The Human Placenta Contains Two Distinct Binding and Immunoreactive Species of Insulin-like Growth Factor-I Receptors*

Two species of insulin-like growth factor-I (IGF-I) receptors in human placenta have been delineated on the basis of their immunoreactivity with an autoanti- serum (B-2) to the insulin receptor. When all the IGF-I binding sites in solubilized human placenta were as- sayed by polyethylene glycol precipitation, a curvilinear Scatchard plot was obtained which could be re- solved into two single classes of binding sites: one immunoprecipitable by B-2 IgG and the other, nonim- munoprecipitable. The B-2 reactive sites bound IGF-I with lower affinity (Kd = 7.1 X 10"O M) than the B-2 nonreactive sites (ICd = 2.1 X 10"O M) and cross-reacted more readily with insulin, the IGF-I/insulin-binding potencies being -120 and -1100, respectively. Both receptor subtypes bound IGF-I with -30-fold higher affinity than multiplication-stimulating activity, and, after affinity cross-linking with '261-IGF-I, migrated as specific reduced bands of M, = 138,000 during sodium dodecyl sulfate-polyacrylamide gel electropho- resis. The subunit sizes of the B-2 reactive IGF-I receptor were similar to those of the insulin receptor. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of '2sI-labeled receptors immunoprecipitated by autoantiserum B-2 or autoantiserum B-10 (which recognizes only insulin receptors) revealed,

The Human Placenta Contains Two Distinct Binding and Immunoreactive Species of Insulin-like Growth Factor-I Receptors* (Received for publication, February 17,1984) Helen A. Jonas  Two species of insulin-like growth factor-I (IGF-I) receptors in human placenta have been delineated on the basis of their immunoreactivity with an autoantiserum (B-2) to the insulin receptor. When all the IGF-I binding sites in solubilized human placenta were assayed by polyethylene glycol precipitation, a curvilinear Scatchard plot was obtained which could be resolved into two single classes of binding sites: one immunoprecipitable by B-2 IgG and the other, nonimmunoprecipitable. The B-2 reactive sites bound IGF-I with lower affinity (Kd = 7.1 X 10"O M) than the B-2 nonreactive sites (ICd = 2.1 X 10"O M) and cross-reacted more readily with insulin, the IGF-I/insulin-binding potencies being -120 and -1100, respectively. Both receptor subtypes bound IGF-I with -30-fold higher affinity than multiplication-stimulating activity, and, after affinity cross-linking with '261-IGF-I, migrated as specific reduced bands of M, = 138,000 during sodium dodecyl sulfate-polyacrylamide gel electrophoresis.
The subunit sizes of the B-2 reactive IGF-I receptor were similar to those of the insulin receptor. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of '2sI-labeled receptors immunoprecipitated by autoantiserum B-2 or autoantiserum B-10 (which recognizes only insulin receptors) revealed, in both cases, specific reduced bands of M, = 130,000 and 90,000; the same bands were also seen after sequential precipitation with B-10 and B-2 antisera to enrich the proportion of IGF-I receptors recovered.
The presence of two distinct binding and immunoreactive species of IGF-I receptors in human placenta raises the possibility that cell-or tissue-specific isotypes of the IGF-I receptor could mediate the different biological actions of IGF-I.
Insulin-like growth factor-I is a growth hormone-dependent polypeptide which shares with insulin a high degree of amino acid sequence homology (1) and biological effects ranging from acute stimulation of metabolism to nucleic acid and protein synthesis and mitogenesis (2). Receptors for IGF-I' * This work was supported by the National Health and Medical Research Council of Australia. It was reported in part at the 2nd International Symposium on Insulin Receptors, Rome, August 31-September 3, 1983. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Each hormone cross-reacts, albeit with considerably lower affinity, with the other's receptors (9). IFG-I also binds with moderate affinity to the structurally distinct Type I1 IGF receptor which binds insulin-like growth factor I1 and multiplication-stimulating activity with high affinity and has little or no affinity for insulin (9). Because IGFs, a t low concentrations, promote growth far more readily than insulin, and because, in tissues containing no IGF-I receptors (e.g. adipocytes (9, ll)), 100-fold higher concentrations of IGFs are required to elicit insulin's metabolic actions, it has generally been considered that insulin and IGF receptors mediate, respectively, the metabolic and mitogenic functions of these hormones (2). On the other hand, in cells and tissues containing IGF-I receptors (e.g. human fibroblasts (121, BC3H-I muscle cells (13), and soleus/cardiac muscle (2, 14)) low concentrations of IGF-I and IGF-I1 can stimulate glucose and amino acid transport, suggesting that these acute metabolic responses may also be mediated by IGF receptors.
It is not known whether the metabolic and mitogenic effects of IGF-I are mediated via IGF-I binding to the same site or to separate sites on its own or related (e.g.  receptors. There has been no evidence to date suggesting the existence of IGF-I receptor subtypes or isotypes. In this paper, we present evidence that the human placenta contains two types of IGF-I receptors with different binding affinities for IGF-I and different cross-reactivities with insulin, which can be delineated on the basis of their immunoreactivity with a human autoantiserum to the insulin receptor. Porcine insulin, used to estimate nonspecific binding, human yglobulin (Cohn fraction II), and Staphylococcus aureus (strain Cowan I) supplied as a formalin-fixed heat-killed 10% (w/v) suspension, were purchased from the Commonwealth Serum Laboratories, Melbourne. Glucose oxidase (Type VI1 from Aspergillus niger), N-aCetyl-D-glUcosamine, a-cellulose, bovine albumin, leupeptin, phenylmethylsulfonyl fluoride, and bacitracin were purchased from Sigma, IODO-GEN and disuccinimidyl suberate from Pierce Chemical Co., Trasylol (aprotinin) from Bayer AG, Germany, lactoperoxidase from Calbiochem, agarose-bound wheat germ agglutinin from Vector Laboratories Inc., bovine albumin ("Pentex") from Miles Laboratories, Inc., and polyethylene glycol 6000 from the Merck Institute. All reagents for SDS-polyacrylamide gel electrophoresis were from Bio-Rad. All other materials were of reagent grade. '251-Insulin was prepared to a specific activity of 100-150 pCi/pg (17); '251-somatomedin-C/IGF-I, kindly donated by Dr. R. Baxter, Sydney, was prepared to a specific activity of approximately 100 pCi/pg and purified by hydrophobic interaction chromatography (16). For the affinity cross-linking experiments, IGF-I (preparation I/4) was labeled with ' ' ' 1 by the IODO-GEN method (18) to a specific activity of approximately 350 pCi/pg and purified by chromatography on a-cellulose (19). Sera containing anti-receptor autoantibodies were obtained from patients B-2 and B-10 with the Type B syndrome of insulin resistance and acanthosis nigricans (20,21). Human placental membranes, solubilized with Triton X-100, were prepared as previously reported (17).
Binding Assays-Solubilized placental membranes (70-140 pg of protein determined by the Bio-Rad assay, using bovine y-globulin as standard) or eluates from wheat germ agglutinin columns (10-20 pg of protein) were incubated with tracer concentrations of '251-insulin (40-70 pg) or '2KI-IGF-I (60-80 pg) for 16-24 h at 4 "C or 1-3 h at 22 "C in a total volume of 0.2 ml in 0.1 M sodium phosphate buffer, pH 7.5, in a final Triton concentration of 0.1 g/100 ml. In competition binding studies, when unlabeled insulin, MSA, or IGF-I were included, incubation mixtures also contained bovine albumin (0.25 g/ 100 ml). Receptor-bound hormone was precipitated by PEG (final concentration, 12.5 g/100 ml) in the presence of carrier human yglobulin (final concentration, 0.05 g/lOO ml) as previously described (3).
To assess specific binding of '=I-insulin or '251-IGF-I to their receptors, the nonspecific binding of radioactivity in the presence of unlabeled insulin (20 pglml) or unlabeled IGF-I (1% pure; 100 fig/ ml) was subtracted from total binding. Data from the competition binding studies were analyzed by the method of Scatchard (22).
Immunoprecipitation of Iodinated Receptors and SDS-Polyacrylamide Gel Electrophoresis-Iodinated proteins were swirled for 19 h at 4 "C with B-2 IgG (which recognizes both insulin and IGF-I receptors (3)), B-10 IgG (which recognizes only insulin receptors (3)), or control nonimmune IgG, each prebound to S. aureus at final dilutions equivalent to 1/50 of serum. To enrich the proportion of IGF-I receptors recovered, iodinated proteins, diluted 5 X with 0.1% Triton-phosphate, were first incubated with B-10 IgG prebound to S. aureus (equivalent final serum dilution, 1/50) for 16 h at 4 "C, and the resultant supernatants, depleted of insulin receptors, were then precipitated with B-2 IgG prebound to S. aureus (equivalent final serum dilution, 1/250) for 4 h at 4 "C.
To monitor the immunoprecipitation procedures, the supernatants obtained after the precipitation of nonradioactive '271-labeled receptors were assayed for residual '251-insulin or '261-IGF-I binding activities, as described under "Binding Assays." To analyze the subunit composition of the '251-labeled receptors, immunoprecipitates were washed twice with 0.1 M sodium phosphate containing 1% Triton and 0.1% SDS, once with 0.1% Triton-phosphate buffer, and then dissociated by boiling for 5 min in 2% SDS, 0.1 M dithiothreitol, 0.01% Bromphenol Blue, and 0.0625 M Tris, pH 6.8.
Affinity Labeling of Placental Membranes-Crude placental membranes (1.3 mg protein/ml) in 0.1 M sodium phosphate buffer, pH 7.5, containing bovine albumin (1 g/100 ml) were incubated at 4 "C for 18 h with '@I-IGF-I (8 X lo6 cpm/ml) in the presence or absence of unlabeled IGF-I (1% pure, 100 pg/ml). The membranes were washed twice with ice-cold phosphate buffer and then resuspended in the original incubation volume in phosphate buffer containing no albumin. Disuccinimidyl suberate, freshly dissolved in dimethyl sulfoxide, was added to a final concentration of 0.03 mM. After 15 min at 0 'C, the reaction was quenched by the addition of 5 volumes of ice-cold 20 mM Tris-HC1, 1 mM EDTA, pH 7.4, and the mixture centrifuged (20 min, 3000 X g). A portion of each pellet (0.4 mg of membrane protein) was boiled for 5 min in SDS-polyacrylamide gel electrophoresis sample buffer. The remainder (1.4 mg of membrane protein) was solubilized in 1 ml of 1% Triton X-I00 in 0.1 M sodium phosphate buffer, pH 7.5, for 1 h at 22 "C and then centrifuged at 100,000 X g for 90 min at 4 "C. Solubilized receptors were immunoprecipitated by serum B-2 (final dilution 1/200, 12 h at 4 "C) and S. aureus (0.25 ml of 10% suspension), and the pellet was boiled with SDS-polyacrylamide gel electrophoresis sample buffer. Samples were subjected to the same conditions of electrophoresis, autoradiography, and densitometry described in the preceding section of "Experimental Procedures."

Immunodepletion of Insulin and IGF-I Receptors by Autoantiserum B-2"
To assess the ability of autoantiserum B-2 to immunodeplete solubilized human placental membranes of insulin and IGF-I receptors, increasing concentrations of B-2 IgG, prebound to S. aureus, were incubated with the solubilized membranes, and the supernatants were analyzed for residual '251-insulin or lZ5I-IGF-I binding activities (Fig. 1). All PEG-precipitable and immunoprecipitable insulin binding  (Fig. L 4 ) and specific lZ51-IGF-I binding (Fig. 1B). PEG X g for 10 min), the supernatants were assayed for specific '261-insulin precipitation (O), as described under "Experimental Procedures"; "Binding Assays," was performed after a 3-h incubation of tracer and supernatant at 22 "C. Immunoprecipitation (O), as described under "Experimental Procedures"; "Immunoprecipitation of Receptors," involved a 1-h incubation of tracer and supernatant at 22 "C, followed by 2 h at 22 "C with serum B-2 at a final assay dilution of 1/200. sites were depleted equally, reaching undetectable levels at a serum dilution of 1/250 (Fig. lA). This was not the case for the IGF-I binding sites. Although the IGF-I binding sites were depleted progressively with increasing concentrations of B-2 IgG, PEG-precipitable '2SI-IGF-I binding remaining in the supernatants always exceeded immunoprecipitable "'1-IGF-I binding. At serum B-2 dilutions of 1/250 or less there was no further decrease in the PEG-precipitable '"I-IGF-I binding activity, and this residual binding activity (usually 50-80% of original binding activity) could not be further precipitated using serum B-2 (Fig. 1B). Thus serum B-2, while capable of immunoprecipitating all the insulin receptors in the solubilized membranes, could only recognize a fraction of the IGF-I binding sites. After the immunodepleted receptor preparation was equilibrated with 1261-IGF-I, the proportion of ' "1-IGF-I specifically bound and precipitated by PEG (18%) was comparable to that recovered in the void volume after gel filtration on a Sephadex G-50 column (20%, Fig, 2). IGF binding proteins in B-2 serum (2) could not account for the nonimmunoprecipitable "'I-IGF-I binding activity, since precipitation was effected by adding washed B-2 IgG-S. aureus complexes.
Binding Characteristics of Immunoprecipitable and Nonimmunoprecipitable IGF-I Receptors-When solubilized human placental membranes, preincubated with '261-IGF-I and increasing concentrations of IGF-I or insulin, were precipitated by PEG and the binding data (Fig. 3B) subjected to Scatchard analysis, a curvilinear plot was obtained (Fig. 3C). However, the "'I-IGF-I binding precipitated with serum B-2 and S. aureus (Fig. 3A) yielded a linear Scatchard plot, consistent with a single class of lower affinity binding sites ( K d = 7.1 x 10"O M, R, , = 0.13 X 10"' mol/mg protein; Fig. 3C). In contrast, the residual nonimmunoprecipitable lZ5I-IGF-I binding, assayed by PEG precipitation (Fig. 4A) and subjected to Scatchard analysis, yielded a single class of higher affinity binding sites ( K d = 2.1 X 1 0 " ' M; R, = 0.077 X 10"' mol/mg protein; Fig. 4C). Curvilinear Scatchard plots were obtained when normal nonimmune globulins prebound to S. aureus were incubated with solubilized placental membranes, and the supernatants were analyzed for PEG-precipitable lZ5I-IGF-I binding (Fig. 4, B and C).
The B-2-immunoreactive low affinity IGF binding sites cross-reacted more readily with insulin (Fig. 3A) than did the nonimmunoreactive high affinity sites (Fig. 4A). Porcine insulin was -120 times less potent than IGF-I in competing with lZ5I-IGF-I for its low affinity site and -1100 times less potent in competing for the high affinity site. In another experiment, both IGF-I binding sites reacted similarly with MSA, their relative IGF-I/MSA potencies being -30 (Table  I).
To determine the size of the IGF-I binding sites, '"I-IGF-I was covalently coupled to placental membranes with the bifunctional reagent disuccinyl suberate. When the membranes were subjected to SDS-polyacrylamide gel electrophoresis under reducing conditions, "'I-IGF-I appeared in M, = 138,000 and M, > 300,000 complexes (Fig. 5). Only the M, = 138,000 band could be detected when membranes were solubilized with Triton X-100 and immunoprecipitated with serum B-2 prior to electrophoresis (Fig. 5). The M, > 300,000 complex may represent IGF-I receptor cross-linked to its own subunits or to other proteins, because this band could also be immunoprecipitated by serum B-2 when higher concentrations (0.2 mM) of cross-linking agent were used.
Immunoprecipitation of 1251-labeled Receptors from Human Placental Membranes-Autoantiserum B-10, which recognizes only insulin receptors and autoantiserum B-2, which recognizes both insulin and IGF-I receptors (3), both precipitated specific '251-labeled proteins of reduced molecular weights 130,000 and 90,000 (Fig. 6). The immunoprecipitation procedures were monitored in parallel experiments using lZ7Ilabeled wheat germ eluates. PEG precipitation of '251-insulin and "'I-IGF-I binding sites in the supernatants after immunoprecipitation of 'T-labeled glycoproteins confirmed that most of the insulin receptors were depleted by precipitation by either sera, since specific '*'I-insulin binding after immunoprecipitation by serum B-10 and serum B-2 was 6 and 0.5%, respectively, compared to 30% for control serum. IGF-I receptors were not removed by serum B-10 but were fractionally removed by serum B-2; specific '251-IGF-I binding after immunoprecipitation by serum B-10 and serum B-2 was 22 and 17%, respectively, compared to 22% for control serum.
To enrich the proportion of IGF-I receptors immunoprecipitated, iodinated proteins were immunoprecipitated sequentially with B-10 and then B-2 antiserum (Fig. 7). '251-insulin binding activities remaining after the first and second precipitation steps were 2.8 and 0.3% respectively, compared to 18% with control serum. On the other hand, residual "'I-IGF-I binding activities were diminished only after the second precipitation step; lZ5I-IGF-I binding activities remaining after the first and second steps were 12 and 7%, respectively, compared to 11% after control serum. Analysis of the immunoprecipitate, depleted of 85% of '251-insulin binding activity, again revealed specific lZ5I-labeled proteins of molecular weights 130,000 and 90,000 (Fig. 7). DISCUSSION We have shown that human placenta contains two types of IGF-I receptors that can be delineated on the basis of their immunoreactivity with an autoantiserum to the insulin receptor (serum B-2). The IGF-I receptor recognized by serum B-2 (Kd = 7.1 X 10"O M) is more related to the insulin receptor than the nonimmunoprecipitable IGF-I receptor ( K d = 2.1 X 1o"O M), in terms of its antigenicity and cross-reactivity with insulin, and its subunits appear to have the same size as those of the insulin receptor. Neither receptor can be defined as a FIG. 3. Immunoprecipitation ver-SUB PEG precipitation of '"I-IGF-Ilabeled receptors. Solubilized placental membranes (115 pgltube) were incubated with lZ6I-IGF-I and increasing concentrations of unlabeled IGF-I or insulin for 24 h at 4 "C as described under "Experimental Procedures"; "Binding Assays." After a further incubation (20 h at 4 "C) with serum B-2 (50 &tube; final dilution, 1/1000) or buffer (0.1% Tritonphosphate; 50 pl/tube), '261-IGF-I-labeled receptors were precipitated with S. aureus (A; see "Experimental Procedures"; "Immunoprecipitation of Receptors") or PEG (E; see "Experimental Procedures"; "Binding Assays"), respectively. Under the assay conditions employed, serum B-2 did not alter the binding of '2SI-IGF-I to its receptors because PEG precipitation in the presence or absence of serum B-2 was identical (C). The binding data obtained with unlabeled IGF-I were subjected to Scatchard analysis (C).

2.
Solubilized placental membranes (115 pglO.1 ml) were swirled for 19 h at 4 "C with serum B-2 or control nonimmune serum IgG prebound to S. aureus (final serum dilution, 1/250), as described in the legend to Fig. 1). The supernatants (0.1-ml aliquots) were incubated with lZ6I-IGF-I and increasing concentrations of unlabeled IGF-I or insulin for 21 h at 4 "C. '2SI-IGF-I binding was measured by PEG precipitation, as described under "Experimentai Procedures"; "Binding Assays" (A and E ) . The binding data obtained with unlabeled IGF-I were subjected to Scatchard analysis (C). 0 \classical IGF-I1 receptor, because affinity-labeling studies with lZ5I-IGF-I and whole membranes have only identified M, = 138,000 (not M , = 260,000) cross-linked proteins on reduced SDS-polyacrylamide gels (9,ll). Furthermore, both receptors demonstrate high affinity for IGF-I, moderate affinity for MSA, and cross-reactivity with insulin (9,11).
Our finding, that antiserum B-2 can delineate two types of placental IGF-I receptors, is supported indirectly by Rechler et al. (25) who showed that B-2 IgG inhibited nearly all (90%) of the binding of '''I-IGF-I to IM-9 lymphoblasts, but, even at high concentrations (>lo0 pg/ml) only partially inhibited lZ6I-IGF-I binding to human placenta (40%). This inhibitory effect of B-2 IgG on IGF-I binding was absent from our binding assays (Fig. 3C) under the incubation conditions employed.
The binding of insulin to its receptor exhibits, universally, a curvilinear Scatchard plot. The experiments of De Meyts and colleagues (26) support the idea that the curvilinear   nunimmunoreactive IGF-I receptors: competition by IGF-I and MSA The experimental format was similar to that described in the legends to Figs Scatchard plot of insulin binding is due to negative cooperativity, although this explanation is disputed (27). The present findings clearly demonstrate that the curvilinear Scatchard plot of IGF-I binding to placenta is explained by two orders of binding affinity related to structurally distinct receptor subtypes. Curvilinear Scatchard plots have been obtained by other workers for IGF-I binding to particulate placental cell membranes (28,29) and to human placental explants in organ culture (30). However, Scatchard analyses of IGF-I binding to cultured cell lines (e.g. human fibroblasts (31), IM-9 lymphocytes (32), rat chondrocytes (33), and mouse muscle Precipitation of insulin and IGF-I receptors from iodinated human placental membranes using antisera B-10 and B-2. Human placental membranes were iodinated with Na'l'I using lactoperoxidase, solubilized in Triton X-100, and partially purified by chromatography on wheat germ agglutinin-agarose (see "Experimental Procedures"; "Iodination of Placental Membranes"). After immunoprecipitation with serum B-10, B-2, or control nonimmune serum, '251-labeled proteins were analyzed by SDS-polyacrylamide gel electrophoresis and autoradiography (see "Experimental Procedures"; "Immunoprecipitation of Iodinated Receptors and SDS-Polyacrylamide Gel Electrophoresis"). Densitometric tracings were performed on each autoradiogram. tors. Thus, the curvilinear plots obtained for human placental membranes could represent binding to two (or more) types of IGF-I receptors from different cell populations present in the placenta. Sara et al. (35) have observed two species of IGF-I receptors at different stages in the development of the human fetal brain, a lower affinity receptor present in high concentrations before 17 weeks gestation and a higher affinity receptor present after 25 weeks gestation. As well, the IGF-I receptor in chicken embryo fibroblasts exhibits markedly higher affinity for IGF-I1 than does the IGF-1 receptor in Erlich-Lettre strain E cells (11). Alternatively, since the "mi- crosomal" placental preparations consist of both endoplasmic reticulum and plasma membranes, curvilinear binding plots could also result from binding to both internal and external IGF-I receptor subtypes within the one cell. The findings of Kull et al. (lo), using monoclonal antibodies to immunoprecipitate the human IGF-I receptor, provide some support for the notion that different IGF-I receptors could exist within the same cell type. They showed that the M, = 95,000 or p subunit of the IGF-I receptor from biosynthetically labeled or lactoperoxidase-iodinated IM-9 lymphoblasts appeared as a broad band on SDS-polyacrylamide gels, and sometimes as a doublet, the faint lower component having a mobility similar to that of the corresponding subunit of the insulin receptor (10). Similar gel patterns were noted for lactoperoxidaselabeled human placental membranes (10).
A number of questions arise. Could the different IGF-I receptors be derived from each other by post-translational processing or are they synthesized by different types of placental cells? Regardless, what biological roles might they play? Although insulin and IGF receptors have clearly distinct hormonal binding sites, their physiological roles are less sharply defined, in that both receptors can mediate metabolic or mitogenic effects, depending on the cell type (25). Are the varying bioeffects mediated by one receptor due to subtypes of that receptor or to different postbinding mechanisms in different cells? Are similar "effector" regions on the insulin and IGF receptors responsible for mediating the metabolic actions of the hormones? Could the IGF-I receptor most closely related to the insulin receptor mediate the acute metabolic actions of IGF-I, and the nonimmunoreactive IGF-I receptor, the growth-promoting effects of IGF-I? Studies with other cell types may resolve this question.