Increases in free, unbound insulin-like growth factor I enhance insulin responsiveness in human hepatoma G2 cells in culture.

Insulin-like growth factor-binding protein (IGFBP)-1 binds to insulin-like growth factor (IGF)-I and -II with high affinity and has been shown to modulate IGF-I actions in vivo and in vitro. The synthesis of IGFBP-1 is suppressed by insulin, and administration of IGFBP-1 to rats results in impaired glucose metabolism. A synthetic peptide (bp1-01) has been shown to have a high affinity and specificity for human IGFBP-1 and to inhibit IGF-I binding. The current studies were undertaken to determine if, after incubation of bp1-01 with IGF-I.IGFBP-1 complexes, anabolic and insulin-like effects of IGF-I could be detected in human hepatoma (HepG2) cell cultures and to determine the receptor subtype(s) through which these effects were mediated. Incubation of HepG2 cells with bp1-01 (200 nm) increased IGF-I-stimulated protein synthesis by 44% and glycogen synthesis by 170% compared with stimulation by IGF-I alone. Incubation with bp1-01 also enhanced IGF-I-stimulated tyrosine phosphorylation of the IGF-I/insulin hybrid receptor and insulin receptor substrate 1. Exposure of the cells to bp1-01 alone enhanced glycogen synthesis and phosphorylation of IGF-I/insulin hybrid receptors. This was not a direct effect of bp1-01 because it did not bind to the receptor and did not activate tyrosine kinase activity in the presence of an anti-IGF-I receptor antibody. The addition of bp1-01 (200 nm) plus insulin to HepG2 cell culture medium resulted in increased tyrosine phosphorylation of the hybrid receptor, insulin receptor substrate 1, and the glycogen synthesis response compared with the effects of insulin alone. This enhancement of hybrid receptor phosphorylation and glycogen synthesis by bp1-01 peptide was diminished by preincubation with an inhibitory antibody for the alpha subunit of IGF-I receptor (alphaIR3). bp1-01 stimulated the hybrid receptor phosphorylation response to IGF-I, and this effect was inhibited by prior incubation of the cells with alphaIR3. In conclusion, bp1-01 competes with IGF-I for binding to IGFBP-1, which leads to release of free IGF-I from IGF-I.IGFBP-1 complexes. This released IGF-I stimulates biologic actions that are mediated predominantly through the IGF-I/insulin hybrid receptor.

Insulin-like growth factor (IGF) 1 -I is a potent mitogen that has anabolic actions and insulin-like actions in vivo and in vitro (1). IGF-I actions are determined by the capacity of free (unbound) IGF-I to interact with IGF-I receptors (2). Most of the IGF-I in the extracellular fluids is bound to specific IGF-binding proteins (IGFBPs) that can extend the half-life of IGF-I and have been shown to modulate the IGF-I actions (3,4). One of the binding proteins, IGFBP-1, is not saturated in serum, and its levels fluctuate acutely after meals (5). IGFBP-1 is produced by the liver. Plasma IGFBP-1 concentrations are suppressed by insulin (5)(6)(7) and increased by corticosteroids that stimulate its hepatic synthesis and secretion (8 -10). The circulating levels of IGFBP-1 are elevated in type I diabetes and are lowered after the administration of insulin (11). This increased IGFBP-1 is believed to contribute to the low free IGF-I concentrations that are present in poorly controlled diabetes (12). Administration of IGF-I to patients with type I or type II diabetes results in increased free IGF-I levels and improvement in insulin sensitivity (13,14). Conversely, administration of excess IGFBP-1 to normal rats results in an increase in blood glucose levels (15), and IGFBP-1 transgenic mice that overexpress rat IGFBP-1 show glucose intolerance (16,17). Thus, the concentration of IGF-I that has unrestricted access to receptors appears to be a determinant of glucose metabolism, and in diabetes the plasma insulin levels partially regulate free IGF-I by control of IGFBP-1 concentrations.
In previous studies, investigators used phage-display libraries to identify a disulfide-containing peptide (bp1-01, CRAG-PLQWLCEKYF) that bound to the IGF-I binding site on human IGFBP-1 (18). This peptide had a high affinity for human IGFBP-1, and it blocked human IGF-I binding to IGFBP-1 with an IC 50 of 180 nM. It did not bind to forms of IGFBPs that are present in other species such as rat IGFBP-1, and it did not bind to human IGFBP-3. Based on these observations, we proposed that after binding of this peptide to IGFBP-1 in extracellular fluids, IGF-I would be released from the IGF-I⅐IGFBP-1 complexes and that if the appropriate receptors were present in the target cell type, the released IGF-I would stimulate IGF-I-like biological actions including enhancement of insulin-like actions.
The current studies were undertaken using human heptoma (HepG2 cells). These cells synthesize IGFBP-1 and IGF-I and express both insulin and IGF-I receptors on their surface (19). In addition, these cells express IGF-I/insulin hybrid receptors. These receptors are composed of one ␣ and ␤ insulin receptor subunit and one ␣ and ␤ subunit from the IGF-I receptor. These two half receptors are linked by a disulfide bond to form a hybrid heterotetramer (20). This hybrid receptor has a relatively high affinity for IGF-I that is similar to the affinity of the IGF-I receptor homodimer and a much lower affinity for insulin (21). Hybrid receptors are widely distributed in human tissues including liver, spleen, skeletal muscle, and placenta (22). These hybrid receptors have been proposed to mediate IGF-I signaling in tissues where they represent the predominant receptor subtype (23). Recent studies have shown increases in hybrid receptor abundance in skeletal muscle of patients with type II diabetes (24), and gene targeting experiments that disrupt the normal function of these receptors in skeletal muscle lead to the development of insulin resistance (25). The major aims of this study were to determine whether this peptide could compete with IGF-I for binding to IGFBP-1 in HepG2 cell conditioned medium, to determine whether release of IGF-I from IGFBP-1 would result in alteration in insulin and/or IGF-I actions, and to compare the relative importance of each receptor subtype in mediating these effects.

125
I-IGF-I Binding Assay-The capacity of IGFBP-1 in HepG2 conditioned medium to bind to 125 I-IGF-I and the effect of increasing concentrations of bp1-01 on binding were determined using the polyethylene glycol precipitation method as described previously (28). Duplicate tubes containing 125 I-IGF-I (25,000 cpm/tube) were incubated with 5 l of serum-free DMEM that had been conditioned by HepG2 cells for 30 h and 0 -1 M bp1-01 peptide for 2 h at room temperature. Bound IGF-I and free IGF-I were separated by precipitation using 12.5% polyethylene glycol (M r 8,000). The pellets were washed in 6.25% polyethylene glycol, and bound 125 I-IGF-I was determined by gamma counting.
[ 35 S]Methionine Incorporation into HepG2 Cells-HepG2 cells were grown to 80% confluence (6 ϫ 10 4 cells/cm 2 ) on 24-well culture plates. The cultures were rinsed twice with serum-free DMEM-H and incubated for 30 h in 0.25 ml of low-methionine (10 Ϫ7 M) DMEM (a mixture of 10% DMEM-H and 90% methionine-free DMEM-H). The cultures then received 0 or 200 nM bp1-01, various concentrations of IGF-I (0 -150 ng/ml), and 0.05 Ci/well [ 35 S]methionine (ICN Biochemical Inc.) (specific activity, 1,206 Ci/mmol). The incubation was continued for 4 h. The plates were placed on ice, washed twice with ice-cold phosphate-buffered saline, and then exposed to 5% trichloroacetic acid for 10 min twice. The trichloroacetic acid-precipitable radioactivity was solubilized in 0.2 ml of 1% SDS, 0.1 N NaOH, and 3 ml of scintillation mixture (ScintiSafe TM Econo2; Fischer Scientific, Fair Lawn, NJ) and then counted in a liquid scintillation counter (Beckman Instruments) (29). Ci of D-[ 3 H]glucose (specific activity, 34.0 Ci/mmol) were added directly, and the incubation was continued for 2 h. The medium was replaced with 200 l of 0.2 N KOH containing 1 mg of glycogen, and cells were incubated for 40 min at 60°C. The cell lysates were transferred to 12 ϫ 75-mm tubes, and the incubation was continued for 1 h at 60°C. The tubes were placed on ice, and 1 ml of cold ethanol (Ϫ20°C) was added. Glycogen was separated by centrifugation at 5,000 ϫ g for 5 min, and the pellets were washed twice with 1 ml of cold ethanol containing 0.1% LiBr and 0.2 N KOH. The final precipitable radioactivity was solubilized in 0.2 ml of 0.2 N HCl and 2 ml of scintillation fluid and counted in a liquid scintillation counter. In some experiments, either the antibody for the IGFI-R (␣IR3; 10 nM) and/or the insulin receptor (␣IR (clone 47-9); 50 nM) was incubated with the cultures for 30 min before the addition of bp1-01, insulin, and [ 3 H]glucose.
Immunoprecipitation and Immunoblotting-HepG2 cells were grown to 80% confluence on 10-cm tissue culture dishes. The cells were rinsed three times with serum-free DMEM-H and incubated with the same medium for 30 h, and then various concentrations of bp1-01 (0 -500 nM) were added directly, and the incubation was continued for 1.5 h. IGF-I (0 or 10 ng/ml) was added, and the incubation was continued for 10 min. The cultures were washed once with ice-cold phosphate-buffered saline and solubilized in lysis buffer (1% Nonidet P-40, 0.25% sodium deoxycholate, 1 mM EGTA, 150 mM NaCl, 50 mM Hepes, pH 7.5, 100 mM sodium fluoride, 10 mM sodium pyrophosphate, 2 mM sodium vanadate, 1 mM phenylmethylsulfonyl fluoride, 1 g/ml pepstatin A, and 1 g/ml leupeptin). The insoluble material was removed by centrifugation at 15,000 ϫ g for 10 min, and the supernatant (900 g of protein) was incubated with either anti-␤-IGFI-R antibody, anti-IRS-1 antibody, or anti-␤-IR antibody (1:300 dilution) overnight at 4°C. The immune complexes were precipitated by incubation with protein A-Sepharose for 2 h at 4°C followed by centrifugation at 7,000 ϫ g for 1 min and washed with lysis buffer without phosphatase inhibitors four times. The proteins were resuspended in Laemmli sample buffer, separated by SDS-PAGE (7.5%), and transferred to a polyvinylidene difluoride membrane (0.45-m pore size). The membrane was probed with a 1:1,000 dilution of anti-phosphotyrosine antibody (PY99) and visualized with enhanced chemiluminescence (Super-Signal CL-H substrate system; Pierce) and then exposed to Kodak X-AR film (Eastman Kodak Co.). The signal intensities were quantified by scanning densitometry and analyzed using NIH Image. In the experiments where the effect of insulin was to be determined, the cells that had been exposed to serum-free medium for 30 h were incubated with bp1-01 for 1.5 h and then exposed to insulin (0 -1 nM) for 5 min. Further processing was the same as that described for IGF-I stimulation. Total protein content was measured using the bicinchoninic acid protein assay (Pierce).
Separation of IGF-I/Insulin Hybrid Receptor and IGFI-R or IR Homodimers-To identify IGF-I/insulin hybrid receptor and IGFI-R homodimers, HepG2 cells were solubilized in 1.0 ml of lysis buffer without a nonionic detergent. The insoluble material was removed by centrifugation, and supernatant was incubated with anti-␣IR monoclonal antibody (1.0 g) for 18 h at 4°C and then incubated with anti-mouse IgG-agarose for 3 h at 4°C. The immobilized anti-mouse IgG-agarose was sedimented by centrifugation. The supernatant (1.0 ml) was incubated with an anti-␤-IGFI-R polyclonal antibody (1 g) for 18 h at 4°C. Immunoprecipitation and immunoblotting were conducted as described previously. To separate IGF-I/insulin hybrid receptors and the IR homodimer, the cells were lysed as described above, and then, after the removal of the insoluble material, the supernatant was incubated with anti-IGFI-R monoclonal antibody (␣IR3) (1.0 g) for 18 h at 4°C, exposed to anti-mouse IgG-agarose for 3 h at 4°C, and centrifuged. The supernatant was incubated with anti-␤IR polyclonal antibody (1:300 dilution) for 18 h at 4°C and then immunoprecipitated and immunoblotted as described previously.
Statistical Analysis-A paired Student's t test was used to compare difference between the control and the test groups.

RESULTS
The Effect of bp1-01 Peptide on IGF-I/IGFBP-1 Binding-To determine whether bp1-01 peptide could compete with IGF-I for binding to the IGFBP-1 in HepG2 cell conditioned medium, 125 I-IGF-I binding was quantified in the presence of increasing concentrations of the bp1-01 peptide. Increasing concentrations of bp1-01 peptide inhibited IGF-I binding to IGFBP-1 in HepG2 conditioned medium, and half-maximal inhibition occurred with 140 nM, which is similar to the concentration reported by Lowman et al. (18) (e.g. 160 nM) using pure IGFBP-1 (Fig. 1).

IGFBP-1 Binding Peptide Increases IGF-I-stimulated [ 35 S]Methionine Incorporation into Protein by HepG2 Cells-
HepG2 cells were grown to 80% confluence in 24-well culture plates, and the ability of IGF-I to stimulate [ 35 S]methionine incorporation into cellular protein in the presence of the bp1-01 peptide was determined. IGF-I stimulated the incorporation of [ 35 S]methionine into HepG2 cells by 42 Ϯ 2.1% above the basal level. Coincubation with bp1-01 (200 nM) increased this response to 83 Ϯ 5.5% above the basal level (p Ͻ 0.05 compared with the effect of IGF-I alone) (Fig. 2). bp1-01 alone (200 nM) had no significant effect. Thus, the addition of bp1-01 peptide increased the IGF-I-stimulated protein synthesis presumably by inhibiting IGF-I binding to IGFBP-1, thus allowing more IGF-I to be available to interact with the IGF-I receptor.

IGFBP-1 Binding Peptide Increases IGF-I-stimulated D-[ 3 H]Glucose Incorporation into Glycogen by HepG2
Cells-IGF-I stimulated [ 3 H]glucose incorporation into HepG2 cells by 162 Ϯ 18% above control (Fig. 3A). The addition of bp1-01 (200 nM) with IGF-I enhanced the cell glycogen synthesis response to each concentration of IGF-I that was tested. The maximal response was a 317 Ϯ 18% increase (p Ͻ 0.01) (Fig. 3A). Because HepG2 cells secrete IGF-I into culture medium, the ability of increasing concentrations of bp1-01 to stimulate glycogen synthesis in the absence of exogenously added IGF-I was determined. HepG2 cells were grown to 80% confluence in 24-well culture plates, and the [ 3 H]glucose incorporation response to increasing concentrations of bp1-01 was measured. bp1-01 stimulated [ 3 H]glucose incorporation into HepG2 cells up to 71.8 Ϯ 8.0% above the basal level without the addition of IGF-I or insulin (Fig. 3B). To exclude the possibility that this was a direct effect of the bp1-01 peptide, the identical peptide concentrations were tested in the presence of the anti-IGFI-R antibody, ␣IR3 (an antibody that blocks IGF-I binding to IGFI-R). This antibody completely blocked the glycogen synthesis response to bp1-01 (p Ͻ 0.01). The antibody that blocks binding of insulin to the insulin receptor (␣IR) partially blocked the glycogen synthesis response (p Ͻ 0.05), but it was not as effective as ␣IR3. To determine that bp1-01 was not binding to the IGFI-R or IR, we quantified 125 I-bp1-01 binding and the effect of each anti-receptor antibody in competition for binding. Minimal binding (e.g. Ͻ0.1%) was detected, and it was not inhibited by either antibody (data not shown). These results are consist-ent with the conclusion that bp1-01 stimulated the dissociation of IGF-I from the IGF-I⅐IGFBP-1 complex and that the released IGF-I enhanced glycogen synthesis.
Ligand Inhibitor for IGFBP-1 Enhances Insulin Actions-Because the bp1-01 peptide enhanced the effect of IGF-I on glycogen and protein synthesis, we determined whether it could alter insulin-stimulated glycogen synthesis. When 1 nM insulin was incubated with HepG2 cells, it stimulated [ 3 H]glucose incorporation by 54 Ϯ 1.6% above the basal level (Fig. 4). The addition of bp1-01 increased the response to insulin stimulation further to 121 Ϯ 6.8%. bp1-01 alone increased glucose incorporation by 48 Ϯ 3.6%. This enhancement of the insulinstimulated effect by bp1-01 was completely inhibited by preincubation of the cultures with ␣IR3 or ␣IR. The combination ␣IR3 plus ␣IR completely inhibited the response to insulin plus bp1-01 (Fig. 4).
The IGFBP-1 Binding Peptide Increases IGF-I/Insulin Hybrid Receptor Phosphorylation and Enhances the Response of This Receptor Subtype to Insulin or IGF-I-IGF-I can bind to both IGFI-R and IR; however, the affinity of the IR for IGF-I is substantially lower, therefore we used a relatively low concentration of IGF-I (e.g. 10 ng/ml) to analyze IGF-I receptor and IRS-1 activation. Immunoprecipitation of IGFI-R, IRS-1, and IR from HepG2 cells that had been stimulated by the addition of 10 ng/ml IGF-I followed by immunoblotting for phosphotyrosine revealed that IGF-I increased tyrosine phosphorylation of IGFI-R, IRS-1, and IR. The addition of bp1-01 to the culture medium during the incubation with IGF-I enhanced tyrosine phosphorylation of IGFI-R (Fig. 5A), IRS-1 (Fig. 5B), and IR (Fig. 5C). In contrast, the addition of bp1-01 without IGF-I had no significant effect on IGFI-R phosphorylation, and it had no detectable effect on IR or IRS-1 phosphorylation.
It should be noted that in this experiment, the effect of IGF-I on IR phosphorylation cannot be clearly separated from an effect on the hybrid IGF-I/IR receptor. Because only 10 ng/ml IGF-I was added, and this concentration is not adequate to compete with insulin for binding to insulin receptors, it is probable that most of the effect of IGF-I alone noted in Fig. 5C is due to stimulation of hybrid receptor phosphorylation. Similarly, the enhancement of the response to IGF-I noted in Fig.   FIG. 1. The effect of bp1-01  5, A and B, could be due to enhanced hybrid receptor phosphorylation. To distinguish between these possibilities, HepG2 cells were exposed to IGF-I or bp1-01, and then the cell lysates were immunoprecipitated with ␣IR3. The remaining supernatant was immunoprecipitated with the ␤IR antibody. Fig. 6 shows that IGF-I and bp1-01 stimulated phosphorylation of hybrid receptors (top panel, lanes 3, 5, and 7), but they had no effect on IR phosphorylation (top panel, lanes 4, 6, and 8). Furthermore, the effects of IGF-I and bp1-01 were additive. This result supports the conclusion that the result shown in Fig. 5C is due to phosphorylation of hybrid receptors and not insulin receptors. To determine the extent to which the response shown in Fig. 5A was due to hybrid receptor as compared with IGFI-R phosphorylation, the experiment was repeated as described in Fig. 5A, except that the lysates were immunoprecipitated with ␣IR3 first, and the remaining supernatant was immunoprecipitated with anti-␤-IGFI-R. Similar to the results obtained in the experiment shown in Fig. 6, IGF-I and bp1-01 stimulated hybrid receptor phosphorylation but had a minimal effect on the IGFI-R phosphorylation (data not shown).
Because the addition of bp1-01 enhanced the cellular response to insulin as well as IGF-I, we determined its effect on insulin-stimulated IR and IRS-1 phosphorylation. Insulin (1.0 nM) stimulated tyrosine phosphorylation of IR and IRS-1, and the addition of bp1-01 stimulated an additional increase in tyrosine phosphorylation of both IR (Fig. 7A) and IRS-1 (Fig.  7B) compared with insulin alone. In contrast, bp1-01 alone had no effect. When the experiment was repeated three times, scanning densitometry showed that the intensities of the phosphorylated IR or IRS-1 bands from the cultures that received bp1-01 followed by stimulation with 1 nM insulin were increased by 65 Ϯ 5.9% and 120 Ϯ 18%, respectively, above the level that was detected with insulin stimulation alone.
Because bp1-01 enhanced the effect of insulin but had no effect when added alone, these results strongly suggest that peptide inhibitor increased the cellular responsiveness to insulin by increasing the interactions between free IGF-I and either IGFI-R or the IGF-I/insulin hybrid receptor.
To determine which receptor(s) was mediating the effects of bp1-01, phosphorylation of IGF-I/insulin hybrid receptor and the insulin receptor were analyzed after insulin or bp1-01 exposure. Stimulation of HepG2 cells with 200 nM bp1-01 peptide alone showed enhanced phosphorylation of a band corresponding to the IGF-I/insulin hybrid receptor plus the insulin receptor (Fig. 8A, a, lane 3, 1). In contrast, IGFI-R phosphorylation was not increased (Fig. 8A, a, lane 4, 2). Insulin alone also stimulated phosphorylation of this band (Fig. 8A, a, lane 5, 1). The combination of bp1-01 plus insulin stimulated an addi-

FIG. 3. Effects of bp1-01 on [ 3 H]glucose incorporation into glycogen in HepG2 cells after stimulation with IGF-I.
A, HepG2 cells were seeded at a density of 6 ϫ 10 4 cells/cm 2 in 24-well tissue culture plates, and [ 3 H]glucose incorporation into glycogen was determined as described under "Experimental Procedures" after stimulation with IGF-I in the presence (E) or absence (q) of bp1-01 (200 nM). B, HepG2 cells were grown to 80% confluence on 24-well tissue culture plates, and the effect of bp1-01 on glycogen synthesis in HepG2 cells in the absence of added IGF-I was determined (E). The results are expressed as the percentage increase over control cultures that were not exposed to bp1-01. Additional cultures were exposed to DMEM-L containing the same concentrations of bp1-01 in the presence of ␣IR (q) or ␣IR3 (OE) without IGF-I. Each value is the means Ϯ S.E. of triplicate determinations.

FIG. 4. Effects of bp1-01 on [ 3 H]glucose incorporation into HepG2 cells after stimulation with insulin.
HepG2 cells were seeded at a density of 6 ϫ 10 4 cells/cm 2 in 24-well tissue culture plates and grown for 24 h. The medium was replaced with serum-free DMEM-L, and the incubation was continued for 24 h. Then 0 or 200 nM bp1-01, 1 nM insulin, and 0.05 Ci/well [ 3 H]glucose were added directly, and the incubation was continued for 2 h. The incorporation of [ 3 H]glucose into glycogen was measured as described under "Experimental Procedures." In some experiments, ␣IR and/or ␣IR3 were added for 1 h before the addition of bp1-01, insulin, and [ 3 H]glucose. The results are expressed as the percentage of increase over control cultures that were incubated with DMEM-L without insulin or bp1-01. Each value is the mean Ϯ S.E. of triplicate determinations. tional increase (Fig. 8A, a, lane 7, 1). To distinguish between the portion of this effect that was due to stimulation of the hybrid IGF-I/insulin receptors (as compared with the insulin receptors), cells were exposed to ␣IR3 for 30 min before insulin addition. The addition of ␣IR3 completely inhibited the increase in hybrid receptor phosphorylation that was induced by the addition of bp1-01 (Fig. 8A, b, lane 3, 1). In contrast, the effect of insulin alone was not inhibited (Fig. 8A, b, lane 5, 1), and the combination of bp1-01 plus insulin had no effect over that of insulin alone (Fig. 8A, b, lane 7, 1) in cells that were pretreated with ␣IR3. Because ␣IR3 does not block ligand binding to insulin receptors, this result supports the conclusion that the effect of bp1-01 is due solely to hybrid receptor stimulation by IGF-I that is released from IGFBP-1. This result further supports the conclusion that the result noted with the bp1-01 peptide in Fig. 5C was due to hybrid receptor phosphorylation. To further confirm this finding, the experiment was repeated, and several time points were examined. When bp1-01 (400 nM) alone was incubated with HepG2 cells, only the IGF-I/insulin hybrid receptor was phosphorylated, and phosphorylation of this receptor was completely inhibited by preincubation with ␣IR3 (Fig. 8B, a and b). To confirm that hybrid receptor phosphorylation resulted in downstream signaling, phosphorylation of IRS-1 was analyzed. Insulin stimulated

FIG. 6. Effects of bp1-01 and/or IGF-I stimulation on the phosphorylation of IGF-I/insulin hybrid receptors or IGF-I receptor homodimers.
HepG2 cells were grown to 80% confluence as described under "Experimental Procedures." The media were replaced with serum-free DMEM-H, and the incubation was continued for 30 h. Then 0 or 200 nM bp1-01 was added directly, and the incubation was continued for 1.5 h. The cells were stimulated with IGF-I (10 ng/ml) for 10 min and then solubilized in lysis buffer without ionic detergent and immunoprecipitated with ␣IR3 (lanes 1, 3, 5 and 7). The supernatants were immunoprecipitated with anti-␤-IR (lanes 2, 4, 6, and 8 IRS-1 phosphorylation, and this effect was enhanced by bp1-01. The response to bp1-01 was inhibited by preincubation with ␣IR3 (Fig. 8C). These results are consistent with the conclusion that bp1-01 disassociates the endogenous IGF-I from the IGF-I⅐IGFBP-1 complex and that the released IGF-I is mediating a significant part of these effects through an interaction with IGF-I/insulin hybrid receptor. DISCUSSION Six forms of IGFBPs have high affinities for IGF-I, and the affinity of each protein for IGF-I is greater than that of the IGF-I receptor. IGFBP-1 has been shown to inhibit cellular responses to IGF-I by preventing IGF-I binding to its receptor; therefore, dissociation of IGF-I from the IGF-I⅐IGFBP-1 complexes represents a potential mechanism for increasing free IGF-I and receptor activation. Recently, Loddick et al. (31) demonstrated that displacement of IGFs from several forms of IGFBPs, using a compound that has high affinity for IGFBPs and no affinity for IGF-I receptors, increased free IGF-I in cerebrospinal fluid, and this resulted in increased IGF-I actions. However, that compound is not specific for a single form of IGFBP. Because HepG2 cells synthesize and secrete IG-FBP-1, -2 and -4, in the current study we utilized bp1-01, a molecular mimic that binds specifically to IGFBP-1 and not to other forms of IGFBPs, such as IGFBP-3 (18).
In the previous studies, bp1-01 was shown to release sufficient IGF-I from the IGF-I⅐IGFBP-1 complexes to activate IGF-I receptor phosphorylation in MCF-7 cells (18). The current study demonstrates conclusively that the IGFBP-1 binding peptide (bp1-01) binds to IGFBP-1 and competes with IGF-I for binding. The results also show that bp1-01 does not bind to the IGF-I receptor. Because ␣IR3, the anti-IGF-I receptor blocking antibody, inhibited the effects of bp1-01 on glycogen synthesis, and it inhibited IGF-I/insulin hybrid receptor activation by bp1-01, we conclude that activation of this receptor in HepG2 cells is due to release of bound IGF-I from IGFBP-1. Although bp1-01 binding to IGFBP-1 could directly modulate IGFBP-1 binding to HepG2 cell surfaces, a direct action of IGFBP-1 would not be inhibited by ␣IR3. Therefore, we conclude that the effects demonstrated with bp1-01 are due to inhibition of IGF-I binding to IGFBP-1, leading to increased association of free IGF-I with the hybrid receptors.
IGF-I has been shown to stimulate glycogen synthesis through IGFI-R and IR, and it can stimulate tyrosine phosphorylation of both receptors (19). Although insulin has been reported to stimulate glycogen synthesis through both IR and IGFI-R in HepG2 cells, it has not been shown to stimulate tyrosine phosphorylation of IGFI-R (32). Activation of either receptor in HepG2 cells has been shown to evoke quantitatively similar biologic responses (19). Therefore, we used HepG2 cells as an in vitro model to investigate the effects of increasing free IGF-I concentrations on both the insulin-like and anabolic actions of IGF-I. When added alone, bp1-01 stimulated glycogen FIG. 8. Effect of bp1-01 and/or insulin stimulation on the phosphorylation of IGF-I/insulin hybrid receptors, insulin receptor homodimers, and IRS-1. A, HepG2 cells were grown to 80% confluence as described under "Experimental Procedures." The media were replaced with serum-free DMEM-H, and the cells were incubated for 30 h. At that time, some cultures (b) received 10 nM ␣IR3, and some cultures did not (a). The incubation was continued for 1.5 h, and then some cultures were stimulated with 1 nM insulin for 5 min. The cells were solubilized in lysis buffer without ionic detergent and immunoprecipitated with anti-␣IR (1). The supernatants were immunoprecipitated with anti-␣-IGFI-R (2), and both sets of precipitates were analyzed by immunoblotting using an anti-phosphotyrosine antibody. a, scanning densitometry values for the bands shown in top panel were (from the left) 0, 0, 2,005, 0, 5,844, 94, 7,829, and 113; for the middle panel, they were 7,480, 7,952, 8,003, 8,674, 8,483, 8,128, 7,779, and 7,532; and for the bottom panel, they were 2,929, 0, 5,309, 0, 7,090, 0, 5,465, and 0. b, scanning densitometry values for the bands shown in top panel were (from the left): 0, 0, 0, 0, 9,898, 0, 10,062, and 0; for the middle panel, they were 8,552, 6,859, 9,314, 6,364, 7,470, 6,989, 6,612, and 6,087; and for the bottom panel, they were 9,618, 0, 9,063, 0, 1,035, 0, 11,848, and 0. B, HepG2 cells serum-starved for 30 h were incubated in the absence (a) or presence (b) of 10 nM ␣IR3 for 1 h, and then bp1-01 (400 nM) was added, and the incubation was continued for the indicated time periods. The cells were solubilized and immunoprecipitated with anti-␣-IR (1), and the supernatants were re-immunoprecipitated with anti-␣-IGFI-R (2). Both sets of precipitates were analyzed as described  11,220, 11,876, 10,698, 12,487, 12,190, 12,043, 11,268, and 12, synthesis, and its effect was inhibited by both ␣IR and ␣IR3. Similarly, it enhanced the glycogen synthesis response to insulin or to IGF-I. Inhibition of the bp1-01 response by ␣IR3 suggests that this portion of the response was due to blocking binding of the released IGF-I to IGF-I or hybrid IGF-I/insulin receptors. However, the inhibitory effect of ␣IR can only be explained by assuming that a fraction of the total response is mediated through enhancement of insulin action or the ability of ␣IR to block IGF-I binding to hybrid receptors.
To determine which receptor might be mediating the effect of bp1-01 on glycogen synthesis, several different types of experiments were performed. Because the affinity of IGF-I for the insulin receptor is 100-fold lower than that for the IGF-I receptor, concentrations of free IGF-I that are high enough to activate IR are not likely to be generated by bp1-01 addition. Our results are consistent with this conclusion because the addition of bp1-01 alone did not stimulate IR phosphorylation. This requirement of very high IGF-I concentrations to stimulate IR has led other investigators to propose that the insulin-like effects of IGF-I are mediated by activation of IGFI-R or IGF-I/ insulin hybrid receptors (33,34). Hybrid IGF-I/insulin receptors were activated by the addition of bp1-01 alone, and this response could be blocked by ␣IR3. In contrast, bp1-01 induced no detectable activation of IGFI-R when hybrid receptors were immunoprecipitated first, and the residual IGFI-R homodimers were analyzed (Fig. 8). Taken together with the results shown in Figs. 3 and 4, the results presented in Figs. 6 and 8 strongly suggest that activation of hybrid receptors by IGF-I is the principle means by which bp1-01 stimulates glycogen synthesis in these cells. However, the effect of the anti-insulin receptor antibody in partially blocking bp1-01-stimulated glycogen synthesis suggests either that ␣IR is blocking the binding of released IGF-I to hybrid receptors or that insulin activation of IR is contributing to the glycogen synthesis response obtained after the addition of insulin plus bp1-01. Our results show that addition of bp1-01 to conditioned medium of HepG2 cells did not increase tyrosine phosphorylation of IR, and it did not enhance the effect of insulin on hybrid receptor phosphorylation after preincubation with ␣IR3. This finding further supports the conclusion that bp1-01 is stimulating glucogen synthesis primarily though the activation of hybrid receptors by IGF-I that has been released from IGFBP-1 and that the inhibitory effect of ␣IR may partially inhibit IGF-I/hybrid receptor activation. This inhibition of hybrid receptor activation could result in blocking the activation of downstream signaling components in the insulin signaling pathway because signal transfer through the IR ␤ subunit kinase of the hybrid receptor that is activated by transphosphorylation after binding of IGF-I to the hybrid receptor ␣ subunits results in activation of downstream signaling components that are also activated by IR (19,32). Previous studies have reported that 40 -50% of IGFI-Rs form hybrid receptors in HepG2 cells (35). bp1-01 enhancement of insulin-stimulated glycogen synthesis is completely inhibited by ␣IR3, which further suggests that binding IGF-I to the ␣ subunit of the hybrid receptor is required. Furthermore, bp1-01 enhanced the increase in glycogen synthesis that was stimulated with insulin, which strongly suggests that this response is due to increased free IGF-I because the response to bp1-01 under these conditions was inhibited by ␣IR3. Therefore, taken together, the results support the conclusion that the predominant mechanism by which bp1-01 functions is inhibition of IGF-I binding to IGFBP-1, thus resulting in enhancement of the activation of hybrid receptors.
Although bp1-01 alone stimulated glycogen synthesis, when added alone, it had no effect on protein synthesis (Fig. 2). We conclude that this difference is due to the difference in sensi-tivity of these two processes to stimulation by IGF-I. When the results of Figs. 2 and 3 are compared, the lowest concentration of IGF-I caused a significant increase in glycogen synthesis (81 Ϯ 7%; p Ͻ 0.01), whereas it had no effect on protein synthesis (18 Ϯ 4%; p, nonsignificant). Because the concentration of IGF-I that is released from IGFBP-1 by the addition of bp1-01 is low (e.g. probably Ͻ10 ng/ml), differential sensitivity of these two responses to IGF-I stimulation probably explains the difference in these responses to bp1-01.
Numerous studies have shown decreased IGF-I and increased IGFBP-1 levels in patients with catabolic states such as poorly controlled diabetes (7,36), hyperthyroidism (37), and starvation (38) and in mothers with intrauterine growth-retarded fetuses (39). In these states, insulin resistance that is characterized by hyperinsulinemia is frequently observed. In patients with insulin resistance, IGF-I administration has been shown to reduce the elevated glucose and insulin levels and to restore insulin sensitivity (13,14). On the other hand, patients with other types of insulin-resistance such as Cushing disease (40), acromegaly (41), polycystic ovarian syndrome (42), and premature adrenarche (43) have low plasma IGFBP-1 concentrations, suggesting that it plays no role in the development of their insulin resistance. In most patients with poorly controlled type I or type II diabetes, IGFBP-1 concentrations are increased, presumably due to impaired insulin action in the liver. Our studies demonstrate that the dissociation of the IGF-I⅐IGFBP-1 complex could lead to increased free IGF-I that is available to bind to IGF-I receptor and IGF-I/insulin hybrid receptor, and as a result of increased IGF-I/insulin hybrid receptor interaction, there could be a subsequent increase insulin sensitivity. These results suggest that this may be an important mechanism for enhancing insulin sensitivity in cells that possess abundant IGF-I/insulin hybrid receptors.