To elucidate the mechanism of action of the anticonvulsant gabapentin (GBP), we compared its effects on K+-evoked [3H]-noradrenaline ([3H]-NA) release from rat hippocampal and human neocortical slices with those of the KATP channel opener pinacidil and the Na+ channel blockers phenytoin, carbamazepine and lamotrigine. Rat hippocampal and human neocortical slices were loaded with [3H]-NA and superfused. [3H]-NA release was evoked by increasing the extracellular [K+] from 3 to 15 mM. GBP decreased [3H]-NA release from rat hippocampal with a pIC50 of 5.59 and a maximum inhibition of 44%. Concentration-dependent inhibition was also seen in human neocortical slices (39% inhibition with 100 microM GBP). These inhibitory effects were antagonized by the KATP channel antagonist glibenclamide, yielding a pA2 of 7.50 in the rat. The KATP channel opener pinacidil (10 microM), like GBP, decreased [3H]-NA release from rat hippocampal slices by 27% and this effect was also antagonized by glibenclamide. In human neocortical slices the inhibition by pinacidil (10 microM) was 31%. Although phenytoin (10 microM), carbamazepine (100 microM) and lamotrigine (10 microM) also decreased [3H]-NA release (by 25%, 57% and 22%, respectively), glibenclamide did not antagonize the effects of these classical Na+ channel blockers. These findings suggest that GBP inhibits K+-evoked [3H]-NA release through activation of KATP channels. To establish whether the KATP channels under investigation were located on noradrenergic nerve terminals or on other neuronal elements, the effects of GBP were compared in the absence and in the presence of tetrodotoxin (TTX 0.32 microM) throughout superfusion. Since the functional elimination of the perikarya of interneurons by TTX reduced the inhibitory effect of GBP, the KATP channels mediating the effect of GBP may be located on nerve terminals, probably on both noradrenergic and glutamatergic nerve endings.