Levetiracetam has no acute effects on brain γ-aminobutyric acid levels

Abstract

Objective

The mechanism of action of levetiracetam (LEV), an antiepileptic drug, is related to a novel binding site, SV2, but LEV acts on GABA-A receptors. The objective of the study described here was to determine if LEV modulates brain GABA in vivo.

Methods

Concentrations of cerebral GABA and serum LEV were obtained in seven healthy individuals using ¹H magnetic resonance spectroscopy at baseline and 3 and 6 hours following oral administration of 1 g of LEV.

Results

Brain cerebral GABA acutely concentrations did not change from baseline.

Conclusion

The results indicate that LEV docs not increase human cerebral GABA concentrations acutely in healthy individuals.

1. Introduction

γ-Aminobutyric acid (GABA) is the major inhibitory neurotransmitter in the brain. Several antiepileptic drugs (AEDs) enhance GABA function, resulting in an enhanced anticonvulsant effect [1]. GABA-agonistic AEDs, including barbiturates, benzodiazepines, tiagabine, vigabatrin, gabapentin, and topiramate, increase GABA activity and/or concentrations by different mechanisms. Levetiracetam (LEV) is a novel AED approved for the treatment of partial onset, myoclonic and primary generalized tonic clonic seizures [2]. Multiple mechanisms of action have been proposed for LEV, with attention most recently focused on SV2 receptor binding [3]. This study reports our findings on the effect of LEV on cerebral GABA concentrations in healthy individuals.

2. Methods

We studied seven healthy subjects (three women) of mean age 28.9 years (range, 28–34 years). All subjects were screened for neurological, systemic, or psychiatric disorders. Subjects received physical and laboratory examinations including a toxicology screen. Subjects were asked to refrain from drinking alcoholic beverages, smoking, or taking any medications for at least 48 hours before the experiment. All subjects gave full consent for the study, which was approved by the New York University institutional review board.

2.1. Experimental design

The group underwent a baseline GABA magnetic resonance spectroscopy (MRS) study prior to the acute dosing experiment, Each subject then received a single oral dose of 1 g of LEV. After ingesting the LEV, the subjects were under continuous medical observation. Three hours after ingestion, the individuals were placed in the magnet, and GABA measurements were obtained. At termination, the subjects were removed from the scanner and allowed to rest. The same procedure was repeated 6 hours after administration of the initial dose. Serum LEV levels were obtained at 3 and 6 hours, concurrently with each MRS study. LEV levels were measured with a validated chromatographic method.

2.2. MRS study

GABA was measured using a Varian INOVA 4-T whole-body imaging system with a ¹H TEM volume head coil. Volume localized spectra were obtained using a PRESS sequence and CHESS water suppression with TE = 68 ms and TR = 2000 ms. A 4 × 4 × 2-cm³ volume (2 cm in the interior–superior direction) was positioned within the occipital lobe (Fig. 1), To suppress overlap with other resonances, such as those of creatine, glutamate, and glutamine, single-quantum spectral editing was created by application of a 20-ms DANTE inversion pulse shaped with a Gaussian window to the GABA C-2 resonance (1.9 ppm) and a position symmetrically disposed about the 1.7-ppm macromolecule resonance. Subtraction of the resulting spectra results in a GABA edited spectrum or the C-4 GABA resonance at 3.0 ppm. As described by Henry et al. [4], macromolecule contributions [5] to the edited spectrum are eliminated by symmetrical placement of the editing pulse. Residual differences in the refocusing of creatine and choline due lo the inversion pulse were eliminated by use of a postacquisition phase correction (Fig. 1). The acquisitions were interleaved on a scan-by-scan basis to minimize subtraction errors and stored in separate bins [6]. Two or three 8.5-minute acquisitions (17 or 25 minutes, respectively) were averaged to generate a final spectrum.

Fig. 1.

Scout image showing placement of the voxel. (a, b) Spectra acquired with the editing pulse applied at 1.5 and 1.9 ppm, respectively. (c) Difference spectrum displayed using a vertical scaling of ×8. (d) Expansion of the spectral region about the GAB A resonance at a vertical scale of ×32. Data in this example were acquired from the average of two 8.5-minute acquisitions.

To minimize the effect of study variation in signal amplitude resulting from small differences in subject loading, the GABA levels were referenced to the creatine area determined from that study. This was determined by fitting the creatine resonance in the spectral domain. The GABA resonance area was modeled as a doublet (outer lines of the triplet) with a 14-Hz separation. To convert the measured areas to millimolar values, a creatine concentration of 7.1 mM, corresponding to a 50:50 mixture of gray and white matter, was assumed along with equivalent T2 losses for creatine and GABA. To our knowledge, the T2 of GABA at 4 T in human brain has not been measured. However, al 4 T in the human brain, typical metabolite T2 values range from 140 ms (creatine) to 233 ms for NAA [7]. Assuming a T2 of 233 ms would result in a 19% overestimate of the GABA concentration.

2.3. Statistical analysis

The millimolar values of GABA for each subject were tabulated at baseline and at 3 and 6 hours after administration of LEV. The percentage increase in GABA level was determined for each individual, and the mean and SD of the group were reported. Statistical significance was determined by comparing the millimolar values at baseline with levels at 3 and 6 hours using a single-tailed t test.

3. Results

No serious side effects were observed. Side effects were reported by five individuals and consisted of drowsiness. On follow-up, subjects reported no side effects 24 hours after the test dose.

Data from one patient were not used due to suboptimal-quality spectra in one study. Fig. 2 illustrates the mean group cerebral GABA concentrations at baseline and 3 and 6 hours following administration of 1 g of LEV orally to six individuals. The group mean baseline concentration of cerebral GABA was 0.93 ± 0.09 mM. No significant increase was observed at 3 and 6 hours. As expected, significant increases in serum LEV concentrations were observed at 3 hours (mean; 22 μg/mL) and 6 hours (mean: 17.4 μg/mL).

Fig. 2.

Mean and SD brain GABA and LEV concentrations for the group at baseline and after administration of LEV 1 g orally. Mean GABA and SD at 3 and 6 hours indicate no statistically significant increases in brain GABA concentrations.

4. Discussion

Improvements in nuclear magnetic resonance techniques have made it possible to measure brain neurotransmitter pools in vivo. Previous studies by our group and others have demonstrated significant increases in GABA concentrations within 2–3 hours of a single oral dose of different AEDs, including topiramate, gabapentin, vigabatrin, and lamotrigine [8–10]. Petroff et al. have suggested that these changes are due to inhibition of GABA transaminase or an increase in nonvesicular GABA release [11].

LEV is a novel anticonvulsant drug with multiple mechanisms of action. Although a stereoselective binding site, SV2, appears to be crucial to its effect, other data suggest that LEV has other mechanisms of action. LEV does not modulate neuronal voltage-gated sodium channels, T-type calcium currents, or glutamate channels. Studies have, however, indicated that LEV promotes inhibition by reducing negative allosteric effects of β-carbolines on GABA-A receptors [12]. As our group and others have previously demonstrated, it is possible to fail to detect any in vitro or in vivo effects on GABA in different animal models while observing definitive GABA changes in humans [8,13].

This is the first in vivo human study indicating that LEV does not increase human cerebral GABA concentrations in healthy subjects within 3 or 6 hours of administration of 1 g. Although, no significant change was observed in the occipital lobe, in contrast to what is observed with other AEDs (topiramate, lamotrigine, and gabapentin, reflecting a general effect on brain GABA), the possibility that other brain regions may have been specifically affected is not precluded. For example, LEV is currently used to treat patients with seizures involving thalamocortical circuits. However, we have no reason to believe that GABA concentrations, if elevated with LEV, involve specific cortical regions. Interestingly, LEV has been reported to increase homocarnosine, a dipeptide of GABA and histidine, while not increasing GABA in patients with epilepsy [14]. In this study, the total GABA resonance was measured, combining both homocarnosine and free GABA. Thus, reciprocal changes in the two compounds (increased homocarnosine and decreased free GABA) would not have been detectable, Whether an increase in brain GABA occurs with chronic administration of LEV is also not known. Recently, Stefan et al. [15] reported that the time to effect of an acute dose is 2 days, such that secondary effects due to drug administration, including upregulation of different enzymatic pathways, may result in delayed increases. Regardless, contrary to gabapentin, vigabatrin and topiramate, no acute increase was observed in LEV levels.