To examine the mechanism of the satiety-producing effect of cholecystokinin (CCK) in the central nervous system, we compared the potency of intraperitoneally (ip) or intracerebroventricularly (icv) administered CCK-8 and its analogues on food intake in fasted mice. The icv administration of a small dose of CCK-8 (0.03 nmol/brain) or of Suc-(Thr28, Leu29, MePhe33)-CCK-7 (0.001 nmol/brain) suppressed food intake for 20 min, whereas CCK-8 (1 nmol/kg, which is equivalent to 0.03 nmol/brain) or Suc-(Thr28, Leu29, MePhe33)-CCK-7 (1 nmol/kg) had satiety effect after ip administration. Dose-response studies indicated the following rank order of potency: Suc-CCK-7 > or = Suc-(Thr28, Leu29, MePhe33)-CCK-7 > or = CCK-8 > or = (Nle28,31)-CCK-8 >> desulfated CCK-8 = CCK-4 = 0 in the case of ip administration and Suc-(Thr28, Leu29, MePhe33)-CCK-7 >> Suc-CCK-7 > or = CCK-8 > or = (Nle28,31)-CCK-8 >> desulfated CCK-8 = CCK-4 = 0 in the case of icv administration. The selective CCK-A receptor antagonist MK-329 reversed the inhibitory effect of the centrally as well as peripherally administered CCK-8, or of Suc-(Thr28, Leu29, MePhe33)-CCK-7, whereas the selective CCK-B receptor antagonist L-365260 did not. The icv administered CCK-8 did not appear in the peripheral circulation. These findings suggest the participation of CCK-A receptors in the brain in mediating the satiety effect of CCK and the difference in CCK-A receptors in the brain and peripheral tissues.