An improved radioimmunological method for serum free insulin and antibody-bound insulin

Abstract

A practical, sensitive, and specific method for the assay of serum free insulin and antibody-bound insulin in insulin-treated diabetic patients has been developed and clinically evaluated. Clinical evaluation was carried out at the Youngstown Hospital Association in cooperation with the Department of Laboratories and several staff physicians. The method for free insulin involves the use of aqueous polyethylene glycol which causes the precipitation of the free antibody and the antibody-bound insulin with little or no precipitation of the free insulin. The free insulin in the upper phase was then measured in a simplified, but sensitive radioimmunological system.

Free insulin was found to be completely soluble in a 12% w/v final concentration of 6000 molecular weight polyethylene glycol. Two other polyethylene glycol solutions with average molecular weights of 1,600 and 20,000 were studied to determine their ability to extract the free insulin and precipitate the insulin antibody complex. Only the 6000 molecular weight compound was found to be suitable. Particular attention was focused on the use of the polyethylene glycol extract containing the free insulin in the radioimmunoassay (RIA ) system. The pH of the extract was found to have a range of 8.32 to 8 .52; however when introduced into the RIA system, employing a phosphates-saline buffer, the pH was 7.40 - 0.02. This pH range was considered optimum for the RIA system. Studies were conducted on normal serum where insulin antibodies were absent to determine the correlation between the free insulin method ·and the "Phadebas Insulin" reference method. Results of the methods were compared at many different insulin levels. In 40 determinations, standard deviation, coefficients of variation, correlation coefficient, and t test values were determined and found to be within acceptable limits. The mean value for the Reference method at all insulin levels on these 40 samples was 61 μ U/ml.In the same 40 determinations and for the same insulin levels, the mean values for the free method were 60 μU/ml. The method for total insulin requires prior treatment of serum with 1 N HCl to break the conjugation of the insulin antibody complex. The total insulin is then extracted with polyethylene glycol and assayed in the RIA system. The antibody-bound insulin is the total insulin minus the free insulin. The polyethylene glycol extract in the total insulin method after neutralization has a pH range of 8.20 to 8. 80. When the extract is introduced into the RIA system employing the phosphate, saline buffer, its pH was 7.40 - 0.02. Studies were conducted on the non-insulin-treated patients to determine the correlation between the total insulin method and the "Phadebas Insulin" reference method. In addition, correlation experiments to compare the total method and the free insulin method were performed. In 40 determinations, standard deviation, coefficients of variation, correlation coefficient, and t test values were determined at all levels of insulin concentration. In the same series of 40 measurements described for the reference and free insulin methods, the total insulin method gave a mean value of 59 μU/ml . Recovery studies of added insulin were performed for both the free and total insulin methods. The free method gave an average recovery of 85% at all levels with better recovery at the 50-100 μU/ml range. The total method gave an average recovery of 87% at all levels with better recovery at the 100-300 range. Clinical studies on insulin-treated diabetic patients as well as non-insulin-treated diabetic patients were conducted. In the non-insulin-treated patients, good correlation was ' found by all three methods (reference, free, and total). Mean values for these patients by the three methods were 22.7, 21.8, and 22.1 μU/ml respectively.

In the insulin- treated patients , where insulin had been taken for 30 days or more , antibody- bound insulin l evels were found to be present in varying amounts depending, in part, on length of insulin therapy and amount of insulin taken. The range of all insulin-treated patients studied for free insulin was 10 to 440 with a mean of 46.7 μU/ml. The range for the same patients for total insulin was 67 to 17,920 with a mean of 2,718 μU/ml. Two patients studied with bound insulin levels of 4992 and 17,480 μU/ml and who were difficult to control, were immediately started on pork insulin therapy in an effort to reduce their immunological response to the exogenous insulin. The methods provide a practical approach to following the dynamics of insulin therapy in the presence of insulin antibody and can be used to quantitate the level of antibody present in terms of bound insulin. In addition, the Free method could be used to determine the amount of endogenous insulin produced by the pancreas in insulin-treated diabetic children where remission in insulin requirements has occurred.