Fig. I. Purification of damaged insuun-I1'1 by cellulose column adsorption. Top: Chromato-elec-trophoretograms of beef insulin-I"1, with specific activity about 300 mc per mg, after 30 minutes' dialysis. Bottom: Same preparation after elution from cellulose column with control (nonimmune) plasma. At significantly lower specific activities the preparations appear as in the bottom figure without purification.
into a test tube beneath a layer of a few drops of water (to prevent loss of I"1 into the air) and assayed for radioactivity in any low-sensitivity counting device. If much less than one-half of the starting radioactivity has been extracted, a second extraction with 0.2 to 0.3 ml chloroform is performed. The total amount of chloroform should be kept as small as possible to facilitate the subsequent extraction of iodine into the aqueous protein solution. The volume of the aqueous phase in the separatory funnel also should be kept small to favor the initial extraction of iodine into the chloroform. The chloroform-iodine mixture is added to 0.5 ml of 0.2 M borate buffer, pH 8, containing 20 /tg of crystalline beef insulin in a 50 ml centrifuge tube, which provides for a broad interface between the two phases. The tube is shaken briskly but not violently for not more than 2 to 3 minutes following which an additional 1.0 ml borate buffer is mixed into the contents. A barely visible flocculate appears occasionally and should be allowed to settle, whereupon the top 0.5 ml (one-third of total) of the water phase is quickly removed and dialyzed against 2 L of distilled water.1 Owing to the high concentration of radioactivity and low concentration of protein, the insulin is very susceptible to radiation damage (18, 19) ; therefore, exposure to l1" at this stage should be as brief as possible, not more than 5 to 10 minutes elapsing between addition of I™ to the insulin and the start of dialysis. Of the total radioactivity in the dialysis hag, approximately 65 to 80 per cent represents unbound 1"' which is reduced to less than 1 per cent of the Ini bound to insulin after 2 hours of dialysis. Between 20 and 60 per cent of the insulin-Iul is adsorbed to the dialysis membrane during this time so that the procedure yields approximately 3 to 5 fig insulin labeled with about 1.0 to 2.0 mc I"1. Considerable sacrifices in total yield are made to expedite the procuring of a highly labeled preparation which usually contains no more than 4 to 6 per cent damaged components. We have the impression that the addition of 10~* M KI or phenol (as radical scavengers) to the dialyzing solution may help to minimize radiation damage, but this has been difficult to establish since other factors are also responsible for damage to the protein during the procedure. Distilled water is used in the last dialysis following which 1 drop of human serum albumin (250 mg per ml) is added to the insulin-!131 solution to prevent losses of labeled insulin by adsorption to glassware (20, 21) and to minimize any further irradiation damage (18, 19), Solutions are kept frozen when not in use.
If the insulin-I1" solution is surveyed for radioactivity at completion of dialysis, the specific activity of the in-sulin-I"1 may be estimated approximately. If the yield of labeled insulin has been sufficient to produce a specific activity in excess of 150 mc per mg, it can be anticipated that damage will be significantly in excess of 4 to 6 per cent, and at 300 mc per mg may be as great as 15 to 18 per cent. It is then necessary to effect partial purification of the insulin-l111. Since the damaged components do not adsorb to paper but are observed to migrate with serum proteins on paper strip chromatography or electrophoresis (22), it is possible to use a cellulose column for the purification procedure as follows: The dialyzed insulin-I1*1 solution is added to 0.1 ml control (nonimmune) serum and the mixture is then passed through a column packed
1 Removal of unbound iodide1" by anionic exchange resins is usually unsatisfactory because much of the in sulin-I111 at this low concentration is lost by adsorption to the resin.
IMMUNOASSAY OF ENDOGENOUS PLASMA INSULIN IN MAN 1159
with a cellulose powder 2 about 1 ml in volume following which the column is washed 3 or 4 times with 1 ml of veronal buffer, 0.1 ionic strength. Most of the damaged components pass through the column with the serum while the undamaged insulin remains adsorbed to the cellulose in the column and can now be eluted slowly with undiluted control serum or plasma. Usually 3 to 4 elu-ates (each 0.5 ml of plasma) are collected and diluted immediately 1:20 to 1:100 with veronal buffer containing 0.025 per cent serum albumin to prevent further damage to the insulin by the concentrated plasma. Although the elution of insulin-I131 from the column is far from complete, adequate amounts are obtained for almost any number of insulin assays. Most of the damaged fraction is removed by this procedure (Figure 1).
Principles of immunoassay. The basis of the technique resides in the ability of human insulin to react strongly with the insulin-binding antibodies present in guinea pig antibeef insulin serum (11, 12, 14), and by so doing, to inhibit competitively the binding of crystalline beef insulin-I131 to antibody. The assay of human insulin in unknown solutions is accomplished by comparison with known concentrations of human insulin. The use of I131-labeled animal insulin as a tracer is necessitated by the lack of a crystalline preparation of human insulin.
The determination of antibody-bound insulin-I181 and free insulin-Im by paper chromato-electrophoresis has been described previously (22). Briefly, the separation of antibody-bound insulin from unbound insulin in plasma results from the adsorption of all free insulin (when present in amounts less than 1 to 5 /ig) to the paper at the site of application ("origin"), while the antibody-bound insulin migrates toward the anode with the intern-globulins. Thus, in the presence of insulin-Im there appear two separate peaks of radioactivity; measurement of the areas beneath the two peaks (by planimetry) yields the relative proportion of bound insulin-Im (migrating with serum globulins) and free insulin-Im (remaining at origin). The ratio of bound insulin-I131 to free insulin-I1S1 (B/F) is a function of the concentration of insulin-binding antibodies, of both insulin concentrations, and of the characteristic kinetic and thermodynamic constants for the reactions between the insulins and the particular antiserum (23). Selection of an antiserum for purposes of this assay is determined primarily by the desirability of obtaining a relatively marked decrease in B/F ratio with small increments in the concentration of human insulin. Although the antibody concentration is of only secondary importance, it should be high enough to permit at least 1: 100 dilution of the antiserum (preferably 1:1,000 dilution or greater). On the basis of preliminary tests the antiserum is diluted appropriately to yield an initial B/F ratio between 2 and 4 for tracer beef insulin-I"1 alone, in the absence of added human insulin. Provided that the amount of the beef insu!in-Im used is truly a tracer quantity, the initial B/F ratio is inversely proportional to the dilution factor
2 Genuine Whatman Cellulose Powder, W & R Balston Ltd., England.
(23). In the presence of human insulin, the B/F ratio decreases progressively with increase in insulin concentration; with sensitive antisera the B/F ratio is reduced by about 50 per cent in the presence of 15 /aU per ml human insulin.
Standard curves. Two preparations of human insulin were employed as standards. The first ("Tietze human insulin")3 is reported (24) to have a potency of 1.8 U per mg crude preparation; the second ("Fisher human insulin"),3 was assayed at 6.8 U per mg in 1956 (25), but it was believed that the activity of the latter preparation might have decreased slightly since its initial preparation (25). A tentative value of 6 U per mg for the Fisher insulin was assigned. However, since a value as low as 22 U per mg could be placed on a crystalline sample of the latter preparation (25), whereas the Tietze insulin was assayed relative to a standard of 27 to 29 U per mg, we have regarded the Tietze crude insulin preparation as 1.8/28 X 100 = 6.45 per cent pure insulin by weight, and the Fisher insulin powder preparation as 6/22 X 100 = 28.2 per cent pure insulin by weight. When compared on this basis, no consistent differences in potencies of the two preparations were observed in the immunoassay procedure and the value of 6 U per mg for the Fisher preparation was accepted as the correct value. Since the Fisher preparation is the more highly purified, it was employed as standard in most of the studies.
All dilutions of insulin and antiserum are prepared in 0.1 ionic strength veronal buffer containing 0.25 per cent human serum albumin to prevent adsorption of reac-tants to glassware. (There is no detectable insulin in commercial supplies of human serum albumin.) Standard solutions each contain identical concentrations of tracer beef insulin-Im (about 0.05 to 0.15 mug per ml but differing in different runs) and antiserum, but varying concentrations of human insulin ranging from 0.05 to 5.0 mfig per ml (calculated as "pure" human insulin). The antiserum is added last in all cases. Mixtures are refrigerated at 4° C for 4 days. These conditions provide sufficient time to reach equilibrium between bound and free insulin. The mixtures are then subjected to chromato-electrophoresis (22) in a cold room at 4° C (Whatman 3 MM paper, veronal buffer, 0.1 ionic strength, pH 8.6, constant voltage 20 to 25 v per cm, cover of apparatus open), which produces a satisfactory separation of the peaks of bound and free insulin-I131 in about 1 to 1.5 hours. Earlier immunoassays (10) were performed after prolonged incubation at 37° C. However, it has since been shown (23) that the standard free energy change of the reaction in the direction of antigen-antibody complex formation is increased considerably at 4° C, which results in an approximately twofold greater slope in the B/F versus insulin concentration curves at low insulin concentrations. Just prior to chromato-
3 We are greatly indebted to Dr. F. Tietze of the National Institutes of Health and Dr. A. M. Fisher of the Connaught Laboratories, Toronto, Canada, for these preparations.
Was this article helpful?