Glucose-dependent insulinotropic polypeptide [GIP-(1-42)] is degraded by dipeptidyl peptidase IV (DPP IV), forming GIP-(3-42). In mice, high concentrations of synthetic GIP-(3-42) may function as a GIP receptor antagonist, but it is unclear whether this occurs at physiological concentrations. In COS-7 cells transiently transfected with the human GIP receptor, GIP-(1-42) and -(3-42) bind with affinities (IC(50)) of 5.2 and 22 nM, respectively. GIP-(1-42) was a potent agonist, stimulating cAMP accumulation (EC(50), 13.5 pM); GIP-(3-42) alone had no effect. When incubated together with native GIP, GIP-(3-42) behaved as a weak antagonist (IC(50), 92 and 731 nM for inhibition of cAMP accumulation elicited by 10 pM and 1 nM native GIP, respectively). In the isolated perfused rat pancreas, GIP-(3-42) alone had no effect on insulin output and only reduced the response to GIP (1 nM) when coinfused in >50-fold molar excess (IC(50), 138 nM). The ability of GIP-(3-42) to affect the antihyperglycemic or insulinotropic actions of GIP-(1-42) was examined in chloralose-anesthetized pigs given intravenous glucose. Endogenous DPP IV activity was inhibited to reduce degradation of the infused GIP-(1-42), which was infused alone and together with GIP-(3-42), at rates sufficient to mimic postprandial concentrations of each peptide. Glucose, insulin, and glucagon responses were identical irrespective of whether GIP-(1-42) was infused alone or together with GIP-(3-42). We conclude that, although GIP-(3-42) can weakly antagonize cAMP accumulation and insulin output in vitro, it does not behave as a physiological antagonist in vivo.