To avoid the confounding effect of metabolic degradation, the stable mu-opioid peptide agonist [D-Arg2,Lys4]-dermorphin analogue (DALDA) was used to quantitate blood-brain barrier (BBB) permeability by intravenous injection and internal carotid artery perfusion techniques. With intravenous injection, the BBB permeability-surface area products for [3H]DALDA (0.84 +/- 0.13 microliters.min-1.g-1) and [14C]sucrose (0.39 +/- 0.05 microliters.min-1.g-1) correlated with the lipid solubility of the two molecules: the 1-octanol-Ringer partition coefficient for DALDA was approximately 2 log orders greater than that for sucrose. The brain delivery of [3H]DALDA at 30 min after intravenous administration was 0.019 +/- 0.002% of the injected dose per gram, and analgesia was induced with a 5-mg/kg dose administered systemically. In contrast to the result after intravenous injection, the BBB permeability-surface area product for DALDA estimated with the internal carotid artery perfusion technique was manyfold greater. This was due to nonspecific absorption of the peptide into the cerebral microvasculature, which precluded use of the capillary depletion technique to study transcytosis through the BBB after internal carotid artery perfusion. The present studies show that the brain delivery of a metabolically stable peptide, such as DALDA, is comparable to that for sucrose, correlates with lipid solubility, and is mediated by a nonsaturable mechanism, probably free diffusion.