This Article Abstract Full Text (PDF) All Versions of this Article: 147/7/3484    most recent Author Manuscript (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Burrows, C. Articles by Perks, C. M. Search for Related Content PubMed PubMed Citation Articles by Burrows, C. Articles by Perks, C. M.

Insulin-Like Growth Factor Binding Protein 3 Has Opposing Actions on Malignant and Nonmalignant Breast Epithelial Cells that Are Each Reversible and Dependent upon Cholesterol-Stabilized Integrin Receptor Complexes

Department of Clinical Sciences at North Bristol, IGFs and Metabolic Endocrinology Group, University of Bristol, The Medical School, Southmead Hospital, Westbury-on-Trym, Bristol BS10 5NB, United Kingdom

Address all correspondence and requests for reprints to: Claire Perks, Department of Clinical Sciences at North Bristol, IGFs and Metabolic Endocrinology Group, University of Bristol, The Medical School, Southmead Hospital, Westbury-on-Trym, Bristol BS10 5NB, United Kingdom. E-mail: Claire.M.Perks{at}bristol.ac.uk.

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
IGF-binding protein (IGFBP)-3 is generally considered to have actions that counterbalance those of IGFs and is therefore being developed as a cancer treatment. In breast tumors, however, high levels are associated with aggressive tumors and poor prognosis. Consistent with this we have demonstrated that although IGFBP-3 and a non-IGF-binding fragment (serine phosphorylation domain peptide) reduced attachment and enhanced apoptosis of Hs578T breast cancer cells cultured on collagen or laminin, it promoted their attachment and survival on fibronectin, which is abundant in the matrix of aggressive tumors. We have now examined the factors that determine whether IGFBP-3 has positive or negative actions on breast epithelial cells. IGFBP-3 also promoted survival of Hs578T cells in the presence of an antibody to the ß1-integrin subunit or when cholesterol-stabilized complexes were disrupted. These actions were blocked by IGF-I or a MAPK inhibitor. Serine phosphorylation domain peptide had similar actions on MCF-7 cells that were again reversed on fibronectin or with disruption of cholesterol-stabilized complexes and blocked by the ß1-integrin antibody. In contrast, IGFBP-3 promoted growth and survival for nonmalignant MCF-10A cells, but these effects were again reversed on fibronectin and blocked by the ß1 antibody or a MAPK inhibitor or by disruption of cholesterol-stabilized complexes. On Hs578T cells, IGFBP-3 bound to caveolin-1 and ß1-integrins, enhancing their aggregation, the recruitment of focal adhesion kinase, and the activation of MAPK. In summary, with three breast epithelial cell lines, IGFBP-3 had positive or negative effects on growth and survival dependent upon the status of cholesterol-stabilized integrin receptor complexes.

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
THE IGFs (IGF-I AND -II) ARE POTENT cell survival factors and regulators of migration and proliferation. They play an important role in normal mammary gland development and have been heavily implicated as contributing to neoplastic growth in breast and several other cancers (1). The IGFs are present throughout the body almost entirely in association with six specific, high-affinity IGF-binding proteins (IGFBP-1 to -6), which are critical determinants of IGF availability and actions. In addition to acting as the main circulating carrier protein, IGFBP-3 is produced in many tissues where it has multiple effects on cell functions both via its ability to modulate IGF actions and also because of direct intrinsic actions (2). The production of IGFBP-3 by epithelial cells has been reported to be increased by many agents that are associated with growth inhibition and increased apoptosis including p53, tumor necrosis factor-, TGF-ß, retinoids, vitamin D, antiestrogens, silibinin, butyrate, DNA damage, and hypoxia (3). Furthermore, there have been many reports that IGFBP-3 can accentuate apoptosis of prostate (4), breast (5), colorectal (6), and esophageal (7) epithelial cells. These cumulative reports have promoted a general impression that IGFBP-3 has actions that would counterbalance those of IGFs with negative effects on cell growth and survival (8). With the expectation that IGFBP-3 will have such negative effects, the stimulation of IGFBP-3 expression by epidermal growth factor (EGF) has been described as aberrant (9), and circumstances where IGFBP-3 does not inhibit have been described as acquiring resistance (10). The accumulation of reports of negative actions has culminated in proposals for IGFBP-3 to be developed as an anticancer therapeutic (11). There have, however, also been many other reports that IGFBP-3 can positively stimulate the proliferation and survival of various cells (12, 13, 14). In addition, increased in vivo expression of IGFBP-3 has been reported in colonic and esophageal tumors (9, 12), and there have been several reports that in breast tumors, the expression of IGFBP-3 is positively associated with large, highly proliferative tumors and poor prognostic markers (15, 16). Furthermore, we have previously reported that in contrast to its inhibitory effects on breast cancer cells, IGFBP-3 promoted the proliferation and survival of the relatively normal, nonmalignant, anchorage-dependent MCF10A cells (17). We went on to show that although IGFBP-3 could reduce cell attachment and enhance apoptosis of Hs578T breast cancer cells when these were cultured on plastic, collagen, or laminin, when the same cells were cultured on fibronectin, then IGFBP-3 had the opposite effects and increased cell attachment and acted as a cell survival factor (18). These reports challenge the widely held view that IGFBP-3 normally has inhibitory actions and suggest that its actions may depend not just on cell type but also on cell context. Before interventions using IGFBP-3 to treat human cancers, it will be important to understand better the situations where it might have beneficiary effects and when it could actually promote tumor progression. In this study, we have examined further the factors that determine whether IGFBP-3 has positive or negative actions on breast epithelial cells.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
Materials
Recombinant human nonglycosylated IGFBP-3 (ngIGFBP-3) was a kind gift from Dr. C. Maack, Celtrix (Santa Clara, CA). ngIGFBP-5 and human IGF-I were bought from GroPep (Adelaide, Australia). Serine phosphorylation domain peptide (SPD) is a 15-amino-acid peptide sequence spanning the two mid-region serines of IGFBP-3 that we have previously demonstrated to mimic the intrinsic actions of IGFBP-3 but has no interaction with IGF-I (19). It was synthesized at the microchemical facility of the Babraham Institute (Cambridge, UK). The MAPK inhibitor PD98059 and C2-ceramide were purchased from Calbiochem (Nottingham, UK). Protein A-Sepharose beads were bought from Oncogene Research Products (Cambridge, MA). The anti-caveolin-1 and anti-focal adhesion kinase (anti-FAK) antibodies were obtained from Transduction Laboratories (Lexington, KY). The ß1 blocking antibody and the ß1 Western blotting antibody were purchased from Chemicon (Hampshire, UK), and the control mouse IgG antibody was purchased from Dako (Cambridgeshire, UK). The anti-ERK-2 antibody was bought from Santa Cruz Biotechnology (Santa Cruz, CA). The anti-phospho-MAPK antibody was purchased from Promega (Southampton, UK). All other chemicals including antimycin A, an RGD-containing fibronectin fragment (-Gly-Arg-Gly-Asp-Thr-Pro-), extracellular matrix (ECM) gel, and agents used to deplete cholesterol, filipin, nystatin, and methyl-ß-cyclodextrin (MßCD) were bought from Sigma (Dorset, UK). Tissue culture plastics were obtained from Greiner Labortechnik Ltd. (Tyne and Wear, UK). The enhanced chemiluminescence reagents were bought from Amersham International (Little Chalfont, UK). The BCA protein assay reagent kit was purchased from Pierce (Rockford, IL).

Cell cultures
Human breast cancer cells Hs578T and MCF-7 were purchased from ECACC (Porton Down, Wiltshire, UK) and grown in a humidified 5% carbon dioxide atmosphere at 37 C. The cells were cultured in growth medium (GM) containing DMEM (BioWhittaker, Verviers, Belgium) supplemented with 10% fetal bovine serum (Life Technologies, Inc., Paisley, UK), 50 IU/ml penicillin (Britannia Pharmaceuticals, Redhill, UK), 50 µg/ml streptomycin (Celltech Pharmaceuticals, Slough, UK), and 2 mM L-glutamine (Sigma). The relatively normal MCF10A cell line was purchased from the American Type Culture Collection (Manassas, VA). This is a spontaneously immortalized breast epithelial cell line that maintains a relatively normal phenotype as determined by 1) lack of tumorigenicity in nude mice, 2) three-dimensional growth in collagen, 3) growth controlled by hormones and growth factors, 4) lack of anchorage-independent growth, and 5) formation of domes in confluent cultures (20). The MCF10A cells were maintained in a 1:1 mixture of Ham’s F12 medium and DMEM with 2.5 mM L-glutamine (Life Technologies). This was supplemented with 5% horse serum (Life Technologies), penicillin, and streptomycin (as above), 20 ng/ml EGF (Calbiochem), 100 ng/ml cholera toxin (Sigma), 10 µg/ml insulin (Novo Nordisk, West Sussex, UK), and 0.5 µg/ml hydrocortisone (Sigma). Experiments were performed in serum-free medium (SFM) containing phenol red- and serum-free DMEM and Hams Nutrient Mix F12 supplemented with penicillin and streptomycin (as above), 0.12% sodium bicarbonate (Sigma), 0.2 mg/ml BSA (Sigma), and 0.01 mg/ml transferrin (Sigma) (SFM).

Immunoprecipitation and Western immunoblotting
Cells (0.3 x 10⁶ to 0.5 x 10⁶) were grown to 80% confluency in T25 flasks and then washed twice with PBS. The GM was replaced with SFM for 24 h. Cells were treated with either ngIGFBP-3 (0–200 ng/ml) or 500,000 cpm [¹²⁵I]IGFBP-3 prepared using the chloramine-T method of iodination (specific activity of 3.7 MBq/µg IGFBP-3), incubated at 37 C for 30 min, and then lysed on ice for 10 min (1 ml containing 10 mM Tris-HCl, 5 mM EDTA, 50 mM NaCl, 30 mM sodium pyrophosphate, 50 mM sodium fluoride, 100 µM sodium orthovanadate, 1% Triton, and 1 mM phenylmethylsulfonyl fluoride, pH 7.6). Lysates were then centrifuged at 14,000 x g for 15 min at 4 C. The protein content of each sample was determined using a BCA protein assay reagent kit, and equivalent amounts of protein used were then immunoprecipitated or immunoblotted.

Samples were immunoprecipitated by incubation at 4 C with antibodies to caveolin-1 (2 µg) or ß1-integrin (4 µg) for 2 h and then 25 µl protein A-Sepharose beads for 1 h. The samples were washed three times with lysis buffer (500 µl) centrifuged at 2500 x g for 3 min and the supernatant removed. Laemmli sample buffer was added to the beads, which were then boiled for 5 min before centrifugation at 2500 x g for 3 min. Proteins were separated by SDS-PAGE and then transferred onto a nitrocellulose membrane followed by immunoblotting. Nonspecific binding sites on the nitrocellulose membranes were blocked overnight with 5% milk in Tris-buffered saline (TBS)/2% Tween for probing with anti-FAK, anti-ß1, anti-caveolin-1, or ERK-2 (all at 1:1000) or blocked with 3% BSA for probing with phospho-MAPK (1:5000). After the removal of excess unbound antibody, an antimouse antibody (1:2000 for FAK, ß1, or caveolin-1), antigoat (1:5000 for ERK-2), or antirabbit (1:10,000 for MAPK) conjugated to peroxidase was added for 1 h. Binding of the peroxidase was visualized by enhanced chemiluminescence according to the manufacturer’s instructions. In experiments with [¹²⁵I]IGFBP-3, the samples were run on the same gel. The sections of the gel corresponding to the ß1 or caveolin-1 were treated as above, and the section corresponding to [¹²⁵I]IGFBP-3 was directly exposed to x-ray film. OD measurements were determined using a scanning densitometer (Bio-Rad, Hemel Hempstead, UK) and analyzed using Totalab (Nonlinear Dynamics Ltd., Newcastle upon Tyne, UK).

Preparation of membrane fraction
Cells were grown to 80% confluency in T25 flasks. The GM was replaced with SFM for 24 h before dosing with IGFBP-3 (100 ng/ml) for 30 min at 37 C. The SFM was removed, and the cells were washed with ice-cold PBS before the addition of hypotonic lysis buffer (10 mM Tris-HCl, 10 mM EDTA, and 1 mM phenylmethylsulfonyl fluoride, pH 7.4) for 10 min. Cells were scraped from the flask, needled (0.38 mm), and then centrifuged at 1000 x g for 30 min. The supernatant was centrifuged at 55,000 x g for 30 min; the pellet was resuspended and designated the membrane fraction. Proteins were assessed by Western immunoblotting as described in the previous section.

Determination of the distribution of proteins between the Triton-soluble and -insoluble fractions
Cells were incubated in SFM for 24 h before the addition of filipin (0–5 µg/ml) for 24 h and then lysed at 4 C in ice-cold 0.2% Triton X-100 lysis buffer [50 mM HEPES (pH 7.5), 150 mM NaCl, 1 mM sodium orthovanadate, 1 mM sodium fluoride, 1 mM phenylmethylsulfonyl fluoride, 1 µg/ml pepstatin A, 1 µg/ml leupeptin, and 1 µg/ml aprotinin]. The lysates were clarified by centrifugation at 14,000 x g for 10 min at 4 C. The resultant supernatant was termed the Triton-soluble fraction. The remaining pellet was resuspended in buffer [50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1% Nonidet P-40, 0.25% sodium deoxycholate, 1 mM EGTA, 1 mM sodium orthovanadate, 1 mM sodium fluoride, 1 mM phenylmethylsulfonyl fluoride, 1 µg/ml pepstatin A, 1 µg/ml leupeptin, and 1 µg/ml aprotinin]. The resuspended pellet was recentrifuged at 14,000 x g for an additional 10 min at 4 C, and the resulting supernatant was considered to be the Triton-insoluble fraction of the cell, excluding the insoluble cytoskeletal fraction. A 10-µl aliquot of Laemmli buffer was added to 50 µl of each supernatant and heated at 65 C for 10 min. The proteins were separated by SDS-PAGE (8% gel) and then visualized by Western immunoblotting as described above.

Immunocytochemistry and confocal microscopy
Cells (0.1 x 10⁶) were grown to 80% confluency in slide chambers. The GM was replaced with SFM for 24 h before dosing with IGFBP-3 (100 ng/ml) for 30 min at 37 C. The SFM was removed, and the cells were then washed twice with PBS, air dried, and fixed in acetone for 10 min. Cells were then washed in TBS for 5 min followed by incubation with 1% BSA in TBS. This blocking solution was removed, and an antibody to FAK (1:100) was applied for 1 h at room temperature. Cells were then washed for 5 min (three times) followed by incubation with biotinylated rabbit antimouse (1:300 in TBS) for 30 min at room temperature. The cells were then washed with TBS for 5 min (three times) and incubated with streptavidin fluorescein isothiocyanate (1:200 in TBS) for 30 min at room temperature in the dark. Cells were washed again in TBS for 5 min (three times), and a coverslip was applied with Vectashield mount (Vector Laboratories, Burlingame, CA). The localization of FAK was then assessed using confocal microscopy (Leica Microsystems, Wetzlar, Germany).

Trypan blue dye exclusion
Floating cells were collected and mixed with adherent cells after trypsinization, and the resulting cell suspension was loaded onto a hemocytometer (1:1) with the dye trypan blue, which is excluded by viable cells. Both viable and dead cells were counted, from which the percentage of dead cells or the percentage of cells attached relative to control was calculated.

Internalization assay
We assessed IGFBP-3 uptake by the cell using radiolabeled [¹²⁵I]IGFBP-3. Cells were seeded at 0.12 x 10⁶ cells per well in six-well plates in GM before being switched to SFM for 24 h before dosing with nystatin (50 µg/ml) for 3 h followed by the addition of 500,000 cpm [¹²⁵I]IGFBP-3 (specific activity of 3.7 MBq/µg IGFBP-3) for 30 min at either 4 C or room temperature. After treatment, the medium was aspirated and the cells incubated with 200 µl of 0.1 M acetic acid at 4 C for 30 min to remove cell surface-associated [¹²⁵I]IGFBP-3. The cells were then lysed by incubation with 200 µl lysis buffer for 10 min at 4 C, and 180 µl was removed for quantification of internalized [¹²⁵I]IGFBP-3. Samples were analyzed using a Berthold LB2111 -counter and data recorded as counts per minute (cpm). Any internalization at 4 C, following the same protocol, was subtracted from that at room temperature as an assessment of active internalization.

Dosing protocol for apoptosis assay
Hs578T and MCF-7 cells were seeded at 0.1 x 10⁶ cells per well and MCF-10A cells at 0.2 x 10⁶ cells per well in six-well plates in GM for 24 h and then switched to SFM for 24 h before predosing with IGFBP-3 (100 ng/ml) or SPD (5 ng/ml) for 24 h with or without ß1 blocking antibody or control mouse IgG (200 ng/ml) or an inhibitor of MAPK PD98059 (1 µM), nystatin (50 µg/ml), filipin (1 µg/ml), or fibronectin (0.25 µg/ml) for 24 h or MßCD (4 mM) for 30 min. The cells were then redosed with IGFBP-3 (100 ng/ml) or SPD (5 ng/ml) for 24 h with the same additions in the presence or absence of an apoptotic dose of either antimycin A or C2-ceramide (doses chosen to give approximately 40–60% cell death). We have shown previously that the amount of apoptosis as quantified by flow cytometry is directly comparable to the amount of cell death measured by Trypan blue cell counts in these models of cell death (5, 21).

Adhesion assay protocol
Cell adhesion assays were undertaken as we described previously (18). Briefly, Hs578T cells were grown to confluency in T75 flasks in GM and switched to SFM 24 h before dosing. The 24-well plates were coated in 500 µl of either ECM solution (30 µg/ml; additive-free DMEM) or fibronectin (0.25 µg/ml; PBS) for 1 h at 37 C. Wells were then washed in PBS before nonspecific binding was blocked with 500 µl PBS containing 0.1% BSA for at least 2 h at 37 C. Meanwhile, cells were trypsinized and collected in SFM. Pellets were resuspended in 1 ml SFM, and 50 µl of the cell solution was counted to determine cell number. Cells were further diluted, using SFM to 0.3 x 10⁶ cells/1.5 ml, to which the treatments as described in Results were added. The cells were placed on a shaker and incubated for 1 h at room temperature. Wells were washed twice with PBS before control, and pretreated cells were applied at 0.1 x 10⁶ cells per well and incubated at 37 C for 30 min. Unattached cells were collected, and the wells were washed with PBS. Cell pellets were collected and resuspended in 100 µl PBS. Adherent cells were trypsinized and collected. Cell pellets were again resuspended in 100 µl PBS. A 50-µl volume of each solution was counted after trypan blue cell staining, from which the percentage of cells attached was determined.

Dosing protocol for growth assay
MCF10A cells were seeded at 5 x 10⁴ cells per well in 24-well plates in GM before being switched to SFM for 24–48 h. Cells were predosed with MßCD (4 mM) for 30 min followed by treatment with IGFBP-3 (100 ng/ml), SPD (10 ng/ml), IGF-I (100 and 200 ng/ml), or filipin (5 µg/ml) for 24 h or EGF (1 and 25 ng/ml) for 48 h. For the involvement of ß1 in IGFBP-3- and SPD-mediated growth, MCF10A cells were dosed with IGFBP-3 (100 ng/ml), SPD (10 ng/ml), and the ß1-integrin receptor antibody (200 ng/ml) either alone or in combination for 24 h. For the involvement of MAPK in IGFBP-3/SPD growth, the cells were dosed with IGFBP-3 (100 ng/ml), SPD (10 ng/ml), or the MAPK inhibitor PD98059 (1 µM) alone or in combination for 48 h. The cells were incubated with 0.1 µCi [³H]thymidine per well for the final 4 h of the dosing time period. After the removal of the supernatant, cells were then incubated with 500 µl of 5% trichloroacetic acid (Merck Ltd., Middlesex, UK) at 4 C for 10 min followed by incubation with 400 µl of 1 M NaOH (Fisher Scientific Ltd., Leicestershire, UK) for 1 h at room temperature. The resulting suspension was placed into individual scintillation vials, and 3 ml of scintillation fluid was added. Samples were analyzed using a Beckman Scintillation Counter LS6500. Data were recorded as disintegrations per minute.

Statistical analysis
The data were analyzed with the Microsoft Excel version 5.0a software package using ANOVA followed by least-significant difference post hoc test. A statistically significant difference was considered to be present at P < 0.05.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
A ß1-integrin receptor antibody reverses the effects of IGFBP-3 on C2-induced apoptosis of Hs578T cells
We have shown previously that IGFBP-3-enhanced apoptosis and reduced attachment of Hs578T breast cancer cells grown on culture plates was unaffected if the cells were plated onto laminin or collagen, but these effects were blocked in the presence of a disintegrin (an RGD-containing fibronectin fragment) and switched to opposite effects when the cells were plated onto fibronectin (18). The most prevalent fibronectin receptor on these cells is 5ß1. We therefore investigated whether a ß1-containing fibronectin receptor played a role in IGFBP-3-mediated enhancement of apop- tosis by the use of a specific ß1 antibody (Fig. 1A). C2-ceramide-induced apoptosis was accentuated further in the presence of IGFBP-3. Alone, IGFBP-3 or the ß1-integrin antibody had no effect on cell death. In the presence of the ß1-integrin antibody, the C2-induced death was unaffected; however, the accentuating actions of IGFBP-3 were reversed, such that IGFBP-3 then conferred significant cell survival, reducing C2-induced apoptosis. This was a specific effect of the ß1 antibody because IGFBP-3 was still able to accentuate C2-induced cell death in the presence of a control mouse IgG (Control = 4.9%, IGFBP-3 = 3.9%, C2 = 41.6%, and C2 plus IGFBP-3 = 47.2%). We also found that the ß1-integrin antibody did not affect IGFBP-3 binding to the cells (data not shown). These data were consistent with the reversal of IGFBP-3 actions by fibronectin being mediated by a ß1- containing integrin receptor.

View larger version (19K): [in this window] [in a new window]   FIG. 1. Effects of a ß1-integrin receptor antibody, disruption of cholesterol-stabilized complexes, and inhibition of MAPK on the intrinsic actions of IGFBP-3 on Hs578T cells. A–C, Hs578T cells were predosed for 24 h with either IGFBP-3 (100 ng/ml) with or without a ß1-integrin receptor blocking antibody (200 ng/ml) (A) or filipin (1 µg/ml) (B) or IGFBP-5 (100 ng/ml) with or without nystatin (50 µg/ml) (C). Cells were then redosed in the same manner with or without an apoptotic dose of C2-ceramide for another 24 h. D, Hs578T cells were incubated with nystatin (50 µg/ml) and then dosed with approximately 500,000 cpm of [125I]IGFBP-3 (specific activity, 3.7 MBq/µg IGFBP-3) for 30 min at either 4 C or room temperature. Membrane-associated [125I]IGFBP-3 was removed by acetic acid washing, and the cellular [125I]IGFBP-3 collected by cell lysis. Data represent active internalisation. E, Hs578T cells were predosed for 24 h with IGFBP-3 (100 ng/ml) with or without PD98059 (1 µM) and then redosed for another 24 h in the same manner with or without an apoptotic dose of antimycin A. Cell death was measured using trypan blue cell counting. The graphs show the mean of three experiments, each repeated in triplicate. F and G, Hs578T cells were treated with RGD (10 µg/ml: used as positive control), IGFBP-3 (100 ng/ml), PD98059 (1 µM), or combinations of each for 1 h before plating on either ECM gel (30 µg/ml) (F) or on fibronectin (0.25 µg/ml) (G) for another 30 min. The graphs show the percentage of cells attached and represent the mean of three experiments, each repeated in triplicate. (Figure continues on next page.)

View larger version (27K): [in this window] [in a new window]   FIG. 1A. Continued

View larger version (27K): [in this window] [in a new window]   FIG. 3. The presence of a ß1-integrin antibody or fibronectin, the disruption of cholesterol-stabilized complexes, and MAPK inhibition modulates the effects of IGFBP-3 in relatively normal MCF-10A breast epithelial cells. Cells were placed in SFM for 48 h before dosing for 24 h with the following: A, IGFBP-3 (100 ng/ml) or SPD (10 ng/ml) with or without ß1-integrin antibody (200 ng/ml); B, IGFBP-3 (100 ng/ml) in the presence or absence of fibronectin (0.25 µg/ml); C, inset, filipin (0–5 µg/ml) after which Triton-soluble and Triton-insoluble membrane fractions were prepared. Western blotting was performed on lysates and the membrane probed for caveolin-1. The blot is representative of experiments repeated three times. C, IGFBP-3 (100 ng/ml) or SPD (10 ng/ml); D, IGF-I (100 and 200 ng/ml); E, EGF (1 and 25 ng/ml) either with or without filipin (5 µg/ml) or after a 30-min preincubation with MßCD (4 mM); F, IGFBP-3 (100 ng/ml) for 15 min (inset, blot for p-MAPK and ERK-2) or IGFBP-3 (100 ng/ml) and SPD (10 ng/ml) with or without PD98059 (1 µM) for 48 h; G and H, IGFBP-3 (100 ng/ml) (G) with or without PD98059 (1 µM) or with or without fibronectin (0.25 µg/ml) (H) for 24 h followed by redosing in the same manner with or without an apoptotic dose of C2-ceramide. Proliferation was assessed using either cell counting or [3H]thymidine incorporation. Cell death was assessed using Trypan blue cell counting. Graphs show the mean of three experiments, each repeated in triplicate.

View larger version (34K): [in this window] [in a new window]   FIG. 3A. Continued (see panels G and H on next page)

View larger version (18K): [in this window] [in a new window]   FIG. 3B. Continued

Involvement of transferrin and caveolin-1 in IGFBP-3 actions
Together these findings suggested that IGFBP-3 actions on breast epithelial cells were determined by integrin receptor complexes and associated downstream signaling pathways. It has recently been reported that IGFBP-3 binds to ß1-integrin subunits and to the associated membrane proteins, caveolin-1 and the transferrin/transferrin receptor complex (24, 25). We therefore went on to investigate their role in the actions of IGFBP-3 on breast epithelial cells. Initially, we examined whether transferrin was required for IGFBP-3 to enhance apoptosis of Hs578T cells. Removal of transferrin from the SFM had no effect on cell death either basally or on C2-induced cell death. In the presence of transferrin, IGFBP-3 significantly increased C2-induced cell death. In the absence of transferrin, however, IGFBP-3 no longer enhanced C2-induced apoptosis (Fig. 2A). Despite this lack of activity, we found that removing transferrin did not reduce IGFBP-3 internalization (data not shown). We then examined whether the ability of transferrin to bind directly to IGFBP-3 was required for it to facilitate IGFBP-3 actions, by repeating this experiment using the SPD, which lacks the region that interacts with transferrin (19). In Fig. 2B, treatment with C2 induced cell death in the presence or absence of transferrin, and in contrast to IGFBP-3, SPD increased C2-induced cell death in either the presence or absence of transferrin. This suggests that transferrin can affect IGFBP-3 actions only if it can bind directly to IGFBP-3 and also that transferrin is required for the actions of full-length IGFBP-3 but not for fragments of IGFBP-3 that lack the transferrin-binding domain. To investigate whether caveolin-1 was required for IGFBP-3 actions, we repeated these experiments in MCF-7 breast cancer cells, which do not express caveolin-1. In Fig. 2C (inset), we confirm that the MCF-7 cells have no detectable caveolin-1, unlike the Hs578T cells. We have previously shown that IGFBP-3 has no effect on ceramide-induced cell death of MCF-7 cells (2); however, having observed that SPD, but not full-length IGFBP-3 could still act on cells without transferrin, we then investigated whether SPD could also still act on MCF-7 cells despite the absence of caveolin-1. With MCF-7 cells, ceramide-induced cell death was enhanced in the presence of SPD (Fig. 2C). We have observed similar results with plasmin-cleaved IGFBP-3 (all fragments of IGFBP-3 produced by plasmin cleavage were added), also accentuating C2-induced cell death in the MCF-7 cells (data not shown) and in the Hs578T cells (26). In Fig. 2C, it can also be seen that in the presence of MßCD, the action of SPD was reversed such that it now acted as a survival factor reducing C2-induced apoptosis. The actions of IGFBP-3 on MCF-7 cells in the absence of caveolin-1 are therefore very similar to that on Hs578T cells in the absence of transferrin. In Fig. 2D, we show that the removal of transferrin from MCF-7 cells also had no effect on SPD enhancement of ceramide-induced cell death. Together these observations suggest that caveolin-1 and transferrin may facilitate the actions of full-length IGFBP-3, but fragments of IGFBP-3, which lack the overlapping domains that interact with these proteins, can still have exactly the same actions in the absence of either transferrin or caveolin-1. We also investigated whether the action of this IGFBP-3 fragment on MCF-7 cells was similarly affected by the presence of fibronectin. In Fig. 2E, we demonstrate that when MCF-7 cells were plated onto fibronectin, the effect of SPD on C2-induced apoptosis was reversed, conferring cell survival. We also found that the ß1-blocking antibody prevented the ability of SPD to enhance C2-induced apoptosis of MCF-7 cells, but MAPK was not involved in this effect and the SPD did not affect MAPK activation (data not shown).

View larger version (20K): [in this window] [in a new window]   FIG. 2. The presence of transferrin affects the actions of IGFBP-3 but not those of SPD on Hs578T and MCF-7 cells. The presence of fibronectin or the disruption of cholesterol-stabilized complexes modulates the effects of SPD in MCF-7 breast cancer cells. A and B, After 24 h in SFM with or without transferrin (0.01 mg/ml), Hs578T cells were predosed with either IGFBP-3 (100 ng/ml) (A) or SPD (10 ng/ml) (B) for 24 h in SFM with or without transferrin (0.01 mg/ml). Cells were redosed in the same manner with or without an apoptotic dose of ceramide in SFM with or without transferrin (0.01 mg/ml) for another 24 h. C, MCF-7 cells were serum starved for 24 h before being lysed and Western blotting performed to determine caveolin-1 levels. MCF-7 cells at three different passages were examined (inset). MCF-7 cells were dosed with MßCD (4 mM) or fresh SFM for 30 min. The medium was then removed and replaced with fresh SFM with or without SPD (5 ng/ml) for 24 h. Cells were then redosed in the same manner with or without an apoptotic dose of ceramide for another 24 h. D, After 24 h in SFM with or without transferrin (0.01 mg/ml), MCF-7 cells were predosed with SPD (5 ng/ml) for 24 h in SFM with or without transferrin (0.01 mg/ml). Cells were reincubated in the same way with or without an apoptotic dose of ceramide with or without transferrin (0.01 mg/ml) for another 24 h. E, After 24 h in SFM, MCF-7 cells plated on either plastic or fibronectin (0.25 µg/ml) were predosed with SPD (5 ng/ml) for 24 h in SFM. Cells were redosed in the same manner with or without an apoptotic dose of ceramide for another 24 h. Cell death was measured by trypan blue dye exclusion. The graphs show the mean of three experiments, each repeated in triplicate.

View larger version (24K): [in this window] [in a new window]   FIG. 2A. Continued

In Fig. 3C, depletion of membrane cholesterol with either MßCD or filipin had no effect on basal levels of cell proliferation. However, in the presence of either MßCD or filipin treatment, the ability of either IGFBP-3 or SPD to increase cell number was completely blocked. In the inset in Fig. 3C, we show, as an example of one of these agents, that the dose of filipin was effective at depleting cholesterol from the membrane of MCF-10A cells. With increasing doses of filipin, there was a dose-dependent shift in the localization of caveolin-1 from the Triton-insoluble to the Triton-soluble fraction of the membrane with 5 µg/ml causing a complete relocalization of caveolin-1 from the Triton-insoluble to the non-raft-associated fraction of the membrane.

This block in proliferation was specific to IGFBP-3, because disruption of cholesterol-stabilized complexes had very different effects on classical mitogens. In Fig. 3D, MCF-10A cells treated with IGF-I at 100 and 200 ng/ml resulted in a significant increase in DNA synthesis, and in the presence of MßCD, these responses were doubled. In the same cells treated with EGF at 1 and 25 ng/ml, the increases in cell proliferation, unlike the response to IGF-I, were unaltered in the presence of MßCD (Fig. 3E). We also examined whether MAPK was involved in these actions of IGFBP-3 on MCF-10A cells. We demonstrate in Fig. 3F that the ability of SPD and IGFBP-3 to significantly increase MCF-10A cell growth was blocked in the presence of PD98059. We also showed that IGFBP-3 was able to cause a rapid increase in the activation of MAPK in the MCF10A cells (Fig. 3F, inset). However, despite the mitogenic effect of IGFBP-3 being MAPK dependent, in Fig. 3G, we show that the ability of IGFBP-3 to reduce C2-induced apoptosis of the MCF-10A cells was unaffected in the presence of the MAPK inhibitor PD98059. In the presence of fibronectin, however, the actions of IGFBP-3 on apoptosis were reversed (Fig. 3 H). Treatment of MCF10A cells with IGFBP-3 alone had no effect on basal cell death. Ceramide increased apoptosis, but in contrast to the survival actions of IGFBP-3 on ceramide-induced cell death on plastic, on fibronectin, apoptosis was accentuated in the presence of IGFBP-3.

IGFBP-3 modifies integrin receptor complexes on the cell surface of Hs578T cells
Having demonstrated that IGFBP-3 actions on all three breast epithelial cell lines can be reversed by manipulations of integrin receptor complexes, we then sought to confirm that IGFBP-3 was directly interacting with these complexes and affecting downstream intracellular signaling pathways. As with growth factor receptors, it is now evident that integrin receptors can interact with a large number of ancillary proteins that enable activation of multiple intracellular signaling pathways. Different cells express different complements of integrin receptors and ancillary proteins that interact specifically in a differential manner dependent upon which integrins are activated by the surrounding ECM. The elucidation of all pathways affected by IGFBP-3 on different cells will take considerable time to fully characterize. We therefore aimed to initially establish that IGFBP-3 was directly affecting the association of such complexes and initiating downstream signaling events in our original model of Hs578T cells, which do not respond to IGF. It has been recently reported that IGFBP-3 can bind to ß1-integrin subunits and caveolin-1 on HEK293 cells (24, 25). We therefore examined whether IGFBP-3 would also bind to these on Hs578T cells. We confirmed using immunoprecipitation with anti-caveolin-1 and anti-ß1-integrin subunit antibodies that IGFBP-3 can associate with both caveolin-1 and ß1 on the Hs578T cells and that IGF-I reduced IGFBP-3 binding to caveolin-1 by 68.2 ± 9.8% (P < 0.001) (mean densitometry readings from three immunoblots) and reduced IGFBP-3 binding to ß1 by 43.6 ± 15.8% (P < 0.05) (Fig. 4, A and B). The binding of IGFBP-3 to the cell surface and its actions have been shown previously to be reduced when IGF-I is bound (26, 27). We next examined whether IGFBP-3 binding to caveolin-1 and the integrin receptor altered the extent to which these two proteins associated with each other; we found that IGFBP-3 significantly increased the association of caveolin-1 with ß1 by 39.1 ± 21% (P < 0.05) (Fig. 4C). Having shown that IGFBP-3 is able to associate with both caveolin-1 and ß1 and has the ability to enhance their association, we looked at how disrupting cholesterol-stabilized complexes altered the ability of IGFBP-3 to associate with these two components. As expected, we observed that disruption of cholesterol-stabilized complexes by the use of nystatin resulted in a decrease in the association of ß1 with caveolin-1 (Fig. 4D). After disruption of cholesterol-stabilized complexes, IGFBP-3 treatment, rather than enhancing complex formation, then reduced this association by 36.3 ± 8.6% (P < 0.001) (Fig. 4D). The ability of IGFBP-3 to associate individually with either caveolin-1 or ß1 was, however, markedly increased upon disruption of cholesterol-stabilized complexes by 115 ± 19.9% (P < 0.001) and by 104 ± 39.6% (P < 0.05), respectively (Fig. 4, E and F). This indicates that in conditions where IGFBP-3 accentuated ceramide-induced apoptosis, it increased the association of ß1 with caveolin-1, but in conditions where it reduced the apoptosis, it had the opposite effect and decreased this association. This also suggests that the actions of IGFBP-3 are not a result of the isolated association of IGFBP-3 with either ß1 or caveolin-1 but are related to its ability to alter their association together in membrane complexes. The formation of such complexes is recognized to be the critical determinant of signaling from integrin receptors, which lack any intrinsic kinase activity.

View larger version (55K): [in this window] [in a new window]   FIG. 4. IGFBP-3 modifies integrin receptor complexes at the cell surface. Hs578T cells were treated for 30 min with approximately 500,000 cpm of [125I]IGFBP-3 (specific activity, 3.7 MBq/µg IGFBP-3) with or without IGF-I (100 ng/ml) (A and B) or nystatin (50 µg/ml) (E and F) or transferrin (0.01 mg/ml) (G and H) and then lysed and immunoprecipitated (IP) with antibodies to either caveolin-1 (A, E, and G) or ß1 (B, F, and H) or ngIGFBP-3 (100 ng/ml) was added to the cells either alone (C) or with or without nystatin (D) followed by immunoprecipitation with the ß1-integrin antibody. Lysates were separated on a gel that was either exposed to x-ray film or transferred onto a nitrocellulose membrane and immunoblotted (WIB) for caveolin-1 or ß1 to show comparable levels of protein immunoprecipitated. In these experiments, we used a nonspecific mouse IgG antibody as a control and showed that the immunoprecipitation was specific to the antibodies used. I, Representative example where 500,000 cpm of [125I]IGFBP-3 was added in the presence or absence of IGF-I (100 ng/ml) or transferrin (TR) (0.01 mg/ml) and immunoprecipitated with antibodies to either caveolin-1 or a control mouse IgG. The blots are representative of experiments repeated three times.

IGFBP-3 recruits FAK to the membrane, increases its association with ß1, and enhances activation of MAPK in Hs578T cells
Having shown a role for integrins in IGFBP-3 function, we investigated whether IGFBP-3 was able to modulate the classical integrin-mediated signaling pathway of membrane recruitment of FAK and subsequent activation of MAPK. Treatment of Hs578T cells with IGFBP-3 resulted in a significant increase in the association of FAK with ß1 by 75.8 ± 21.1 (P < 0.01) (Fig. 5A). We also confirmed that IGFBP-3 altered the localization of FAK within the cell; after 30 min exposure to IGFBP-3, there was an increase in the amount of FAK localized to the cellular membrane by 96.4 ±4.2 (P < 0.05) (Fig. 5B). This translocation of FAK to the membrane after IGFBP-3 treatment was also visualized using confocal microscopy (Fig. 5C). These images illustrated that in the controls, FAK was primarily localized in the cytoplasm, and after treatment with IGFBP-3, there was a dramatic shift in the localization of FAK to the focal adhesions in the cell membrane, entirely consistent with an increase in the association of FAK with ß1 (shown in Fig. 5A). Having determined that IGFBP-3 interacted with integrin signaling complexes and promoted the recruitment of FAK, we then showed that the addition of IGFBP-3 caused a dose-dependent increase in the phosphorylation of MAPK (Fig. 5D) with 200 ng/ml IGFBP-3 giving a 3-fold increase in MAPK activation. At 100 ng/ml, there was a 26 ±12.6 (P < 0.05) increase in MAPK activation, which was blocked in the presence of the MAPK inhibitor PD98059 (Fig. 5E).

View larger version (35K): [in this window] [in a new window]   FIG. 5. IGFBP-3 recruits FAK to the membrane, increases its association with ß1, and enhances activation of MAPK in Hs578T cells. A, Hs578T cells were dosed with ngIGFBP-3 (100 ng/ml) for 30 min and then lysed followed by immunoprecipitation (IP) with the ß1-integrin receptor antibody and then immunoblotting (WIB) for FAK. The membrane was also probed for ß1 to show comparable levels of protein immunoprecipitated. B, Cells were treated with ngIGFBP-3 (100 ng/ml) for 30 min, and then membrane fractionation of whole-cell lysates was performed, which were then immunoblotted for FAK. Blots are representative of experiments repeated three times. C, Images focusing on the top, middle, and bottom of Hs578T cells using confocal microscopy. The images demonstrate that in the controls, FAK is primarily found in the cytoplasm, and after treatment with ngIGFBP-3 (100 ng/ml) for 30 min, FAK shifts to the focal adhesions of the cell membrane, mainly seen at the bottom where the cell is attached to the slide. D and E, Hs578T cells treated with IGFBP-3 (0–200 ng/ml) for 30 min (D) or IGFBP-3 (100 ng/ml) with or without PD98059 (1 µM) for 30 min (E) before lysing. Samples were then immunoblotted for p-MAPK and ERK-2. The blots are representative of experiments repeated at least three times. In these experiments, we used a nonspecific control IgG antibody as a control and showed that the immunoprecipitation was specific to the antibodies used (see Fig. 4I).

View larger version (57K): [in this window] [in a new window]   FIG. 6. Disrupting cholesterol-stabilized complexes and the presence of transferrin modulate the ability of IGFBP-3 to enhance activation of MAPK in Hs578T cells. Cells were dosed with either IGFBP-3 (100 ng/ml) with or without nystatin (50 µg/ml) (A) or IGFBP-3 (100 ng/ml) with or without transferrin (0.01 mg/ml) (B) for 30 min minutes before lysing. Samples were then immunoblotted for p-MAPK and ERK-2. Blots are representative of experiments repeated three times.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

For breast epithelial cells to survive and proliferate, they require cues from both soluble growth factors and also from the solid ECM within which they reside (29). In the normal breast, the epithelial cells form monolayers within ductal structures surrounded by a basement membrane composed mainly of laminin and collagen. The local appearance of increased fibronectin within such glandular structures clearly has a marked effect on epithelial cell function, resulting in ductal branching (30). In breast tumors, disruption of the well-defined tissue architecture results in exposure of the epithelial cells to fibronectin in the stroma. Increased expression of fibronectin in breast cancers is associated with poor prognosis and metastasis (31, 32). Our results suggest that exposure to fibronectin reverses the actions of IGFBP-3. This could explain the apparent contradiction between the many reports of growth-inhibitory and proapoptotic actions of IGFBP-3 and the data from in vivo studies indicating that IGFBP-3 within breast tumors is positively associated with large, highly proliferative tumors and poor prognostic markers (15, 16). The increased exposure to fibronectin with tumor progression would switch IGFBP-3 from having inhibitory actions to enhancing tumor cell survival and proliferation.

The major class of cell surface adhesion receptors are the integrin receptors, a large family of glycoproteins that form heterodimeric receptors comprising one each of a number of - and ß-subunits noncovalently linked together. The combination of - and ß-subunits determine specificity for particular components of the ECM. The integrins congregate within sphingolipid-rich, cholesterol-stabilized microdomains within the plasma membrane that form platforms for the clustering of cell signaling components (33). ECM ligation of integrins induces receptor clustering, their association with other membrane proteins such as caveolin-1 (34), and the recruitment of signaling molecules such as FAK, which can then transduce the signal via intracellular pathways such as MAPK (35) affecting proliferation and survival.

There have been many reports of IGF-independent actions of IGFBP-3, although the mechanism of such actions remains unclear. A putative cell surface IGFBP-3 receptor has been reported (36), and there have been separate reports that IGFBP-3 can modulate different signaling pathways including SMADs (37), Stat1, (38) and MAPK (10). IGFBP-3 can also be internalized and translocated into the nucleus where it can associate with the retinoid receptor (RXR) (39) and potentially act via modifying gene transcription. There has, however, been limited linkage between these potential mechanisms and the many described actions of IGFBP-3. Several of the actions of IGFBP-3 occur very rapidly, which excludes a nuclear transcriptional mode of action in these instances. We have shown that IGFBP-3 has effects on adhesion of breast epithelial cells within 1 h and effects on growth and apoptosis that are preceded by changes in association of membrane complexes and activation of downstream signaling pathways within minutes (18). The rapid effect on cell adhesion implicated integrin receptors in the intrinsic actions of IGFBP-3, because cell adhesion is the classical indicator of integrin receptor function. We observed that IGFBP-3 decreased the adhesion of these cells to collagen or laminin but promoted their adhesion to fibronectin (18). We also demonstrated that IGFBP-3 accentuated apoptosis of these cells on collagen or laminin but promoted survival on fibronectin, and its ability to modulate apoptosis or cell attachment was blocked by a disintegrin (18). We have now extended these observations, demonstrating that fibronectin can reverse the effects of IGFBP-3 on growth and apoptosis of two additional human breast epithelial cell lines. We have also demonstrated the involvement of the ß1-integrin receptor subunit, which is part of the main 5ß1 fibronectin receptor found on these cells.

We have confirmed previous reports that IGFBP-3 associates with the ß1-integrin and associated proteins on the cell membrane. These associations were previously proposed as a mechanism for internalization of IGFBP-3 as an initial step in facilitating nuclear actions (24, 25). Our data, however, are more consistent with these interactions playing a more direct role in IGFBP-3 actions. On Hs578T cells, IGFBP-3 induced a rapid increase in the association of ß1-integrin subunits with caveolin-1 in the plasma membrane, and rapid MAPK activation and the subsequent enhancement of apoptosis was prevented when MAPK was inhibited. Depletion of cholesterol from the plasma membrane decreased the association of ß1/caveolin-1 complexes and prevented IGFBP-3 from activating MAPK. This did not, however, prevent IGFBP-3 internalization; indeed, there was more IGFBP-3 associated individually with ß1 and caveolin-1 and more internalized into the cell. Despite increased uptake, IGFBP-3 then had the opposite effect and reduced further the association of ß1 with caveolin-1 and no longer enhanced apoptosis but conferred survival, reducing apoptosis. Therefore, the actions of IGFBP-3 did not correlate with its association with ß1-integrins or caveolin-1 or its internalization, but it did correlate with the effect of IGFBP-3 on the association of ß1 with caveolin-1.

Although caveolin-1 and the transferrin/transferrin receptor appear to be involved in the actions of full-length IGFBP-3 on Hs578T cells, these integrin ancillary proteins do not appear to be required for IGFBP-3 actions. The SPD does not contain the caveolin-1 binding sequence (25) or the sequence thought to result in binding to transferrin, but it still retains all of the actions of IGFBP-3, which could still be reversed by fibronectin or by cholesterol depletion. The SPD and plasmin-cleaved IGFBP-3 also had the same actions on MCF-7 cells, which lack caveolin-1, confirming that this was not required for these actions. Transferrin also had no effect on SPD actions on MCF-7 cells. These ancillary membrane proteins, therefore, appear to be modulators of the response to IGFBP-3 rather than directly involved in its actions. Full-length IGFBP-3 was without effect in the absence of either transferrin or caveolin-1, but fragments of IGFBP-3 lacking the domain responsible for their binding retained full activity. The domain of IGFBP-3 that interacts with transferrin and caveolin-1 is a sequence of very highly charged amino acids that also interacts with many other proteins including proteoglycans on the cell surface and ECM. With Hs578T cells, removing transferrin caused a small decrease in IGFBP-3 binding to the ß1-integrin and caveolin-1 but completely negated its actions. Depleting cholesterol from these cells actually increased the association of IGFBP-3 with the ß1-integrin and caveolin-1, but it then had the opposite effect on their association and opposite effects on cell function. These data suggest that IGFBP-3 interacts with several cell surface proteins and affects cell function by modifying their association within complexes and that full-length IGFBP-3 binding to transferrin and caveolin-1 can facilitate these interactions. The data also clearly indicate that fragments of IGFBP-3, as could be generated by proteolysis, which do not interact with transferrin or caveolin-1, still retain the same actions that are still dependent upon cholesterol-stabilized complexes but are not modulated by transferrin or caveolin-1. In contrast, the fibronectin receptor was implicated in the actions of IGFBP-3 on all of the three cell lines studied. The fibronectin receptor appears to be of particular importance in the normal development of the mammary gland and is hormonally regulated (40). The ß1 subunit has been reported to have differential effects on the growth and survival of normal and malignant human breast epithelial cells (41). This would be consistent with our observed differential effects of IGFBP-3 on malignant and nonmalignant breast epithelial cells (17). It has also been reported that 5ß1-integrin attenuated the growth of human colon carcinoma cells, but the same receptor enhanced proliferation when they were cultured on fibronectin (42). This is similar to our reversal of the actions of IGFBP-3 in all of the breast epithelial cells examined when they were cultured on fibronectin, although fibronectin alone did not alter their growth or apoptosis. The main fibronectin receptor on breast epithelial cells is the 5ß1-integrin, but there are many other potential integrin receptor subunit combinations that IGFBP-3 could potentially interact with, and future work will be required to establish all of the interactions in different cell types.

We and others have previously demonstrated that IGF-I is able to sequester IGFBP-3 away from the cell surface and prevent its intrinsic actions (26, 27). Consistent with these previous reports, we have now demonstrated that IGF-I can prevent the specific interactions of IGFBP-3 with ß1 and caveolin-1. We have also shown that the actions of SPD and plasmin-cleaved IGFBP-3, which lack the caveolin-1 and the transferrin-binding domains, are unaffected by caveolin-1 and transferrin. Together these observations add considerably to the implications of IGFBP-3 proteolysis. Originally described as a phenomenon limited to pregnancy, we then demonstrated that it was a general occurrence with increased proteolysis of IGFBP-3 in a number of clinical conditions (43). This has subsequently become accepted as a general mechanism to control the availability of the large reservoir of IGF that is present in association with IGFBP-3 (44). Proteolysis of IGFBP-3 that reduces the affinity for binding IGF will not just shift the equilibrium altering the availability of IGF for receptor binding, but it would also reduce the ability for IGF to restrict the intrinsic action of IGFBP-3. The main fragment from in vivo proteolysis contains the N-terminal region and most of the midregion (containing the SPD sequence) but not the sequences involved in binding to caveolin-1 and transferrin. Proteolysis could therefore also free IGFBP-3 from the modulation by these additional interactions. The intact IGFBP-3 would bind IGFs and limit their actions, and the IGFs would limit the actions of IGFBP-3; proteolysis would then increase the availability of IGFs for receptor interactions but also liberate IGFBP-3 from inhibition by IGFs and from its requirement for interactions with caveolin-1 and transferrin.

In the Hs578T cells, we also observed rapid IGFBP-3-induced recruitment of FAK to ß1-integrins in the cell membrane and activation of MAPK. Although MAPK has conventionally been associated with a proliferative and survival response, we demonstrated that ceramide-induced death of Hs578T cells was dependent upon MAPK activation. We had previously described a similar requirement for MAPK activation in ceramide-induced apoptosis of murine fibroblasts (45), and this has also been described in several other models. Activation of MAPK was also involved, in a more classical role, in the mitogenic response of the MCF10A cells to IGFBP-3. It was not, however, implicated in the effects of IGFBP-3 on the apoptosis response of MCF-10A or MCF-7 cells. It is now clear that integrins have a large number of potential linkages to multiple signaling pathways (46). In the Hs578T cells, IGFBP-3 interacts with the fibronectin receptor and alters its association with caveolin-1, and this affects signaling pathways linked to these complexes. In the MCF-7 cells that lack caveolin-1, the fibronectin receptor was still integral to the IGFBP-3 response but clearly not via integrin/caveolin-1 complexes and presumably through one of the many other potential complexes that can be formed with these receptors, which could then modify alternative signaling pathways rather than MAPK. The different integrin complexes formed in different cells will determine the signaling pathway affected by IGFBP-3. This is entirely analogous to conventional growth factors, such as IGF-I, which activates several intracellular signaling pathways, and although MAPK is conventionally considered to mediate its mitogenic actions, this is not the case in all cell types, for instance IGF-I-induced proliferation of MCF-7 cells is phosphatidylinositol-3-kinase and not MAPK mediated (47).

Our findings clearly indicate that in all three breast epithelial cells examined, IGFBP-3 could have either positive or negative effects on cell functions. That IGFBP-3 can both promote and inhibit growth is not dissimilar from conventional growth factors such as TGF-ß (12, 48). In each of the cell models that we examined, the common feature that was a critical determinant of whether IGFBP-3 had positive or negative effects on cell function was the status of cholesterol-stabilized integrin receptor complexes. Two of the IGFBPs, IGFBP-1 and -2, have classical integrin recognition RGD sequences and have been shown to induce dephosphorylation of FAK, cell detachment, and apoptosis (49). Although IGFBP-3 does not possess a classical integrin recognition sequence, it has been reported to bind with high affinity to many other proteins that interact directly with integrins, including fibronectin (50), fibrin and fibrinogen (51), and caveolin-1 and the transferrin receptor (24, 25). It is possible that IGFBP-3 interacts with integrin receptors via one of these intermediates or directly via a nonclassical integrin recognition sequence. The IGFBPs are closely related to a superfamily of CCN proteins, including NOV, mac25, CTGF, and CYR61; these are all cysteine-rich proteins with many pleiotropic actions on cell functions similar to the intrinsic actions of IGFBP-3 (52). No specific receptors have been described for CCN proteins, but it is now recognized that they generally act via integrin receptors with which they interact through nonclassical recognition sequences (53).

Many of the tumor inhibitors that work, at least in part, via IGFBP-3 are limited in clinical practice because of de novo or acquired resistance. We believe that our new data help explain many apparent contradictions in relation to previous reports of IGFBP-3 actions and could have important implications for understanding therapeutic responses to agents that act in concert with IGFBP-3 and whose actions change with tumor progression.

	Acknowledgments

 
We thank the Association for International Cancer Research and The Needham Cooper Trust for supporting this work.

	Footnotes

 
Disclosure Summary: C.B., J.H., N.L., E.V., J.C., M.C., J.M., Z.W., and C.P. have nothing to declare.

First Published Online April 13, 2006

Abbreviations: ECM, Extracellular matrix; EGF, epidermal growth factor; FAK, focal adhesion kinase; GM, growth medium; IGFBP, IGF-binding protein; MßCD, methyl-ß-cyclodextrin; ngIGFBP-3, nonglycosylated IGFBP-3; SFM, serum-free medium; SPD, serine phosphorylation domain peptide; TBS, Tris-buffered saline.

Received January 3, 2006.

Accepted for publication April 3, 2006.

	References

Top Abstract Introduction Materials and Methods Results Discussion References

 

1. Pollak MN, Schernhammer ES, Hankinson SE 2004 Insulin-like growth factors and neoplasia. Nat Rev Cancer 4:505–518[CrossRef][Medline]
2. Perks CM, McCaig C, Clarke JB, Clemmons DR, Holly JM 2002 A non-IGF binding mutant of IGFBP-3 modulates cell function in breast epithelial cells. Biochem Biophys Res Commun 294:988–994[CrossRef][Medline]
3. Perks C, Holly J 2003 Actions of IGFBP on epithelial cancer cells: potential for new therapeutic targets. Horm Metab Res 35:828–835[Medline]
4. Rajah R, Valentinis B, Cohen P 1997 Insulin-like growth factor (IGF)-binding protein-3 induces apoptosis and mediates the effects of transforming growth factor-ß1 on programmed cell death through a p53- and IGF-independent mechanism. J Biol Chem 272:12181–12188[Abstract/Free Full Text]
5. Gill ZP, Perks CM, Newcomb PV, Holly JM 1997 Insulin-like growth factor-binding protein (IGFBP-3) predisposes breast cancer cells to programmed cell death in a non-IGF-dependent manner. J Biol Chem 272:25602–25607[Abstract/Free Full Text]
6. Williams AC, Collard TJ, Perks CM, Newcomb P, Moorghen M, Holly JM, Paraskeva C 2000 Increased p53-dependent apoptosis by the insulin-like growth factor binding protein IGFBP-3 in human colonic adenoma-derived cells. Cancer Res 60:22–27[Abstract/Free Full Text]
7. Hollowood AD, Lai T, Perks CM, Newcomb PV, Alderson D, Holly JM 2000 IGFBP-3 prolongs the p53 response and enhances apoptosis following UV irradiation. Int J Cancer 88:336–341[CrossRef][Medline]
8. Baxter RC 2001 Signalling pathways involved in antiproliferative effects of IGFBP-3: a review. Mol Pathol 54:145–148[Abstract/Free