Trimebutine [2-dimethylamino-2-phenylbutyl-3,4,5-trimethoxybenzoate hydrogen maleate (TMB)] has been demonstrated to be active for relieving abdominal pain in humans. To better understand its mechanism of action, we have tested TMB; nor-TMB, its main metabolite in humans; and their respective stereoisomers for their affinity toward sodium channels labeled by [3H]batrachotoxin, their effect on sodium, potassium, and calcium currents in rat dorsal root ganglia neurons, and their effect on veratridine-induced glutamate release from rat spinal cord slices. TMB has also been tested in an animal model of local anesthesia. TMB (Ki = 2.66 +/- 0.15 microM) and nor-TMB (Ki = 0.73 +/- 0.02 microM) displaced [3H]batrachotoxin from its binding site with affinities similar to that of bupivacaine (Ki = 7.1 +/- 0.9 microM). nor-TMB was found to block veratridine-induced glutamate release with an IC50 value of 8.5 microM, which is very similar to that of bupivacaine (IC50 = 8.2 microM); the effect of TMB was limited to 50% inhibition at 100 microM. TMB and nor-TMB blocked sodium currents in sensory neurons from rat dorsal root ganglia (IC50 = 0.83 +/- 0.09 and 1.23 +/- 0.19 microM, respectively), whereas no effect was observed on calcium currents at the same concentrations. A limited effect was observed on potassium currents (IC50 = 23 +/- 6 at 10 microM) for TMB. In vivo, when tested in the rabbit corneal reflex, TMB displayed a local anesthetic activity 17-fold more potent than that of lidocaine.