The potency and effectiveness of dynorphin A(1-17), E-2078 (a synthetic dynorphin A(1-8) analog) and non-peptidic kappa-opioid agonists were studied in rhesus monkeys in two assays: 1) a drug discrimination assay with the centrally-penetrating kappa-agonist U69,593 as the training stimulus (n=3) and 2) a prolactin release assay; a neuroendocrine effect thought to be mediated by kappa-receptors located in hypothalamic nuclei outside the blood-brain barrier. The non-peptidic kappa-agonists, U69,593 and bremazocine (0.00032-0.01 mg/kg, s.c.) were dose-dependently generalized by all the subjects trained to discriminate U69,593. U69,593 and bremazocine also caused dose-dependent prolactin release (n=4). By contrast, dynorphin A(1-17) and E-2078 (0.1-1 mg/kg, i.v.) were only generalized by one of the U69,593 discriminating subjects. However, both these dynorphins produced potent and robust prolactin release (0.0032-0.032 mg/kg, i.v.), when tested under an identical time course design as above. Naltrexone (0.1 or 0.32 mg/kg), caused a parallel rightward shift in the dose-effect curves for all the above ligands, consistent with kappa-receptor mediation of this neuroendocrine effect. The peripherally selective antagonist, quaternary naltrexone (0.32 mg/kg, s.c.) partially blocked the neuroendocrine effects of U69,593 and E-2078 (0.0032 mg/kg, s.c. and i.v., respectively). Overall, these findings are consistent with the hypothesis that the dynorphins act as high efficacy, peripherally selective kappa-agonists following systemic administration in primates.