Use of aminophenylboronic acid affinity chromatography to measure glycosylated albumin levels.

A simple technique for the measurement of glycosylated albumin by affinity chromatography on m-aminophenylboronic acid agarose columns is presented. The technique relies on bromcresol green determination of albumin in the nonbound and bound fractions. There is a linear correlation between albumin concentration of the bound fraction and glycohemoglobin values in individuals. A control nondiabetic plasma pool with a glycohemoglobin value of 7.10% +/- 0.05% (mean +/- SEM) had a glycoalbumin value of 1.64% +/- 0.06%, while a diabetic control plasma pool with a glycohemoglobin value of 13.63% +/- 0.07% had a glycoalbumin value of 4.02% +/- 0.12%. Compared with results from the affinity technique, the preponderance of colorimetric reaction determined with the thiobarbituric acid procedure is nonspecific, in that it does not correlate with diabetic status or with values derived by the affinity procedure. The bulk of thiobarbituric acid-reactive material is present in the fraction of albumin that does not bind to aminophenylboronic acid. This nonbound fraction contains plasma glucose, which significantly interferes with thiobarbituric acid determinations but only very slightly interferes with the affinity procedure. Prolonged incubation of plasma with 500 mg/dl glucose dramatically increases affinity-determined glycosylated albumin. Thiobarbituric acid reactivity increases much less, the increase being mainly in the fraction bound to aminophenylboronic acid. The percentage glycosylated albumin determined by the affinity technique in crude plasma samples differs very slightly, if at all, from that determined by purification of the albumin from plasma. The affinity technique appears very promising for eventual clinical applications in the management of diabetes.