Human retinal vascular cells differ from umbilical cells in synthetic functions and their response to glucose

Cell culture systems have commonly been used to study mechanisms implicated in the pathogenesis of diabetic retinopathy, but the great majority of cell preparations used have been either of nonhuman retinal origin or nonretinal human origin. Because of questions of species and organ specificity in the function of cells of vascular origin, in this study, cultured microvascular endothelial cells (HREC), pericytes (HRPC), and pigment epithelial cells from the postmortem human retina, and endothelial cells from human umbilical vein (HUVEC) were evaluated with respect to cell proliferation, and secretory products potentially important in diabetic retinopathy, i.e., prostaglandins (PG) and plasminogen activators (PA), normalized to DNA content/well, under both basal (5 mM) and high (25 mM) glucose conditions. Glucose (25 mM) reduced DNA content similarly in both types of endothelial cells, had a lesser effect on HRPC, and did not significantly alter the proliferation of pigment epithelial cells. Basal secretion of PGI2 (measured as 6-keto-PGF1 alpha) was in the order HRPC much greater than HREC greater than HUVEC, whereas PGE2 secretion was in the order HREC much greater than HRPC greater than HUVEC. Glucose (25 mM) stimulated PGI2 secretion by HRPC, but not by either type of endothelial cell, and enhanced PGE2 secretion by HREC, but not by HUVEC or HRPC. Release of plasminogen activator activity differed between HUVEC and HREC under basal conditions and addition of 25 mM glucose stimulated release only from HREC. Glucose (25 mM) stimulated PA secretion by HREC, but not by HUVEC. These findings provide evidence that human retinal pericytes are an important source of prostacyclin, and that there are differences between HREC and HUVEC with respect to secretory functions and their modulation by glucose, indicating regional specificity of these functions. Extrapolation to human retinal vascular cells from experiments using cells from heterologous vascular beds to draw inferences about the pathophysiology of diabetic retinopathy are not valid for these cellular functions.