Previously, we demonstrated that peripheral infusion of pancreatic polypeptide (PP) inhibits insulin response to several stimuli through vagal innervation. Since PP is found not only in pancreas but also in brain or cerebrospinal fluid, we studied the effect of intracerebroventricular infusion of PP on insulin secretion before and after vagotomy in dogs. Mongrel dogs were settled with a chronic cannula allowing intraventricular infusions into the third (n = 4) or lateral (n = 4) cerebral ventricle. All the experiments were performed one week after the operation in a fully conscious, relaxed state. Porcine PP (pPP, 50 ng or 5 micrograms/dog in 100 microliter saline), which has the same primary structure with that of canine PP, or saline alone was infused into the cerebral ventricles for 5 minutes at the rate of 20 microliter/minutes. As stimuli of insulin secretion, modified sham feeding (MSF; sight and smell of food for 5 minutes), glucose injection (IV-Glucose; 0.5 g/kg/30 seconds, intravenously) and CCK-octapeptide infusion (IV-CCK-8; 0.07 micrograms/kg/5 minutes, intravenously) were applied immediately after (and in some experiments various intervals after) the end of pPP or saline infusion into the ventricles. Immunoreactive PP or insulin was measured by a specific radioimmunoassay. Administration of PP caused significant inhibition of insulin secretion by MSF, IV-Glucose and IV-CCK-8 without affecting basal insulin secretion. The observed effect of the peptide was most potent when infused into the third cerebral ventricle at a dose of 50 ng/dog and not in a dose-related fashion. The integrated insulin responses to MSF, IV-Glucose and IV-CCK-8 were 28, 58 and 30%, respectively, as those of controls. This effect was likely to be of central origin because an overflow of PP to the periphery could not be observed by PP radioimmunoassay. Prior transthoracic bilateral truncal vagotomy abolished the suppressive effect of PP on glucose- and CCK-8-induced insulin secretion. Furthermore, the time course study of CCK-8 suggested that PP could interact with the regions surrounding the third cerebral ventricle. These results suggest that PP affects the central nervous system to control pancreatic insulin secretion via the vagus nerve like other peptides/neuroregulators which modify physiological processes (e.g. insulin release, acid secretion, motility).