Antiepileptic Drugs, 5th Edition



Mechanisms of Action

Doru Georg Margineanu PhD*

Henrik Klitgaard PhD**

* Senior Scientist, Preclinical CNS Research, UCB S.A. Pharma Sector, Braine-l'Alleud, Belgium

** Director, Preclinical CNS Research, UCB S.A. Pharma Sector, Braine-l'Alleud, Belgium

Levetiracetam [LEV; ucb L059; (S)-α-ethyl-2-oxo-pyrrolidine acetamide; Figure 40.1] is the (S)-enantiomer of the ethyl analog of piracetam (Figure 40.1), synthesized during a follow-up chemical program aimed at identifying a second-generation nootropic drug. Consequently, the initial pharmacologic studies with LEV explored its ability to facilitate cholinergic neurotransmission (1). However, in vivo results demonstrated an unexpected potent ability of LEV to suppress seizures in the audiogenic-susceptible mouse (2), unlike piracetam, which was only weakly active. Further testing in the same model has shown that LEV remains active after injection directly into the brain, whereas both its (R)-enantiomer (ucb L060;Figure 40.1) and the main metabolite of LEV (ucb L057; Figure 40.1) lack activity in this model (2,3), suggesting that the observed seizure suppression relates to a specific central action of the parent compound.

Adjunctive therapy with LEV (Keppra, UCB S.A., Braine-l'Allend, Belgium) was recently proven effective and very well tolerated in controlling refractory partial seizures in adults (4). This resulted in a marketing authorization from the U.S. Food and Drug Administration in December 1999 and from the European Medicines Evaluation Agency (EMEA) in September 2000. However, the current under-standing of the antiepileptic mechanisms of LEV is at an early stage because it is a molecule unrelated to established antiepileptic drugs (AEDs).

The prevailing consensus that cell biology provides the ultimate conceptual frame for understanding the physiopathology of disease has stimulated the search at the cellular level for the abnormalities that underlie seizures (5). This obviously is legitimate, but complicated by the fact that epilepsy is a brain pathologic condition of large neuronal populations. The main cellular mechanisms that are thought to account for the antiseizure activities of the established AEDs refer to either (a) facilitation of inhibitory γ-aminobutyric acid (GABA)-ergic neurotransmission; (b) inhibition of excitatory glutamate receptors; or (c) block of voltage-gated Na+ or Ca2+ channels. This chapter reviews the effects of LEV on these and other mechanisms presumed to be of antiepileptic relevance.


FIGURE 40.1. Chemical structures of levetiracetam and of its (R)-enantiomer, ucb L060, along with the pyrrolidine compound, piracetam, and the main metabolite of levetiracetam, ucb L057.



LEV differs from other known AEDs by a lack of anticonvulsant activity in the two classic screening tests for AEDs in mice and rats, the maximal electroshock seizure (MES) and pentylenetetrazol (PTZ) models (3,6). Testing in other seizure paradigms in rodents has confirmed a weak anticonvulsant activity in threshold tests involving acute electrical or chemical stimulation (6), and only modest protection was observed against acute seizures induced by submaximal doses of chemoconvulsants (2). This is in contrast to the potent seizure suppression observed in animal models of chronic epilepsy, involving genetic and kindled animals with spontaneous, recurrent seizures or with seizures showing a phenomenology similar to human epileptic seizures.

A striking example is the potent seizure protection afforded by LEV in corneal electroshock- and PTZ-kindled mice and the absence of anticonvulsant activity against MES and PTZ seizures in normal mice (Table 40.1). This observation is supported further by results with fully amygdala-kindled rats (6), phenytoin-resistant amygdala-kindled rats (7), and hippocampal-kindled rats (8), confirming that LEV provides dose-dependent protection against several seizure parameters in kindled animals, including motor seizure severity and the duration of motor seizures and afterdischarges. Likewise, results with LEV in genetic animal models of epilepsy also revealed a suppression of several seizure parameters, including protection against postural stimulation-induced seizures in epilepsy-like mice (9), seizures induced by acoustic stimulation in mice (2) and rats (10), and spontaneous spike-and-wave discharges (SWDs) in the Genetic Absence Epilepsy Rat from Strasbourg (GAERS) model (10).



ED50 (mgkg i.p.)

Antiepileptic Drug


s.c. PTZ

Electroshock Kindling

PTZ Kindling




7 (2-10)

36 (15-96)


188 (154-216)

106 (64-170)

66 (52-83)

147 (116-189)


12 (9-16)


12 (8-17)

5 (3-7)



0.02 (0.01-0.04)

0.03 (0.02-0.05)

0.03 (0.02-0.04)



126 (78-196)




6 (5-9)


6 (4-10)

17 (8-28)


6 (4-10)


6 (1-16)

38 (22-171)

i.p., intraperitoneally; MES, supramaximal electroshock; s.c. PTZ, subcutaneous pentylenetetrazol.

aValues given are ED50values, i.e. the doses protecting 50% of the animals, with associated 95% confidence intervals. All compounds were administered at their optimal pretreatment times. Protection against clonic convulsions occurred in a significant (p < .05) proportion of the animals. From Klitgaard H, Matagne A, Gobert J, et al. Evidence for a unique profile of levetiracetam in rodent models of seizures and epilepsy. Eur J Pharmacol 1998;353:191-206, with permission from Elsevier Science.

The behavioral alterations induced by LEV do not differ between normal and kindled animals, showing only mild sedative and ataxic properties, and no psychotomimetic effects (2,3,6). Likewise, LEV is devoid of negative impact on cognitive performance in normal and kindled rats (11). Sedation and muscle relaxation occur only at doses above 1,000 mg/kg, showing that LEV has a low adverse effect potential in rodents. Combined with its potent seizure protection, this results in an unusually high safety margin for LEV in animal models mimicking both partial and generalized epilepsy (Figure 40.2).

Among the major AEDs in clinical use, only valproate and phenobarbital appear to suppress kindling acquisition, but do so only at doses associated with adverse effects (12). LEV has been reported to inhibit the development of PTZ kindling in mice (2) and amygdala kindling in rats (13) at doses devoid of adverse effects. Afterdischarges recorded from amygdala in animals previously treated with LEV remained significantly shorter compared with vehicle controls in the latter study, despite cessation of LEV treatment and continued amygdala stimulations. This observation led to several experimental approaches that currently are assessing the antiepileptogenic potential of LEV.

Taken together, these findings suggest that LEV is devoid of anticonvulsant activity in traditional seizure screening models, but reveals a potent broad-spectrum activity with a wide safety margin in animal models of chronic epilepsy reflecting both partial and generalized epilepsy. This preferential action in animal models of chronic epilepsy markedly distinguishes LEV from classical AEDs (Table 40.1).




FIGURE 40.2. Safety margins of levetiracetam versus other antiepileptic drugs in corneally kindled mice (main graph) and versus valproate and ethosuximide in Genetic Absence Epilepsy Rat from Strasbourg (GAERS) animals (inset). The safety margins are expressed as the ratios of the median toxic dose (TD50) value for Rotorod impairment and either the protective median effective dose (ED50) value against motor seizures, in corneally kindled mice, or the minimum active dose (MAD) reducing significantly the duration of spontaneous spike-and-wave discharges in GAERS rats. For topiramate, the protective ED50 value in the maximal electroshock seizure test and the TD50 value for Rotorod impairment in normal mice were used owing to the lack of effect of this drug against corneally kindled seizures. (From Klitgaard H, Matagne A, Gobert J, et al. Evidence for a unique profile of levetiracetam in rodent models of seizures and epilepsy. Eur J Pharmacol 1998;353:191-206, with permission from Elsevier Science.)


Electrophysiologic recordings, performed both in vivo and in vitro, consistently showed an absence of any intrinsic effect of LEV on normal neural responses and neuronal characteristics. An early report of the anticonvulsant profile of LEV in rodents (2) indicated that LEV did not change the baseline electroencephalogram (EEG), unlike clonazepam, which increased the proportion of lower-frequency background activity (in rats injected with PTZ to induce SWDs; see later). Similarly, in the GAERS model of absence seizures, LEV left the baseline EEG trace normal, whereas it markedly suppressed the SWDs (see later) (10). Finally, in urethane-anesthetized rats, LEV did not modify the field potentials recorded in hippocampal CA3 area, in response to commissural stimulation (14).

In agreement with these in vivo observations, it also was reported that LEV, perfused for 20 minutes at a concentration of 10 µmol/L, did not alter basic cell characteristics or normal synaptic transmission in pyramidal neurons in the CA3 area of rat hippocampal slices (15), whereas it suppressed epileptiform activity (see later). That study lists no effects of LEV on membrane potential, input resistance, amplitude, duration at half-amplitude and area of the action potential evoked by stimulation of commissural afferents, amplitudes of fast and slow after-hyperpolarizations, amplitudes of excitatory (EPSP) and fast inhibitory (IPSP) postsynaptic potentials evoked by subthreshold stimulation, area of EPSP and amplitudes of fast and slow


IPSPs evoked by suprathreshold stimulation, and amplitudes of baclofen-induced hyperpolarization and (1S,3R)-ACPD-and N-methyl-D-aspartate (NMDA)-induced depolarizations [Table 1 in Birnstiel et al., (15)].

The absence of effect of LEV on normal electrophysiologic responses is in contrast to its clear-cut effects on epileptiform electrophysiologic responses, both in vitro and in vivo. A perfusion of LEV at 10 µmol/L for 20 minutes markedly inhibited the development of epileptiform bursting induced by the GABAA receptor antagonist bicuculline methiodide (BMI) in CA3 pyramidal neurons from rat hippocampal slices (15). LEV (3 to 100 µmol/L) added to the perfusion medium 30 minutes after BMI caused a concentration-dependent decrease in the area of the bursts, without altering the after-hyperpolarization that follows the bursts. LEV (10 µmol/L) significantly reduced the frequency of the spontaneous bursting induced by application of NMDA, 10 or 15 µmol/L, in CA3 pyramidal neurons from rat hippocampal slices in vitro, although no consistent effect was observed on the size of the bursts (15). In another study, LEV (200 and 400 µmol/L) significantly reduced the length of the seizure-like population bursts recorded extracellularly in the CA1 area of rat hippocampal slices that were bathed in a medium with bicuculline and increased potassium (16). These in vitro effects of LEV against epileptiform discharges induced by BMI (or bicuculline) agree with the action of the drug when administered systemically (3.2 to 32 µmol/kg intravenously) to inhibit the epileptiform effect of BMI (applied locally, through the recording microelectrode) on the field potentials recorded in the hippocampal CA3 area of anesthetized rats (17).

We have shown that LEV (32 and 100 µmol/L) consistently diminished the epileptiform field potentials recorded in the CA3 area of rat hippocampal slices that were bathed in an epileptogenic medium containing increased (7.5 mmol/L) potassium and lowered (0.5 mmol/L) calcium (18). LEV, along with the classic AEDs valproate, clonazepam, and carbamazepine, reduced the number of repetitive population spikes evoked by single stimuli when the slices were in the epileptogenic medium, whereas it differed from these reference AEDs in its specific ability to reduce population spike amplitude (i.e., to antagonize neuronal hypersynchronization). In agreement with this, an electrographic study in rats has shown that LEV [17 mg/kg intraperitoneally (i.p.)] also differed from reference AEDs in not interfering with the onset of spike-and-burst discharges in the hippocampus induced by systemic administration of pilocarpine. Instead, LEV selectively inhibited the synchronization necessary for their propagation to the cortex. Likewise, LEV (170 mg/kg i.p.) had no impact on the appearance of kainic acid-induced spike-and-burst discharges in the hippocampus, but it retarded their generalization to the cortex (19).


Animal Model

Electrophysiologic Recording

Reported Action of Levetiracetam


In vivo


Freely moving rats


Inhibition of pentylenetetrazol-induced spike-and-wave discharges



Inhibition of spontaneous spike-and-wave discharges in GAERS rats



Reduction of afterdischarge duration in amygdala-kindled rats

6, 13


Anesthetized rats

Evoked field potentials

Inhibition of bicuculline-induced increases in amplitude of hippocampal population spikes

14, 17

In vitro


Rat hippocampal slices

Evoked field potentials

Reduction of the amplitudes and the number of repetitive population spikes induced by a “high K+-low Ca2+” perfusion fluid



Action potentials (intracellular recordings) in pyramidal neurons

Inhibition of bicuculline-induced bursts of action potentials and of the frequency of bursting induced by N-methyl-D-aspartate


GAERS, genetic absence epilepsy rat from Strasbourg.

In vivo demonstration of LEV's effect on epileptiform EEG activity was observed in several rat models of epilepsy. LEV (17 mg/kg i.p.) reduced the cumulative duration of SWDs induced in rats by a subconvulsant dose of PTZ (2). In GAERS rats, LEV markedly reduced the cumulative duration of spontaneous SWDs to between 15% and 30% of the predrug level, with a dose as low as 5.4 mg/kg i.p. and no further incremental effect at higher doses (10). As mentioned previously, a prominent effect of LEV against electrographic epileptiform discharges in vivo is to reduce the duration of the afterdischarges recorded in amygdala-kindled rats, observed with both acute (13 to 108 mg/kg i.p.) (6) and chronic (13 to 54 mg/kg/day i.p.) (13) administration of the drug.

Table 40.2 summarizes the main currently reported effects of LEV on epileptiform discharges, both in vivo and


in vitro. Several of these effects were not dose or concentration dependent, suggesting that the antiepileptic action of the drug may involve multiple mechanisms.


A specific [3H]LEV binding site (LBS), which is saturable, reversible, and stereoselective, has been reported to exist in rat brain membranes (20, 21, 22). LBS appears to be located in brain structures, with high densities in the cortex, hippocampus, and cerebellum, whereas it was not detected in a range of peripheral tissues (Figure 40.3). [3H]LEV was reported to label a single class of binding sites in hippocampal membranes, with modest affinity (Kd = 780 ± 115 nmol/L) and high binding capacity (Bmax = 9.1 ± 1.2 pmol/mg protein) (22). The rank order of affinity for LBS of the (S)-stereoisomer homologs of LEV appeared well correlated with their anticonvulsant activity against the expression of tonic convulsions in the audiogenic mouse test (Figure 40.3, inset). This correlation suggests a possible functional role for LBS in the antiepileptic mechanisms of LEV.


FIGURE 40.3. Binding of [3H]levetiracetam in rat brain (Hi, hippocampus; Co, cortex; Ce, cerebellum) and peripheral tissues (Li, liver; Lu, lung; Ki, kidney; Sp, spleen; Pa, pancreas; He, heart; Ad, adrenals). The hatched columns indicate the total binding and the open columns indicate the nonspecific binding. The inset shows the correlation of the anticonvulsant activity of (S)-homologs of levetiracetam in the audiogenic mouse with their affinity at the [3H]levetiracetam binding site. Points 1 and 2 on the plot represent piracetam and levetiracetam, respectively. [From Noyer M, Gillard M, Matagne A, et al. The novel antiepileptic drug levetiracetam (ucb L059) appears to act via a specific binding site in CNS membranes. Eur J Pharmacol 1995;286:137-146, with permission from Elsevier Science.]

LEV, up to 10 µmol/L, did not displace known radioligands from a large variety of binding sites presumed to be related to altered neural excitability (22). Thus, no affinity was observed for opiate, adenosine, adrenergic, cholinergic, dopaminergic, serotonergic, histaminergic, glutamatergic, GABAB, and GABAA/benzodiazepine receptors. Monoamine reuptake sites, peptide-specific receptors, second messenger systems such as adenylate cyclase and protein kinase C, and different ion channel proteins likewise were unaffected. Neither phenytoin, carbamazepine, sodium valproate, phenobarbital, nor the benzodiazepines, diazepam and clonazepam, displayed any relevant affinity for the LBS. However, the anticonvulsants ethosuximide and pentobarbital, and the GABA-related convulsants, PTZ and bemegride, were effective at concentrations comparable with active drug concentrationsin vivo (22). Sacaan and Lloyd (21) reported that the T-type calcium channel antagonist amiloride inhibited the binding of [3H]LEV, and the anticonvulsants trimethadione and dimethadione displaced the binding of [3H]LEV with inhibitory concentration 50% (IC50) values close to their anticonvulsant plasma levels. However, the affinity of dimethadione for LBS was shown by Noyer et al. (22) to correspond to a pKi ≤2.0, suggesting a very low affinity for the LBS.



In summary, stereoselective, reversible, low-affinity, highcapacity binding of LEV was identified in rat brain membranes, on a specific site not present in peripheral tissues and distinct from binding sites known to alter neuronal excitability. Intense investigations are in progress at UCB Pharma to isolate this binding site and help to further the understanding of its possible role in the antiepileptic mechanism of LEV.


Ionotropic Inhibitory Receptors

Systemic administration of LEV at doses known to suppress seizures in the rat (1.7 to 170 mg/kg i.p.) was reported to induce alterations in GABA metabolism and turnover in several brain regions (23). LEV, 170 mg/kg i.p., increased the activity of the GABA-degrading enzyme GABA aminotransferase (GABA-T) in 7 of 12 brain regions studied, but the effect was short-lived (~15 minutes) and it was observed only at this relatively high dose. In the striatum, the significant increase in GABA-T activity, observed 15 minutes after LEV 170 mg/kg i.p., was associated with a decrease in activity of the GABA-synthesizing enzyme glutamic acid decarboxylase (GAD), leading to a significant reduction in regional GABA turnover. This was followed, however, by a pronounced increase in striatal GABA turnover, and the activities of both GABA-T and GAD were nearly normalized 60 minutes after LEV, in all brain regions. The reduction of GABAergic activity in the striatum has been reported to be anticonvulsant (24) because a disinhibited striatal output enhances inhibition in the substantia nigra pars reticulata, which receives a strong GABAergic input from the striatum and is known to control seizure propagation (25). Indeed, Löscher et al. (23) recorded decreased spontaneous firing of nondopaminergic, presumably GABAergic neurons in the substantia nigra pars reticulata after LEV 170 mg/kg i.p. administration. However, because LEV induced both increases and decreases in GABA-T and GAD activities and did not alter either GABA-T or GAD activities in vitro, Löscher et al. (23) inferred that the enzyme alterations they found were not direct effects of LEV, but possibly indirect consequences of postsynaptic changes. Moreover, the fact that GABA turnover was normalized 60 minutes after LEV casts serious doubts on the anticonvulsant relevance of the regional alterations in GABA turnover for the antiepileptic action of LEV, because the time of the peak effect of LEV was previously reported from the same laboratory to be 60 minutes in both mice and rats (6). Furthermore, neurochemical studies in mouse brains (26) indicated that LEV (up to 300 mg/kg i.p.; single/multiple doses) had no effect on the concentrations of GABA, glutamate, and glutamine, and on the activities of GAD and GABA-T. Also, the same group reported that a 1-hour exposure to LEV had no effect on GABA transport and metabolism in rat astrocyte culture (27). Accordingly, these authors concluded that it is unlikely that the action of LEV would be mediated through the GABAergic system (26,27).

LEV produced only minor and contrasting effects on GABA-induced currents recorded by whole-cell patch-clamp methods in cultured rat cerebellar granule and hippocampal neurons (28). They consisted of a small reduction in the peak amplitude and a prolongation of the decay phase. However, these effects were significant only at LEV concentrations beyond 100 µmol/L. Interestingly, LEV potently suppressed (median effective dose = 1 - 10 µmol/L) the inhibitory effects of several negative allosteric modulators (Zn2+, β-carbolines, chlorodiazepam) on GABA-gated currents in cultured rat hippocampal and cerebellar granule neurons and glycine-gated currents in spinal cord neurons (28). These results are remarkable in view of the hypothesis that suggests that the hyperexcitability characteristics of the epileptic hippocampus may be associated with circuit and cellular alterations in dentate granule cells (DGCs). These involve changes both in the subunit expression pattern of GABAA receptors, rendering them more sensitive to inhibition by Zn2+, and in sprouting of Zn2+-containing DGC axons (mossy fibers) back onto the inner molecular layer of the dentate gyrus (29). This creates an environment in which excessive Zn2+ is released during repetitive activation of DGCs and is purported to result in a pathologic lowering of inhibition as well as an enhanced seizure propensity in the epileptic hippocampus.

Ionotropic Excitatory Receptors

Single and repeated systemic administrations of LEV (1 to 300 mg/kg i.p.) were reported to have no effect on glutamine synthetase, a key enzyme in the regulation of glutamate neurotransmission, in mouse brain (30).

LEV at high concentrations (IC50 = 268 µmol/L) inhibited the α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA)-gated current in cultured rat hippocampal neurons, although it was without effect on the currents gated by either NMDA or kainate. The inhibition by LEV of α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid-gated currents, however, seems to lack antiseizure relevance in view of the relatively high concentration at which it was observed and because LEV is ineffective in suppressing the clonic convulsions induced in vivo by AMPA, kainic acid, or NMDA (31).

Dopaminergic System

A neurochemical study in the rat (32) has reported that systemic pretreatment with LEV (17 mg/kg i.p.) prevented an increase in the extracellular levels of dopamine and its metabolites (dihydroxyphenylacetic acid and homovanillic acid) induced by a subconvulsant dose of bicuculline (3 mg/kg i.p.). This appears to be an indirect effect of the ability of LEV to inhibit the epileptiform modifications induced by bicuculline


described previously because the same study showed only a minimal effect of LEV against haloperidol-induced enhancement of the release of dopamine and its metabolites.


Na+ Current

LEV (up to 1 mmol/L) did not inhibit or modify the biophysical properties (steady-state activation and inactivation, time to peak, fast kinetics of inactivation, and recovery from steady-state inactivation) of the tetrodotoxin (TTX)-sensitive, inward, voltage-dependent Na+ current, recorded in the whole-cell configuration of the patch-clamp technique, in cultured rat cortical neurons (33).

Ca2+ Currents

LEV (32 and 100 µmol/L) did not inhibit or modify the biophysical properties of the low-voltage-activated (T-type) Ca2+ current, recorded in the whole-cell configuration of the patch-clamp technique, in pyramidal neurons from the CA1 area of rat hippocampal slices (34). On the other hand, LEV (32 µmol/L) depressed high-voltage-activated Ca2+ currents in visually identified pyramidal neurons from rat hippocampal slices (35) (with inhibition becoming significant after a 30-minute perfusion of the drug). The type of LEV-sensitive Ca2+channel and the antiepileptic relevance remain to be established.



Action of Levetiracetam





GABA metabolism or turnover

Local alterations in rat brain


Only at high doses; short lasting; disparate between areas


No effect in mouse brain



No effect in vitro

23, 27


GABA-gated current

Prolongation of the decay


Significant only at ≥100 µmol/L


No consistent effect on the amplitude



GABA- and glycine-gated currents

Suppression of inhibition by negative modulators (e.g., Zn2+, β-carbolines)


EC50 = 1-10 µmol/L



Glutamate metabolism

No effect in mouse brain

26, 30


AMPA-gated current



IC50 = 268 µmol/L, no effect in vivo


N-methyl-D-aspartate-gated current

No effect



Kainate-gated current

No effect



Voltage-gated ion channels


Na+ current

No effect



Low-voltage-gated Ca2+ current

No effect


In hippocampal neurons


High-voltage-gated Ca2+ currents



In hippocampal neurons


K+ currents

Inhibition of the delayed rectifier (IK) current


Maximal effect(≈30%) at 30 µmol/L, IC50 5-45 µmol/L, entailing a broadening of secondary action potentials

Ca2+ -dependent mechanisms,Potentially involving intraneuronal effects

Hypothesized to account for a non-GABAA-related antibicuculline effect of levetiracetam

17,37, 38

No direct proof (only indirect indications) to date

AMPA, α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid; EC50, median effective concentration; GABA, γ-aminobutyric acid, Ik; IC50, inhibitory concentration of 50.

K+ Currents

One study (36) reports that LEV produced a significant inhibition (≈30%) of the delayed rectifier potassium current (IK) in isolated hippocampal neurons from both rat (IC50 5 µmol/L) and guinea pig (IC50 45 µmol/L). The blocking of IK by LEV appeared irreversible and voltage independent. Furthermore, the same study reports a broadening of the secondary action potentials produced by longstep depolarizations, entailing a decrease (≈20%) in the area of the depolarization-induced action potentials. This suggests that a moderate block of the IKcurrents may reduce the generation of repetitive action potentials and thereby contribute to the antiepileptic mechanism(s) of LEV.


The cellular-level effects of LEV on neurons, reported or hypothesized to date, are summarized in Table 40.3. It


appears that two conclusions can be substantiated: (a) no existing data favor ascribing to LEV any conventional modulation of the three main mechanisms currently accepted for the established AEDs (GABAergic facilitation or inhibition of either Na+ currents or low-voltage-activated Ca2+ currents), implying that the antiepileptic mechanism(s) of LEV must be novel; and (b) the actions of LEV on neurons, established to date, appear multiple, mild, and of a modulatory type, rather than expressing a straightforward dose-dependent inhibition/activation of one single cellular effect.

LEV appears able to control pathologic neuronal hyperexcitability through a modulatory inhibition of neuronal high-voltage Ca2+ currents, a prolongation of the hyperpolarizing GABA-gated currents, and an inhibition of the AMPA-gated current. However, although the inhibition of Ca2+ currents was observed at a clinically relevant concentration, the other two effects of LEV appeared only at relatively high concentrations in vitro. Two other cellular effects of LEV, namely, the suppression of the inhibition by Zn2+ and other negative allosteric modulators of both GABA-and glycine-gated currents, and the broadening/delaying of secondary action potentials on inhibition of the IK current, seem of particular potential interest for the basic under-standing of epileptic pathophysiology.

The “sprouted mossy fiber/Zn2+-sensitive GABAA receptor” hypothesis, recently proposed [for reviews, see Coulter (29,39)], postulates that epileptogenesis may involve sprouting of Zn2+-containing mossy fiber terminals, which innervate DGCs containing GABAA receptors with an altered subunit composition, rendering them more sensitive to inhibition by Zn2+. When repetitive stimulation results in enhanced release of Zn2+, this in turn may diffuse to and block these “epileptic” GABAA receptors on DGCs. This would induce a vicious cycle of disinhibition that may result in aberrant activity, triggering epileptiform discharges in the epileptic hippocampus. LEV may reverse this process by its ability to suppress the inhibitory effects of Zn2+ and other negative allosteric modulators on both GABA- and glycine-gated currents (28). These attributes of LEV may represent a novel mechanism that could contribute substantially to its antiseizure action as well as to its potential antiepileptogenic properties, such as those observed in the kindling model (13).

The unexpected possibility of obtaining an antiepileptic effect that involves a reduction of repetitive neuronal firing upon delaying membrane repolarization through the inhibition of a potassium current (36) seems likely to attract the further attention of epileptologists. If confirmed, this would represent a completely novel antiepileptic mechanism.

At this stage, it is not possible to formulate a unified mechanism of action for LEV, especially because the molecular nature of its brain-specific binding site remains to be characterized, and further cellular effects of LEV are likely to be identified. Thus, the robust anti-BMI effect of LEV on hippocampal neurons, which appeared not be associated with the GABAA receptor (17), is mimicked by thapsigargin (38), a blocker of Ca2+-adenosine triphosphatase in the endoplasmic reticulum. This suggested that putative Ca2+-dependent actions of LEV to downregulate neuronal excitability might involve not only the membrane channels, but intraneuronal effects. This possibility needs further exploration. Likewise, it remains to be investigated whether (and, if so, which) nonsynaptic desynchronizing effects might be involved in the distinct ability of LEV to antagonize neuronal hypersynchronization (18). Finally, the nearabsence of an effect of LEV on normal neural responses, together with its antiepileptic efficacy, seem to legitimize the hope that forthcoming progress in understanding the mechanisms of this drug also might shed some light on the pathologic states it treats.


  1. Wülfert E, Hanin I, Verloes R. Facilitation of calcium-dependent cholinergic function by ucb L059, a new “second generation” nootropic agent. Psychopharmacol Bull1989;25:498-502.
  2. Gower A, Noyer M, Verloes R, et al. ucb L059, a novel anticonvulsant drug: pharmacological profile in animals. Eur J Pharmacol1992;222:193-203.
  3. Klitgaard H, Matagne A, Gobert J, et al. Evidence for a unique profile of levetiracetam in rodent models of seizures and epilepsy. Eur J Pharmacol1998;353:191-206.
  4. Cereghino JJ, Biton V, Abou-Khalil B, et al. Levetiracetam for partial seizures: results of a double- blind, randomized clinical trial. Neurology2000;55:236-242.
  5. McNamara JO. Cellular and molecular basis of epilepsy. J Neurosci1994;14:3413-3425.
  6. Löscher W, Hönack D. Profile of ucb L059, a novel anticonvulsant drug, in models of partial and generalized epilepsy in mice and rats. Eur J Pharmacol1993;232:147-158.
  7. Löscher W, Reissmüller E, Ebert U. Anticonvulsant efficacy of gabapentin and levetiracetam in phenytoin-resistant kindled rats. Epilepsy Res2000;40:63-77.
  8. Matagne A, Klitgaard H. Levetiracetam protects against seizures in fully hippocampal-kindled rats. Epilepsia1999;40[Suppl 7]:137.
  9. De Deyn PP, Kabuto H, D'Hooge R, et al. Protective effect of ucb L059 against postural stimulation-induced seizures in EL mice. Neurosciences1992;18[Suppl 2]:187-192.
  10. Gower A, Hirsch E, Boehrer A., et al. Effects of levetiracetam, a novel antiepileptic drug, on convulsant activity in two genetic rat models of epilepsy. Epilepsy Res1995;22:207-213.
  11. Lamberty Y, Margineanu DG, Klitgaard H. Absence of negative impact of levetiracetam on cognitive function and memory in normal and amygdala-kindled rats. Epilepsy Behav2000;1: 333-342.
  12. Silver JM, Shin C, McNamara JO. Antiepileptogenic effects of conventional anticonvulsants in the kindling model of epilepsy. Ann Neurol1991;29:356-363.
  13. Löscher W, Hönack D, Rundfeldt C. Antiepileptogenic effects of the novel anticonvulsant levetiracetam (ucb L059) in the kindling model of temporal lobe epilepsy. J Pharmacol Exp Ther1998;284:474-479.
  14. Margineanu DG, Wülfert E. ucb L059, a novel anticonvulsant, reduces bicuculline-induced hyperexcitability in rat hippocampal CA3 in vivo. Eur J Pharmacol1995;286:321-325.



  1. Birnstiel S, Wülfert E, Beck SG. Levetiracetam (ucb L059) affects in vitro models of epilepsy in CA3 pyramidal neurons without altering normal synaptic transmission. Naunyn Schmiedebergs Arch Pharmacol1997;356:611-618.
  2. Doheny HC, Whittington MA, Jefferys JGR, et al. Levetiracetam inhibits bicuculline/potassium induced seizure like events in vitro. Eur J Neurosci1998;[Suppl 10]:39.
  3. Margineanu DG, Wülfert E. Inhibition by levetiracetam of a non-GABAAreceptor-associated epileptiform effect of bicuculline in rat hippocampus. Br J Pharmacol1997;122:1146-1150.
  4. Margineanu DG, Klitgaard H. Inhibition of neuronal hyper-synchrony in vitro differentiates levetiracetam from classical antiepileptic drugs. Pharmacol Res2000;42:281-285.
  5. Klitgaard H. Levetiracetam counteracts propagation of epileptiform activity induced by systemic administration of pilocarpine and kainic acid to rats in vivo. Epilepsia1999;40[Suppl 7]:138.
  6. Gillard M, Noyer M, Henichart J-P, et al. The novel antiepileptic drug levetiracetam (ucb L059) appears to act via a specific binding site in CNS membranes. Br J Pharmacol1994;112 [Suppl]:615P.
  7. Sacaan AI, Lloyd GK. The binding of the anticonvulsant levetiracetam to rat brain membranes distinguishes a novel site relevant to antiepileptic drug activity.Neuropsychopharmacology1994;10[Suppl, Pt 2]:133S.
  8. Noyer M, Gillard M, Matagne A, et al. The novel antiepileptic drug levetiracetam (ucb L059) appears to act via a specific binding site in CNS membranes. Eur J Pharmacol1995;286: 137-146.
  9. Löscher W, Hönack D, Bloms-Funke P. The novel antiepileptic drug levetiracetam (ucb L059) induces alterations in GABA metabolism and turnover in discrete areas of rat brain and reduces neuronal activity in substantia nigra pars reticulata. Brain Res1996;735:208-216.
  10. Turski L, Cavalheiro EA, Calderazzo-Filho LS, et al. The basal ganglia, the deep piriform cortex, and seizure spread: bicuculline is anticonvulsant in the rat striatum. Proc Natl Acad Sci U S A1989;86:1694-1697.
  11. Gale K. Progression and generalization of seizure discharge: anatomical and neurochemical substrates. Epilepsia1988;29 [Suppl 2]:S15-S34.
  12. Sills GJ, Leach JP, Frazer CM, et al. Neurochemical studies with the novel anticonvulsant levetiracetam in mouse brain. Eur J Pharmacol1997;325:35-40.
  13. Frazer CM, Sills GJ, Butler E, et al. Levetiracetam fails to influence GABA transport and metabolism in rat astrocyte culture. Epilepsia1997;38[Suppl 3]:177.
  14. Rigo JM, Nguyen L, Rocher V, et al. Levetiracetam: novel modulation of ionotropic inhibitory receptors. Epilepsia2000; 41 [suppl]:35.
  15. Coulter DA. Chronic epileptogenic cellular alterations in the limbic system after status epilepticus. Epilepsia1999;40[Suppl 1]:S23-S33.
  16. Frazer CM, Sills GJ, Forrest G, et al. Effects of anti-epileptic drugs on glutamine synthetase activity in mouse brain. Br J Pharmacol1999;126:1634-1638.
  17. Hans G, Rigo JM, Crommen J, et al. Levetiracetam: no relevant effect on ionotropic excitatory glutamate receptors. Epilepsia2000;41 [Suppl]:37.
  18. Zhang X, Wülfert E, Hanin I. Effects of ucb L059, a potential anticonvulsant agent, on release of dopamine and its metabolite from cerebral ventricular perfusate induced by bicuculline and haloperidol in rats. Pharmacologist1993;35(3):187.
  19. Zona C, Marchetti C, Margineanu DG. The novel antiepileptic drug candidate levetiracetam does not modify the biophysical properties of the Na+channel in rat cortical neurons in culture. Soc Neurosci Abstr 1999;25:1868.
  20. Niespodziany I, Klitgaard H, Margineanu DG. Levetiracetam had no effect on low-voltage-activated Ca2+current in rat CA1 hippocampal neurons. Neurology 2000;54[Suppl 3]:A307-A308.
  21. Niespodziany I, Klitgaard H, Margineanu DG. Levetiracetam: modulation of high-voltage-activated Ca2+current in CA1 pyramidal neurons of rat hippocampal slices. Epilepsia2000;31[Suppl]: 347.
  22. Madeja M, Margineanu DG, Klitgaard H. Effect of levetiracetam on voltage-gated potassium channels: a novel antiepileptic mechanism of action? Epilepsia2001;42[Suppl 2]:19.
  23. Margineanu DG, Wülfert E. ucb L059, a novel anticonvulsant, reduces bicuculline-induced neuronal hyperexcitability, possibly through Ca-dependent mechanisms. Pharmacol Res1995;31S: 347.
  24. Wülfert E, Margineanu DG. Thapsigargin inhibits bicuculline-induced epileptiform excitability in rat hippocampal slices. Neurosci Lett1998;243:141-143.
  25. Coulter DA. Mossy fiber zinc and temporal lobe epilepsy: pathological association with altered “epileptic” gamma-aminobutyric acid A receptors in dentate granule cells.Epilepsia2000;41 [Suppl 6]:S96-S99.