Assays of insulin and of insulin-like activity

Hypoglycemic effects

Blood sugar lowering effect in rabbits

A biological assay of insulin preparations in comparison with a stable standard using the blood sugar lowering effect in rabbits has been proposed already in 1925 by Harrison et al. The biological assay of insulin using the blood sugar lowering effect in rabbits has been until recently the official assay in several pharmacopoeias, such as European Pharmacopoeia,;; British Pharmacopoeia; United States Pharmacopoeia  and The National Formulary
                      The rabbit blood glucose bioassay as well as the mouse convulsion assay and the mouse glucose assay were used for establishing international standards for highly purified human, porcine and bovine insulin
                     In several pharmacopoeias, the biological assays have been replaced by chemical methods (British Pharmacopoeia 1999; European Pharmacopoeia, 3rd Edition 1997; but the rabbit blood sugar method is still valid in the United States Pharmacopeia USP 24, 2000.

Four groups of at least 6 randomly distributed rabbits weighing at least 1.8 kg are kept in the laboratory and maintained on a uniform diet for not less than one week before use in the assay. About 24 h before the test each rabbit is provided with an amount of food that will be consumed within 6 h. The same feeding schedule is followed before each test day. During the test all food and water is withheld until the final blood sample has been taken. The rabbits are placed into comfortable restraining cages to avoid undue excitement.
               Immediately before use two solutions of the standard preparation are made, containing 1 unit and 2 units of insulin per ml, respectively, and two dilutions of the preparation being examined which, if the assumption of potency is correct, contain amounts of insulin equivalent to those in the dilutions of the standard preparation. As diluents , a solution is used of 0.1–0.25% w/v of either m-cresol or phenol and 1.4 to 1.8 w/v of glycerol being acidified with hydrochloric acid to a pH between 2.5 and 3.5.
                  Each of the prepared solutions is injected subcutaneously to one group of rabbits, using the same volume, which should usually be between 0.3 and 0.5 ml for each rabbit, the injections being carried out according to a randomized block design. Preferably on the following day, but in any case not more than 1 week later, each solution is administered to a second group of rabbits following a twin crossover design. One hour and 2.5 h after each injection a suitable blood sample is taken from the ear vein of each rabbit. Blood sugar is determined by a suitable method, preferably using glucose oxidase5

 Hypoglycemic seizures in mice
The biological assay of insulin using hypoglycaemic seizures in mice has been suggested already in 1923 by Fraser. The biological standardization of insulin using the mouse convulsion method has been published in detail by the Health Organisation of the League of Nations in 1926 and has been until recently the official assay in several pharmacopoeias, such as European Pharmacopeia, British Pharmacopoeia 1988.
In most pharmacopoeias, the biological assays have been replaced by chemical methods (British Pharmacopoeia 1999; European Pharmacopoeia, 3rd Edition 1997).

Ninety-six mice of either sex (but not of mixed sexes) weighing 20 ±5 g are randomly distributed into 4 groups. The mice are deprived of food 2–20 h immediately preceding the test. Solutions of the insulin standard and of the test preparation containing 30 and 60 milli Units/ml are prepared by diluting the original solution with 0.9% NaCl solution, pH 2.5. 0.5 ml/20 g mouse of these solutions are injected subcutaneously. The mice are kept at a uniform temperature, between 29 and 35 °C, in transparent containers within an air incubator with a transparent front. The mice are observed for 1.5 h and the number of mice is recorded that are dead, convulse or lie still for more than 2 or 3 s when placed on their backs.

The percentage of mice of each group showing the mentioned symptoms is calculated and the relative potency of the test solution calculated using a 2 + 2 point assay11

 Blood sugar determinations in mice
Eneroth and Ahlund recommended a twin crossover method for bio-assay of insulin using blood glucose levels in mice instead of hypoglycaemic seizures giving more precise results. This test was induced into the British Pharmacopoeia 1980 and continued up to 1988.

Non-fasting mice of the same strain and sex are used having body masses such that the difference between the heaviest and lightest mouse is not more than 2 g. The mice are assigned at random to four equal groups of not less than 10 animals. Two dilutions of a solution of the substance or of the preparation to be examined and 2 dilutions of the reference solution are prepared using as diluent 0.9% NaCl solution adjusted to pH 2.5 with 0.1 N hydrochloric acid and containing a suitable protein carrier. In a preliminary experiment, concentrations of 0.02 IU and 0.10 IU are tested. Each of the prepared solutions (0.1 ml/10 g body weight) is injected subcutaneously to one group of mice according to a randomized block design. Not less than 2.5 h later, each solution is administered to a second group of mice following a twin crossover design. Exactly 30 min after each injection, a sample of 50 μl of blood is taken from the orbital venous sinus of each mouse. Blood glucose concentration is determined by a suitable method

The potency is calculated by the usual statistical methods for the twin-cross-over assay10

Binding assays

The first description of an immunoassay of endogenous plasma insulin in man has been given by Yalow and Berson. Yalow et al. providing evidence that the bioassays hitherto being used (isolated rat diaphragm, epididymal fat pad tissue) measure insulin- like activity but not true insulin levels in blood.


Semisynthetic or biosynthetic human insulin is used as immunogen and as standard. Formerly, porcine insulin has been used since Yalow and Berson (1960) and subsequently many other authors have shown that antisera raised against porcine insulin react identically with human and porcine insulin. Guinea pigs weighing 350–450 g are injected subcutaneously with 0.4 ml of an emulsion of 5 mg human insulin dissolved in 1.0 ml 0.01 N HCl and 3.0 ml complete adjuvant. For boostering, 0.2 ml of an identically prepared emulsion is injected in monthly intervals. Fourteen days after the third booster injection, the animals are slightly anesthetized and 8–10 ml blood are withdrawn by cardiac puncture. Boosting is continued at monthly intervals and the animals are bled 2 weeks following each booster injection.

The optimal antiserum titer for use in the radioimmunoassay is determined using conditions identical to those employed in routine immunoassays. The percentage binding of 1μU 125I insulin is determined for dilutions of antisera ranging from 103 to 106 fold. The steepness of the antisera dilution curve is a measure of the affinity of the antiserum and therefore the potential sensitivity of the radioimmunoassay. Antisera with the steepest slopes, but not necessarily the highest titer, are selected for further study. The selected antisera dilutions are then run in an immunoassay using a full range of standards. A reduction in the percent 125I-insulin bound to antibody from 50% (in the absence of unlabeled insulin) to 45% (in the presence of unlabeled insulin) (B/Bo = 0.9) is a reasonable measure of assay sensitivity.

Preparation of 125I-insulin
Most investigators use the “chloramine-T procedure” to iodinate insulin. The reaction is carried out in a 20 ml glass vial in an ice-bath with continuous magnetic stirring. To 2.5 ml 0.05 M phosphate buffer, pH 7.5, 2.0 mCi Na125I, and 15 μl of a 1 mg/ml insulin solution are added. Then, 0.5 ml of a chloramine T (50 mg/ml) solution is added dropwise over the course of 1 min. After 10 min, 0.7 ml of a freshly prepared sodium metabisulfite solution (50 mg/ml in 0.05 M phosphate buffer, pH 7.5) is added. One ml of this reaction mixture is transferred to 10 ml 2% bovine serum albumin for determination of specific activity. In order to absorb unreacted 125I and damaged products 2.0 g 20–50 mesh AG 1X-8 resin are added (equilibrated in 1 ml 0.05 M phosphate buffer, pH 7.5, containing 0.1 mg/ml thiomerosal and 20 mg/ml crystalline bovine serum albumin). The reaction mixture is stirred for 10 min, decanted from the resin and diluted to a concentration of less than 25 μC/ml in a solution of 0.8 M glycine, 0.2 M NaCl, 0.05 M phosphate (pH 7.5), and 2.5 mg/ml crystalline BSA. The final solution is stored in multiple aliquots at –70 °C.

The following procedure is recommended:

·         A buffer is prepared from a solution of 8.25 g boric acid and 2.70 g NaOH dissolved in 1litre water. After dissolving 5.0 g of purified bovine serum albumin, pH is adjusted with concentrated HCl to 8.0.

·         In disposable plastic tubes, 10 * 75 mm, the following volumes are added:

100μl serum or standard

900 μl buffer

100 μl 1 mU 125I-insulin in assay buffer

100 μl guinea pig anti-insulin antiserum diluted in assay buffer (at a concentration to bind  50% of the 125I-insulin in the absence of unlabeled hormone)

  • The mixture is incubated at 4 °C for 72 h. Then, the following solutions are added

100 μl normal guinea pig serum diluted 1 : 400 in the assay buffer

100 μl rabbit anti-guinea pig globulin serum diluted in assay buffer

  • The mixture is again incubated at 4 °C for 72 h and then centrifuged at 4 °C and 2 000 g for 20 min. The supernatant is decanted and radioactivity counted in the precipitate for 5 min.

Counts in the nonspecific binding tubes are subtracted from counts in all other tubes. Data are linearized using an unweighted logit-log transformation .Micro-units insulin in a logarithmic scale are plotted against the ratio B/Bo 125I-insulin on a logit scale. The range of B/Bo between 0.4 and 0.9 is the most suitable for determination of insulin concentration in plasma.

Insulin receptor binding
Insulin receptor binding studies have been performed with various animal tissues and isolated cells as well as with cells of human origin. Human adipocytes can be used to study simultaneously insulin receptor binding and metabolic effects of insulin .The binding tests are of value to characterize newly synthesized insulin derivatives

Subcutaneous adipose tissue (about 4–5 g) is obtained from the abdomen of patients undergoing gastroenterological surgery. Patients suffering from any endocrine or metabolic disorder or taking drugs known to affect metabolism have to be excluded. Other exclusion criteria are impaired glucose tolerance measured by determination of fasting blood glucose and the 2 h value after a 75 g oral glucose load. The adipose tissue is finely chopped and incubated for 90 min at 37 °C in a HEPES buffer (pH 7.4), containing human serum albumin (25 g/l) and collagenase (0.5 g/l). The isolated adipocytes are subsequently washed five times in a HEPES buffer containing 50 g/l human albumin. The diameters of adipocytes are measured at 200-fold magnification using an eyepiece micrometer. Surface and volume are calculated for every cell diameter.
Insulin receptor binding studies with isolated human adipocytes are performed in a 300 μl cell suspension containing about 1 * 105 cells/ml in a HEPES buffer (10 mmol/l HEPES, 50 g/l human serum albumin, (pH 7.4) at 37 °C. The iodine labelled ligand ([125I] TyrA14-monoiodinated insulin, specific activity about 350 mCi/mg) in a final concentration of 20 pmol/l is incubated with increasing amounts of unlabeled human insulin and the insulin derivative to be tested. The reaction is stopped by adding 10 ml of chilled 0.154 mol/l NaCl and subsequent centrifugation with silicone oil
Non-specific binding is measured by incubating tracer in the presence of a large excess of unlabelled insulin. For association studies the 125I-labelled ligand is incubated for various times (1 to 240 min) and the reaction is terminated as described above. At each time point, the non-specific binding is measured and subsequently subtracted from the corresponding data for total binding.
                Dissociation rates are determined by first incubating isolated human adipocytes at 37 °C with either [125I] TyrA14-insulin or the test compound labelled in the same position for 90 min to achieve steady-state binding conditions. Each incubation mixture is then centrifuged for 60s The adipocytes are rapidly washed twice by diluting with buffer to the original volume at 4 °C and the centrifugations and aspirations are repeated. After the third aspiration, the cells are diluted to the original volume with buffer alone or native insulin or the insulin derivative to be tested at a final concentration of 0.2 μmol/l at 22 °C. At this hormone concentration a maximal effect of 125I-insulin dissociation is reported. The reaction is stopped at various times between 10 and 180 min and cell-associated radioactivity is determined.

Results are expressed as percentage specific binding per 10 cm2 plasma membrane surface area 18



Subscribe to Pharmatutor Alerts by Email



Subscribe to RSS headline updates from:
Powered by FeedBurner