Liver and kidney tissue membranes as tissue markers for nonenzymatic glycosylation.

We investigated the relationship of serum protein glycosylation to peripheral tissue membrane glycosylation. We studied 27 Sprague-Dawley rats and induced diabetes in 20 of them. Blood glucose levels were treated in 10 of the diabetic animals with daily subcutaneous insulin. After 8 wk, liver and kidney tissue was removed, purified membranes were prepared, and the percentage of glycosylated membrane protein was determined for the liver and kidney membranes by boronate-affinity methods. The percentage of glycosylated membrane protein for both liver and kidney tissue was found to correlate significantly with the glycemic state of the animal as assessed with glycosylated serum albumin, total glycosylated serum proteins, and glycosylated hemoglobin determinations (P less than 0.001 for each glycosylated protein parameter). In addition, the percentage of glycosylated membrane protein in the liver tissue correlated significantly with the measured level in the corresponding kidney tissue (r = 0.78, P less than 0.001). To identify the nature of the glycosylated membrane proteins, boronate-affinity methods were used to separate the glycosylated and nonglycosylated membrane proteins. It was determined that two major glycosylated protein bands exist for the liver membrane (78,000 and 58,700 Mr) and four for the kidney membranes (ranging from 48,700 to 74,000 Mr). The ultrastructural location and identification of these glycosylated membrane proteins are not known. This study demonstrates that measurement of clinical glycemic state, as reflected in glycosylated blood protein parameters such as glycosylated serum albumin and glycosylated hemoglobin, correlates significantly with ongoing tissue membrane accumulation of glucose.