The role of adenosine in the cerebrovascular response to carbon dioxide inhalation was evaluated in two sets of experiments. The pial circulation was recorded by a Laser-Doppler flow probe placed over a closed cranial window in methoxyflurane anesthetized rats. Topical application of the nonselective adenosine receptor antagonist caffeine (1 mM), the selective A1 receptor antagonist 8-cyclopentyl-1,3-dipropylxanthine (DPCPX,1 microM), or the selective A2A receptor antagonist 4-(2-[7-amino-2-(2-furyl)[1,2,4]triazolo[2,3-a]triazin-5-yl amino]ethyl) phenol (ZM 241385, 1 microM) all failed to affect mean arterial blood pressure, basal cerebral blood flow, or the carbon dioxide-evoked hyperemia. Systemically administered caffeine (20 mg/kg) also had no significant effects. However, following the systemic administration of the nonselective nitric oxide synthase inhibitor, NG-nitro-L-arginine methyl ester (L-NAME, 20 mg/kg), the topical application of both caffeine and ZM 241385 (but not DPCPX) significantly reduced the carbon dioxide-evoked hyperemia. L-NAME (20 mg/kg) administered intravenously, evoked a significant increase in mean arterial blood pressure, a slow progressive decline in cerebral blood flow and, during brief (60-90 s) periods of 7.5% carbon dioxide inhalation, a significant decrease in arterial blood pressure. L-NAME failed to reduce the carbon dioxide-evoked increase in cerebral blood flow as measured by the area under the curve (AUC), although it did reduce the peak flow response. Topically applied L-NAME (1 mM) failed to alter mean arterial blood pressure, basal cerebral blood flow, or the carbon dioxide-evoked increases in cerebral blood flow. In a second series of experiments, we evaluated the ability of 10% carbon dioxide inhalation for 8 min to elicit a release of adenosine from the cerebral cortex. Adenosine levels in the cortical superfusates rose significantly during periods of carbon dioxide inhalation. The data suggest that following the removal of the confounding effects of nitric oxide, which are unlikely to be mediated locally, a significant contribution by adenosine A2A receptor activation to the carbon dioxide-evoked cortical hyperemia was evident.