initially detected in the sera of patients with a rare form of

Transcription

1 Proc. Natl. Acad. Sci. USA Vol. 81, pp , December 1984 Cell Biology Analysis of the insulin receptor by anti-receptor antibodies and flow cytometry (expression and regulation of receptor/effects of enzymes/antigenic determinants of receptor) RUTH MARON*, RICHARD A. JACKSON*, STEVEN JACOBSt, GEORGE EISENBARTH*, AND C. RONALD KAHN* *Research Division, Joslin Diabetes Center and Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA 2215; and tburroughs Wellcome, Research Triangle Park, NC 2779 Communicated by Rachmiel Levine, August 3, 1984 ABSTRACT We characterized insulin receptors on a human lymphoblastoid cell line (IM-9) and studied their regulation using anti-receptor antibodies and fluorescence flow cytometry. The fluorescence intensity distribution of insulin receptors on cells was determined by incubating the cells with one of three different anti-receptor antisera (human serum B-9 containing polyclonal autoantibodies, serum from a rabbit with polyclonal antibodies, and a monoclonal antibody to the receptor produced in mouse hybridomas), followed by incubation with an appropriate fluorescein isothiocyanate-labeled second antibody and analysis on an Epics-V flow cytometer. All three anti-receptor antibodies specifically labeled the insulin receptors. The monoclonal antibody showed the highest level of labeling. Treatment of cells with proteolytic enzymes, such as trypsin or chymotrypsin, produced a dose-dependent loss of 125I-labeled insulin (125I-insulin) binding but a relatively small decrease in the binding of anti-receptor antibodies, suggesting that most antibody binding occurred in domains other than the insulin binding site. Treatment with glycosidic enzymes, such as neuraminidase and fi-galactosidase did not affect the binding of 251-insulin, and fluorescence was actually enhanced by about 2% in the.3-galactosidase-treated cells. Exposure of IM-9 cells to insulin resulted in a reduction in the number of insulin receptors. Analysis of the down-regulated cells by immunofluorescence revealed a complete correlation between the percent binding of 125I-insulin and percent peak fluorescence. In all cases, receptors were lost proportionally from all cells, yielding a single symmetrical peak by fluorescence analysis. Exposure of IM-9 cells to anti-receptor antibodies at 37C for 16 hr also produced a down-regulation in the number of insulin receptors. Incubation with human antiserum B-9 caused a 95% loss of both 2.51-insulin binding and peak fluorescence, while the monoclonal antibody resulted in a 5% loss of receptors. Incubation of cells with anti-receptor antibodies for 2 hr at 4C did not produce any receptor loss; however, the human anti-receptor antisera B-2 and B-9 inhibited the binding of the monoclonal anti-receptor antibody by about 5%, suggesting that these antisera contained autoantibodies directed at the monoclonal antibody binding site. These data indicate that insulin receptors can be regulated by both insulin and anti-receptor antibody and demonstrate the utility of immunofluorescence and flow cytometry as a tool for the study of the insulin receptor. The first step in the action of insulin is binding to a specific glycoprotein receptor on the surface of cells. The primary approach for characterization of the insulin receptor has been through the study of its interaction with 125I-labeled insulin (125I-insulin) (1, 2). Antibodies to the insulin receptor also have provided experimental probes of receptor structure and function (3). Anti-insulin-receptor antibodies were The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C solely to indicate this fact initially detected in the sera of patients with a rare form of insulin-resistant diabetes (4), although subsequently it has been possible to raise such antibodies in rabbits and mice by immunization with partially purified receptor (5) and even to produce monoclonal antibodies to the insulin receptor (6, 7). In addition to binding to the receptor, many of these antibodies block insulin binding (8) and mimic many of insulin's biological effects (9). The use of fluorescence probes to study insulin receptors or action has been limited. Murphy et al. studied the binding and internalization of a fluorescein derivative of insulin to Swiss 3T3 cells by flow cytometry (1, 11). Fluorescein has been coupled to insulin directly or via ovalbumin; using these ligands, Schlessinger et al. have shown that insulin receptors undergo patch-and-cap formation at 37C (12). Subsequently, it was shown that fluorescently-labeled anti-receptor antibody cocapped with the rhodamine-coupled insulin (12) and that the rhodamine-insulin-receptor complex was internalized and degraded (13). In the present study, we used fluorescence flow cytometry for identification of insulin receptors and for studying their turnover. Using this technique, we have been able to study (i) the expression of receptors and their regulation by insulin, (ii) compare the effects of acute and chronic exposure of target cells to antibodies directed against insulin receptors, (iii) demonstrate shared antigenic determinants between polyclonal human autoantibodies to the insulin receptor and the monoclonal anti-receptor antibody, and (iv) study the effects of enzyme treatment on the binding of antibodies to the receptor. MATERIALS AND METHODS Antisera and Cultured Cells. Three types of anti-receptor antisera were used in these studies: ( human sera containing naturally occurring polyclonal autoantibodies to the insulin receptor [these are numbered as described (3)]; (ii) a rabbit polyclonal antiserum (41) raised to insulin receptor purified from rat liver membranes (5); and (iii) a monoclonal antibody (IR-1) raised to insulin receptors purified from human placenta and obtained from the hybridoma cells grown as ascites tumqrs in mice (7). Fluorescein-conjugated goat anti-human IgG, rabbit IgG, or mouse IgG were purchased from Cappel Laboratories (Cochranville, PA). IM-9 cells are a human lymphoblastoid cell line that has been extensively utilized for the study of insulin receptors (14). These cells were maintained at 37C in RPMI 164 medium supplemented with 1% fetal calf serum. Treatment of Cells with Proteolytic and Glycosidic Enzymes. Trypsin (24 units/mg of protein; Worthington), a- chymotrypsin (43 units/mg of protein; Sigma), f3-galactosidase from jack bean (19 units/mg of protein; Sigma), and neuraminidase from Clostridium perfringens (19 units/mg of protein; Sigma) were used to test the protease and glycosidase sensitivity of the receptois.1m-9 cells were treated with each of the enzymes before use -in the I251-insulin bind-

2 Cell Biology: Maron et al ing experiments or flow cytometry analysis. In these experiments, the cells were incubated with trypsin or chymotrypsin at 1-4 pug/ml in Ca2"/Mg2+-free phosphate-buffered saline containing 2% bovine serum albumin for 15 min at 37C. Lima bean trypsin inhibitor was then added to the cells exposed to trypsin (in a concentration equal to that of trypsin on a weight basis), and the cells were. then washed three times in phosphate-buffered saline (Pi/NaCl) containing 2% bovine serum albumin. In separate experiments, cells were treated with either neuraminidase (5 milliunits per 2 x 16 cells) or,-galactosidase (25 milliunits) for 3 min at room temperature and then washed twice with RPMI 164 medium. 125I Insulin Binding and Wash Procedures insulin binding to lymphoid cells was measured at 15'C by using a tracer size of.1 ng/ml as described (15). In experiments in which receptor regulation was studied by preincubating the cells with either insulin or anti-receptor antibody, a wash procedure was used to remove insulin or anti-receptor antibody before the binding or immunofluorescence assay. The wash procedure for insulin was by the method described by Kosmakos and Roth (16), and the acid wash procedure to remove anti-receptor antibodies was used as described by Taylor and Marcus-Samuels (15). Immunofluorescent Analysis. Cultured cells were washed twice in cold RPMI 164 medium without serum. The cell pellet was resuspended in cold RPMI 164 medium supplemented with 1% bovine serum albumin to a concentration of 2-4 x 17 cells per ml; 1,ul of cells were incubated with either the test anti-receptor antiserum (in a dilution of 1:1) or control serum for 1 hr at 4C. The cells then were washed twice with P1/NaCl and incubated for an additional hour at 4 C with 1,l of 1:3 dilution of the fluorescein-conjugated goat anti-human, anti-rabbit, or anti-mouse IgG. Cells were washed twice with P1/NaCl saline and analyzed with a Coulter Epics V flow cytometer (Coulter). After gating on forward-angle light scatter, 1-25 x 13 cells were analyzed by logarithmic integral green fluorescence. As all patterns approximated normal curves, the peak fluorescence was taken as the fluorescence channel with the highest number of cells. Peak fluorescence was then converted from channel units of logarithmic fluorescence to linear units and used to derive percentage of maximal peak fluorescence. RESULTS Characterization of the Insulin Receptor in IM-9 Cells by Indirect Immunofluorescence. The fluorescence intensity distribution of insulin receptors on IM-9 cells, a human lymphoblastoid cell line, was determined by incubating the cells with either human anti-receptor antibodies (B-9), rabbit antisera to the insulin receptor (41), or mouse ascites fluid containing monoclonal antibodies to the insulin receptor (IR-1) (Fig. 1). All three anti-receptor antibodies specifically labeled insulin receptors. In all cases, mean peak fluorescence of the cells treated with the anti-receptor antibodies was at least 3-4 times higher than the peak fluorescence of the cells incubated with nonimmune control sera. The monoclonal anti-receptor antibody showed the greatest specific labeling (about 7 times higher fluorescence than the negative control). Specific fluorescence could be detected with the monoclonal and B-9 antibodies even in a 1:1 dilution (data not shown). The human nonimmune control sera gave a relatively high level of nonspecific fluorescence, probably because it contains antibodies against human cell surface antigens. Preincubation with the human autoantibodies to the insulin receptor has been shown previously to inhibit '25I-insulin binding to IM-9 cells. A close correlation between percent inhibition of binding of 1251-insulin and percent peak fluorescence was observed over a range of dilutions from 1:1 to 1:5 (data not shown). With anti-receptor antibody B-9, Proc. NatL. Acad Sci. USA 81 (1984) 7447 Fluorescence intensity FIG. 1. Immunofluorescence analysis of the insulin receptor in IM-9 cells by anti-receptor (anti-r) antibodies. IM-9 cells (2 x 16 cells per 1 /1) were incubated for 1 hr at 4C with a 1:1 dilution of the different sera. The cells were then washed with P,/NaCl and exposed to a 1:2 dilution of fluorescein-labeled second antibody for 1 hr at 4C. The cells were washed again and suspended in Pi/NaCl for analysis on an Epics-V flow cytometer (2.5 x 14 cells per run). The abcissa represents the channel units of fluorescence intensity; the ordinate, the relative cell number. (A) Fluorescence histograms of IM-9 cells incubated with either p3, a nonspecific (control) monoclonal antibody, or IR-1, a mouse monoclonal anti-r antibody. (B) Fluorescence histograms of IM-9 cells incubated with either normal rabbit serum or a rabbit polyclonal anti-insulin-receptor antiserum (41). (C) Fluorescence histograms of IM-9 cells incubated with either normal human serum or a human polyclonal anti-r antiserum (B-9). maximal peak fluorescence and maximal inhibition of 1251_ insulin binding was observed at the 1:1 dilution. The rabbit polyclonal antiserum and mouse monoclonal antibody do not inhibit insulin binding (5, 7). Effect of Insulin Down-regulation on the Immunofluorescence and 1251-Insulin Binding. Exposure of IM-9 lymphocytes to insulin has been shown to result in a time and concentration reduction in the number of insulin receptors as measured by 1251-insulin binding (14, 16). This has been termed down-regulation. Fig. 2 shows insulin receptor loss from IM-9 cells preincubated for 16 hr at 37 C with 1,uM to.1 nm insulin. The fall of 125I-insulin binding and immunofluorescence was directly related to the concentration of insulin in the medium during the preincubation. Cells downregulated for 16 hr with 1,uM insulin had 39% binding of the control level of insulin binding and 29% of the peak fluorescence, while cells incubated with 1 nm insulin had 85% of maximal binding and 8% of the peak fluorescence. In all cases, the patterns generated by the flow cytometer revealed a uniform loss of receptors from all cells-i.e., the receptor concentration as measured by fluorescence was normally distributed both before and after down-regulation. At no time were two populations of cells detected with respect to receptor concentration. Effects of Enzymatic Treatment on Immunofluorescence

3 7448 Cell Biology: Maron et al ~~~~~~~~ 1 _ Z~~~~~~~~~~~~~ Dr 8' e- t> -o 'R 4- E s2- V~~~~~ lo Insulin, M FIG. 2. Binding of 125I-insulin (o) and monoclonal anti-receptor (anti-r) antibodies (e) to IM-9 cell down-regulated by insulin. IM-9 cells (2 x 16 per ml) were incubated for 16 hr at 37C with various concentrations of unlabeled insulin (1 AM to.1 nm). Thereafter, cells were washed at ph 6. at 3'C to remove the exogenous insulin (16) and assayed for their insulin binding and immunofluorescence. The peak fluorescence was taken as the fluorescence channel with the highest number of cells. and 1251-Insulin Binding in IM-9 Cells. As described (9), both trypsin and chymotrypsin treatment of IM-9 cells produced a dose-dependent loss of 1251-insulin binding (Fig. 3). By contrast, the binding of anti-receptor antibodies measured by indirect immunofluorescence was hardly affected, suggesting that most antibody binding occurred to domains different from that involved in insulin binding. Treatment of IM-9 cells (2 x 16 per ml for 3 min at room temperature) with glycosidic enzymes such as neuroaminidase (5 milliunits) and /3-galactosidase (25 milliunits) did not affect the binding of 1251-insulin, but the peak fluorescence was enhanced by 1 Proc. Natl. Acad Sci. USA 81 (1984) 2-3% of the /3-galactosidase-treated cells (data not shown). Acute Blocking and Chronic Down-regulation of Receptors on IM-9 Cells by Anti-receptor Antibodies. Autoantibodies to the insulin receptor block insulin binding and mimic many effects of insulin (9). To determine if these antibodies could block the binding of the monoclonal antibody or down-regulate the insulin receptor, IM-9 cells were incubated with various anti-receptor antibodies for 2 hr at 4TC or 16 hr at 37C, washed, and then analyzed by flow cytometry. The cell wash was performed either at ph 3.5 to dissociate anti-receptor antibodies from the cell surface or at ph 7.8-a condition that will remove only those antibodies and other serum factors loosely bound to the cell surface. When the cells were preincubated with the monoclonal anti-receptor antibody for 2 hr at 4TC, neither the binding of insulin nor of a subsequent addition of monoclonal antibody was affected (Fig. 4). In contrast, when the cells were preincubated with the human serum B-9 containing autoantibodies to the receptor for 2 hr at 4TC and subjected to a neutral wash, the binding of the monoclonal anti-receptor antibodies, as well as of I-insulin, was reduced by -5%, suggesting that this polyclonal antiserum contained antibodies directed at both the insulin binding site and the antigenic domain recognized by the monoclonal antibody. Preincubation of cells with other human sera containing autoantibodies to the insulin receptor for 2 hr at 4 C reduced the binding of the monoclonal anti-receptor antibody to a variable degree (Fig. 5). Preincubation with human serum B- 2 reduced monoclonal antibody binding by 7%, whereas normal serum resulted in only a 15% reduction. Other antireceptor sera produced a decrease in monoclonal anti-receptor antibody binding by only 15-3%. Inhibition by most sera was completely reversible, with an acid wash designed to remove surface IgG. With sera B-2 and B-9, there was still about 2-3% inhibition of 125I-insulin binding and monoclonal anti-receptor antibody binding after the acid wash procedure, suggesting that either some loss of insulin receptors OD 8- I-,6-4- 8) u C._ No treatment Blocking by Anti-R Monoclonal IR-1 Human B-9 Z 2- E 1 E None Enzyme,,g/ml co *8)'a FIG. 3. Effect of trypsin (Lower) and chymotrypsin (Upper) on insulin binding and immunofluorescence in IM-9-treated cells. IM-9 cells (2 x 16 cells per ml) in Ca2+/Mg2+-free Pi/NaCl with 2% bovine serum albumin were untreated (none) or treated with trypsin or chymotrypsin at 1, 2, or 4 jig/ml for 15 min at 37 C. After the incubation, trypsin inhibitor was added, and the cells were washed three times with Ca2+/Mg2+-free Pi/NaCl containing 2% bovine serum albumin. The cells were then assayed for their insulin binding and anti-receptor antibody binding. Both are expressed as a percentage of the maximal value observed in untreated cells. -E :3. O C O r. ) loou 8j 6k 4f1 2k INS IR-1 B-9 I INS IR-1 B-9 INS Method of analysis I1 IR-1 B-9 FIG. 4. Anti-receptor (anti-r) antibody blocks the insulin receptor in IM-9 cells. IM-9 cells (4 x 17 cells per ml) were incubated with either monoclonal anti-receptor (anti-r) antibodies (IR-1) or human autoantibodies to the insulin receptor (B-9) in a 1:1 dilution for 2 hr at 4 C. Thereafter, cells were washed in either a neutral buffer (ph 7.8) (o) or an acid buffer (ph 3.5) (o) (15), and specific binding of 1251-insulin (INS) or peak fluorescence with anti-r antibodies was measured. Results are expressed as percent binding of maximal 1251-insulin or peak fluorescence observed in the untreated cells.

4 E 8, J., 6' NS 13-5 B-) B-RI B-4 B-9 B-2 Huiman scrum (1:8( dilution) FIG. 5. Blockade of binding of monoclonal anti-receptor antibody (anti-r MAb) by human autoantibodies to the insulin receptor. IM-9 cells (2 x 16 cells per 1 Al) were incubated with different human anti-receptor antisera or normal control serum (NS) in a 1:8 dilution for 2 hr at 4C. Thereafter, the cells were washed at ph 7.8, and specific binding of anti-r MAb was measured by immunofluorescence as described. The results are presented as the percent of the maximal peak fluorescence-i.e., the fluorescence of anti-r MAb bound to IM-9 cells that had not been preincubated with human serum. had occurred at 4C in 2 hr or that the human autoantibodies were tightly bound and could not be completely dissociated. Analysis of the cells using a fluorescein isothiocyanate-labeled anti-human IgG suggested that the latter was the case, since the binding of the anti-human IgG was increased in IM- 9 cells exposed to human antisera containing antibodies to the insulin receptor even after the acid wash. Prolonged exposure of cells to anti-receptor antibodies also induced receptor down-regulation. Incubation of cells with monoclonal anti-receptor antibodies for 16 hr at 37C, followed by neutral or acid wash, revealed a 5% decrease in 1251-insulin binding and a similar decrease in peak fluorescence when assayed with the human autoantibodies (Fig. 6). The inhibition in peak fluorescence with the monoclonal antibodies was even greater (7-9%). Prolonged exposure AD FIG. 6. Cell Biology: No treatment Maron et al Down-regulation by anti-r Monoclonal IR-1 Human B-9 Method of analysis Anti-receptor (anti-r) antibody induced down-regulation of insulin receptors. IM-9 cells were incubated with 1:1 dilution of anti-r antiserum for 16 hr at 37C. Thereafter, cells were washed at either ph 7.8 (i) or ph 3.5 (i), and 1251-insulin binding (INS) or immunofluorescence was measured. Results are expressed as percent maximal binding or peak fluorescence of nontreated cells. The cells were down-regulated with monoclonal anti-r antibodies (IR-1) or with human autoantibodies (B-9) to the insulin receptor. Proc. Natl. Acad Sci. USA 81 (1984) 7449 of IM-9 cells to the human anti-receptor antiserum B-9 resulted in 9% loss of 125I-insulin and monoclonal anti-receptor antibody binding. There was no difference between the cells subjected to the neutral and acid wash, suggesting that the decrease in binding of insulin and antibodies was a reflection of true receptor loss and not simply a blockade of binding. DISCUSSION Autoimmune (4) and xenogeneic (5) polyclonal antisera to the insulin receptor, as well as the more recently produced monoclonal antibodies (6, 7), have provided useful tools for purifying insulin receptors and studying several of their properties. These different types of anti-receptor antibodies, however, exhibit different properties. The polyclonal antireceptor antibodies have little tissue or species specificity, whereas the monoclonal antibodies, thus far available, exhibit strong species specificity and some tissue specificity. The human autoantibodies to the insulin receptor inhibit insulin binding, immunoprecipitate solubilized receptors, and mimic many of the actions of insulin (17), whereas both the polyclonal rabbit antiserum to insulin receptor and the monoclonal antibody (IR-1) immunoprecipitate the insulin receptor, but neither inhibits insulin binding, and only the polyclonal antibody mimics insulin action (5). These results suggest that those antibodies interact with the receptor at sites distinct from the insulin binding site. In this study, we have characterized the insulin receptor on a human lymphocyte cell line (IM-9) and studied its regulation using anti-receptor antibodies and flow cytometry. All anti-receptor antibodies tested specifically labeled the insulin receptor in intact cells, indicating that all bound to extracellular domains on the receptor. The monoclonal anti-receptor antibody raised to human receptor showed the greatest specific labeling. The polyclonal antisera showed less specificity primarily because the nonimmune control sera gave relatively higher levels of fluorescence. In the case of the human sera, this might reflect the presence of antibodies against human cell surface antigens or immunoglobulin receptors on the surface of the cultured lymphocyte cell line, which could bind both nonimmune IgGs and the fluorescein isothiocyanate-labeled second antibody nonspecifically. The higher level specificity of the monoclonal anti-receptor antibody binding was also apparent by the high level of correlation between immunofluorescent labeling and 125I-insulin binding. At low levels of receptor, the human polyclonal antiserum exhibited somewhat higher binding, probably reflecting the presence of other antibodies in the sera, which continue to bind to cells in which insulin receptor content is low. Previous studies have examined the effects of enzymatic treatment on insulin receptor binding and insulin biological responses (18-23). Insulin binding to IM-9 cells and adipocytes is lost after mild trypsin treatment; however, trypsin treatment has relatively little effect on the insulin-like activity of some anti-receptor antibodies (19). In this study, we found that treatment with proteolytic enzymes produced a dose-dependent loss of 125I-insulin binding but had only a small effect on the binding of anti-receptor antibodies measured by indirect immunofluorescence. These results suggest that the monoclonal antibody (and most of the IgGs in the polyclonal antisera) recognize determinants on the receptor molecule that are distinct from those recognized by insulin and are less sensitive to trypsin. This is consistent with previous observations (i) that the monoclonal anti-receptor antibodies and rabbit polyclonal antibodies can immunoprecipitate the receptor but cannot inhibit 125I-insulin binding and (ii) that the human polyclonal antisera often show more insulin-like activity than can be accounted for by

5 745 Cell Biology: Maron et al their titer in binding-inhibition assays. Neuraminidase and f3- galactosidase treatment had little or no effect on 125I-insulin or antibody binding, suggesting that neither terminal sialic acid nor galactose residues is required for recognition by these ligands. Therefore, changes in the biological effect of insulin and anti-receptor antibody observed after treatment with glycosidic enzymes (9, 21) must reflect their activity elsewhere. It has been demonstrated that the concentration of insulin receptors on cells is inversely related to the chronic level of insulin to which the cells are chronically exposed (14). In vivo and in vitro studies have shown that up to 75% of the receptors can be lost in the presence of insulin (24); in IM-9 cells, this appears to be due to increased receptor degradation (16). Antibodies directed against the insulin receptor have been shown to mimic the effect of insulin to down-regulate insulin receptors in IM-9 and 3T3 cells (15, 25, 26); however, in these and most other studies of receptor regulation, the only measure of receptor content has been insulin binding. In the current study, we down-regulated the insulin receptors on IM-9 cells by exposing them to insulin or to anti-receptor antibodies and analyzed the process by measuring immunoreactive receptors with the fluorescence flow cytometry technique. In general, there was a good correlation between percent binding of 125I-insulin and percent peak fluorescence in the cells down-regulated by either insulin or anti-receptor antibodies. Interestingly, although the monoclonal anti-receptor antibodies used were not able to inhibit insulin binding, they were able to down-regulate the insulin receptor by -5% at 37C in 16 hr. The human polyclonal autoantibodies were even more effective in regulation of the receptor. Thus, the human anti-receptor antibodies can produce insulin resistance by both blocking insulin binding and modulating insulin receptor number. A similar effect on receptor regulation has been observed with autoantibodies to the acetylcholine receptor (27). When cells were incubated with monoclonal anti-receptor antiserum at 4C for 2 hr, 125I-insulin binding and anti-receptor antibody peak fluorescence were not affected. In contrast, if the cells were incubated with the human B-9 autoantibodies for 2 hr at 4C, there was a masking effect that was partially reversible by the acid wash. 125I-insulin binding and peak immunofluorescence were reduced by about 3% in comparison to almost 1% inhibition when the cells are incubated for 16 hr at 37C. Most of this appears to be due to residual antibody on the cell; however, some down-regulation may occur, even at this low temperature and short time. Two additional features of down-regulation can be discerned using the fluorescence technique. First, since fluorescence is analyzed individually for each cell, it is possible to ask the question whether a 5% down-regulation of receptors represents a 5% loss of receptors from all cells or a 1% loss of receptors from only 5% of the cells. Analysis using the flow cytometer indicates that all cells are proportionally involved in the receptor loss process. A second feature that can be studied uniquely with the immunofluorescence techniques is the question of whether the washing techniques, in fact, remove all of the surfacebound antibodies before the assay of receptor content is made and whether the antigenic determinants to which the monoclonal anti-receptor antibody is directed differ from those to which the human antisera are directed. Most of the human polyclonal antibodies share almost no antigenic determinants with the monoclonal antibodies; thus, these sera have little or no effect to block monoclonal antibody binding. The two highest-titer human antisera (B-9 and B-2) block almost 7% of the binding of the monoclonal antibodies, suggesting that they do contain antibodies that bind to a site at or near the site of binding of the monoclonal anti- Proc. Natl. Acad Sci. USA 81 (1984) receptor antibody. We also found that some residual human anti-receptor antibodies persist on the cells, even after the acid-wash procedure, indicating the high affinity of their binding and the difficulty in completely removing such antibodies from cells. In summary, insulin receptors can be identified by specific immunofluorescence and flow cytometry. The data indicate that this technique can provide a useful tool to aid in understanding the factors that control expression of insulin receptor and to study variation in expression of the receptor and the role of receptors in insulin action. We thank Ms. Kathy George for her skillful technical assistance with the cell sorter and Ms. Barbara Cornell for typing this manuscript. This work was supported by National Institutes of Health Grants AM 3136 and AM Freychet, P., Roth, J. & Neville, D. M. (1971) Proc. Nail. Acad. Sci. USA 68, Cuatrecasas, P. (1971) Proc. Nail. Acad. Sci. USA 68, Kahn, C. R. & Harrison, L. C. (1981) in Carbohydrate Metabolism and Its Disorders, eds. Randle, P. J., Steiner, D. F. & Whelan, W. J. (Academic, London), Vol. 3, pp Flier, J. S., Kahn, C. R., Roth, J. & Bar, R. S. (1975) Science 19, Jacob, S., Chang, K. J. & Cuatrecasas, P. (1978) Science 2, Roth, R. A., Cassell, D. J., Wong, K. Y., Maddux, B. A. & Goldfine, I. D. (1982) Proc. Natl. Acad. Sci. USA 79, Kull, F. C., Jacobs, S., Ying-Fu, S. & Cuatrecasas, P. (1982) Biochem. Biophys. Res. Commun. 16, Flier, J. S., Kahn, C. R., Jarrett, D. B. & Roth, J. (1977) J. Clin. Invest. 6, Kahn, C. R., Baird, K., Flier, J. S. & Jarrett, D. B. (1977) J. Clin. Invest. 6, Murphy, R. F., Powers, S., Verderame, M., Cantor, C. R. & Pollack, R. (1982) Cytometry 2, Murphy, R. F., Powers, S. & Cantor, C. R. (1984) J. Cell Biol. 98, Schlessinger, J., Van Obberghen, E. & Kahn, C. R. (198) Nature (London) 286, Schlessinger, J., Schechter, Y., Willingham, M. C. & Pastan, I. (1978) Proc. Natl. Acad. Sci. USA 75, Gavin, J. R., Roth, J., Neville, D. M., De Meyts, P. & Buell, D. N. (1974) Proc. Nail. Acad. Sci. USA 71, Taylor, S. I. & Marcus-Samuels, B. (1984) J. Clin. Endocrinol. Metab. 58, Kosmakos, F. C. & Roth, J. (198) J. Biol. Chem. 255, Kahn, C. R., Baird, K., Flier, J. S., Grunfeld, C., Harmon, J. T., Harrison, L. C., Karlsson, F. A., Kasuga, M., King, G. L., Lang, V. C., Podskalny, J. & Van Obberghen, E. (1981) Recent Prog. Horm. Res. 37, Cuatrecasas, P. (1971) J. Biol. Chem. 246, Harrison, L. C., Flier, J. S., Itin, A., Kahn, C. R. & Roth, J. (1979) Science 23, Cuatrecasas, P. (1972) in Insulin Action, ed. Fritz, I. B. (Academic, London), pp Cuatrecasas, P. & Illiano, G. (1971) J. Biol. Chem. 246, Jacobs, S., Hazum, E. & Cuatrecasas, P. (198) Biochem. Biophys. Res. Commun. 16, Maron, R., Kahn, C. R., Jacobs, S. & Fujita-Yamaguchi, Y. (1984) Diabetes 33, Kahn, C. R., Neville, D. M. & Roth, J. (1973) J. Biol. Chem. 248, Roth, R. A., Maddux, B. A., Cassell, D. J. & Goldfine, I. D. (1983) J. Biol. Chem. 258, Grunfeld, C. (1984) Proc. Nail. Acad. Sci. USA 81, Drachman, D. B., Angus, C. W., Adams, R. N., Michelson, J. D. & Hoffman, G. T. (1978) N. Engl. J. Med. 298,