Endogenous GABA Does Not Mediate the Inhibitory Effects of Gabapentin

Home , Semicarbazide

Endogenous GABA Does Not Mediate the Inhibitory Effects of Gabapentin on Spinal Reflexes in Rats Shinobu Shimizu1, Motoko Honda1, Mitsuo Tanabe1, Jun-Ichiro Oka2, and Hideki Ono1,*

1Laboratory of CNS Pharmacology, Graduate School of Pharmaceutical Sciences, Nagoya City University, 3-1 Tanabe-dori, Mizuho-ku, Nagoya 467-8603, Japan 2Laboratory of Pharmacology, Faculty of Pharmaceutical Sciences, Tokyo University of Science, 2641 Yamazaki, Noda, Chiba 278-8510, Japan

Received October 21, 2003; Accepted December 17, 2003

Abstract. The novel antiepileptic drug gabapentin was designed as a structural analog of - aminobutyric acid (GABA). However, its mechanism of action remains unclear. In the present study, we investigated the effect of gabapentin on spinal reflexes in anesthetized rats. The mono- and polysynaptic reflex potentials were recorded from the ipsilateral L5 ventral root after stimulation of the L5 dorsal root. The dorsal root reflex potential, an index of presynaptic inhibition, was recorded from the ipsilateral L4 dorsal root. In non-spinalized (intact) and spinalized rats, intravenously administered gabapentin reduced the mono- and polysynaptic reflex potentials in a dose-dependent manner. These inhibitory effects of gabapentin were not sup- pressed by the GABAA antagonist picrotoxin. Moreover, gabapentin also decreased spinal reflexes in spinalized rats depleted of spinal GABA with semicarbazide, an inhibitor of the GABA-synthesizing enzyme. The dorsal root reflex potentials were not affected by gabapentin. These results suggest that endogenous GABA does not mediate the inhibitory effects of gabapentin on spinal reflexes.

Keywords: spinal reflex, gabapentin, endogenous GABA, dorsal root reflex, pregabalin

Introduction and does not affect GABA uptake and degradation (10, 11). Instead, gabapentin shows specific binding Gabapentin, 1-(aminomethyl)cyclohexaneacetic acid, affinity for 2 subunits of voltage-dependent calcium is a novel anticonvulsant agent that is effective in channels (12). Moreover, similar specific affinity for various animal seizure models (1, 2) and that has been 2 subunits was demonstrated for pregabalin, another shown in several placebo-controlled clinical studies to GABA derivative with antiepileptic and antinociceptive prevent partial seizures and generalized tonic-clonic actions (12). This activity profile may be essential for seizures, both as add-on therapy and monotherapy (3 – some aspects of the pharmacological actions of these 5). Furthermore, it has been shown that gabapentin compounds. exhibits antinociceptive effects on hyperalgesia (6 – 8) It has been demonstrated that gabapentin elevates the and anxiolytic-like effects (8). GABA concentration in human (13) and rat (14) brains The cellular mechanisms of the pharmacological and increases GABA release from rat striatal slices (15). actions of gabapentin have not been completely deter- Similarly, pregabalin increases neuronal GABA content mined. Gabapentin was synthesized as a structural (16). Thus, gabapentin and pregabalin can exert indirect analog of -aminobutyric acid (GABA) and it penetrates GABAergic actions via an increase in the GABA the blood-brain barrier. However, gabapentin does not concentration in the central nervous system, which exhibit affinity for known neurotransmitter binding may underlie their antiepileptic effects. Previous studies sites, including the GABAA and GABAB receptors (9), have shown that the spinal reflexes are susceptible to a change in GABAergic inhibitory tone. The GABAA *Corresponding author. FAX: +81-52-836-3676 agonist muscimol and the GABAB agonist baclofen E-mail: [email protected]

137 138 S Shimizu et al decrease mono- and polysynaptic reflexes (17 – 19). synaptic reflex peak from the monosynaptic reflex peak In addition, the dorsal root reflex, mediated by spinal was 0.8 – 1 ms. The latency of the dorsal root reflex intrinsic GABAergic interneurons, is increased by was 3–4ms. benzodiazepines (20). In the present study we examined the effects of gabapentin (and pregabalin in some Drugs experiments) on spinal reflexes in an attempt to clarify Gabapentin (1-(aminomethyl)cyclohexaneacetic acid) the involvement of indirect GABAergic inhibition in and pregabalin ((S)-(+)-3-(aminomethyl)-5-methyl- the spinal cord. hexanoic acid) were kindly supplied by Pfizer (Ann Arbor, MI, USA). Muscimol, picrotoxin, and urethane Materials and Methods were obtained from Sigma Aldrich (St. Louis, MO, USA). Semicarbazide hydrochloride and -chloralose Surgery were obtained from Tokyo Kasei (Tokyo). Urethane All experimental protocols were approved by the and -chloralose were both dissolved in distilled water. Animal Care and Use Committee of Nagoya City All the test compounds were dissolved in 0.9% w/v University and Tokyo University of Science, and they physiological saline and administered intravenously at were conducted in accordance with the guidelines of 1ml/kg at least 2 h after the C1 transection. Control the National Institutes of Health and the Japanese rats received vehicle at 1 ml/kg. When gabapentin was Pharmacological Society. examined under blockade of GABA synthesis, semicar- Male Wistar rats weighing 200 – 300 g (NRC Haruna, bazide hydrochloride (200 mg/kg, i.v.) was pretreated Gunma) were anesthetized with urethane (1 g/kg, intra- after laminectomy. peritoneally (i.p.)) and -chloralose (25 mg/kg, i.p.). Cannulae were inserted into the trachea and the femoral Statistical analyses vein for artificial respiration and drug administration, The amplitudes of the mono- and polysynaptic reflex respectively. To produce spinalized rats, the vagus and dorsal root reflex potentials recorded after drug nerves were cut bilaterally in the cervical region to administration were expressed as percentages of the eliminate parasympathomimetic effects on the heart, corresponding predrug (time 0) values. All data were and the spinal cord was transected at the C1 level under expressed as the means M S.E.M. Student’s t-test was lidocaine anesthesia (4%, 50 l). A dorsal laminectomy used to compare the data between two groups, and the was performed in the lumbo-sacral region of each rat. t-test with Bonferroni correction following one-way Both the ventral and dorsal roots below L4 were cut analysis of variance (ANOVA) was used for multiple distally at their points of exit from the vertebral column, comparisons of control and treated groups (21). Differ- and the entire exposed surgical area was covered with ences at P0.05 (two-tailed) were considered to be liquid paraffin that was maintained at 36 M 0.5LC by significant. radiant heat. Rectal temperature was maintained at 36 M 0.5LC. Results

Measurement of the spinal reflex potentials Effect of gabapentin on monosynaptic, polysynaptic, and For the recording of mono- and polysynaptic reflexes, dorsal root reflexes the dorsal and ventral roots of the L5 segment were In non-spinalized (intact) rats, gabapentin reduced placed on bipolar Ag-AgCl wire electrodes for stimu- the mono- and polysynaptic reflex potentials in a dose- lation (5 V, 0.2 Hz, 0.05 ms; SEN-7103; Nihon Kohden, dependent manner. The amplitude of the monosynaptic Tokyo) and recording, respectively (inset in Fig. 1A). reflex potential was reduced by gabapentin to 92.9 M The dorsal root reflex was recorded from the L4 dorsal 6.3% (n 5) and 70.9 M 9.7% (n 5) at doses of 10 root with a bipolar Ag-AgCl wire electrode following and 30 mg/kg, i.v., respectively (60 min after admin- stimulation of the ipsilateral L5 dorsal root (5 V, istration) (Fig. 1A). The amplitude of the polysynaptic 0.2 Hz, 0.05 ms). The reflex potentials were amplified reflex potential was significantly reduced by gabapentin (AVB-10, Nihon Kohden), displayed on an oscilloscope to 47.3 M 8.2% and 12.8 M 2.4% at doses of 10 and (VC-10, Nihon Kohden) and averaged eight times by an 30 mg/kg, i.v., respectively (60 min after administra- averager (DAT-1100, Nihon Kohden). The amplitudes tion) (Fig. 1B). In spinalized rats, gabapentin also re- of mono- and polysynaptic reflex potentials and dorsal duced the mono- and polysynaptic reflex potentials in a root reflex potentials were measured (insets in Figs. 1A dose-dependent manner. The amplitude of the mono- and 2A). The latency of the monosynaptic reflex synaptic reflex potential was reduced by gabapentin to potential was 1.5 – 2 ms, and the delay of the poly- 88.5 M 11.4% (n 5) and 65.2 M 5.2% (n 5) at doses of Gabapentin and Spinal Reflexes 139

Fig. 1. Gabapentin suppresses mono- and polysynaptic reflex potentials. Gabapentin was administered intravenously (open circle: control, closed circle: 3 mg/kg, open triangle: 10 mg/kg, and closed triangle down: 30 mg/kg). A and B: effects in intact rats on the mono- and polysynaptic reflex potentials, respectively. C and D: effects in spinalized rats on the mono- and polysynaptic reflex potentials, respectively. Each point represents the mean M S.E.M. of 5 animals. Abscissae: time in minutes after the injection of gabapentin. Ordinate: amplitude of spinal reflexes expressed as a percentage of the corresponding values at time 0. The significance of the differences between the test and control values was determined by the two-tailed Bonferroni-type multiple t-test following one-way analysis of variance (ANOVA) (2 comparisons in 3 groups); *P0.05 and **P0.01 vs control. Inset shows a schematic representation of the experimental situation and a sample trace of the mono- and polysynaptic reflex potentials.

3 and 10 mg/kg, i.v., respectively (60 min after admin- Influence of GABAA-receptor antagonist on gabapentin- istration) (Fig. 1C). The amplitude of the polysynaptic induced effects in spinalized rats reflex potential was significantly reduced by gabapentin In the following studies, we used a higher dose of to 51.7 M 10.1% and 28.2 M 7.2% at doses of 3 and gabapentin (30 mg/kg), since this dose of gabapentin 10 mg/kg, i.v., respectively (60 min after administra- exhibited a more stable reduction of the monosynaptic tion) (Fig. 1D). In both intact and spinalized rats, the reflex than 10 mg/kg. inhibitory effects of gabapentin appeared gradually after In spinalized rats, the GABAA-receptor antagonist administration, and gabapentin reduced the polysynaptic picrotoxin at a dose of 3 mg/kg alone had no effect on reflex more strongly than the monosynaptic reflex. the monosynaptic reflex potential (n 6, Fig. 3A). This Stimulation of the L5 dorsal root produces the dorsal dose of picrotoxin sufficed to block GABAA receptor- root reflex potential in the L4 dorsal root. Gabapentin mediated responses, as shown in Fig. 3B. Muscimol (10 mg/kg, i.v.) did not affect the dorsal root reflex (3 mg/kg) gradually decreased the monosynaptic reflex (Fig. 2). potential to 53.5 M 7.0% at 60 min after administration, 140 S Shimizu et al

Fig. 2. Gabapentin does not change the dorsal root reflex potential in spinalized rats. Each point represents the mean M S.E.M. of 5 animals. Abscissae: time in minutes after the injection of gabapentin. Ordinate: amplitude of spinal reflexes expressed as a percentage of the corresponding values at time 0. The significance of the differences between the test (closed circle: 10 mg/kg, i.v.) and control (open circle) values was determined by the two-tailed Student’s t-test; *P0.05 vs control. Inset shows a schematic representation of the experimental situation and a sample trace of the dorsal root reflex potential. which was partially reversed by picrotoxin injected 30 min after muscimol (inhibition to 74.52 M 6.7% at 30 min after picrotoxin, n 7). However, picrotoxin failed to reverse the gabapentin-induced decrease of monosynaptic reflex potential (inhibition to 58.5 M 10.8% and 63.7 M 14.5% at 60 min after gabapentin in the control and picrotoxin-treated groups, respectively, n 6; Fig. 3C).

Influence of semicarbazide on gabapentin-induced effects in spinalized rats GABA is synthesized from glutamate by glutamic acid decarboxylase (GAD) at GABAergic terminals. Systemic administration of semicarbazide hydrochloride (200 mg/kg, i.v.), an inhibitor of GAD, was shown to deplete GABA to about 15% in the cat spinal cord in 3 h (22). In the present study, the same dose of semi- Fig. 3. The GABAA-receptor antagonist picrotoxin does not counteract the inhibition of the monosynaptic reflex potential by carbazide hydrochloride resulted in a gradual increase gabapentin in spinalized rats. A: picrotoxin (open circle: control, in the monosynaptic reflex to 142.4 M 3.6% in 3 h, and closed circle: 3 mg/kg, i.v.) alone did not alter the monosynaptic this enhancement was maintained for at least 1 h reflex. B: the inhibitory effect of muscimol (3 mg/kg, i.v.) was (n 4, Fig. 4A), consistent with a marked decrease in partially reversed by picrotoxin (open triangle: muscimol, closed triangle: +picrotoxin). C: the inhibitory effect of gabapentin (30 mg GABA in the spinal cord (22). Nevertheless, gabapentin /kg, i.v.) was not affected by picrotoxin (open triangle down: still decreased the spinal reflex (Fig. 4B). gabapentin, closed triangle down: +picrotoxin). Picrotoxin was administered 30 min after muscimol or gabapentin. Each point M Effect of pregabalin on the mono- and polysynaptic represents the mean S.E.M. of 5 – 7 animals. Abscissae: time in minutes. Ordinate: amplitude of spinal reflexes expressed as a reflexes percentage of the corresponding values at time 0. The significance In spinalized rats, pregabalin reduced the mono- and of the differences between the test and control values was determined polysynaptic reflex potentials in a dose-dependent by the two-tailed Student’s t-test; *P0.05 vs control. Gabapentin and Spinal Reflexes 141

Fig. 4. The GABA-synthesis inhibitor semicarbazide does not Fig. 5. Pregabalin suppresses spinal reflex potentials in spinalized impair the effect of gabapentin on the monosynaptic reflex in rats. Pregabalin was administered intravenously (open circle: control, spinalized rats. A: semicarbazide hydrochloride (open circle: control, closed circle: 3 mg/kg, and open triangle: 10 mg/kg). A: effects on closed circle: 200 mg/kg, i.v.) gradually increased the monosynaptic the monosynaptic reflex. B: effects on the polysynaptic reflex. reflex. B: semicarbazide 3 h before gabapentin (open triangle: Each point represents the mean M S.E.M. of 5 animals. Abscissae: 30 mg/kg, closed triangle: +semicarbazide hydrochloride 200 mg time in minutes after the injection of pregabalin. Ordinate: amplitude /kg, i.v.) did not affect the inhibitory effect of gabapentin. Each of spinal reflexes expressed as a percentage of the corresponding point represents the mean M S.E.M. of 4 – 5 animals per groups. values at time 0. The significance of the differences between the Abscissae: time in min. Ordinate: amplitude of spinal reflexes test and control values was determined by the two-tailed Bonferroni- expressed as a percentage of the corresponding values at time 0. type multiple t-test following one-way analysis of variance The significance of the differences between the test and control (ANOVA) (2 comparisons in 3 groups), *P0.05 and **P0.01 vs values was determined by the two-tailed Student’s t-test; *P0.05 control. and **P0.01 vs control.

manner. The amplitude of the monosynaptic reflex the polysynaptic reflex more strongly than the mono- potential was reduced by pregabalin to 89.6 M 5.2% and synaptic reflex. 76.6 M 5.1% at doses of 3 and 10 mg/kg, i.v., respectively (60 min after administration) (n 5, Fig. 5A). Discussion The amplitude of the polysynaptic reflex potential was significantly reduced by pregabalin to 50.8 M 11.6% and The spinal reflex employed in the present study 24.4 M 3.5% at doses of 3 and 10 mg/kg, i.v., respec- consists of the mono- and polysynaptic reflexes. tively (60 min after administration) (n 5, Fig. 5B). Glutamatergic monosynaptic excitation of motoneurons The inhibitory effects of pregabalin appeared gradually (23) is evoked by excitation of group Ia primary afferent after intravenous administration, and pregabalin reduced fibers originating from muscle spindles located in 142 S Shimizu et al skeletal muscles. The polysynaptic reflex measured in transmission. this study is evoked by disynaptic excitation of moto- In conclusion, we showed that gabapentin decreased neurons via one interneuron. There is GABAergic syn- the spinal reflexes in intact and spinalized rats by a aptic inhibition on these spinal reflex transmissions (24). process not mediated by endogenous GABA. We have shown that gabapentin decreased the mono- and polysynaptic reflexes in intact and spinalized rats Acknowledgment (Fig. 1: A – D). While studies have been focused on the well-known effect on chronic pain at the spinal cord We would like to acknowledge the assistance of (25 – 27), we now for the first time revealed that Pfizer, who supplied gabapentin and pregabalin. gabapentin can reduce motor output from the spinal cord at the spinal level. Although it is controversial whether gabapentin References directly activates GABA receptors, indirect GABAergic 1 Dalby NO, Nielsen EB. Comparison of the preclinical anti- inhibition using endogenous GABA (13 – 15) might convulsant profiles of tiagabine, lamotrigine, gabapentin and mediate some pharmacological actions elicited by vigabatrin. Epilepsy Res. 1997;28:63–72. gabapentin including depression of the spinal reflex in 2 Xiong ZQ, Stringer JL. Effects of felbamate, gabapentin and this study. The observations reported here, however, lamotrigine on seizure parameters and excitability in the rat exclude this possibility. First, blockade of GABA syn- hippocampus. Epilepsy Res. 1997;27:187–194. thesis by the GAD inhibitor semicarbazide did not affect 3 Bergey GK, Morris HH, Rosenfeld W, et al. Gabapentin monotherapy: I. An 8-day, double-blind, dose-controlled, multicenter the action of gabapentin (Fig. 4B). Secondly, the dorsal study in hospitalized patients with refractory complex partial or root reflex, which is mediated by spinal intrinsic secondarily generalized seizures. Neurology. 1997;49:739–745. GABAergic interneurons (24), was not affected by 4 Beydoun A, Fischer J, Labar DR, et al. Gabapentin mono- gabapentin (Fig. 2). Finally, depression of the monotherapy: II. A 26-week, double-blind, dose-controlled, multi- synaptic reflex by gabapentin was not reversed by the center study of conversion from polytherapy in outpatients with GABAA-receptor antagonist picrotoxin (Fig. 3C). refractory complex partial or secondarily generalized seizures. Furthermore, the lack of change in the dorsal root reflex Neurology. 1997;49:746–752. 5 Chadwick D, Leiderman DB, Sauermann W, Alexander J, reveals that GABAB receptors do not mediate the actions Garofalo E. Gabapentin in generalized seizures. Epilepsy Res. of gabapentin, since the selective GABAB-receptor 1996;25:191–197. agonist baclofen strongly decreases the dorsal root 6 Field MJ, McCleary S, Hughes J, Singh L. Gabapentin and reflex (17). In addition, the preferential effects of pregabalin, but not morphine and amitriptyline, block both static baclofen on monosynaptic reflex rather than poly- and dynamic components of mechanical allodynia induced by synaptic reflex transmission support this conclusion streptozocin in the rat. Pain. 1999;80:391–398. (17, 19). However, presently, there exist several lines of 7 Jun JH, Yaksh TL. The effect of intrathecal gabapentin and 3- evidence that argue for (28, 29) and against (30) the isobutyl -aminobutyric acid on the hyperalgesia observed after thermal injury in the rat. Anesth Analg. 1998;86:348–354. mediation of GABAB receptors in the action of 8 Singh L, Field MJ, Ferris P, et al. The antiepileptic agent gabapentin. Moreover, gabapentin may reveal a specific gabapentin (Neurontin) possesses anxiolytic-like and anti- agonism of the GABAB-receptor subtypes (29). Further nociceptive actions that are reversed by D-serine. Psycho- studies will be required to clarify the involvement of pharmacology (Berl). 1996;127:1–9. GABAB receptors in the inhibition of spinal reflexes by 9 Taylor CP. Gabapentin: Mechanism of action. In: Levy RH, gabapentin. Mattson RH, Meldrum BS, editors. Antiepileptic drugs. New The inhibitory effects of pregabalin on spinal reflexes York: Raven Press; 1995. p. 829–841. 10 Goldlust A, Su TZ, Welty DF, Taylor CP, Oxender DL. Effects were similar to those of gabapentin; both drugs produced of anticonvulsant drug gabapentin on the enzymes in metabolic stronger effects on the polysynaptic reflex than on the pathways of glutamate and GABA. Epilepsy Res. 1995;22:1–11. monosynaptic reflex, and the onsets of effects were 11 Su TZ, Lunney E, Campbell G, Oxender DL. Transport of slow (Figs. 1 and 5). Moreover, there is evidence that gabapentin, a -amino acid drug, by system L -amino acid pregabalin increases neuronal GABA content (16) and transporters: a comparative study in astrocytes, synaptosomes, that it selectively binds to the 2 subunit of calcium and CHO cells. J Neurochem. 1995;64:2125–2131. channels with high affinity in the nanomolar range (12). 12 Gee NS, Brown JP, Dissanayake VU, Offord J, Thurlow R, Woodruff GN. The novel anticonvulsant drug, gabapentin Together with the present results, these observations (Neurontin), binds to the 2 subunit of a calcium channel. suggest that gabapentin and pregabalin may inhibit J Biol Chem. 1996;271:5768–5776. spinal reflex transmission via binding to the 2 subunit. 13 Petroff OA, Rothman DL, Behar KL, Lamoureux D, Mattson Further studies are required to clarify the detailed RH. The effect of gabapentin on brain gamma-aminobutyric acid mechanisms of action of gabapentin on spinal reflex in patients with epilepsy. Ann Neurol. 1996;39:95–99. Gabapentin and Spinal Reflexes 143

14 Löscher W, Hönack D, Taylor CP. Gabapentin increases reflex potentials in rats: an in vivo study. Br J Pharmacol. aminooxyacetic acid-induced GABA accumulation in several 1995;114:1193–1205. regions of rat brain. Neurosci Lett. 1991;128:150–154. 24 Eccles JC, Schmidt R, Willis WD. Pharmacological studies on 15 Götz E, Feuerstein TJ, Lais A, Meyer DK. Effects of gabapentin presynaptic inhibition. J Physiol. 1963;168:500–530. on release of -aminobutyric acid from slices of rat neostriatum. 25 Luo ZD, Calcutt NA, Higuera ES, et al. Injury type-specific Arzneimittelforschung. 1993;43:636–638. calcium channel 2-1 subunit up-regulation in rat neuropathic 16 Whitworth TL, Quick MW. Upregulation of -aminobutyric acid pain models correlates with antiallodynic effects of gabapentin. transporter expression: role of alkylated -aminobutyric acid J Pharmacol Exp Ther. 2002;303:1199–1205. derivatives. Biochem Soc Trans. 2001;29:736–741. 26 Luo ZD, Chaplan SR, Higuera ES, et al. Upregulation of dorsal 17 Fukuda H, Kudo Y, Ono H. Effects of -(p-chlorophenyl)- root ganglion 2 calcium channel subunit and its correlation GABA (baclofen) on spinal synaptic activity. Eur J Pharmacol. with allodynia in spinal nerve-injured rats. J Neurosci. 2001; 1977;44:17–24. 21:1868–1875. 18 Kaneko T, Ono H, Fukuda H. Decrease in effects of muscimol 27 Shimoyama M, Shimoyama N, Hori Y. Gabapentin affects on the dorsal root after repeated administration of diazepam in glutamatergic excitatory neurotransmission in the rat dorsal rats. J Pharmacobiodyn. 1981;4:947–951. horn. Pain. 2000;85:405–414. 19 Ono H, Fukuda H, Kudo Y. Mechanisms of depressant action of 28 Bertrand S, Ng GY, Purisai MG, et al. The anticonvulsant, baclofen on the spinal reflex in the rat. Neuropharmacology. antihyperalgesic agent gabapentin is an agonist at brain - 1979;18:647–653. aminobutyric acid type B receptors negatively coupled to 20 Goto M, Ono H, Matsumoto K, Kondo M, Fukuda H. Effects voltage-dependent calcium channels. J Pharmacol Exp Ther. of zopiclone and benzodiazepines on spinal reflexes, anemic 2001;298:15–24. decerebrate rigidity and benzodiazepine binding. Jpn J Pharma- 29 Ng GY, Bertrand S, Sullivan R, et al. -Aminobutyric acid col. 1983;33:1241–1246. type B receptors with specific heterodimer composition and 21 Wallenstein SV, Zucker CL, Fleiss JL. Some statistical methods postsynaptic actions in hippocampal neurons are targets of useful in circulation research. Circ Res. 1980;47:1–9. anticonvulsant gabapentin action. Mol Pharmacol. 2001;59:144– 22 Bell JA, Anderson EG. The influence of semicarbazide-induced 152. depletion of -aminobutyric acid on presynaptic inhibition. 30 Jensen AA, Mosbacher J, Elg S, et al. The anticonvulsant Brain Res. 1972;43:161–169. gabapentin (Neurontin) does not act through -aminobutyric 23 Farkas S, Ono H. Participation of NMDA and non-NMDA acid-B receptors. Mol Pharmacol. 2002;61:1377–1384. excitatory amino acid receptors in the mediation of spinal