The glucagon content of crystalline insulin preparations.

Shortly after the discovery of insulin, investigators reported that intravenous administration of this hormone leads to a transient hyperglycemia. Kimball and Murlin (1) suggested that this response was caused by the presence of a second physiologically active substance which they named glucagon. Some workers (2, 3), however, suggested that this effect might be ascribed to insulin, or possibly to an artifact created during the manufacture of insulin (4). Considerable evidence accumulated during the last decade favors the view that the short initial hyperglycemia observed following administration of insulin preparations, is due to the presence of small quantities of a second active substance, probably secreted by the a-cells of the pancreatic islets. The successful isolation of the hyperglycemic principle in highly purified and crystalline form (5) furnished conclusive proof that the factor is indeed an entity different from insulin. The amino acid composition of the crystalline material differs significantly from that of insulin (6). Available evidence indicates that the two proteins also differ in regard to their physico-chemical properties. Hence, it appears unlikely that difficulties in separating glucagon from insulin are due to similar chemical or physico-chemical properties as concluded by Burger and Brandt (7). Although most of the hyperglycemic principle is removed from insulin in the course of its purification, small amounts can be found in most purified preparations. Many of these have been described as homogeneous on the basis of physico-chemical evidence (8). However, more recent work, based on counter current distribution (9), electrophoresis (10), and electrophoresis convection ( 11 ), have suggested the presence of some heterogeneity in crystalline insulin preparations. Small amounts of glucagon might partly account for these findings. Several authors have attempted to estimate the glucagon content of various crystalline insulin preparations (10, 12-14). These estimates, ranging from 1 to 10 per cent, have been obtained by indirect or at best semi-quantitative methods and must be regarded as tentative. The uncertainty of these results prompted the present quantitative study of the glucagon content of crystalline insulin samples. For this purpose the hyperglycemic activity of a number of different insulin preparations was compared with that of reference samples consisting of glucagon-free insulin to which known amounts of crystalline glucagon had been added. The results of these experiments are described in this paper.