The mechanism by which the recreational drug (+/-)-3,4-methylenedioxymethamphetamine (MDMA) destroys brain serotonin (5-HT) axon terminals is not understood. Recent studies have implicated MDMA metabolites, but their precise role remains unclear. To further evaluate the relative importance of metabolites versus the parent compound in neurotoxicity, we explored the relationship between pharmacokinetic parameters of MDMA, 3,4-methylenedioxyamphetamine (MDA), 3,4-dihydroxymethamphetamine (HHMA), and 4-hydroxy-3-methoxymethamphetamine (HMMA) and indexes of serotonergic neurotoxicity in the same animals. We also further evaluated the neurotoxic potential of 5-(N-acetylcystein-S-yl)-HHMA (5-NAC-HHMA), an MDMA metabolite recently implicated in 5-HT neurotoxicity. Lasting serotonergic deficits correlated strongly with pharmacokinetic parameters of MDMA (C(max) and area under the concentration-time curve), more weakly with those of MDA, and not at all with those of HHMA or HMMA (total amounts of the free analytes obtained after conjugate cleavage). HHMA and HMMA could not be detected in the brains of animals with high brain MDMA concentrations and high plasma HHMA and HMMA concentrations, suggesting that HHMA and HMMA do not readily penetrate the blood-brain barrier (either in their free form or as sulfate or glucuronic conjugates) and that little or no MDMA is metabolized to HHMA or HMMA in the brain. Repeated intraparenchymal administration of 5-NAC-HHMA did not produce significant lasting serotonergic deficits in the rat brain. Taken together, these results indicate that MDMA and, possibly, MDA are more important determinants of brain 5-HT neurotoxicity in the rat than HHMA and HMMA and bring into question the role of metabolites (including 5-NAC-HHMA) in MDMA neurotoxicity.