The pathogenesis of diabetic nephropathy remains poorly defined, and animal models that represent the human disease have been lacking. It was demonstrated recently that the severe endothelial dysfunction that accompanies a diabetic state may cause an uncoupling of the vascular endothelial growth factor (VEGF)-endothelial nitric oxide (eNO) axis, resulting in increased levels of VEGF and excessive endothelial cell proliferation. It was hypothesized further that VEGF-NO uncoupling could be a major contributory mechanism that leads to diabetic vasculopathy. For testing of this hypothesis, diabetes was induced in eNO synthase knockout mice (eNOS KO) and C57BL6 controls. Diabetic eNOS KO mice developed hypertension, albuminuria, and renal insufficiency with arteriolar hyalinosis, mesangial matrix expansion, mesangiolysis with microaneurysms, and Kimmelstiel-Wilson nodules. Glomerular and peritubular capillaries were increased with endothelial proliferation and VEGF expression. Diabetic eNOS KO mice showed increased mortality at 5 mo. All of the functional and histologic changes were improved with insulin therapy. Inhibition of eNO predisposes mice to classic diabetic nephropathy. The mechanism likely is due to VEGF-NO uncoupling with excessive endothelial cell proliferation coupled with altered autoregulation consequent to the development of preglomerular arteriolar disease. Endothelial dysfunction in human diabetes is common, secondary to effects of glucose, advanced glycation end products, C-reactive protein, uric acid, and oxidants. It was postulated that endothelial dysfunction should predict nephropathy and that correction of the dysfunction may prevent these important complications.