining the maximal response of a cell to GABA 1 mM application and normalizing this response to cellular capacitance. Despite the marked GABA efficacy differences noted above, the potency of GABA in activating chloride currents was not significantly different when the responses to control and "epileptic" pyramidal cells to multiple concentrations of GABA were compared. The EC50 for GABA was 4.5 +/- 0.2 (mean +/- SE) for control neurons and 3.5 +/- 0.4 microM, 5.2 +/- 0.5 microM, 3.7 +/- 0.3 microM, and 4.6 +/- 0.3 microM for epileptic neurons 2 h, 2 days, 3 days, and 8 days after low Mg2+ pretreatment, respectively. Modulation of GABA responses by the benzodiazepine, clonazepam, was significantly reduced in epileptic neurons compared with controls. The kinetically determined clonazepam 100 nM GABA augmentation efficacy decreased from 44.1% in control neurons to 9.3% augmentation in neurons recorded from cultures 10 days posttreatment. The kinetics of GABA current block by the noncompetitive antagonist picrotoxin were determined in hippocampal cultured neurons, and an IC50 of 14 microM determined. Bath application of picrotoxin at half of the IC50 concentration (7 microM) induced epileptiform activity in control cultures and this activity appeared very similar to the epileptiform activity induced by prior low Mg2+ treatment. This concentration of picrotoxin was determined experimentally to block 30% of the GABA(A)-mediated receptor responses in these cultures, and this level of block was sufficient to trigger spontaneous epileptiform activity. The 50% reduction of GABA responses induced as a permanent consequence of low Mg2+ treatment therefore was determined to be sufficient in and of itself to induce the spontaneous epileptiform activity, which was also a consequence of this treatment.

Anticonvulsant drug effects on spontaneous thalamocortical rhythms in vitro: ethosuximide, trimethadione, and dimethadione

Spontaneous generalized epileptiform discharges were elicited in rodent thalamocortical slices by perfusion with a medium containing no added Mg2+. In multiple-channel extracellular field potential recordings in thalamus and cortex, several distinct types of discharges were recorded, with two principal variants bearing marked similarity to spike-wave and generalized tonic-clonic seizure discharges recorded in patients with generalized seizure disorders. These discharges were termed sTBCs and cTBCs, respectively, for simple and complex thalamocortical burst complexes. The sensitivity of these discharges to the generalized absence anticonvulsants ethosuximide, trimethadione and dimethadione (the active metabolite of trimethadione) was studied. sTBCs were reduced or blocked by ethosuximide and dimethadione, when these drugs were applied in clinically relevant concentrations. The order of effectiveness of these agents was dimethadione > or = ethosuximide >> trimethadione. This paralleled the relative efficacy of these drugs in blocking T current in thalamic neurons. cTBCs were unaffected or exacerbated by these drugs. Structural control drugs including succinimide, the behaviorally inactive ring base of ethosuximide, and alpha, alpha-dimethyl-beta-methylsuccinimide, a convulsant succinimide, were inactive or exacerbated either sTBCs or cTBCs, respectively. These spontaneous generalized thalamocortical discharges in rodent thalamocortical slices may represent a potentially valuable in vitro model of generalized seizure discharges, with marked pharmacological and physiological similarities to various forms of clinical epileptic seizure activity.

Anticonvulsant drug effects on spontaneous thalamocortical rhythms in vitro: valproic acid, clonazepam, and alpha-methyl-alpha-phenylsuccinimide

Spontaneous thalamocortical epileptiform activity was elicited in rodent thalamocortical slices by a medium containing no added Mg2+. Multiple varieties of activity were generated in these slices, including simple thalamocortical burst complex (sTBC) activity that resembled the spike-wave discharges of generalized absence epilepsy, and complex thalamocortical burst complex (cTBC) activity that resembled generalized tonic-clonic seizure discharges. In a further pharmacological characterization of this activity, the effects of the broad-spectrum anticonvulsants valproic acid, alpha-methyl-alpha-phenylsuccinimide (the active metabolite of methsuximide) and clonazepam were studied. All three drugs were found to be effective in controlling both sTBC and cTBC activity when applied in clinically relevant concentration ranges. The effectiveness of valproic acid against spontaneous rhythms in vitro was not due to augmentation of GABAergic inhibition. No effect of valproic acid on GABA-activated chloride currents was evident in patch-clamp recordings of acutely isolated thalamic or cortical neurons. The equivalent general clinical and experimental spectrum of action of broadly effective anticonvulsants provided an additional correlation between the clinical efficacy of anticonvulsant drugs and their effects against epileptiform discharges in rodent thalamocortical slices. This further validates spontaneous generalized low-Mg2+ thalamocortical activity as a potentially valuable in vitro model of the primary generalized epilepsies, in which the cellular mechanisms underlying generation and control of these seizure discharges can be studied.