Pulmonary hyperoxic injury manifests as widespread alveolar-epithelial and microvascular endothelial cell necrosis, resolution of which requires angiogenesis. We investigated the hypothesis that inhaled nitric oxide (iNO) and hyperoxia each decreases lung vascular endothelial growth factor (VEGF) expression but increases endostatin and that concurrent administration of both gases will show a greater effect. Piglets were randomized to breathe for 5 d room air (RA); RA + NO (RA + 50 ppm NO), O(2) (hyperoxia, F(I)O(2) >0.96), O(2) + NO, or O(2) + NO + REC (O(2) + NO plus recovery in 50% O(2) for 72 h. After the piglets were killed, we measured lung capillary leak, VEGF mRNA, VEGF, and endostatin protein in homogenates, plasma, and lavage. VEGF mRNA decreased significantly with O(2) and O(2) + NO compared with breathing RA (p < or = 0.05). VEGF protein declined in the experimental groups with a significant reduction in the recovery group compared with the RA group (p < or = 0.05). Similar but more dramatic, endostatin declined in all groups relative to the RA group (p < 0.001). Lavage fluid VEGF protein and lung capillary leak rose significantly with O(2) and O(2) + NO compared with RA, but endostatin was unchanged. At 72 h of recovery from hyperoxia, VEGF mRNA and lavage fluid VEGF but not lung VEGF protein had normalized. Hyperoxia and iNO suppresses lung endostatin expression, but iNO unlike hyperoxia alone does not alter lung VEGF production. Hyperoxia paradoxically raises lavageable VEGF levels. This latter effect and that on VEGF mRNA level but not protein is abrogated by recovery in reduced F(I)O(2) for 72 h.