Elevated cellular sorbitol levels resulting from conversion of increased glucose by aldose reductase might deplete cellular myoinositol content, which could then lower inositol phosphates (InsPs) and diacylglycerol levels, key regulators of protein kinase C (PKC). Secondary to altered PKC activity, other cellular enzymes such as (NaK)-ATPase could be affected. To test this hypothesis we examined the association between PKC activity, (Na,K)-ATPase activity, and sorbitol, myoinositol, and InsP levels in cultured bovine retinal capillary endothelial cells, a cell type prominently involved in diabetic retinopathy.Elevating glucose concentration in culture media from 100 to 400 mg/dl led to a 100% increase in sorbitol levels, which could be inhibited completely by sorbinil, an aldose reductase inhibitor. In contrast, no changes were observed in myoinositol or InsP levels. Subfractionated PKC activities showed a 100% increase in the membranous pool with a parallel decrease in the cytosolic fraction. Adding sorbinil did not affect PKC activity, whereas the PKC agonist, phorbol myristate acetate (PMA), stimulated translocation of PKC. Ouabain-inhibitable (NaK)-ATPase activity was decreased 70% by elevated glucose levels. This decrease could be prevented by adding either PMA or sorbinil. Thus, in retinal capillary endothelial cells elevated glucose concentration can affect PKC and (Na,K)-ATPase activities, probably via different mechanisms.

Insulin actions and receptors were studied in capillary endothelial cells cultured from diabetic BB rats and their nondiabetic colony mates. The endothelial cells from diabetic rats of 2 mo duration had persistent biological and biochemical defects in culture. Compared with normal rats, endothelial cells from diabetic rats grew 44% more slowly. Binding studies of insulin and insulin-like growth factor I (IGF-I) showed that cells from diabetic rats had 50% decrease of insulin receptor binding (nondiabetic: 4.6±0.7; diabetic: 2.6±0.4% per milligram protein, P < 0.01), which was caused by a 50% decrease in the number of binding sites per milligram protein, whereas IGF-I binding was not changed. Insulin stimulation of 2-deoxy-glucose uptake and a-aminoisobutyric acid uptake were also severely impaired with a 80-90% decrease in maximal stimulation, in parallel with a 62% decrease of insulin-stimulated autophosphorylation (P < 0.05). 125I-insulin cross-linking revealed an 140-kD a subunit of the insulin receptor similar to that in cells from nondiabetic rats, although bands at > 200 kD were also detected. The molecular weight of the insulin receptor,8 subunit (by SDS-PAGE) was smaller in cells from diabetic than from normal rats (88-90 vs. 95 kD). Neuraminadase treatment of the partially purified insulin receptors decreased the molecular weight of the insulin receptors from nondiabetic rats to a greater degree than its diabetic counterpart. In contrast, Northern blot analysis of insulin receptor mRNAs using human cDNA probes revealed two species of 9.4 and 7.2 kb with no difference in mRNA abundance between cells from diabetic and nondiabetic rats. We conclude that the exposure of capillary endothelial cells to a diabetic milieu in vivo can cause specific and persistent changes in the insulin receptor and insulin action.