Diabetic retinopathy (DR) is one of the major complications of diabetes mellitus. It is characterized by retinal microvascular changes caused by chronic exposure to hyperglycemia, leading to low tissue oxygenation and ultimately to neovascularization. Laser photocoagulation and vitrectomy are the most efficient treatments for DR, but display severe side effects such as the destruction of the healthy retina. Another clinical approach uses antiangiogenic agents to prevent and delay progression of neovascularization, but these require recurrent local administrations that increase the possibility of retinal detachment, vitreous hemorrhage, and cataract formation. Studies in human diabetic retinas have revealed an imbalance between proangiogenic factors such as the vascular endothelial growth factor (VEGF) and antiangiogenic factors, such as pigment epithelial-derived factor (PEDF). This imbalance favors pathological angiogenesis contributing to DR, and can constitute a therapeutic target. Gene therapy was recently shown to be an adequate intervention for long-term treatment of several retinal pathologies. We have previously shown the newly engineered episomal vector pEPito to be able of sustained gene expression in the mouse retina. We here show that pEPito was able to overexpress PEDF for up to three months, both in in vitro cultures of human retinal pigment epithelial cells and in the retina of diabetic mice after a single subretinal injection. In vivo, in parallel with the increase in PEDF we observed a decrease in VEGF levels in injected compared with noninjected eyes and a significant effect on two hallmarks of DR: reduction of glucose transport (by glucose transporter GLUT1), and reduction of inflammation by decreased reactivity of microglia. Jointly, these results point to a significant therapeutic potential of gene therapy with pEPito-PEDF for the treatment of DR.