The anti-epileptic drug levetiracetam reverses the inhibition by negative allosteric modulators of neuronal GABA- and glycine-gated currents

Abstract

1. In this study in vitro and in vivo approaches were combined in order to investigate if the anti-epileptic mechanism(s) of action of levetiracetam (LEV; Keppra) may involve modulation of inhibitory neurotransmission. 2. GABA- and glycine-gated currents were studied in vitro using whole-cell patch-clamp techniques applied on cultured cerebellar granule, hippocampal and spinal neurons. Protection against clonic convulsions was assessed in vivo in sound-susceptible mice. The effect of LEV was compared with reference anti-epileptic drugs (AEDs): carbamazepine, phenytoin, valproate, clonazepam, phenobarbital and ethosuximide. 3. LEV contrasted the reference AEDs by an absence of any direct effect on glycine-gated currents. At high concentrations, beyond therapeutic relevance, it induced a small reduction in the peak amplitude and a prolongation of the decay phase of GABA-gated currents. A similar action on GABA-elicited currents was observed with the reference AEDs, except ethosuximide. 4. These minor direct effects contrasted with a potent ability of LEV (EC(50)=1 - 10 microM) to reverse the inhibitory effects of the negative allosteric modulators zinc and beta-carbolines on both GABA(A) and glycine receptor-mediated responses. 5. Clonazepam, phenobarbital and valproate showed a similar ability to reverse the inhibition of beta-carbolines on GABA-gated currents. Blockade of zinc inhibition of GABA responses was observed with clonazepam and ethosuximide. Phenytoin was the only AED together with LEV that inhibited the antagonism of zinc on glycine-gated currents and only clonazepam and phenobarbital inhibited the action of DMCM. 6. LEV (17 mg kg(-1)) produced a potent suppression of sound-induced clonic convulsions in mice. This protective effect was significantly abolished by co-administration of the beta-carboline FG 7142, from a dose of 5 mg kg(-1). In contrast, the benzodiazepine receptor antagonist flumazenil (up to 10 mg kg(-1)) was without any effect on the protection afforded by LEV. 7. The results of the present study suggest that a novel ability to oppose the action of negative modulators on the two main inhibitory ionotropic receptors may be of relevance for the anti-epileptic mechanism(s) of action of LEV.

Figures

Figure 1

Effect of levetiracetam (LEV) on GABA- and glycine-induced currents. Left panel: Three to twenty-three days in vitro (DIV) cerebellar granule neurons (A), 7 – 16 DIV hippocampal neurons (C and G) and 3 – 5 DIV spinal cord neurons (E) were perfused for 10 s with GABA (20 or 50 μM) or glycine (100 μM) as indicated, alone or in combination with LEV (25 s, 100 μM). A 60 s period was allowed for the washout of drugs. For each recording, a curve fitting procedure was applied to the desensitization phase using a single exponential (see text) and allowing to calculate a decay time constant (τ). Right panel: Currents evoked by 50 μM GABA for cerebellar granule neurons (B) and by 20 μM for hippocampal neurons (D) or by 100 μM glycine for spinal cord neurons (F) or for hippocampal neurons (H) (concentrations which elicit a half-maximal response in each cell types) were recorded in the presence of increasing LEV concentrations. Results are expressed as percentage of inhibitory amino acid-induced currents peak amplitudes or decay time constants in the absence of LEV (mean±s.e.mean, n=7 – 135 for A, 6 – 71 for D, 5 – 20 for F and 6 – 17 for H). ***P<0.001 and *P<0.05 using ANOVA followed by Dunnett's multiple comparisons post-tests.

Figure 2

Effect of reference anti-epileptic drugs (AEDs) on GABA- and glycine-induced currents. Currents evoked by 20 μM GABA for hippocampal neurons (A and B) or by 100 μM glycine for spinal cord neurons (C) were recorded in the presence of reference AEDs (100 μM levetiracetam, 10 μM carbamazepine, 1 μM clonazepam, 100 μM phenobarbital, 50 μM phenytoin, 1 mM valproate and 1 mM ethosuximide). Results are expressed as percentage of inhibitory amino acid-induced currents peak amplitudes (A and C) or decay time constants (B) in the absence of AEDs (mean±s.e.mean, n=5 – 12). ***P<0.001 and *P<0.05 using Student's t-test analysis. It has to be mentioned that the marginal significance (P<0.05) reported in (C) for LEV, which contrasts with Figures 1G and H (P>0.05), is due to the use of a different statistical test.

Figure 3

Interaction between LEV and negative allosteric modulators of the GABAA receptor (hippocampal neurons). Left panel: Seven to sixteen DIV hippocampal neurons were perfused for 10 s with 20 μM GABA alone (left traces) or in combination with LEV (right traces) (25 s, 100 μM) (controls) and negative allosteric modulators of the GABAA receptor (25 s, half-maximal inhibitory concentrations: 3 μM bicuculline (BIC), 30 μM picrotoxin (PIC), 15 μM DMCM, 15 μM FG7142 and 15 μM zinc). A 60 s period was allowed for the washout of drugs (90 s in the case of picrotoxin or β-carbolines). Right panel: Currents elicited by 20 μM GABA in the presence of the indicated negative allosteric modulator of the GABAA receptor were recorded for increasing LEV concentrations. Results are expressed as percentage of GABA-induced currents peak amplitudes in the absence of any drug (solid horizontal line) (mean±s.e.mean, n=4 – 16 for B, 5 – 41 for D, 5 – 34 for F, 16 – 33 for H and 6 – 36 for J). The horizontal dashed line represents the level of the GABA response in the presence of the GABAA receptor inhibitor alone. The curve fitting procedure yielded a half-maximal effective LEV concentration of 3.1±0.2 μM for DMCM- and 20.1±19.8 μM for zinc-induced inhibitions of GABA currents. ***P<0.01 and *P<0.05 using ANOVA followed by Dunnett's multiple comparisons post-tests (Student's t-test analysis was used for FG7142).

Figure 4

Interaction between LEV and negative allosteric modulators of the GABAA receptor (cerebellar granule neurons). Left panel: Three to twenty-three DIV cerebellar granule neurons were perfused for 10 s with 50 μM GABA alone (left traces) or in combination with LEV (right traces) (25 s, 100 μM) (controls) and negative allosteric modulators of the GABAA receptor (25 s, half-maximal inhibitory concentrations: 10 μM bicuculline (BIC), 10 μM picrotoxin (PIC), 10 μM FG7142 and 1 μM zinc). A 60 s period was allowed for the washout of drugs (90 s in the case of picrotoxin or FG7142). Right panel: Currents evoked by 50 μM GABA in the presence of the indicated negative allosteric modulator of the GABAA receptor were recorded for increasing LEV concentrations. Results are expressed as in Figure 3 (mean±s.e.mean; n=3 – 6 for B, 6 – 13 for D, 4 – 26 for F and 4 – 6 for H). ***P<0.001 using ANOVA followed by Dunnett's multiple comparisons post-tests.

Figure 5

Comparison of the potencies of reference AEDs in reversing the inhibitory effects of DMCM (A) and zinc (B) on GABA-induced currents in 7 – 14 DIV cultured hippocampal neurons. Currents evoked by 20 μM GABA in the presence of 15 μM DMCM or 15 μM ZnCl2 were recorded with the different reference AEDs (10 μM LEV, 10 μM carbamazepine, 1 μM clonazepam, 100 μM phenobarbital, 50 μM phenytoin, 1 mM valproate and 1 mM ethosuximide). Results are expressed as in Figure 3 (mean±s.e.mean; n=5 – 20). AEDs, which exhibit significant reversal ability towards DMCM- or zinc-induced inhibition of GABA-gated currents are represented by solid bars. ***P<0.001, **P<0.01 and *P<0.05 using Student's t-test analysis.

Figure 6

Interaction between LEV and negative allosteric modulators of the glycine receptor. Left panel: Three to five DIV spinal cord neurons (A, C and E) and 4 – 21 DIV hippocampal neurons (G) were perfused for 10 s with 100 μM glycine alone (left traces) or in combination with LEV (25 s, 10 μM) (right traces) (controls) and negative allosteric modulators of the glycine receptor (25 s, half-maximal inhibitory concentrations: 1 μM strychnine (STR), 15 μM DMCM and 15 or 500 μM zinc). A 60 s period was allowed for the washout of drugs (90 s in the case of DMCM). Right panel: Currents evoked by 100 μM glycine in cultured spinal cord or hippocampal neurons in the presence of the indicated negative allosteric modulator of the glycine receptor were recorded for increasing LEV concentrations. Results are expressed as in Figure 3 (mean±s.e.mean; n=4 – 20 for B, 5 – 33 for D, 5 – 33 for F and 5 – 44 for H). The curve fitting procedure yielded a half-maximal effective LEV concentration of 0.7±0.1 μM in spinal cord neurons and 0.04±0.02 μM in hippocampal neurons for zinc-induced inhibitions of glycine currents. ***P<0.001 and *P<0.05 using ANOVA followed by Dunnett's multiple comparisons post-tests.

Figure 7

Comparison of the potencies of reference AEDs in reversing the inhibitory effects of DMCM (A) and zinc (B) on glycine-induced currents in 3 – 5 DIV cultured spinal cord neurons. Currents evoked by 100 μM glycine in the presence of 15 μM DMCM or 15 μM ZnCl2 were recorded with the different AEDs (same concentrations as in Figure 4). Results are expressed and represented as described in Figure 4 (mean±s.e.mean; n=5 – 19).

Figure 8

Effect of FG 7142 (A) and flumazenil (B) on the protection afforded by LEV and reference AEDs against clonic convulsions in sound-sensitive mice. LEV (17 mg kg−1), carbamazepine (23.6 mg kg−1), clonazepam (0.032 mg kg−1), phenobarbital (14.2 mg kg−1), phenytoin (45.4 mg kg−1), valproate (230 mg kg−1) and ethosuximide (254 mg kg−1) were all administered i.p. 30 min before testing. The effect of FG 7142 (10 mg kg−1 i.p., 15 min before testing) and flumazenil (5 mg kg−1 p.o., 15 min before testing) was determined on the seizure protection afforded by LEV and the reference AEDs against clonic convulsions in audiogenic-susceptible mice. The protection given by LEV and reference AEDs administered alone was (%): (A) LEV (90), carbamazepine (90), clonazepam (80), phenobarbital (100), phenytoin (70), valproate (100) and ethosuximide (80); (B) LEV (90), carbamazepine (90), clonazepam (90), phenobarbital (100), phenytoin (60), valproate (100) and ethosuximide (70). To allow a direct comparison between the different drugs, the results from the interaction studies were expressed as percentage of a maximal protective effect (100%) obtained with each AEDs and indicated by the horizontal line. *P<0.05 using a Fisher test versus AED administered alone. (a) a higher dose FG 7142 (50 mg kg−1) was used for carbamazepine and phenytoin.