To determine whether angiotensin receptor blockade decreases vascular tone in heart failure by improving endothelial-dependent vasorelaxation and increasing nitric oxide (NO) bioavailability, we treated infarcted adult male Sprague-Dawley rats with candesartan for 7 days or 8 weeks (10 mg/kg/day in drinking water). Candesartan, at both time points, lowered left ventricular (LV) systolic pressure (P < 0.05) (122 +/- 22 versus 74 +/- 16 and 73 +/- 10 mm Hg) and LV dP/dt (5914 +/- 1294 versus 2857 +/- 1672 versus 3175 +/- 769 mm Hg/s), but lowered LV end-diastolic pressure only at 8 weeks (16.9 +/- 9.7 versus 11.2 +/- 5.7 versus 6.9 +/- 5.3 mm Hg). The vasorelaxation response to acetylcholine (ACh) in thoracic aortic segments was decreased with infarction (P < 0.05), remained unchanged with 1 week of candesartan, but increased 84 and 86% at 10-4 and 10-5 M ACh (P < 0.05) at 8 weeks. The enhanced candesartan-induced vasorelaxation at 8 weeks was abolished with NG-nitro-l-arginine methyl ester (200 microM). In bovine pulmonary endothelial cells, 20 microM candesartan increased endothelial nitric-oxide synthase (eNOS) protein levels (P < 0.05) (28.9 +/- 2.6 versus 16.1 +/- 3.7 intensity units/microg of protein); the increased eNOS was abolished by a specific angiotensin subtype 2 (AT2) receptor antagonist, PD 123319. These data suggest that AT1 receptor blockade enhances vasorelaxation in heart failure by increasing NO bioavailability, in part via an AT2 receptor-mediated up-regulation of eNOS protein.