Summary. A precise, reproducible radioreceptor assay has been developed for insulin receptors on human erythrocytes. The mean specific binding in 11 normal.
AJEBAK 59 (Pt. 4) 439-448 (1981)
©THE ERYTHROCYTE INSULIN RECEPTOR by A. McELDUFF AND C. J. EASTMAN (From the Endocrine Unit, Department of Medicine, Westmead Hospital, Westmead, N.S.W., Australia, 2145) (Accepted for publication March 5, 1981.) Summary. A precise, reproducible radioreceptor assay has been developed for insulin receptors on human erythrocytes. The mean specific binding in 11 normal volunteers was 7-7% ± 1 6 (S.D.) per 2 25 X lOO erythrocytes. The intra and inter assay coefficients of variation are 4%. The erythrocyte insulin receptor is specific for insulin and recognises insulin analogues in proportion to their respective biological activities. The pH and temperature dependence, the negative cooperativity and the inverse relationship between specific binding and fasting plasma Insulin levels suggest it is similar to insulin receptors elsewhere. The erythrocyte insulin receptor offers a model for the study of insulin/insulin receptor interaction in man.
INTRODUCTION Insulin receptors on circulating monocytes reflect the status of insulin receptors on other insulin sensitive tissues (Olefsky, 1976) and have provided a model for dynamic studies in humans (Bar and Roth, 1977). This has provided valuable information on the role of receptors in modulating insulin action. There are, however, several problems associated with their use: for example, moderately large volumes of blood (60 to 120 ml) are required to harvest cells for binding studies and repeated venesection alters the binding characteristics of the remaining circulating monocytes (Beck-Nielsen and Pedersen, 1978). For these reasons, serial studies requiring frequent blood sampling are difficult to perform. Insulin receptors have been demonstrated recently on human erythrocytes (RBC) (Gambhir, Archer and Carter, 1977) and may provide a useful tool for serial studies in man. We present in this communication modifications of this radioreceptor assay which improve precision, reproducibility and ease of performance, together with data which suggest that the erythrocyte receptor is similar to insulin receptors elsewhere. MATERIALS AND METHODS Erythrocyte preparation Ten ml of heparinised blood wa.s centrifuged at 120 .£,' for 10 min to remove plasma. The precipitated cells were resuspended in 10 ml of incubation buffer and layered on to a Ficoll-Hypaque gradient in the ralio 3:2 by volume (Boyum, 1968). This preparation was
440
A. McELDUFF AND C. J. EASTMAN
centrifuged at 400 g for 20 min before removal of plasma, wbite cells and the Ficoll-Hypaqtie to leave the RBC pellet untouched (Boyum. 1968). The pellet was resuspended in six to sevenfold excess (by volume) incubation buffer, cenirifuged at 1000 a for 5 min and the supernatant removed. The RBCs were resuspended in incub;ition buffer before use. RBC viability was invariably greater than 98% as assessed by ihe exclusion of trypan blue. Tbe working RBC solutions were counted on a Coulter counter and contained bolb reticulocytes and mature red blood cells. Contaminaiion witb granulocytes and platelets occurred in the ratio of one granulocyte and one platelet per 2000 red cells. Methods Human insulin was iodinated by the method of Hunter and Greenwood (1962) with modifications of De Meyts) (1976a). Ten ul of I 4 X lO-'M chloramine T were added lo 10 /
50
O
CO
40 Z
30
tfT*
-
20 10 15
30
45
60
75
90
TIME mm Fig. 5. Illustrates the dissociation of ^-H insulin from the receptor in response to dilution by a hundredfold excess diluting buffer (upper unbroken line) :ind dilution in the presence of 1-5 X lO-* Mg/ml of unlabelled insulin (lower broken line). The accelerated dissociation in the presence of unlabelled insulin is characteristic of negative coopcraiiviiy.
THE ERYTHROCYTE INSULIN RECEPTOR
445
plateau in specific binding was reached by 150 min. At this temperature and time both insulin degradation, as assessed by TCA precipitability or rebinding experiments, and receptor degradation, as assessed by preincubation of the cells for 150 min. were both less than 5% of control values (De Meyts, 1976b). The appropriate control experiments were performed for both ^^''I insulin degradation and receptor degradation. Thus, the further increase in binding observed at 4° (Fig. 4) could not be explained by inhibition of receptor or insulin degradation. Because insulin receptor and insulin degradation are both less than 5% of control, at 150 min at 15°, and because the binding has plateaued, conditions at this time approximate equilibrium conditions and the data can be used in Scatchard analysis. Binding was optimal at pH 8 and decreased significantly witb minor changes in pH; for example, at pH 7 binding was reduced by 50%. The influence of unlabelled insulin on the dissociation of ^-•'''I insulin from the receptor is shown in Fig. 5. The presence of unlabelled insulin in an incubation dilution (1/100) which approximates an infinite dilution (De Meyts, Bianco and Roth, 1976) accelerates the dissociation of ^-''I insulin from its receptor compared with that of dilution alone. Specific binding of ^-"'I insulin to the erythrocyte receptor, from cells collected in the fasting state, increased as fasting plasma insulin decreased, although this relationship failed to reach statistical significance, (r = 0 54, 0 1 > p > 0 05).
DISCUSSION The binding of insulin to erythrocytes was first reported in 1954, albeit in minute amounts compared with its binding to leucocytes (Haugaard, Haugaard and Stadie, 1954). Although this finding was confirmed by more sophisticated methods developed for studying insulin receptors on cell membranes, high nonspecific binding obscured the nature of the interaction of insulin with its receptor (Gavin et al. 1972). Subsequently, attention was focussed upon insulin binding to circulating mononuclear cells as a model for studying insulin receptor interaction in man. Recently, Gambhir and coworkers reported a modification of the monocytc insulin radioreceptor assay permitting measurement of specific binding of insulin to human erythrocytes (Gambhir et al, 1977; Gambhir, Archer and Bradley. 1978). Our modifications of Gambhir's method result in improvement in three areas. First, improved reproducibility and precision: second, lower and more constant nonspecific binding, and finally a simplification of the assay procedures. In the published reports of Gambhir et al (1977, 1978) they emphasise the problems inherent in the dibutyl phthalate separation method wherein bubble trapping results in variable misclassification errors with resultant poor reproducibility and precision. Our wash step overcomes this problem and is in fact easier to perform.
446
A. McELDUFF AND C . J. EASTMAN
A problem wiiich became apparent while developing the wash separation, and probably contributes to the relatively high NSB and difficulty with reproducibility using Gambhir's method, was the marked variability of nonspecific binding using an apparently constant technique. We have found major and variable changes in nonspecific binding with different incubation tubes and, in several types of tubes, variations in NSB among tubes of the same batch. Similarily, if cells are allowed to sediment during incubation, nonspecific binding varies markedly. By carefully selecting and screening incubation tubes, by incubating in a shaking water bath, and by using the wash separation technique, NSB is reduced to less than I 9h of added tracer and remains constant. The problems associated with misclassification errors are well known and the reduction in both the absolute degree of and variations in these errors is advantageous. However, the problems of variations in NSB are more subtle but equally important. An explanation of these problems involves an understanding of the analysis of the curvilinear Scatchard plots obtained when insulin-insulin receptor interactions are examined. The method of analysis is not uniformly agreed upon. The method we have chosen, that of De Meyts and Roth (1975), relies on the principle that the insulin receptor displays negative cooperativity; that is, the affinity of the receptor for insulin decreases as insulin binds to the receptor, analogous to the positive cooperative binding of oxygen to haemoglobin. The critical experiment supporting negative cooperativity demonstrates that, after preincubation of ^-^'I insulin wifh tbe receptor, the addition of unlabelled insulin accelerates the dissociation of ^^''I insulin when effectively at an infinite dilution. This basic experiment is shown in Fig. 5. A full discussion of negative cooperativity can be found in the work of De Meyts (1976b), and has recently been critically reviewed by Rodbard (1979). The Scatchard plot of insulin binding is curvilinear, and minor variations in NSB alter the position of the terminal portion of the curve and therefore alter the projection of this curve to the abscissa. The subsequent calculations of both receptor concentration and afiinity depend upon this x-intercept and can thus be significantly influenced by minor variations in NSB. It is therefore apparent that the calculated receptor concentration and affinity will be more reproducible if variations in NSB are kept to a minimum. It is worthwhile at this point to consider the variation in erythrocyte receptor concentration and affinity which have been reported using Gambhir's methodology. The definition of NSB, i.e., binding not due to receptor binding and estimated as residual 'binding' occurring in the presence of an excess of unlabelled hormone, is critical to the derivation of affinity and capacity by Scatchard analysis of the primary displacement data. The amount of excess unlabelled insulin is arbitrarily chosen and varies in radioreceptor assays from one laboratory to another. The concentration chosen influences not only the calculation of binding and the concentration of bound, but also limits the maximum amount bound. Consequently, this markedly affects the projected
THE ERYTHROCYTE INSULIN RECEPTOR
447
X axis intercept, because the changes are more pronounced as the projection to the abscissa now begins at a significantly different point on the curve. Thus. calculated values of receptor aRinity and capacity vary with the concentralion of excess unlabelled insulin used in the assay system. Consequently, these parameters should be reported and examined within the context of assay conditions employed to derive them and are useful for comparative purposes only. The data we report using 100 /^g/nil of excess unlabelled insulin define NSB as similar to that of Gambhir et al. (1978) but markedly different from concentrations used in monocyte radioreceptor assays (Olefsky. 1976; Bar and Roth, 1977). The present study demonstrates specific receptors for insulin on human erythrocytes which bind insulin with high affinity and low capacity and bind insulin analogues in proportion to their known insulinomimetic activities. Since the insulin receptor has not been isolated in pure form, it is necessary to define the receptor by its binding characteristics. The known physicochemical features of insulin binding to insulin sensitive tissues (Freychet. 1976). such as pH optima, temperature dependence and negative cooperativity, are also characteristics of insulin binding to human erythrocytes. Thus, it appears that the insulin binding site on the erythrocyte membrane is functionally similar if not identical to insulin binding sites on other tissues. The physiological significance of insulin receptors on the erythrocyte remains unclear, as there is no evidence that insulin stimulates glucose uptake into mature erythrocytes. The absence of a specific biological effect in response to insulin binding to the cell membrane raises questions regarding the function of this receptor site. Regardless of whether it is a true hormone receptor site, coupled to a specific biological response mechanism, the physicochemical properties of the erythrocyte insulin receptor are similar, if not identical, to receptors on known insulin sensitive tissues (Freychet, 1976). We have described a radioreceptor assay for insulin binding to human erythrocytes which is easy to perform and which is specific, precise and reproducible. It has minimal nonspecific binding and can be performed on less than 10 ml of whole blood in contrast to established receptor assays on monocytes which require large volumes of blood. Preliminary results from physiological studies reveal that it may offer a model for dynamic studies of insulin receptors in man. REFERENCES BAR, R. S., and ROTH, J. (1977): 'Insulin receptor status in disease of man.' Arch. Intern. Med., 137. 474-481. BHCK-NIELSEN,
H..
and
PEDERSEN,
cyles from human Wood.' Scand. J. Clin. i^b. Invest., 21, Suppl. 97, 51-76.
O.
DE MEYTS. P. (1976a): 'Tnsiilin and growth
(1978): 'Diurnal variation in in?;ulin binding to human monocytes.' /. Clin. Endocrinot. Metab., 47, 385-390. BovuM, A. (1968): 'A one-stage procedure for isolation of granulocytes and lynipho-
hormone receptors in human culiured lymphocytes and peripheral blood monocytes." In "Methods in Receptor Research." Marcel Dekker. N.Y., Blecher Med., p. 328.
448
A. McELDUFF AND C. J. EASTMAN
DE MEVTS, P . (1976b): 'Insulin and growth hormone receptors in human cultured lymphocytes and peripheral blood monocytes.' In "Methods in Receptor Research.'* Marce! Dekker, N.Y., Blecher Med., p. 356. DE MEYTS. P., and ROTH, J. (1975): 'Co-
operativity in ligand binding: a new graphic analysis.' Biochem. Biophys. Res. Commtot., 66, 1118-1126. D E MEYTS. P., BIANCO, A. R., and ROTH,
J. (1976): 'Site-site interactions among insulin receptors. Characterization of the negative cooperativity.' /. Biol. Chem., 251, 1877-1888. FREYCHLT, P . (1976): Interactions of polypeplidc hormones with cell membrane specific receptors: studies with insulin and glucagon.' Diabetologia, 12, 83-100. GAMBHIR,
K.
K.,
ARCHER,
J.
A.,
and
BR.ADLEY. C . J. (1978): 'Characteristics of human erythrocyte insulin receptors.' Diabetes, 27, 701-708. GAMHHIR,
K. K . , ARCtlER,
J.
A., and
CARTER, L. (1977): 'Insulin radioreceptor assay for human erythrocytes.' Clin. Chem.. 23, 1590-1595.
GAVtN, J. R., ROTH,
I., JEN, P.. and
FREYCHET, P. (1972): 'Insulin receptors in human circulating cells and fibroblasts.' Proc. Nat. Acad. Sci., 69, 747. HAUGAARD,
N.,
HAUGAARD,
E.
J.,
and
STADIE, W . C . (1954): 'Combination of insulin with cells.' /. Biol. Chem.. 211, 289-295. HUNTER, W . M . . and GREENWOOD, F . C .
(1962): 'Preparation of iodine—131 labelled growth hormone of high specific activity.' Nattdrc, 194, 495-496. OLEFSKY, J. M. (1976): 'Decreased insulin binding to adipocytes and monocytes from obese subjects.' /. Clin. Invest.. 57, 1165-1172. RODBARD, D. (1979); 'Negative cooperativity: a positive finding?' Am. J. Physiot., 6, E2O3. SCATCHARD, G . (1949): 'The attraction of proteins for small molecules and ions.' Ann. N.Y. Acad. Sci.. 51. 660-672. SUNDBY, F . , and
MARKUSSEN, J.
(1970):
'Preparative method for the isolation of C-peptides from ox and pork pancreas.' Horm. Metab. Res.. 2, 17-20.
