RESUMEN
Retinal pigment epithelial cells exert an important supporting role in the eye and develop adaptive responses to oxidative stress or high glucose levels, as observed during diabetes. Endogenous antioxidant defences are mainly regulated by Nrf2, a transcription factor that is activated by naturally-derived and electrophilic compounds. Here we investigated the effect of the Nrf2 activators dimethylfumarate (DMF) and carnosol on antioxidant pathways, oxygen consumption rate and wound healing in human retinal pigment epithelial cells (ARPE-19) cultured in medium containing normal (NG, 5mM) or high (HG, 25 mM) glucose levels. We also assessed wound healing using an in vivo corneal epithelial injury model. We found that Nrf2 nuclear translocation and heme oxygenase activity increased in ARPE cells treated with 10 µM DMF or carnosol irrespective of glucose culture conditions. However, HG rendered retinal cells more sensitive to regulators of glutathione synthesis or inhibition and caused a decrease of both cellular and mitochondrial reactive oxygen species. Culture in HG also reduced ATP production and mitochondrial function as measured with the Seahorse XF analyzer and electron microscopy analysis revealed morphologically damaged mitochondria. Acute treatment with DMF or carnosol did not restore mitochondrial function in HG cells; conversely, the compounds reduced cellular maximal respiratory and reserve capacity, which were completely prevented by N-acetylcysteine thus suggesting the involvement of thiols in this effect. Interestingly, the scratch assay showed that wound closure was faster in cells cultured in HG than NG and was accelerated by carnosol. This effect was reversed by an inhibitor of heme oxygenase activity. Moreover, topical application of carnosol to the cornea of diabetic rats significantly accelerated wound healing. In summary, these data indicate that culture of retinal epithelial cells in HG does not affect the activation of the Nrf2/heme oxygenase axis but influences other crucial oxidative and mitochondrial-dependent cellular functions. The additional effect on wound closure suggests that results obtained in in vitro experimental settings need to be carefully evaluated in the context of the glucose concentrations used in cell culture.