In vitro studies in isolated pancreas and islets have shown that glucagon-like peptide-1 (GLP-1) promotes insulin release in a typical concentration-dependent manner. In contrast, the relationship between plasma GLP-1 and insulin concentrations in vivo is complicated, because GLP-1-promoted insulin release lowers blood glucose, which influences glucose-dependent insulinotropic ability of GLP-1. GLP-1 also stimulates insulin release via hepatoportal neuronal mechanism. Hence, the dynamic relationship between plasma active GLP-1 vs. insulin and glucose concentrations is obscure. In this study, we aimed to determine in vivo relationships between these parameters in rats. To mimic postprandial state, intraduodenal glucose challenge in anesthetized rats was performed, which can minimize the release of endogenous GLP-1. The glucose challenge induced the 1st phase and 2nd phase insulin release. GLP-1 infusion from jugular vein significantly and concentration-dependently enhanced area under the curve (AUC) of the 1st phase insulin, in which the minimum effective active GLP-1 concentration was 6.6 pmol/l. In contrast, bell-shaped dose responses were observed for both the 2nd phase and total insulin AUCs, in which a significant increase was obtained only with 11 pmol/l of active GLP-1 for total insulin AUC. A statistically significant reduction in the plasma glucose AUC was observed when active GLP-1 concentration was 11 pmol/l and 21 pmol/l. These results indicate that GLP-1 markedly enhances the 1st phase insulin release while less potently the 2nd phase insulin release, possibly due to a negative feedback regulation of β-cells via reduced plasma glucose levels by the enhanced 1st phase insulin release.