To better understand the in vivo pharmacological effects of morphine 3-glucuronide (M3G, a weak opioid antagonist) and morphine 6-glucuronide (M6G, a potent opioid agonist), the permeability of the blood-brain barrier (BBB) for these metabolites was compared with morphine. Tracers were prepared by enzymatic glucuronidation of [N-methyl-3H]morphine. Brain uptake in rats was measured by the internal carotid perfusion technique and after intravenous bolus injections. In the perfusion experiments morphine showed a permeability-surface area product (PS) of 3.52 +/- 0.61 microliter min-1 g-1. Uptake seemed to be mediated by passive diffusion and was not saturable by 100 microM morphine in the perfusate. The BBB permeability of [3H]M3G and [3H]M6G was too low to be quantified after 5 min of perfusion. Brain uptake of [3H]M3G and [3H]M6G 60 min after i.v. bolus injection reached 0.0060 +/- 0.0003% and 0.0030 +/- 0.0005% injected dose per g, respectively. From these brain concentrations and the corresponding plasma concentration-time curves, BBB PS values of 0.14 +/- 0.02 microliter min-1 g-1 and 0.11 +/- 0.01 microliter min-1 g-1, respectively, were calculated. The ratio of BBB PS values is complementary to the analgesic potencies of morphine and M6G after different routes of administration. The low PS of M6G explains why it is approximately equipotent to morphine after systemic injection, although it is about 2 orders of magnitude more potent than morphine after administration directly into the central nervous system.