Insulin-Like Growth Factor (IGF) Binding Protein 2 Functions Coordinately with Receptor Protein Tyrosine Phosphatase β and the IGF-I Receptor To Regulate IGF-I-Stimulated Signaling

Abstract

Insulin-like growth factor I (IGF-I) is a mitogen for vascular smooth muscle cells (VSMC) and has been implicated in the development and progression of atherosclerosis. IGF binding proteins (IGFBPs) modify IGF-I actions independently of IGF binding, but a receptor-based mechanism by which they function has not been elucidated. We investigated the role of IGFBP-2 and receptor protein tyrosine phosphatase β (RPTPβ) in regulating IGF-I signaling and cellular proliferation. IGFBP-2 bound RPTPβ, which led to its dimerization and inactivation. This enhanced PTEN tyrosine phosphorylation and inhibited PTEN activity. Utilization of substrate trapping and phosphatase-dead mutants showed that RPTPβ bound specifically to PTEN and dephosphorylated it. IGFBP-2 knockdown led to decreased PTEN tyrosine phosphorylation and decreased AKT Ser473 activation. IGFBP-2 enhanced IGF-I-stimulated VSMC migration and proliferation. Analysis of aortas obtained from IGFBP-2^−/− mice showed that RPTPβ was activated, and this was associated with inhibition of IGF-I stimulated AKT Ser473 phosphorylation and VSMC proliferation. These changes were rescued following administration of IGFBP-2. These findings present a novel mechanism for coordinate regulation of IGFBP-2 and IGF-I signaling functions that lead to stimulation of VSMC proliferation. The results have important implications for understanding how IGFBPs modulate the cellular response to IGF-I.

INTRODUCTION

IGFBP-2 is a member of a highly conserved family of six insulin-like growth factor (IGF) binding proteins (IGFBPs) that modulate IGF-I actions, and they play an important role in the regulation of several cellular processes (13). IGFBP-2 is the second most abundant IGFBP and is expressed in several tissues, including blood vessels and the skeleton (19). IGFBP-2 can prevent IGF-I binding to its receptor (13, 22), but it also modulates cellular functions independently of IGF-I binding (16). These include regulation of growth hormone (GH)-stimulated growth (15), cell proliferation and adhesion (43). Although these effects do not require IGFBP-2 binding to IGF-I, earlier studies did not exclude the possibility that IGF-I receptor activation was required for IGFBP-2 to alter these events. Previous work had shown that IGFBP-2 binds to cell surface proteoglycans (PGs) and glycosaminoglycans, but a functionally active cell surface IGFBP-2 receptor that links IGFBP-2 binding to a signaling pathway has not been identified (21, 42). Furthermore, a functionally active receptor has not been identified for any of the members of the IGFBP family.

Recently, an association between IGFBP-2 expression and tensin homolog deleted on chromosome 10 (PTEN) has been reported (9). PTEN is a lipid phosphatase that dephosphorylates phosphatidylinositol-3,4,5-triphosphate (PIP3), thereby preventing AKT activation. Global deletion of IGFBP-2 in mice resulted in increased levels of PTEN in aorta and osteoblasts. IGFBP-2^−/− osteoblasts had reduced cell growth and cell survival. The addition of IGFBP-2 to cells derived from these mice suppressed PTEN (9), and this response was unique to IGFBP-2. The activation of PTEN is regulated by protein-protein interactions, cellular localization, and posttranslational modifications (25, 38, 44). Most studies have analyzed the effect of serine/threonine phosphorylation on PTEN activity, but some have reported that tyrosine phosphorylation downregulates PTEN activity (5, 25, 38).

Receptor-type protein tyrosine phosphatase β (RPTPβ) is a proteoglycan that is localized in the cell surface (47). RPTPβ is inactivated via ligand-induced dimerization that occurs when it binds to heparin-binding domain (HBD)-containing growth factors, such as midkine and pleiotrophin (PTN) (1, 3, 26). Following PTN binding to RPTPβ, its phosphatase activity is inhibited (1, 26, 30). IGFBP-2 contains two putative hHBDs: a C-terminal HBD which shares sequence similarity with HBDs in other HBD-containing IGFBPs, and a unique HBD (referred to hereafter as uHBD) which is not present in other IGFBPs. Since IGFBP-2 contains HBDs, we determined if IGFBP-2 could bind to RPTPβ, if binding was mediated through the uHBD, and if it led to inactivation of its catalytic activity. We further investigated the mechanism by which this response was linked to enhanced IGF-I signaling and its biological actions.

MATERIALS AND METHODS

Human IGF-I was a gift from Genentech (South San Francisco, CA). Immobilon-P membranes and the β-actin antibody were purchased from Millipore Corp. (Bedford, MA). Dulbecco's modified Eagle medium (DMEM) containing 4,500 mg glucose per liter (25 mM), streptomycin, and penicillin were purchased from Gibco (Grand Island, NY). PQ401 was purchased from Tocris Bioscience (Ellisville, MO). Antibodies against PTEN, phospho-AKT, and total AKT were from Cell Signaling Technology Inc. (Beverly, MA). The antiphosphotyrosine (anti-pY99) antibody was purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). The Ki67 antibody was from Abcam (Cambridge, MA). The horseradish peroxidase (HRP)-conjugated mouse anti-rabbit, goat anti-mouse, and mouse anti-rabbit light-chain-specific antibodies were purchased from Jackson ImmunoResearch Laboratories (West Grove, PA). IGFBP-2 antiserum was prepared as previously described (6). All other reagents were purchased from Sigma Chemical Company (St. Louis, MO) unless otherwise stated.

Generation of HBD peptide.

The synthetic peptide containing the heparin-binding domain of IGFBP-2 (¹⁸⁸KHLSLEEPKKLRP²⁰⁰) (referred to as uHBD peptide) or heparin-binding domain of IGFBP-5 (²⁰¹RKGFYKRKQCKPSRGRKR²¹⁸) and the C-terminal heparin-binding domain of IGFBP-2 (²⁴³KHGRYNLKQCKMSLNGQR²⁶⁰) were synthesized by the Protein Chemistry Core Facility at the University of North Carolina. Purity and the sequences were confirmed by mass spectrometry.

Cell culture.

Vascular smooth muscle cells (VSMC) were isolated from aortic explants obtained from 3-week-old pigs and were maintained as described previously (37). Cells were maintained in DMEM high-glucose (25 mM) growth medium with 10% fetal bovine serum (FBS) (HyClone, Logan, UT), 100 μg/ml streptomycin, and 100 U/ml penicillin. CHO-K1 cells were maintained in serum-free Dulbecco's modified Eagle minimal essential medium (α-MEM) with 10% FBS but without phenol red. They were used between passages 5 and 14.

Generation of pLenti-IGFBP-2 wild type (WT) and two mutant forms.

Mouse IGFBP-2 cDNA was amplified from mouse pCMV-SPORT6 (ATCC, Manassas, VA) using a 5′ primer sequence corresponding to nucleotides 89 to 110 of mouse IGFBP-2 (5′-ATGCTGCCGAGATTGGGCGGCC-3′) and a 3′ primer sequence complementary to nucleotides 981 to 1003 (5′-GGGCCCATGCCCAAAGTGTGCAG-3′). After DNA sequencing to confirm that the correct sequence had been amplified, the PCR product was subcloned into pENTR/D-TOPO vector and subsequently transferred into the pLenti6-V5 DEST expression vector using the LR Clonase reaction and following the manufacturer's instructions (Invitrogen).

The wild-type IGFBP-2 inserted into the pENTR/D-TOPO vector was used as a template to make the substitution mutant. The two IGFBP-2 mutants incorporated substitutions of amino acids within the unique heparin-binding domain of IGFBP-2 (MT1) containing the sequence ¹⁸⁸KHLSLEEPKKLR¹⁹⁹. The IGF-I-binding domain mutant substitutions were within the sequence ¹⁰³LPLKALV¹⁰⁹ (MT2). The substitutions, highlighted in bold, were as follows: AALSLEEPAALA and AAAKAAA, respectively. The QuikChange site-directed mutagenesis kit by Stratagene (Agilent Technologies, Santa Clara, CA) was used to incorporate the base changes needed to encode these substitutions (shown in bold below). To generate the MT1 construct, the following primer was used: 5′-AAGGGTGCCGCAGCCCTCAGTCTGGAGGAGCCCCGCGGCGTTGGCCCCGCCTCCC-3′. To generate the MT2 construct, the following primer was used: 5′-GGCTCCGAGGCG GCC GCG AAGGCG GCT GCC ACAGGCGCG-3′. After selection of the correct clone based on sequence analysis, the cDNAs encoding the two mutated forms of IGFBP-2 were transferred from pENTR/D-TOPO vector into pLenti6-V5 DEST vector using the LR Clonase reaction according to the manufacturer's instructions (Invitrogen). The constructs were expressed in CHO-K1 cells by following the procedure described previously (23).

Purification of wild-type IGFBP-2 and the two mutant forms.

Conditioned medium was collected from confluent CHO-K1 cells expressing WT IGFBP2 or the MT1 and MT2 mutants that had been maintained in serum-free α-MEM for 48 h. The media were brought to 75% saturation by adding ammonium sulfate and centrifuged at 23,000 × g for 60 min after standing overnight at 4°C. The precipitate was resuspended in 40 ml of 50 mM NaH₂PO₄ and 4 mM EDTA (pH 7.2) with 1 mM 4-(2-aminoethyl)benzenesulfonyl fluoride (AEBSF) and stored at −20°C. The salt content of the resuspended pellet was increased to 1 M NaCl before the sample was loaded onto a phenyl-Sepharose column (2.5 by 12 cm) previously equilibrated with 50 mM NaH₂PO₄, 4 mM EDTA, and 1.2 M NaCl (pH 7.2). After loading, the column was washed with equilibrating buffer until absorbance returned to baseline, and the column was subsequently eluted in two steps with 50 mM NaH₂PO₄, 4 mM EDTA, and 0.1 M NaCl (pH 7.2) and 80% 50 mM NaH₂PO₄ and 4 mM EDTA (pH 7.2) with 20% acetonitrile. The fractions containing IGFBP-2, MT1, or MT2 were concentrated by lyophilization before they were applied to an IGF-I affinity column (1.5 by 10 cm) preequilibrated with 50 mM NaH₂PO₄, 4 mM EDTA, and 0.1 M NaCl (pH 7.2). After loading, the column was washed with 10 to 15 column volumes of equilibrating buffer and subsequently eluted with 0.5 M acetic acid and stored at −80°C until use. For WT and MT1 purification, 2-ml aliquots of IGF affinity-purified IGFBP-2 or MT1 fractions were further purified using a Vydac C-4 high-performance liquid chromatography (HPLC) column. The major protein peak, containing IGFBP-2 or MT1, was lyophilized and reconstituted with H₂O and stored at −80°C. For MT2 purification, the flowthrough material that did not adhere to the affinity IGF-I column was directly applied to a heparin-Sepharose column (1 by 7.5 cm) preequilibrated with 50 mM NaH₂PO₄, 2 mM EDTA, and 25 mM NaCl (pH 7.2). After sample loading, the column was washed with equilibrating buffer and subsequently eluted with 50 mM PO₄, 2 mM EDTA, and 1 M NaCl (pH 7.2). The eluted fractions were pooled, and 1- to 2-ml aliquots (adjusted to a pH less than 6.0 with trifluoroacetic acid [TFA]) were further purified using a Vydac C-4 high-performance liquid chromatograph (HPLC). The major protein peak, containing MT2, was lyophilized and reconstituted with H₂O and stored at −80°C until use.

Generation of pLenti-HA-RPTPβ/wild type, pLenti-HA-RPTPβ/D1870A, and pLenti-HA-RPTPβ/C1904S.

Full-length RPTPβ/WT, RPTPβ/D1870A, and RPTPβ/C1904S were PCR amplified using pShuttle RPTPβ/WT, pShuttle RPTPβ/D1870A, and pShuttle RPTP β/C1904S constructs, kindly provided by Gary J. Fisher and Yiru Xu at University of Michigan, and cloned into the pENTR/D-TOPO Gateway entry vector according to the manufacturer's instructions (Invitrogen). The forward and reverse primers used to generate the PCR product were as follows: forward primer, 5′-CACC ATG CTG AGC CAT GGA GCC GGG TTG-3′, and reverse primer, 5′-TTA AGC GTA ATC TGG AAC ATC GTA TGG GTA ATG CCT TGA ATA GAC TGG ATC-3′. The forward primer includes an ATG start site. The reverse primer contained the sequence encoding a hemagglutinin (HA) epitope followed by the stop codon (bold). After selection of the correct clones based on sequencing, the cDNAs encoding the wild-type and mutant proteins were transferred from the entry vector into pLenti6/V5-DEST Gateway vector using the LR Clonase reaction by following the manufacturer's instructions (Invitrogen). Similarly, VSMC expressing these cDNAs were prepared as described previously (24).

Construction of a plasmid containing an shRNA template for IGFBP-2 and RPTPβ silencing.

Based on Invitrogen website design tools, sequences containing 21 oligonucleotides (IGFBP-2, GG AGT TCT GAC ATG CGT ATT T, and RPTPβ, GC AGG ACT GGA ACC TTT ATT G) were used to construct short hairpin RNA (shRNA) template plasmids in order to knock down IGFBP-2 and RPTPβ, respectively. Two-nucleotide substitutions (underlined; for IGFBP-2, GGAGTTCTGTGATGCGTATT, and for RPTPβ, GCAGGTGTGGAACCTTTATTGT) were inserted as control (Ctrl) shRNA. BLAST analyses were used to verify that they were not complementary to any target sequence in the pig genome. The expression vector of the shRNA template for LacZ was also used as a control. The expression plasmids encoding these shRNAs were prepared as described previously (46).

Generation of virus stocks and establishment of SMC and porcine endothelial cells expressing pLenti constructs.

293FT cells (Invitrogen, Carlsbad, CA) were plated at 5 × 10⁶ cells/10-cm plate and then transfected (24). After 3 days in culture, virus stocks were prepared, purified, and transfected into VSMC and porcine endothelial cells to obtain cells expressing IGFBP-2 shRNA, RPTPβ shRNA, and corresponding control shRNA as described previously (46, 52). The effectiveness of each shRNA for inhibiting IGFBP-2 or RPTPβ expression was determined by immunoblotting using anti-IGFBP-2 and anti-RPTPβ antibodies and comparing the results to those for smooth muscle cells (SMC) and porcine endothelial cells expressing their controls, respectively.

Transient transfection with siRNA targeting mouse RPTPβ.

Small interfering RNAs (siRNAs) targeting mouse RPTP (AUA UUC CUU GGAUAG AAG GTT) and a control siRNA (scrambled siRNA) were purchased from Invitrogen. MC3T3 cells (ATCC, Manassas, VA) were transfected at a concentration of 40 nM using the PepMute Plus reagent (SignaGen Laboratories, MD) in growth medium for 16 h, followed by replacement with fresh growth medium. Experiments were initiated 72 h after transfection.

Immunoprecipitation and immunoblotting.

The immunoprecipitation and immunoblotting procedures were performed as described previously (45). The proteins were detected using enhanced chemiluminescence (Pierce Chemical Co., Rockford, IL), and the images were captured using autoradiographic film (Denville Scientific Inc., South Plainfield, NJ). For some experiments, multiple exposures were taken so that the signals that were used for quantification were in the linear range of the film. Signal intensity was quantified using scanning densitometry (40). In order to control for loading and transfer differences, the band intensities in the loading control lanes were quantified. Each value was divided by the lowest value to obtain a ratio. To calculate the scan units for the experimental determinations, the scan values were multiplied by the ratio that was determined for difference in total protein loaded. The values were then expressed as normalized units.

In vitro IGFBP-2 binding assay.

The cells were starved with serum-free medium overnight and then exposed to 0 or 50 ng/ml IGF-I for 5 min. They were suspended in 100 μl of phosphate-buffered saline (PBS) (pH 7.4) and incubated with biotinylated IGFBP-2 (wild type or the two mutant forms) (28 nM) and/or polyethylene glycol (PEG)-uHBD (140 nM) for 4 h at 4°C. The cells were washed three times with PBS and lysed with radioimmunoprecipitation assay buffer. The cell lysates were centrifuged at 14,000 × g for 10 min at 4°C. The proteins from supernatants were separated by SDS-PAGE and detected using HRP-conjugated avidin.

Chemical cross-linking.

The cells were starved with serum-free medium overnight and then exposed to 0 or 50 ng/ml IGF-I for 5 min. They were suspended in 100 μl of PBS (pH 7.4) and incubated with IGFBP-2 (wild type or the two mutant forms) (200 ng/ml) for 4 h at 4°C. Cells were washed three times with PBS and then incubated with 2 mg/ml bis[sulfosuccinimidyl]suberate (BS3; Pierce, Rockford, IL) in PBS for 1 h on ice. Cross-linking was terminated by adding 50 mM Tris for 15 min. The cells were lysed and 25 μl of lysate was loaded onto a 5% SDS-PAGE gel, blotted, and probed using the anti-RPTPβ antibody.

PTEN lipid phosphatase assay.

VSMC were lysed in 1 ml of 1× PBS, 1% Nonidet P-40, 0.5% sodium deoxycholate, and 0.1% SDS containing 10 μg/ml aprotinin, 1 μg/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride, and 1 μg/ml pepstatin. Lysates were centrifuged (14, 000 × g), and supernatant protein was determined (23). Two micrograms of PTEN antibody was added with 500 μg of total cellular protein, incubated for 4 h at 4°C, and then precipitated using protein A-Sepharose. The precipitates were washed twice in lysis buffer and twice in 10 mM Tris (pH 8) and 50 mM sodium chloride. The immunoprecipitated PTEN was added to reaction buffer (50 mM Tris [pH 8], 50 mM sodium chloride, 10 mM dithiothreitol [DTT], 10 mM magnesium chloride) with 10 μM phospholipid vesicles in 50 μl. After a 45-min incubation at 37°C, the reaction was terminated with 100 μl of BIOMOL Green reagent (48). The amount of phosphate present was determined by reading the absorbance of the samples at 630 nm after 30 min of incubation at room temperature (RT). Phosphate concentrations were estimated, in triplicate experiments, by comparison to phosphate standards diluted in the reaction buffer. PTEN activity is expressed as pmol of phosphate released in 30 min. The data shown are the means of triplicate determinations from three separate experiments.

Phospholipid vesicles.

Phospholipid vesicles were prepared using 0.1 mM diC16PIP3 (Echelon, Salt Lake City, UT; catalog number P3916) and 0.5 mM DOPS (Sigma; catalog number P-1060) in 20 mM HEPES (pH 7.4) and 1 mM EGTA in the original diC16PIP3 glass vial and sonicated for 30 to 45 min on ice.

RPTPβ tyrosine phosphatase activity assay.

Cell extracts were prepared as described above. RPTPβ antibody (1.5 μg) was added to 500 μg of total protein and incubated for 4 h at 4°C. The immunoprecipitates were washed in lysis buffer 3 times and equilibrated in buffer containing 20 mM imidazole (pH 7.2) and 0.1 mg/ml bovine serum albumin (BSA). The immunoprecipitated RPTPβ was utilized to determine phosphatase activity in vitro using a tyrosine phosphatase assay kit (Promega) (1). Tyrosine-phosphopeptide [END(pY)INASL] at 100 μM was used as a substrate and the assay carried out according to the manufacturer's instructions. Serial dilutions of a 1 mM KH₂PO₄ standard solution were used to generate a standard curve. Phosphatase activity is expressed as pmol of phosphate released per 10 min. The data shown are the means of triplicate determinations from three separate experiments.

In vitro direct dephosphorylation of PTEN by RPTPβ.

VSMC expressing RPTPβ shRNA were serum starved overnight and then exposed to 50 ng/ml IGF-I for 10 min. Phosphorylated PTEN as input was prepared for use as a substrate by immunoprecipitation (IP) using an anti-PTEN antibody. The immune complexes were released from beads with 50 μl of SDS buffer (20 mM Tris [pH 7.5], 50 mM NaCl, 4% SDS) containing 1 mM DTT. The resuspended supernatant was dialyzed against a buffer containing 20 mM imidazole (pH 7.2) and 50 mM NaCl for 2 h. The dialyzed sample was concentrated to 60 μl with an Ultrafree centrifugal filter device (molecular weight [MW] cutoff of 10,000 [10K]; Millipore). Quiescent VSMC expressing IGFBP-2 shRNA were used to prepare RPTPβ by IP with an anti-RPTPβ antibody. Normal mouse IgG was used as a control. RPTPβ-containing immune complexes were also obtained from cells incubated with IGFBP-2 for 2 h prior to IGF-I stimulation and were used as a negative control. The immunoprecipitates were washed in lysis buffer 3 times and equilibrated with 20 μl of 1× phosphatase buffer containing 20 mM imidazole (pH 7.2) and 0.1 mg/ml BSA. Twenty microliters of each IP eluate was mixed and incubated for various times at 30°C. Fifteen microliters of 4× SDS sample buffer was added, and the final volume of 55 μl was loaded onto a 9% SDS-PAGE gel and then analyzed by immunoblotting using an anti-pY99 antibody. The blot was stripped and reprobed with an anti-PTEN antibody as a loading control. That equal amounts of RPTPβ were precipitated was verified by Western blotting.

Isolation of cytoplasmic and membrane fraction proteins.

Isolation of cytoplasmic and membrane fraction proteins was performed as described previously (52).

Cell proliferation and migration assays.

Assessment of VSMC proliferation was performed as described previously (33). For cell migration, wounding was performed as described previously (20). In both assays, IGFBP-2 (WT or the two mutant forms) (200 ng/ml) was added for 16 h before IGF-I was added. Each treatment was analyzed in triplicate. The results represent mean values of four independent experiments.

In vivo treatment of IGFBP-2^−/− mice with IGFBP-2.

IGFBP-2^−/− mice or wild-type mice (both strains were C57BL/6) at 6 or 8 weeks of age were administered PBS or IGFBP-2 (100 μg/day) intraperitoneally (i.p.) twice (48 h and 24 h) before sample collection. Similarly, the mice were injected with 1 mg/kg (of body weight) of IGF-I or PBS twice, at 24 h and 15 min before sample collection.

Preparation of aortas for analysis.

The mice were euthanized by injection with pentobarbital sodium (Nembutal; 100 mg/kg) i.p. The aortas were harvested and placed into ice-cold PBS. Connective tissue and endothelium were removed prior to protein extraction or fixation. For biochemical analyses, the aortas were homogenized in ice-cold buffer consisting of 20 mM Tris, 150 mM sodium chloride (pH 7.4), 2 mM EDTA, and 0.05% Triton X-100 using a glass tissue grinder.

Immunohistochemistry.

The aortas were sectioned and then incubated with anti-Ki67 antibody using a procedure described previously (53). The number of SMC staining positive for Ki67 was determined as a percentage of the total number of VSMC in each section (Ki67-positive nuclei/total cell nuclei). There were a total of 3 or 4 mice in each treatment group. The data shown are the means ± standard errors (SEs) for Ki67-positive cells from all four sections from each of the mice, expressed as percentages.

Statistical analysis.

The results that are shown for all experiments are representative of at least three independent experiments and expressed as the means ± SEs. The Student t test was used to compare differences between treatments for in vitro experiments. One-way analysis of variance (ANOVA) was applied for all data obtained from in vivo studies. A P value of <0.05 was considered statistically significant.

RESULTS

IGFBP-2 binds to the specific cell surface receptor RPTPβ via its uHBD.

To determine if IGFBP-2 bound to the VSMC surface and whether this was altered by IGF-I, we prepared stable transfectants expressing IGFBP-2 shRNA to avoid interference from endogenous IGFBP-2. IGFBP-2 expression was reduced by 90% ± 8% (Fig. 1A). In the absence of IGF-I, there was minimal IGFBP-2 binding to the cell surface (Fig. 1B, lane 1), and addition of IGF-I resulted in a major increase (Fig. 1B, lane 4). To determine the region of IGFBP-2 that was required, we utilized two IGFBP-2 mutants. One mutant (MT1) had substitutions for charged amino acids in the uHBD region, and the second (MT2) had substitutions for hydrophobic residues in the IGF-I binding domain (17). IGF-I stimulated binding of the MT2 mutant to the cellular membrane (Fig. 1B, lanes 3 and 6), but the MT1 mutant did not bind (Fig. 1B, lanes 2 and 5). Furthermore, a peptide that contained the uHBD sequence significantly inhibited IGFBP-2 binding (Fig. 1B). In contrast, a peptide containing the C-terminal HBD in IGFBP-2 and one containing the C-terminal HBD in IGFBP-5 had no effect on IGFBP-2 binding (Fig. 1C). Since RPTPβ had been shown to bind to growth regulatory peptides through their HBDs (1, 3, 26), we determined if IGFBP-2 bound to RPTPβ and if this was mediated through the uHBD. After allowing VSMC to secrete IGFBP-2 for 16 to 18 h, immunoprecipitation of IGFBP-2 followed by immunoblotting for RPTPβ showed that IGFBP-2 associated with RPTPβ and that IGF-I induced an increase of 4.9-fold ± 0.8-fold (mean ± SE) (Fig. 1D). The uHBD peptide significantly disrupted association of IGFBP-2 with RPTPβ (Fig. 1D). Similarly, addition of biotinylated WT IGFBP-2 to cells expressing the IGFBP-2 shRNA showed that it bound to RPTPβ in response to IGF-I, but the biotinylated MT1 mutant did not bind (Fig. 1E). The MT2 mutant also bound to RPTPβ in response to IGF-I (Fig. 1E). This result indicates that IGF-I does not mediate the increase in association of IGFBP-2 with RPTPβ through binding to IGFBP-2.

Fig 1.

IGF-I induces the binding of IGFBP-2 to RPTPβ through its unique HBD. (A) VSMC were transduced with control (Ctrl) or IGFBP-2 shRNA template plasmids, and conditioned media were analyzed by immunoblotting for IGFBP-2. Cell lysates were immunoblotted with anti-β-actin antibody as a loading control. (B) VSMC expressing the IGFBP-2 shRNA were incubated with or without IGF-I. The cells were suspended and then incubated with biotinylated IGFBP-2 and the treatments listed, and then binding was determined as described in Materials and Methods. (C) Quiescent VSMC expressing the IGFBP-2 shRNA were used. A 10-μg/ml concentration of either uHBD, cHBD, or IGFBP-5HBD was incubated with the cells as described above and biotinylated IGFBP-2 binding determined. (D) Nontransfected VSMC were serum starved overnight and incubated with (+) or without (−) the uHBD peptide (10 μg/ml) for 2 h prior to IGF-I treatment. The cell lysates were immunoprecipitated with an anti-IGFBP-2 antibody, followed by immunoblotting for RPTPβ. The membrane was reprobed with an anti-IGFBP-2 antibody. (E) VSMC expressing IGFBP-2 shRNA were incubated with biotinylated IGFBP-2 (WT), the uHBD mutant IGFBP-2 (MT1), or the IGF-I-binding mutant form of IGFBP-2 (MT2) for 2 h prior to IGF-I addition for 5 min. The cell lysates were incubated with streptavidin beads for 2 h at 4°C. The precipitated proteins were separated by SDS-PAGE, and biotinylated IGFBP-2 was detected using avidin-HRP. uHBD, unique IGFBP-2HBD; cHBD, C-terminal IGFBP-2HBD; IB, immunoblotted. IGFBP-2 is abbreviated as BP-2 in all figures.

IGFBP-2 binding to RPTPβ leads to RPTPβ dimerization and inactivation.

Since RPTPβ phosphatase activity is inactivated by dimerization (3, 26), we investigated whether association of IGFBP-2 with RPTPβ led to RPTPβ dimerization. IGF-I addition to control cells stimulated RPTPβ dimerization, and knockdown of IGFBP-2 prevented this response (Fig. 2A). IGFBP-2 knockdown had no effect on RPTPβ protein levels (data not shown). The addition of intact WT IGFBP-2 to the IGFBP-2 shRNA cultures restored RPTPβ dimerization, but the uHBD mutant (MT1) had no effect (Fig. 2B). To determine if dimerization reduced RPTPβ phosphatase activity, we immunoprecipitated RPTPβ from lysates of IGFBP-2 shRNA-expressing and control cells and determined its capacity to dephosphorylate a tyrosine-phosphorylated peptide substrate. IGF-I treatment of control cells resulted in a 64.5% ± 5.2% decrease in phosphatase activity, whereas knockdown of IGFBP-2 prevented this IGF-I-induced change (8.5% ± 2.5% decrease; n = 3; P < 0.01) (Fig. 2C). The addition of intact IGFBP-2 (WT) to the knockdown cultures rescued the response to IGF-I (60.2% ± 5.5% decrease), and addition of the IGF-I binding mutant (MT2) resulted in a 65.4% ± 4.2% decrease. In contrast, the uHBD mutant (MT1) had no effect (7.1% ± 3.8% decrease; P value, not significant [NS]), whereas the IGF-binding domain mutant (MT2) was fully active (Fig. 2D). To determine the role of IGF-I binding to the IGF-I receptor in facilitating RPTPβ dimerization, we inhibited IGF-I receptor kinase activation. The results show that PQ401, an IGF-I receptor kinase inhibitor, decreased RPTPβ dimerization in a dose-dependent manner (Fig. 2E). Control culture showed that PQ401 was functioning to inhibit IGF-I receptor autophosphorylation (Fig. 2F) (39).

Fig 2.

Knockdown of IGFBP-2 suppresses RPTPβ dimerization and inactivation, which is reversed by IGFBP-2 but not by the uHBD mutant (MT1). (A) VSMC expressing IGFBP-2 or control (Ctrl) shRNA were incubated with (+) or without (−) IGF-I (50 ng/ml) for 5 min, washed, and then exposed to the non-cell-permeable cross-linker bis[sulfosuccinimidyl]suberate (BS3) (+) or left untreated (−) as described in Materials and Methods. The cell lysates were separated using 5% SDS-PAGE, transferred, and immunoblotted with an anti-RPTPβ antibody. (B) VSMC expressing IGFBP-2 shRNA with (+) or without (−) IGF-I for 5 min were incubated with WT IGFBP-2 or the MT1 mutant (200 ng/ml) for 4 h at 4°C, the cross-linker was added, and the cell lysates were analyzed as for panel A. (C) VSMC expressing IGFBP-2 shRNA or control IGFBP-2 shRNA (Ctrl) were incubated with IGF-I (50 ng/ml). Cell lysates were immunoprecipitated with an anti-RPTPβ antibody, and then the immunoprecipitates were used to determine RPTPβ enzymatic activity as described in Materials and Methods. Phosphatase (ptase) activity is expressed as pmol of phosphate released per 10 min. The values are the means ± SEs from three independent experiments. A P value of <0.01 indicates a significant difference between two different treatments or two cell types; NS, not significant. (D) VSMC expressing IGFBP-2 shRNA were incubated with IGF-I (50 ng/ml) and the WT or MT1 or MT2 mutant form of IGFBP-2. The cell lysates were immunoprecipitated with an anti-RPTPβ antibody, and RPTPβ enzymatic activity was determined. The data are the means ± SEs from three independent experiments. A P value of <0.05 indicates a significant difference between two different treatments; NS, not significant. (E) Confluent cultures were serum starved overnight and incubated with the IGF-I receptor tyrosine kinase inhibitor PQ401 or the same volume of dimethyl sulfoxide (DMSO) for 1 h prior to IGF-I stimulation for 5 min, washed, and then exposed to BS3 (+) or left untreated (−), as for panel A. The cell lysates were immunoblotted with an anti-RPTPβ antibody. (F) Confluent cultures were prepared and treated as for panel E. The cell lysates were immunoprecipitated with anti-IGF-I receptor antibody, followed by immunoblotting for pY99. The blot was stripped and reprobed with an anti-β-actin antibody as a loading control. #, monomer; ##, dimer.

To determine whether IGFBP-2 bound to the extracellular domain of RPTPβ, the amount of IGFBP-2 in medium was manipulated by changing fresh serum-free medium. IGF-I-stimulated association of IGFBP-2 with RPTPβ correlated with the amount of IGFBP-2 in the medium (Fig. 3A). Since RPTPβ contains a fibronectin motif (Fn3 motif) in its extracellular region and this domain can bind to glycosaminoglycans (7), we investigated whether preincubation with an anti-Fn3 antibody could prevent association of IGFBP-2 with RPTPβ. Our results indicated that this antibody recognized a band which has a similar MW estimate to RPTPβ, and that band intensity was reduced in cells expressing RPTPβ shRNA (Fig. 3B). Exposure to this antibody significantly impaired IGF-I-stimulated IGFBP-2 binding to RPTPβ compared to incubation with normal mouse IgG (Fig. 3C). These results show that secreted IGFBP-2 interacts with the extracellular domain of RPTPβ.

Fig 3.

The secreted IGFBP-2 interacts with the extracellular domain of RPTPβ. (A) Quiescent cells (serum starved for 18 h) were changed to fresh serum-free medium (I) or kept in the original serum-free medium (II) and incubated for an additional 2 h prior to addition of IGF-I (50 ng/ml) for 10 min. (Top) Cell lysates were immunoprecipitated (IP) with an anti-RPTPβ antibody, followed by immunoblotting for IGFBP-2. To control for loading, the blot was stripped and reprobed with anti-RPTPβ antibody. (Bottom) Conditioned medium was immunoblotted for IGFBP-2. (B) Twenty micrograms of cell lysate from cultures expressing RPTPβ shRNA was analyzed by immunoblotting with anti-Fn3 or anti-β-actin antibody. (C) Confluent cultures expressing RPTPβ shRNA were serum starved overnight and incubated with normal mouse IgG (NM) or Fn3 antibody (1:50) for 90 min. IGF-I was added for 5 min. The cell lysates were immunoprecipitated with anti-IGFBP-2 antibody, followed by immunoblotting for RPTPβ. That equal amounts of IGFBP-2 were precipitated was verified by immunoblotting.

IGF-I stimulates the tyrosine phosphorylation of PTEN, and stimulation requires the presence of IGFBP-2.

To identify a substrate of RPTPβ whose change in tyrosine phosphorylation could alter IGF-I-stimulated AKT activation (21, 30), we focused on PTEN, since enhancement of tyrosine phosphorylation of PTEN decreases its enzymatic activity (38). IGF-I stimulated a 2.01-fold ± 0.35-fold (mean ± SE) increase in PTEN tyrosine phosphorylation that was further enhanced 1.5-fold ± 0.3-fold by addition of IGFBP-2 with IGF-I (Fig. 4A). Enhanced tyrosine phosphorylation of PTEN was associated with reduced PTEN enzymatic activity (a 69.3% ± 4.5% decrease with IGF-I plus IGFBP-2 and a 35.4% ± 3.1% decrease with IGF-I; n = 3; P < 0.05) (Fig. 4B). These changes were accompanied by a significant increase in AKT Ser473 phosphorylation in response to IGF-I (81.5% ± 8.7% increase; n = 3; P < 0.05) (Fig. 4C).

Fig 4.

IGF-I-stimulated tyrosine phosphorylation of PTEN and its activation are enhanced by addition of IGFBP-2. Nontransfected VSMC cultures were serum starved overnight and then incubated with 200 ng/ml of IGFBP-2 for 2 h prior to IGF-I stimulation. The cell lysates were immunoprecipitated (IP) with an anti-PTEN antibody, followed by immunoblotting for pY99 (A) or measurement of PTEN enzymatic activity (B) as described in Materials and Methods. The blot was stripped and reprobed with an anti-PTEN antibody. The bar graph (right side of panel A) shows the means ± SEs (n = 3), expressed as normalized phospho-PTEN units, which were calculated as described in Materials and Methods. PTEN activity is expressed as pmol of phosphate released in 30 min. The data are expressed as the means ± SEs (n = 3). (C) Twenty micrograms of cell lysate was used for detection of phospho-AKT. The blot was stripped and reprobed with an anti-AKT antibody as a loading control.

To determine if IGFBP-2 was absolutely required, we used cells expressing IGFBP-2 shRNA. The addition of IGF-I had minimal increased PTEN tyrosine phosphorylation in cells expressing IGFBP-2 shRNA, whereas in control cells, there was a 6.7-fold ± 1.8-fold increase (P < 0.01) (Fig. 5A). The response of the IGFBP-2 shRNA cells was rescued by adding IGFBP-2 (4.8-fold ± 1.5-fold increase; P < 0.05). Knockdown of IGFBP-2 also inhibited the ability of IGF-I to inhibit PTEN enzymatic activity compared to control cells (a 10.6% ± 2.1% decrease compared to a 62.2% ± 4.9% decrease; n = 3; P < 0.01) (Fig. 5B), and it resulted in a reduced AKT phosphorylation in response to IGF-I (increase of 2.6-fold ± 0.5-fold compared to 10.9-fold ± 2.4-fold; n = 3; P < 0.01). This was reversed by adding IGFBP-2 (Fig. 5C).

Fig 5.

Knockdown of IGFBP-2 suppresses tyrosine phosphorylation of PTEN which is reversed by adding IGFBP-2. (A, left side) VSMC were transduced with control (Ctrl) or IGFBP-2 shRNA. Conditioned media were analyzed by immunoblotting for IGFBP-2. Cell lysates (with or with IGF-I for 10 min) were immunoprecipitated (IP) with an anti-PTEN antibody followed by immunoblotting for pY99 (A) or measurement of PTEN enzymatic activity (B). The data are expressed as the means ± SEs (n = 3). The blot was stripped and reprobed with anti-PTEN antibody. The bar graph on the right side of panel A shows the means ± SEs (n = 3) expressed as normalized phospho-PTEN units. (C) Twenty micrograms of the above-mentioned cell lysates (with or with IGF-I for 10 min) were used for detection of phospho-AKT. The blot was stripped and reprobed with an anti-AKT antibody. (D) Membrane proteins were isolated as described in Materials and Methods and immunoblotted (IB) with anti-AKT or anti-pAKT (S473). The blots were stripped and reprobed with anticaveolin as a membrane fraction marker and anti-14-3-3β as a cytoplasmic fraction marker.

PTEN is a lipid phosphatase that dephosphorylates phosphatidylinositol-3,4,5-triphosphate (PIP3), thereby preventing membrane localization of Akt (49). Therefore, we determined if inhibiting PTEN activity altered AKT membrane recruitment in response to IGF-I stimulation. SMC expressing IGFBP-2 shRNA significantly impaired AKT membrane localization compared to control shRNA cells in response to IGF-I (Fig. 5D). Similarly, knockdown of IGFBP-2 significantly prevented AKT Ser473 phosphorylation in the membrane fraction after IGF-I stimulation (Fig. 5D).

IGFBP-2 regulation of PTEN is mediated via its uHBD and does not require IGF-I binding to IGFBP-2.

Although addition of WT IGFBP-2 to cultures expressing IGFBP-2 shRNA rescued PTEN tyrosine phosphorylation (Fig. 6A), PTEN enzymatic activity (Fig. 6B) and phosphorylation of AKT Ser473 (Fig. 6C) in response to IGF-I and the uHBD mutant (MT1) had a minimal effect (Fig. 6A to D). In contrast, the IGF-binding mutant (MT2) had the same activity as WT IGFBP-2 (Fig. 6A to C). Consistent with the signaling results, the cell migration response to IGF-I in IGFBP-2-silenced cells was significantly decreased (1.24-fold ± 0.07-fold increase compared to a 1.92-fold ± 0.14-fold increase; n = 4; P < 0.05). Addition of IGFBP-2 rescued the response (1.81-fold ± 0.14-fold increase; n = 4; P < 0.05), but the uHBD mutant (MT1) did not (1.28-fold ± 0.05-fold increase; n = 4; P value, NS) (Fig. 6E). Similarly, cell proliferation in response to IGF-I in IGFBP-2-silenced cells was also significantly decreased (1.39-fold ± 0.08-fold increase versus a 2.04-fold ± 0.18-fold increase in control cells; n = 4; P < 0.05), and addition of IGFBP-2 led to a 1.82-fold ± 0.04-fold increase (n = 4; P value was NS compared to control cells). The uHBD mutant (MT1) had no effect (1.27-fold ± 0.05-fold increase; n = 4; P value, NS) (Fig. 6F).

Fig 6.

Knockdown of IGFBP-2 suppresses tyrosine phosphorylation of PTEN, which is reversed by wild-type IGFBP-2 but not by the uHBD mutant (MT1). (A, left side) VSMC were transduced with control (Ctrl) or IGFBP-2 shRNA. Conditioned media were analyzed by immunoblotting (IB) for IGFBP-2. VSMC expressing IGFBP-2 shRNA were incubated with or without wild-type IGFBP-2 (WT) or two mutant forms (MT1 and MT2) for 2 h prior to IGF-I stimulation. Cell lysates were immunoprecipitated (IP) with an anti-PTEN antibody, followed by immunoblotting for pY99 (A) or measurement of PTEN enzymatic activity (B). The data are expressed as the means ± SEs (n = 3). The blot was stripped and reprobed with an anti-PTEN antibody. The graph on the right side of panel A shows the means ± SEs (n = 3), expressed as normalized phospho-PTEN units. (C and D) Twenty micrograms of cell lysate was used for detection of phospho-AKT. The blot was stripped and reprobed with an anti-AKT antibody. Cell migration (E) and proliferation (F) were determined as described in Materials and Methods. The data shown are the means of triplicate determinations from four separate experiments ± SEs.

RPTPβ plays an essential role in mediating IGF-I-dependent PTEN inactivation and AKT activation.

To definitively determine that the SMC response to IGFBP-2 was mediated through RPTPβ, we prepared SMC expressing RPTPβ shRNA. We achieved an 86% ± 5% reduction in RPTPβ (Fig. 7A). In RPTPβ-silenced cells, IGF-I-stimulated tyrosine phosphorylation of PTEN was significantly enhanced and sustained compared to that in control cells, suggesting that RPTPβ dephosphorylated PTEN (7.6-fold ± 1.1-fold increase [30 min]; n = 3; P < 0.01) (Fig. 7B). Knockdown of RPTPβ also enhanced the ability of IGF-I to inhibit PTEN enzymatic activity compared to control cells (a 61.1% ± 5.1% decrease compared to a 13.1% ± 1.9% decrease [30 min]; n = 3; P < 0.01) (Fig. 7C). This resulted in an increase in AKT phosphorylation in response to IGF-I (e.g., 4.5-fold ± 0.7-fold increase [30 min]; n = 3; P < 0.01) (Fig. 7D). When we added intact IGFBP-2 to RPTPβ-silenced cells, tyrosine phosphorylation of PTEN as well as AKT phosphorylation was not further increased in response to IGF-I (Fig. 7E and F). In addition, we determined whether the suppression of RPTPβ could lead to the alterations of IGF-I biological actions. Our results showed that cell migration in response to IGF-I was enhanced in RPTPβ-silenced cells (a 2.42-fold ± 0.09-fold increase in RPTPβ-silenced cells and a 1.99-fold ± 0.1-fold change in control cells; n = 6; P < 0.05) (Fig. 7G). Similarly, IGF-I-stimulated cell proliferation was enhanced in RPTPβ-silenced cells compared to control cells (a 2.39-fold ± 0.1-fold change in RPTPβ-silenced cells and a 1.81-fold ± 0.14-fold increase in control cells; n = 6; P < 0.05) (Fig. 7H). To determine whether the proposed mechanism is unique for porcine SMC (pSMC) or is applicable to other cell types, we examined the key events in the porcine endothelial cells and preosteoblasts (MC3T3 cells) since they express IGFBP-2 and RPTPβ. Similar to the case with pSMC, knockdown of RPTPβ resulted in a 2.08-fold ± 0.16-fold (P < 0.01) increase in IGF-I-dependent PTEN tyrosine phosphorylation and a 1.8-fold ± 0.1-fold (P < 0.01) increase in AKT activation in both cell types (Fig. 8).

Fig 7.

Knockdown of RPTPβ enhances tyrosine phosphorylation of PTEN and AKT activation in response to IGF-I. (A) VSMC that had been transduced with control (Ctrl) or RPTPβ shRNA were analyzed by immunoblotting (IB) for RPTPβ and β-actin, respectively. (B) VSMC expressing RPTPβ or control (Ctrl) shRNA were incubated with IGF-I (50 ng/ml) for the indicated times. Cell lysates were immunoprecipitated (IP) with an anti-PTEN antibody, followed by immunoblotting for pY99. The blot was stripped and reprobed with an anti-PTEN antibody. The graph shows the means ± SEs (n = 3), expressed as normalized phospho-PTEN units (lower portion of panel B). (C) The cell lysates were immunoprecipitated with an anti-PTEN antibody, followed by measurement of PTEN enzymatic activity. The data are expressed as the means ± SEs (n = 3). (D) Twenty micrograms of lysate obtained from the cultures analyzed for panel B was used for detection of phospho-AKT. The blot was stripped and reprobed with an anti-AKT antibody. The graph shows the means ± SEs (n = 3), expressed as normalized phospho-AKT scan units (right portion of panel D). (E) VSMC expressing RPTPβ shRNA were incubated with fresh SFM containing 0 or 200 ng/ml of IGFBP-2 for 2 h prior to IGF-I stimulation. The cell lysates were immunoprecipitated with an anti-PTEN antibody, followed by immunoblotting for pY99. The blot was stripped and reprobed with an anti-PTEN antibody. (F) Twenty micrograms of cell lysate obtained from the cultures analyzed for panel E was used for detection of phospho-AKT. The blot was stripped and reprobed with anti-AKT antibody. Cell migration (G) and proliferation (H) were determined as described in Materials and Methods. The data shown are the means of triplicate determinations from four separate experiments ± SEs.

Fig 8.

Knockdown of RPTPβ enhances PTEN tyrosine phosphorylation and AKT activation in porcine endothelial cells and MC3T3 preosteoblasts. (A) For the top portion, porcine endothelial cells were transduced with control (Ctrl) or RPTPβ shRNA template plasmid. Cell lysates were analyzed by immunoblotting (IB) for RPTPβ and β-actin, respectively. For the middle portion, porcine endothelial cells expressing RPTPβ or control (Ctrl) shRNA were incubated with IGF-I (50 ng/ml) for 10 min. Cell lysates were immunoprecipitated (IP) with an anti-pY99 antibody, followed by immunoblotting for PTEN. The PTEN input was measured by immunoblotting. For the bottom portion, 20 μg of lysate obtained from the cultures analyzed for the middle portion was used for detection of phospho-AKT. The blot was stripped and reprobed for AKT. (B) For the top portion, MC3T3 cells were transducted with siRNAs targeting mouse RPTP or a control siRNA (scrambled siRNA). Cell lysates were analyzed by immunoblotting for RPTPβ and β-actin, respectively. For the middle portion, MC3T3 cells expressing RPTPβ or control siRNAs were incubated with IGF-I (50 ng/ml) for 10 min. Cell lysates were immunoprecipitated with an anti-pY99 antibody, followed by immunoblotting for PTEN. The PTEN input was measured by immunoblotting. For the bottom portion, 20 μg of lysate obtained from the cultures analyzed for the middle panel was used for detection of phospho-AKT. The blot was reprobed with an anti-AKT antibody.

To definitively determine whether RPTPβ dephosphorylated PTEN directly, several approaches were used. An RPTPβ substrate-trapping mutant (D1870A) was utilized to determine if PTEN is a physiological substrate. The results showed that PTEN was able to bind to substrate-trapping mutant RPTPβ but not the WT or the catalytically inactive mutant (C1904S) (Fig. 9A). Our in vitro dephosphorylation assays showed that RPTPβ dephosphorylated tyrosine-phosphorylated PTEN directly in a time-dependent manner (Fig. 9B) and that exogenous addition of IGFBP-2 inhibited dephosphorylation (Fig. 9C, lane 2 versus lane 3). In contrast, the catalytically inactive mutant form of RPTPβ C1904S had no effect on PTEN tyrosine phosphorylation (Fig. 9D). Consistently, overexpression of wild-type RPTPβ decreased IGF-I-stimulated PTEN tyrosine phosphorylation, whereas overexpression of the catalytically inactive mutant form of RPTPβ C1904S was unable to decrease PTEN tyrosine phosphorylation in intact cells (Fig. 9E). To exclude the possible involvement of other structurally similar PTPs, we examined the RPTPα, a receptor-type PTP, and PTP1B, a nontransmembrane PTP, since they are expressed in SMC. RPTPβ knockdown had no effect on RPTPα and PTP1B protein levels (Fig. 9F). Unlike RPTPβ, RPTPα and PTP1B had no ability to dephosphorylate PTEN (Fig. 9G). Taken together, these findings clearly demonstrate that RPTPβ inhibits IGF-I-stimulated AKT activation via PTEN dephosphorylation and that IGFBP-2 functions through suppression of this pathway.

Fig 9.

RPTPβ dephosphorylates PTEN. (A) For the left side, SMC expressing hemagglutinin (HA)-tagged WT RPTPβ or D1870A and C1904S mutants were lysed and the lysates immunoblotted (IB) for HA to detect expression of WT RPTPβ or the two mutants. The blot was probed with an anti-β-actin antibody. For the right side, association of PTEN with a substrate-trapping mutant was investigated. Cell lysates from quiescent VSMC expressing the RPTPβ wild type (WT), a substrate-trapping mutant (D1870A), or a catalytically inactive mutant (C1904S) were immunoprecipitated (IP) with an anti-PTEN antibody. RPTPβ association with PTEN was detected by immunoblotting using an anti-HA antibody. The blot was stripped and reprobed with an anti-PTEN antibody. A representative immunoblot from three independent experiments is shown. (B and C) VSMC expressing RPTPβ shRNA were serum starved overnight and then exposed to 50 ng/ml IGF-I for 10 min. Phosphorylated PTEN as input was prepared for use as a substrate by immunoprecipitation with an anti-PTEN antibody as described in Materials and Methods. Quiescent VSMC expressing IGFBP-2 shRNA were used to prepare RPTPβ by immunoprecipitation with an anti-RPTPβ antibody. Normal mouse IgG (NM) was used as a control. RPTPβ-containing immune complexes were also obtained from cells incubated with IGFBP-2 for 2 h prior to IGF-I stimulation and were used as a negative control (panel C, lane 3). Twenty microliters of each IP eluate was mixed and incubated for the indicated times (B) or for 60 min (C) at 30°C. The samples were analyzed by immunoblotting with an anti-pY99 antibody. The blot was stripped and reprobed with an anti-PTEN antibody. That equal amounts of RPTPβ were precipitated was verified by immunoblotting. (D) Phosphorylated PTEN as input was prepared as for panel B. Quiescent VSMC expressing WT RPTPβ or the C1904S mutant were used to prepare active or inactive RPTPβ by immunoprecipitation with an anti-HA antibody. The in vitro dephosphorylation assay was as described for panel B. (E) VSMC expressing WT RPTPβ or the C1904S mutant were serum starved overnight and then exposed to 50 ng/ml IGF-I for the indicated times. Cell lysates were immunoprecipitated with an anti-PTEN antibody, followed by immunoblotting for pY99. The blot was reprobed with anti-PTEN antibody. A representative immunoblot from three independent experiments is shown. (F) Effect of RPTPβ activation on the expression of RPTPα and PTP1B. VSMC were transduced with the control RPTPβ shRNA (Ctrl) or the RPTPβ shRNA template plasmid and analyzed for expression of RPTPβ, RPTPα, and PTP1B by immunoblotting with the indicated antibodies. The blot was stripped and reprobed with an anti-β-actin antibody. (G) Phosphorylated PTEN as input was prepared for use as a substrate by immunoprecipitation with an anti-PTEN antibody as described for panel B. Quiescent VSMC expressing IGFBP-2 shRNA were used to prepare active RPTPβ, RPTPα, or PTP1B by immunoprecipitation with the indicated antibodies. Phospho-PTEN was incubated with each phosphatase for 30 min as described in Materials and Methods.

IGFBP-2 expression is required for IGF-I to induce RPTPβ and PTEN inactivation as well as AKT activation in vivo.

To verify that IGFBP-2 functions through RPTPβ to regulate these responses in vivo, IGFBP-2^−/− mice were utilized (9). In the absence of IGFBP-2, IGF-I did not induce RPTPβ inactivation (5.8% ± 2.9% decrease) in the aorta compared to results with control mice (36.5% ± 4.5% decrease). This response was rescued by injection of IGFBP-2 (59.4% ± 6.1% decrease) (Fig. 10A). Similarly, the injection of IGF-I into the IGFBP-2^−/− mice induced a minor increase in PTEN tyrosine phosphorylation and AKT phosphorylation, whereas in control mice, there was a major increase in phospho-PTEN and phospho-AKT (Fig. 10B and C, lanes 2 and 4). These responses were rescued by injection of exogenous IGFBP-2 in the presence of IGF-I (9.5-fold ± 2.4-fold increase in phospho-PTEN [P < 0.01] and a 5.1-fold ± 1.2-fold increase in phospho-AKT [P < 0.01]) (Fig. 10B and C, lanes 4 and 6). These increases were greater than the increases in control mice (Fig. 10B and C, lanes 2 and 6). To examine an IGF-I-stimulated biological action, Ki67 labeling (an indicator of cell proliferation) was utilized. IGFBP-2 knockout was associated with impairment in Ki67 labeling in response to IGF-I compared to that in control mice (4.8% ± 0.7% versus 9.2% ± 1.3%; P < 0.05). IGF-I failed to stimulate an increase in Ki67 labeling in the IGFBP-2^−/− mice (4.8% ± 0.7% compared to the result with PBS, 4.2% ± 0.9%; P value, NS), but when IGFBP-2^−/− mice were injected with IGFBP-2, IGF-I increased the percentage of SMC positive for Ki67 from 8.8% ± 1.2% to 13.2% ± 1.2% (P < 0.05). Interestingly, in the mice that did not receive IGF-I, administration of IGFBP-2 increased the percentage of Ki67-positive cells from 4.2% ± 0.9% to 8.8% ± 1.2% (P < 0.05) (Fig. 10D).

Fig 10.

IGFBP-2 enhances IGF-I-stimulated PTEN tyrosine phosphorylation due to a decreased RPTPβ activation in vivo. In vivo treatment of IGFBP-2^−/− mice or wild-type mice (IGFBP-2^+/+) (Ctrl) with or without IGFBP-2 and/or IGF-I was carried out and aortic extracts were prepared as described in Materials and Methods. (A) Aortic extracts (n = 3 per group) were immunoprecipitated with an anti-RPTPβ antibody, and RPTPβ enzymatic activity was determined. (B) Aortic extracts (n = 3 per group) were immunoprecipitated (IP) with an anti-PTEN antibody, followed by immunoblotting (IB) for pY99. The blot was stripped and reprobed with an anti-PTEN antibody. The graph shows the means ± SEs, expressed as normalized phospho-PTEN units (lower portion). (C) Ten micrograms of extract was used for detection of phospho-AKT. The blot was stripped and reprobed with an anti-AKT antibody as a loading control. The graph shows the means ± SEs, expressed as normalized phospho-PTEN units (lower portion). (D) The Ki67 sections were prepared and the number of proliferating cells was determined as described in Materials and Methods. The mean values ± SEs from 3 to 4 mice per treatment group (four sections measured/mouse) are shown graphically. A P value of <0.05 indicates a significant difference between two different treatments or two animal types; NS, not significant. The data in all panels are expressed as the means of three separate experiments ± SEs.

DISCUSSION

IGF-I stimulates multiple anabolic functions, and many of its actions are modulated by IGF binding proteins that are present in all extracellular fluids (13, 19). High concentrations of IGFBPs can inhibit IGF-I-stimulated actions (16) by blocking IGF-I binding to its receptor (13, 15). In contrast, IGFBP-2 concentrations that do not exceed those of IGF-I can enhance the ability of IGF-I to stimulate VSMC proliferation (4, 22). Knockdown of IGFBP-2 in zebra fish resulted in attenuation of vascular development, but IGFBP-2 overexpression inhibited cell growth (12). IGFBP-2 also stimulates endometrial epithelial cell proliferation, and both IGF-dependent and independent mechanisms have been postulated (50). IGFBP-2 stimulates cell migration and differentiation (14, 29), and the ability of IGFBP-2 to bind extracellular matrix proteins is related to its ability to stimulate cell migration (8). However, a major limitation of these studies has been the inability to identify a cell surface transmembrane receptor that is linked to a specific signaling pathway by which IGFBP-2 or other forms of IGFBPs function to enhance IGF-I actions.

Since IGFBP-2 contains an uHBD which binds to cell surface proteoglycans (41, 42) and our recent studies had shown that this domain was required for IGFBP-2 to enhance the effect of IGF-I on osteoblast proliferation (21), we focused on a cell surface protein proteoglycan that could bind to HBDs within growth factors and which had been shown to transduce a signal. Based on these criteria we selected RPTPβ for analysis, since its binding to the mitogen pleiotrophin (26) leads to increased proliferation in endothelial cells, stromal fibroblasts, and bone marrow precursor cells (30, 36). IGFBP-2 bound to RPTPβ through its uHBD, and this led to RPTPβ dimerization and inhibition of RPTPβ phosphatase activity. That this interaction was mediated through the uHBD was shown by mutagenesis. The IGFBP-2 uHBD mutant did not bind to RPTPβ, and it did not induce dimerization. Interestingly, this effect also required the addition of IGF-I, and association of IGFBP-2 with RPTPβ and dimerization of RPTPβ were not detected unless IGF-I was present. Inhibition of IGF-I binding to IGFBP-2 had no effect, which led us to conclude that the IGF-I receptor was required. This was confirmed by demonstrating that inhibition of IGF-I receptor tyrosine kinase activity inhibited RPTPβ dimerization. Importantly, VSMC expressing this mutant form of IGFBP-2 had a marked reduction in IGF-I-stimulated cell migration and proliferation. These findings strongly suggest that the responses to IGFBP-2 and IGF-I are coordinated and that IGF-I binding to its receptor is required for IGFBP-2-induced RPTPβ dimerization.

To determine the mechanism by which IGFBP-2 binding to RPTPβ altered IGF-I actions, we focused on RPTPβ tyrosine phosphatase activity (3, 26). IGFBP-2- and IGF-I-stimulated RPTPβ dimerization was associated with inhibition of its phosphatase activity. To determine the signaling protein whose function was likely to be altered in response to RPTPβ inhibition, we focused on PTEN, since cells derived from IGFBP-2^−/− mice had increased PTEN and decreased IGF-I-stimulated AKT activation (21). Inhibition of RPTPβ was associated with increased PTEN tyrosine phosphorylation and decreased PTEN activity. That these changes in response to IGFBP-2 were mediated through RPTPβ was confirmed by showing that RPTPβ knockdown eliminated the ability of IGFBP-2 to enhance IGF-I-stimulated PTEN tyrosine phosphorylation and AKT activation. These findings support the conclusion that the effect of IGFBP-2 to enhance IGF-I-stimulated AKT activation in VSMC is mediated directly through inhibition of RPTPβ dephosphorylation of PTEN. To prove that PTEN is a physiological substrate of RPTPβ, an RPTPβ substrate-trapping mutant was constructed and expressed in VSMC. Cells expressing this mutant showed RPTPβ-PTEN association, which was undetectable in the cells expressing wild-type RPTPβ. In vitro dephosphorylation experiments showed that RPTPβ, but not RPTPα or PTPB1, dephosphorylated phospho-PTEN. Similarly, knockdown of RPTPβ or expression of the inactive RPTPβ mutant abrogated the ability of IGF-I and IGFBP-2 to regulate tyrosine phosphorylation of PTEN and AKT activation. These results clearly demonstrate that PTEN is a substrate of RPTPβ. RPTPβ has other intracellular targets, including β-catenin and β-adducens, the tyrosine kinase Fyn, anaplastic lymphoma kinase, and membrane-associated guanylate kinase PDZ domain-containing protein (MAGI) (11). Whether any of these targets are involved in mediating the effect of IGF-I on VSMC was not investigated in this study. The importance of these observations for VSMC function was documented in vivo by demonstrating that IGFBP-2^−/− mice had increased RPTPβ activity, decreased PTEN tyrosine phosphorylation, impaired AKT activation, and attenuated IGF-I-stimulated VSMC replication in the aorta. These changes were reversed by administration of IGFBP-2. These results strongly suggest that coordination of these molecular events is essential for IGF-I to stimulate an optimal proliferative response in arterial smooth muscle. Our results also showed that this mechanism is also applicable in other cell types, including porcine endothelial cells and murine preosteoblasts.

Our results clearly indicate the importance of the uHBD for mediating the association of IGFBP-2 with RPTPβ. Other investigators have shown that IGFBP-2 promotes IGF-I-stimulated neuroblastoma cell replication through the uHBD (42). These investigators attributed this change to the ability of IGFBP-2 to bind to extracellular matrix (ECM) components. Other forms of IGFBPs contain an HBD in their C terminus, but IGFBP-2 contains 2 HBDs, and the HBD within the linker region is unique to IGFBP-2 (13). Our results show that the uHBD accounts for the ability of IGFBP-2 to bind to RPTPβ and that a mutant form of IGFBP-2 in which the C-terminal HBD was altered had no reduction in RPTPβ association. However, IGFBP-2 binding to RPTPβ through the uHBD alone was not adequate to stimulate RPTPβ dimerization, which also required IGF-I. Since our IGFBP-2 mutant that did not bind IGF-I had no change in its ability to induce RPTPβ dimerization and inhibition of IGF-I receptor kinase activity impaired RPTPβ dimerization, we conclude that the IGF-I receptor stimulates an event that leads to conformational change in RPTPβ and that this is required for optimal IGFBP-2 binding and stimulation of dimerization.

PTEN is a central negative regulator of the phosphatidylinositol-3-kinase (PI3K) signal transduction cascade. It is known that the predominant lipid phosphatase substrate is PIP3, which can activate AKT signaling, leading to increased cell migration, proliferation, and motility. A previous study concluded that enhanced PI3K/Akt signaling induced by sodium orthovanadate treatment was attributable to inactivation of PTEN mediated by increased PTEN tyrosine phosphorylation (51). In addition, inhibition of PTEN tyrosine phosphorylation has been correlated with PTEN activation (32, 38). Our current study obtained similar results which showed that IGF-I induced a significant increase in PTEN tyrosine phosphorylation that correlates with a reduction in PTEN activity in SMC, porcine endothelial cells, and preosteoblasts (MC3T3 cells). Our studies did not identify the tyrosine kinase that phosphorylates PTEN, and the role of tyrosine kinases in regulating PTEN activity requires further investigation. We have identified Src kinases as important for mediating IGF-I's effects in SMC (23), and future studies should investigate the possibility that Src or a Src family kinase can directly tyrosine phosphorylate PTEN (35).

Our findings with mice support a role for IGFBP-2 in enhancing the effects of IGF-I on VSMC proliferation. IGF-I is a potent VSMC mitogen, and it stimulates VSMC growth in response to vascular injury (34). Strategies that target IGF-I have successfully inhibited the VSMC proliferation response in experimental animals (27, 54). Recent studies have emphasized the role of PTEN in regulating VSMC proliferation (18, 31). Our findings suggest that inhibition of interaction between IGFBP-2 and RPTPβ could provide an important mechanism for enhancing PTEN expression in VSMC, thus limiting VSMC proliferation and atherosclerotic lesion progression.

IGFBP-2 contains multiple domains that may function coordinately to enhance IGF-I actions. In studies published recently, we demonstrated that IGF-I binding to the IGF binding domain within IGFBP-2 is required for osteoclastogenesis even though it is not required for osteoblast or VSMC proliferation (10). Our results show that a peptide containing the sequence of the C-terminal HBD domain had no effect on IGFBP-2 binding to RPTPβ, leading us to conclude that this domain does not influence interaction of IGFBP-2 with RPTPβ. This HBD is present in four members of the IGFBP family, and it can bind to cell surface proteoglycans. Therefore, it is possible that this HBD could bind to a distinct receptor and thereby modulate the responses of other cell types to IGF-I. Based on these observations, future studies should focus on cell-type-specific differences in the response to IGFBP-2 and on how these multidomain interactions are coordinated to modulate IGF-I actions.

Our findings may have major importance for understanding the role of other forms of IGFBPs in modulating IGF-I actions. All of the IGFBPs are multidomain proteins, and several contain domains other than the IGF-binding domain that have been shown to have specific functions. A recent report showed that IGFBP-2 nuclear localization regulated vascular endothelial growth factor (VEGF) synthesis (2). Our studies have shown that IGF-I can stimulate VEGF synthesis (28), and it is possible that IGF-I modulation of IGFBP-2 activity participates in this process. It is clear that the multiplicity of members of this family, each of which contains multiple domains, affords the opportunity to determine if other members of the IGFBP family mediate alterations in IGF-I actions that are distinct from those that are mediated by IGFBP-2. The results of this study suggest that the identification of specific receptors for other forms of IGFBPs would greatly enhance the ability of investigators to determine the mechanism by which these proteins function to alter tissue responses to IGF-I.