Insulin-like Growth Factor-binding Protein-3 Potentiates Epidermal Growth Factor Action in MCF-10A Mammary Epithelial Cells

Insulin-like growth factor-binding protein-3 (IGFBP-3) is inhibitory to the growth of many breast cancer cells in vitro; however, a high level of expression of IGFBP-3 in breast tumors correlates with poor prognosis, suggesting that IGFBP-3 may be associated with growth stimulation in some breast cancers. We have shown previously in MCF-10A breast epithelial cells that chronic activation of Ras-p44/42 mitogen-activated protein (MAP) kinase confers resistance to the growth-inhibitory effects of IGFBP-3 (Martin, J. L., and Baxter, R. C. (1999) J. Biol. Chem. 274, 16407–16411). Here we show that, in the same cell line, IGFBP-3 potentiates DNA synthesis and cell proliferation stimulated by epidermal growth factor (EGF), a potent activator of Ras. A mutant of IGFBP-3, which fails to translocate to the nucleus and has reduced ability to cell-associate, similarly enhanced EGF action in these cells. By contrast, the structurally related IGFBP-5, which shares many functional features with IGFBP-3, was slightly inhibitory to DNA synthesis in the presence of EGF. IGFBP-3 primes MCF-10A cells to respond to EGF because pre-incubation caused a similar degree of EGF potentiation as co-incubation. In IGFBP-3-primed cells, EGF-stimulated EGF receptor phosphorylation at Tyr-1068 was increased relative to unprimed cells, as was phosphorylation and activity of p44/42 and p38 MAP kinases, but not Akt/PKB. Partial blockade of the p44/42 and p38 MAP kinase pathways abolished the potentiation by IGFBP-3 of EGF-stimulated DNA synthesis. Collectively, these findings indicate that IGFBP-3 enhances EGF signaling and proliferative effects in breast epithelial cells via increased EGF receptor phosphorylation and activation of p44/42 and p38 MAP kinase signaling pathways.

Insulin-like growth factor-binding protein-3 (IGFBP-3), 1 a 45-kDa glycoprotein abundant in the circulation and extracellular environment, is a key regulator of the peptide hormones IGF-I and IGF-II (1). By virtue of its high affinity for these growth factors, IGFBP-3 competes for ligand binding with the receptor primarily responsible for mediating the actions of IGF-I and -II, the type I IGF receptor (IGFR1) (2), and thereby blocks mitogenic and anti-apoptotic signaling initiated by its activation. An important role for IGFBP-3 in modulating the proliferative effects of IGFs in many cell types is well recognized, and both exogenous and endogenous IGFBP-3 have been shown to block IGF action in breast cancer cells in vitro (3)(4)(5)(6).
An additional role for IGFBP-3 also exists as a growth modulator with intrinsic bioactivity in breast cancer cells and other cells in vitro. The antiproliferative effects of a number of antitumor agents including transforming growth factor-␤ (TGF-␤), vitamin D, and retinoic acid in breast cancer cells appear to be mediated, at least in part, by IGFBP-3 acting independently of IGF sequestration (7,8). The IGF independence of IGFBP-3 actions is inferred from an inability of other IGFBPs to mimic the effect, persistence of the growth-inhibitory effect of IGF-BP-3 in the presence of insulin or IGF analogs, which activate the IGFR1 but which do not bind IGFBP-3, and no clear evidence of IGFs being present in the system under investigation (7). In addition to its anti-mitogenic effects, IGFBP-3 exhibits pro-apoptotic activity in vitro. IGFBP-3 may sensitize breast cancer cells to apoptotic inducers such as ionizing radiation (9) and ceramide (10), and directly effect apoptosis via induction of pro-apoptotic proteins such as Bax and Bad (9). Collectively, these observations suggest that IGFBP-3 is an important antiproliferative agent in breast cancer cells, acting both through IGF-independent and IGF-modulatory pathways.
By contrast with these findings, however, IGFBP-3 may also be growth-stimulatory in vitro. In MCF-7 breast cancer cells, IGFBP-3 enhanced IGF-stimulated DNA synthesis (11) similar to its effects in fibroblasts reported previously (12). More recent studies have shown that proliferation of airway smooth muscle cells is stimulated by IGFBP-3 in the presence of serum (13), and IGFBP-3 increases proliferation of LNCaP prostate cancer cells in the absence of serum or IGFs (14). In another cell model, T47D breast cancer cells transfected with IGFBP-3 cDNA were initially growth-inhibited by the expressed protein, but with increasing passage number became resistant to its growth-inhibitory effects, and were instead growth-stimulated by the endogenous IGFBP-3 (15).
Such observations of growth-stimulation by IGFBP-3 in breast cancer cells, although difficult to reconcile with the many reports of its antiproliferative actions, are consistent with clinical data, which indicate that IGFBP-3 may be associated with indicators of poor outcome. Thus, IGFBP-3 in breast tumors correlates inversely with estrogen receptor expression, and is positively associated with aneuploidy, S-phase fraction, and tumor size (16 -19). Although the significance of these associations is unclear, they imply that there may be changes in sensitivity to the effects of IGFBP-3 on breast epithelial cells in vivo, losing growth-inhibitory bioactivity and perhaps even accelerating tumor growth.
While investigating mechanisms that may underlie IGFBP-3 insensitivity in breast cancer cells, we found that, whereas IGFBP-3 inhibits DNA synthesis in MCF-10A breast epithelial cells, this response is lost when the cells have undergone malignant transformation via expression of oncogenic Ras (20). Blockade of p44/42 MAP kinase-activation downstream of Ras restored IGFBP-3 sensitivity, implicating this pathway in the development of resistance to IGFBP-3. The present study was initiated to determine whether activation of Ras and the p44/42 MAP kinase pathway by epidermal growth factor (EGF), a potent mitogen for normal breast epithelial cells and many breast cancer cell lines, could similarly induce resistance to the inhibitory effects of IGFBP-3 in MCF-10A cells. We now report that IGFBP-3 enhanced the growth stimulatory effects of EGF in this cell line, and that the p44/42 and p38 MAP kinase pathways appear to be involved in this potentiation.

EXPERIMENTAL PROCEDURES
Reagents-Tissue culture reagents and plasticware were from Trace Biosciences (North Ryde, New South Wales, Australia) and Nunc (Roskilde, Denmark). Bovine serum albumin (BSA), bovine insulin, hydrocortisone, and EGF were purchased from Sigma, and cholera enterotoxin was from ICN Biomedicals Australasia (Seven Hills, New South Wales, Australia). Transforming growth factor-␣ (TGF-␣) and heregulin (heregulin-␤, EGF domain) were from Upstate Biotechnology, Inc. (Lake Placid, NY). Signaling pathway inhibitors were purchased from Calbiochem-Novabiochem (Alexandria, New South Wales, Australia): MEK inhibitor PD98059, PI 3-kinase inhibitor LY294002, and p38 MAP kinase inhibitor SB203580. Inhibitors were made up as 50 mM stock solutions in dimethyl sulfoxide, and stored at Ϫ20°C. The following antibodies for Western blotting were purchased from Cell Signaling (Beverly, MA): phospho-Thr-202/Tyr-204 and total p44/42 MAP kinase; phospho-Ser-473 and total Akt; phospho-Thr-180/Tyr-182 and total p38 MAP kinase; phospho-Tyr-1068 EGFR and total EGFR. Natural human IGFBP-3 was purified from Cohn fraction IV of human plasma, as reported previously (21). Recombinant human IGFBP-3 and IGFBP-5 were expressed in human 911 retinoblastoma cells using an adenoviral expression system, and purified by IGF-I affinity chromatography and reverse-phase high performance liquid chromatography (22,23). Electrophoresis and ECL reagents were purchased from Bio-Rad, Amrad-Pharmacia (Ryde, New South Wales, Australia), and Pierce. EGF and protein A were radiolabeled with 125 I (ICN) using chloramine T.
DNA Synthesis and Cell Proliferation Assays-DNA synthesis was assessed by incorporation of [methyl-3 H]thymidine (Amersham, Bucks, United Kingdom). MCF-10A cells were dispensed into 48-well plates at 1 ϫ 10 5 cells/well in growth medium, and incubated for 48 h. Cells were changed to serum-free medium (Dulbecco's modified Eagle's medium/ F-12 containing 1 g/liter BSA) and incubated for 48 h prior to treatment. Test reagents (e.g. EGF, IGFBP-3) were added in 0.2 ml/well serum-free medium, and incubation was continued at 37°C for another 24 h. For the final 4 h of this incubation period, 1 Ci/well [ 3 H]thymidine was added in 25 l of serum-free medium. Monolayers were rinsed with cold saline (9 g/liter NaCl) and fixed in cold methanol:acetic acid (3:1) for 2 h at 4°C. Cells were solubilized in 0.4 ml of 1 M NaOH, and lysates were mixed with 3 ml of OptimaGold (Packard Instrument Co., Meriden, CT) before counting for 1 min in a Hewlett-Packard ␤-counter.
To measure cell proliferation, cells were dispensed into six-well plates at 5 ϫ 10 5 cells/well in growth medium, allowed to attach for 24 h, then changed to serum-free medium as above. Reagents were added, and incubation continued for 7 days. Cells were trypsinized and counted using a hemacytometer.
Analysis of Cell-associated IGFBP-3-Cell-associated IGFBP-3 was determined immunologically, as described previously (25). Briefly, media were removed from cell monolayers after treatment, and cells were washed twice with serum-free medium. Anti-IGFBP-3 antibody R30 was added to each well (final concentration, 1/5000 in serum-free medium), and incubation was continued overnight at 22°C. Monolayers were washed as before, and then 125 I-protein A was added to each well (20,000 cpm in 200 l of serum-free medium) for 2 h at room temperature. Cells were washed and solubilized with 5 g/liter SDS before counting in a ␥-counter.
EGF Binding Assays-Binding of 125 I-EGF to monolayers was carried out at 4°C. Monolayers (with and without IGFBP-3 pretreatment) were washed, and 50,000 cpm 125 I-EGF was added per well in 200 l of serum-free medium, in the absence or presence of 100 ng/ml unlabeled EGF (to determine nonspecific binding). Incubation at 4°C was continued for 2 h, after which cells were washed twice with ice-cold saline, solubilized in 5 g/liter SDS, and counted. Full displacement curves were generated by co-incubation of cells with 125 I-EGF in the presence of 0.1-100 ng/ml unlabeled EGF under the same conditions.
Analysis of Signaling Intermediates-Cells were plated and grown in six-well plates as described above for cell proliferation assays. Preincubation with IGFBP-3 with or without inhibitors was performed when required for the second 24-h period of the 48-h serum-free incubation. Media were then removed and replaced by serum-free medium (1 ml/well) without or with EGF. Incubation was continued at 37°C for 8 min, after which media were removed, and the monolayers were washed immediately in ice-cold saline, then solubilized in 500 l of 2ϫ SDS-PAGE sample buffer (62.5 mM Tris-HCl, pH 6.8, containing 20 g/liter SDS, 100 ml/liter glycerol, 1 g/liter bromphenol blue, and 50 mM dithiothreitol) at 4°C for 10 min. Lysates were scraped into cold Eppendorf tubes and frozen immediately at Ϫ80°C. Prior to SDS-PAGE analysis, thawed lysates were sonicated for 15 s, heated at 90°C for 10 min, cooled on ice, and then centrifuged at 15,000 rpm for 2 min.
SDS-PAGE and Western Analysis-Prepared lysates were subjected to SDS-PAGE (7.5% separating gel for EGFR, 12% separating gel for p44/42, p38, and Akt) and electrophoretic transfer, as described previously (20). After transfer, filters were blocked in either 50 g/liter skim milk powder (for p44/42 and p38 MAPK) or 50 g/liter BSA (for EGFR and Akt) in Tris-buffered saline with Tween 20 (10 mM Tris, 150 mM NaCl, pH 7.4 containing 1 ml/liter Tween 20). Primary antibody incubation was carried out at 4°C for 16 h, with antibodies diluted 1:1000 in blocking buffer. Filters were washed three times for 10 min each in cold Tris-buffered saline with Tween 20, and then incubated with the appropriate horseradish peroxidase-linked secondary antibody for 1-2 h at room temperature. Filters were washed as before and developed by enhanced chemiluminescence using Pierce reagents.
Statistical Analyses-All experiments were performed a minimum of two times and are shown as data pooled from the two experiments, unless indicated otherwise. Significance on pooled data was determined by analysis of variance using the Statview package for Macintosh.

RESULTS
We have shown previously that, in the absence of exogenous growth factors, IGFBP-3 is inhibitory to DNA synthesis in MCF-10A breast epithelial cells over 24 h, but that constitutive up-regulation of Ras-MAPK signaling abolishes the growthinhibitory effect of IGFBP-3 (20). Because EGF stimulates activation of the Ras-MAPK pathway in these cells (data not shown), we therefore investigated whether it similarly affected their growth response to IGFBP-3. As shown in Fig. 1A, EGF is a potent stimulator of DNA synthesis in MCF-10A cells over 24 h, with a significant effect apparent with 0.1 ng/ml EGF. In the absence of EGF, natural human IGFBP-3 inhibited DNA synthesis in MCF-10A cells, as reported previously (20) (data not shown). However, IGFBP-3 enhanced the stimulatory effect of low concentrations (0.1 and 1 ng/ml) of EGF (Fig. 1A), with a significant and maximal effect at 10 ng/ml IGFBP-3. The potentiating effect of IGFBP-3 was lost at high (10 ng/ml) EGF concentrations (Fig. 1A).
The effects of IGFBP-3 and EGF on cell proliferation were determined over 7 days in the absence of serum or other exogenous growth factors. EGF alone (1 ng/ml) increased cell number by ϳ30% compared with untreated cultures (p Ͻ 0.05, Fig.  1B), and this was enhanced in a dose-dependent manner by 10 and 100 ng/ml IGFBP-3. Final cell numbers in MCF-10A cultures treated with EGF and 100 ng/ml IGFBP-3 were increased by 40% relative to control (p Ͻ 0.001).
IGFBP-3 shares many structural and functional features with another IGF-binding protein, IGFBP-5 (26). Therefore, we investigated whether IGFBP-5 was also able to potentiate EGF action in MCF-10A cells, using purified recombinant human IGFBP-5 expressed in an adenoviral expression system. As shown in Fig. 2A, adenovirus-derived recombinant human IGF-BP-3 (adIGFBP-3) enhanced EGF-stimulated DNA synthesis similarly to plasma-derived IGFBP-3, with a significant effect at 1 ng/ml, and maximal enhancement at 10 ng/ml. In the absence of EGF, adIGFBP-5 had no significant effect on DNA synthesis in MCF-10A cells (data not shown). By contrast with adIGFBP-3, adIGFBP-5 did not potentiate EGF-stimulated DNA synthesis, and in fact was slightly inhibitory at low concentrations in the presence of EGF (Fig. 2B).
We then investigated whether a recombinant human IGF-BP-3 mutant that exhibits decreased cell binding and nuclear import (27,28) retains EGF potentiating activity. This mutant, IGFBP-3(mut), has 5 residues in the basic domain in the Cterminal region of IGFBP-3 substituted with the analogous sequence of IGFBP-1, i.e. 228 KGRKR 3 MDGEA. Exogenous IGFBP-3(mut) showed significantly decreased cell association compared with wild type IGFBP-3 in MCF-10A cells (Fig. 2C), similar to that shown previously for endogenous mutant protein in Chinese hamster ovary cells (27). However, in MCF-10A cells some binding was evident at high concentrations of IGF-BP-3(mut), with a significant increase in cell-associated IGF-BP-3 detected with 1000 ng/ml IGFBP-3(mut), comparable to the binding of 100 ng/ml wild type IGFBP-3 (Fig. 2C). In the presence of EGF, IGFBP-3(mut) enhanced DNA synthesis to a degree similar to that for adIGFBP-3 (Fig. 2D), suggesting that the potentiating effect of IGFBP-3 occurs in the absence of its nuclear localization and under conditions where its cell association is markedly reduced. The ability of IGFBP-3 to enhance the effects of other growth factors in MCF-10A cells was then investigated. TGF-␣, which binds and activates the EGF receptor (also known as ErbB1), stimulated DNA synthesis in MCF-10A cells (Fig. 3A). The stimulation caused by 1 ng/ml TGF-␣ (ϳ40%) was increased to ϳ120% in the presence 100 ng/ml IGFBP-3 (Fig. 3B). As observed for EGF, the level of potentiation by IGFBP-3 decreased with increasing concentrations of TGF-␣, so that with 10 ng/ml TGF-␣, DNA synthesis was increased only an additional 20% by IGFBP-3 (Fig. 3C).
To examine whether IGFBP-3 enhanced the effects of hormones that act through other members of the EGF receptor/ ErbB family, we tested its effects on DNA synthesis stimulated by heregulin, which binds the EGF receptor family members ErbB3 and ErbB4. In the absence of IGFBP-3, heregulin stimulated DNA synthesis in MCF-10A cells (Fig. 3D), with the maximum dose tested, 25 ng/ml, inducing a 12-fold increase in DNA synthesis. The stimulatory effects of low (0.3 ng/ml, Fig.  3E) or high (10 ng/ml, Fig. 3F) heregulin were not enhanced significantly by IGFBP-3. IGFBP-3 had no effect on DNA synthesis stimulated by platelet-derived growth factor or long-[Arg 3 ]IGF-I, an IGF analog with affinity for IGF receptors but not IGFBPs (data not shown).
To determine whether co-incubation of IGFBP-3 and EGF was necessary for potentiation, MCF-10A cells were preincubated with IGFBP-3 for 24 h prior to its removal, then EGF was added for an additional 24 h. As shown in Fig. 4, pre-exposure to IGFBP-3 resulted in a similar degree of enhancement of EGF-stimulated DNA synthesis to that seen when EGF and IGFBP-3 were added together for 24 h, suggesting that preincubation with IGFBP-3 sensitizes MCF-10A cells to EGF.
We then investigated the effect of IGFBP-3 preincubation on subsequent binding of EGF to cell monolayers. Cells were preincubated with or without IGFBP-3 for 16 h, and then binding of 125 I-EGF to cells over 2 h at 4°C was determined. As shown in Fig. 5A, the amount of 125 I-EGF bound to MCF-10A cells was not changed by pre-exposing the cells to IGFBP-3, and displacement curves generated by incubating cells with tracer in the presence of unlabeled EGF were superimposable in IGFBP-3pretreated and untreated cells. These data suggest that exposure to IGFBP-3 does not alter EGF receptor affinity or number.
The effects of IGFBP-3 on EGFR activation were then examined by assessing Tyr-1068-phosphorylated EGFR in cells pretreated with IGFBP-3 and then exposed to EGF for 5, 15, or 60 min. In the absence of IGFBP-3 pretreatment, EGF-stimulated phosphorylation of EGFR was apparent within 5 min (Fig. 5B); there was no effect of IGFBP-3 on the degree of receptor phosphorylation at this time point. However in cells pretreated for 24 h with either 10 or 100 ng/ml IGFBP-3, EGFR phosphorylation was increased relative to non-preincubated cells within 15 min of exposure to EGF (Fig. 5B). The enhancing effect of IGFBP-3 on receptor phosphorylation was transitory, being lost within 60 min after addition of EGF. Analysis of data from two similar experiments confirmed that pre-exposure to IGF-BP-3 significantly enhanced EGF-stimulated phosphorylation of the EGFR at Tyr-1068 (Fig. 5C). IGFBP-3 tested over similar doses, but without subsequent exposure to EGF, did not stimulate phospho-EGFR (data not shown).
Next, we examined whether IGFBP-3 modulates EGF activation of intracellular signaling, targeting particularly some pathways downstream of the EGFR. MCF-10A cells were preincubated with IGFBP-3 for 24 h and then exposed to EGF for 8 min. Cell lysates were analyzed by Western blot using phosphospecific antibodies directed against p44/42 MAPK, Akt, and p38 MAPK. In the absence of EGF, there was no detectable phosphorylation of p44/42 or Akt, whereas a low level of phosphorylation of p38 MAPK was apparent (Fig. 6A). Phosphorylation of each of these signaling intermediates was stimulated within 8 min of treatment with 1 ng/ml EGF. Pre-incubation with IGFBP-3 had no affect on phosphorylation of Akt in response to EGF, but EGF-induced phosphorylation of p44/42 MAPK and p38 MAPK was enhanced by pre-incubation with IGFBP-3 with a maximal effect apparent between 8 and 15 min of EGF stimulation. Densitometric analysis of these data (Fig.  6B) indicated that IGFBP-3 enhanced EGF-stimulated phosphorylation of p44/42 MAPK and p38 MAPK by ϳ2-fold and 30%, respectively. Analysis of lysates from cells stimulated for 1, 4, 8, 12, 18, and 24 h with EGF indicated that, although IGFBP-3 preincubation increased the magnitude of p44/42 and p38 MAP kinase phosphorylation stimulated by EGF at early time points, this was not sustained. Within 1 h of EGF treatment, phosphorylation of p44/42 and p38 MAP kinases had returned to basal levels, regardless of whether there had been pre-incubation with IGFBP-3 (data not shown). A second smaller peak of phosphorylation of p38 MAP kinase, but not p44/42 MAP kinase, was apparent 18 h after addition of EGF, but this was not enhanced in IGFBP-3-preincubated cells.
The effect of IGFBP-3 preincubation on EGF-stimulated p44/42 and p38 MAP kinase activity was then assessed using in vitro phosphorylation of substrates for these enzymes, Elk-1 and Atf-2, as markers of kinase activity. As shown in Fig. 7, lysates prepared from MCF-10A cells treated for 10 min with EGF showed increased activity of both p44/42 and p38 MAP kinases. The activity of both kinases was enhanced further in lysates from cells incubated with 10 or 100 ng/ml IGFBP-3 prior to exposure to EGF, indicating that IGFBP-3 potentiates EGF-stimulated p44/42 and p38 MAP kinase activity.
To determine the involvement of these pathways in the growth potentiating effects of IGFBP-3 on EGF action, we examined whether attenuation of the p44/42 MAPK or p38 MAPK signaling pathways altered the ability of IGFBP-3 to enhance the effects of EGF. MCF-10A cells were co-incubated for 24 h with IGFBP-3 and EGF in the presence of inhibitors of MEK upstream of p44/42 MAP kinase (PD98059), Akt (LY294002), or p38 MAPK (SB203580). DNA synthesis was determined after 24 h. The dose of inhibitors used in these experiments was intentionally submaximal, because concentrations high enough to block signaling fully through these pathways were cytotoxic over 24 h. As shown in Table I, the doses used were sufficient to abolish the stimulatory effect of EGF on DNA synthesis, and Western blot analysis confirmed that these concentrations were sufficient to reduce the IGFBP-3-induced enhancement of EGF-stimulated phosphorylation of p44/42 and p38 MAPK (data not shown). As shown in Fig. 8 Panels B and C, MCF-10A cells were incubated with or without 10 or 100 ng/ml IGFBP-3 for 24 h, washed, and exposed to EGF (1 ng/ml) for 5, 15, or 60 min, as indicated. Cell lysates were analyzed by Western blotting for EGFR phosphorylation at Tyr-1068 using a phosphospecific antibody as described under "Experimental Procedures." Blots were analyzed using NIH Image Software, and pooled data from two experiments are shown in panel C, where increased band intensity relative to stimulation with EGF in the absence of IGFBP-3 preincubation is shown as mean Ϯ S.E. *, p Ͻ 0.05; **, p Ͻ 0.001 compared with EGF.  1 and 2) or presence (lane 3) of 10 ng/ml IGFBP-3 for 24 h. Media were removed and replaced by serum-free medium (lane 1) or medium containing 1 ng/ml EGF (lanes 2 and 3) for 8 min. Cell lysates were prepared as described under "Experimental Procedures" and then subjected to 12% SDS-PAGE and Western blotting for the indicated proteins using phosphospecific antibodies, followed by ECL detection. Panel B shows densitometric quantitation of the blots, where the EGF and EGF ϩ IGFBP-3 bands have been expressed as a percentage increase above control (no EGF or IGFBP) for each protein.
DNA synthesis in the presence of EGF when p44/42 MAPK activation was blocked (p Ͻ 0.05). Similarly, inhibition of the p38 MAPK pathway using SB203580 (10 M) abolished the potentiating effect of IGFBP-3 on DNA synthesis in MCF-10A cells (Fig. 8B). Consistent with its lack of effect on Akt phosphorylation shown in Fig. 6, IGFBP-3 was still able to potentiate DNA synthesis stimulated by EGF when the PI 3-kinase pathway was blocked using LY294002 (10 M, Fig. 8C). Collectively, these data suggest that IGFBP-3 potentiates EGF signaling in MCF-10A cells by enhancing its activation of p44/42 and p38 MAPK signaling pathways.

DISCUSSION
Numerous studies have shown that IGFBP-3 exerts growthinhibitory effects in a variety of cell types, either through blockade of IGF-stimulated mitogenesis and cell survival (1) or via antiproliferative activity unrelated to its ability to bind IGFs (7,9,10,20,29); however, growth-stimulatory effects of IGFBP-3 are less well documented. Early studies in fibroblasts (12,30) and MCF-7 breast cancer cells (11) suggested that ligand interaction was involved in the stimulatory effect of IGFBP-3 on IGF activity, with its mechanism of action thought to involve either prevention of ligand-induced down-regulation of the IGFR1 (11) or processing of the IGFBP-3 to forms with altered affinity for IGFs (30). A study in airway smooth muscle cells also indicated a growth-stimulatory role for IGFBP-3 in the presence of fetal calf serum (13). In a bovine mammary epithelial cell line transfected to express IGFBP-3, DNA synthesis was increased in response to IGF-I, insulin, and long-[Arg 3 ]IGF-I (31), implying that IGF-IGFBP-3 interaction was not required for a potentiating effect. However, growth-stimulatory interactions between IGFBP-3 and systems other than the IGF axis have not been reported previously.
In this study we have shown that, in MCF-10A breast epithelial cells, IGFBP-3 enhances the potent growth-promoting effects of members of the EGF system, which has been implicated in the development and progression of malignant disease, by priming cells to respond to EGF and TGF-␣. Although the underlying mechanism by which IGFBP-3 potentiates EGF action in these cells remains unclear, we made the novel and significant observation that in cells pre-exposed to IGFBP-3, EGF-stimulated phosphorylation of the EGFR at Tyr-1068 was increased. Phosphorylation of this residue is crucial to linking EGFR activation with the Ras-MAPK signaling pathway via Grb2 and Sos (32) and, although we did not explicitly determine whether EGFR kinase activity was increased, our observation of increased phosphorylation and activation of p44/42 and p38 MAPK signaling intermediates downstream of Ras in IGFBP-3-primed cells is consistent with increased Ras activation occurring consequent to increased Tyr-1068 phosphorylation of the EGFR.
At present it is unclear how IGFBP-3 might be bringing about an increase in EGFR phosphorylation. Preincubation with IGFBP-3 did not markedly affect steady-state binding of EGF to cells, suggesting that an overall increase in EGF⅐EGFR interaction is not involved in the potentiating effect of IGF-BP-3, although more dynamic effects on receptor availability arising from changes in internalization or heterodimerization (33) were not investigated. It is possible that IGFBP-3 is modulating the expression or activity of molecules involved in regulating receptor interaction with binding partners such as the Ras exchange factor Sos1 (34), or EGFR dephosphorylation, such as phosphotyrosine phosphatases. IGFBP-3 activation of a phosphotyrosine phosphatase that dephosphorylates the IGFR1 was recently proposed (35), although the IGF independence of IGFBP-3 action in this model has not been clearly demonstrated. Increased phosphorylation of the EGFR, as in the present study, would suggest decreased activity of phosphotyrosine phosphatases rather than increased, and the possibility that this is involved in IGFBP-3-enhancement of receptor phosphorylation is currently under investigation.
In MCF-10A cells, IGFBP-3 enhanced activation of the p44/42 and p38 MAP kinase signaling pathways, but not the PI 3-kinase pathway, in response to EGF stimulation. Involvement of the p44/42 and p38 MAP kinase pathways in progestin priming of breast cancer cells to respond to EGF has also been demonstrated in T47D breast cancer cells (36), associated with induction and activation of multiple proteins, including EGF receptor family members and Stat (signal transducers and activators of transcription) proteins (36,37). In experiments not presented in this report, there was no clear change in expression of total Stat1, Stat5a or Stat5b protein in IGFBP-3-preincubated cells, suggesting that progestin and IGFBP-3 priming of cells to respond to EGF may involve distinct pathways and intermediates.
Although the structural elements of IGFBP-3 required for its interaction with EGF signaling have not been elucidated, our experiments suggest that its interaction with major cell-surface moieties is not required. The 228 KGRKR motif of IGFBP-3 has been implicated in its cell association (27), and, indeed, we found that the mutant with this region replaced with the corresponding region of IGFBP-1, IGFBP-3(mut), exhibited greatly reduced binding to MCF-10A cells compared with plasma-derived IGFBP-3. Despite this, mutant IGFBP-3 enhanced EGF activity with similar potency to wild type IGFBP-3, implying that IGFBP-3-priming of cells to respond to EGF occurs in the absence of IGFBP-3 binding to a major cell-surface component, although not ruling out the possibility of interaction with a low abundance cell-surface protein. Although a clear relationship between cell association of IGFBP-3 and its biological activity has not been demonstrated, the existence of, and requirement for, an IGFBP-3 receptor capable of mediating a growth-inhibitory signal has been inferred from studies indicating a correlation between IGFBP-3 interaction with cell surfaces and effects on cellular growth (38,39). We have shown that IGFBP-3 can initiate inhibitory signaling in breast cancer cells through the TGF-␤ receptor/Smad pathway (40,41); however, the primary site of cell interaction of IGFBP-3 in this system was not identified. Indeed, a plasma membrane receptor with IGFBP-3 signal transduction capability, either for growth stimulation or growth inhibition, has not yet been identified in any cell system.
Nuclear localization of IGFBP-3 has been identified by a number of research groups (28,42,43); however, the function of IGFBP-3 in the nucleus remains unknown. It has been suggested that interaction between IGFBP-3 and the retinoid X receptor-␣ and regulation of its transcriptional activity in the nucleus is essential for the apoptotic effects of IGFBP-3 in prostate cancer cells (43). The findings of the present study indicate that intranuclear interactions between IGFBP-3 and transcriptional regulators are not required for its growth-stimulatory interaction with EGF signaling, because the C-terminal basic motif of IGFBP-3 that is required for its nuclear localization (28), 228 KGRKR, is not necessary for enhancement of EGF signaling by IGFBP-3. Consistent with this, IGFBP-5, which has a similar basic motif and shares a common pathway of nuclear transport with IGFBP-3 (28), did not enhance EGF action, and was in fact slightly inhibitory in the presence of EGF.
Other studies from our laboratory have also shown that neither significant cell binding nor nuclear localization of IGF-BP-3 is required for its bioactivity. Exogenous IGFBP-3(mut) induced phosphorylation of TGF-␤ signaling intermediates similar to plasma-derived IGFBP-3 in a T47D cell line (41), and IGFBP-3(mut) overexpression inhibited growth and induced apoptosis in T47D breast cancer cells (44). In the latter case, endogenous mutant IGFBP-3 may have direct intracellular effects, thereby overcoming the need for secretion and re-uptake, which might require cell-surface binding. Bioactivity of exogenous mutant IGFBP-3 as in the present study and that of Fanayan et al. (41) is somewhat more difficult to explain, although, as we demonstrated, cell binding of IGFBP-3 was not completely abolished by substitution of the basic residues, and residual binding of IGFBP-3(mut), perhaps to low abundance cell-surface moieties, may be sufficient to elicit a response. Notably, N-terminal fragments of IGFBP-3 that lack this domain are biologically active in a number of systems (6,45,46), frequently with increased potency compared with intact wild type IGFBP-3. Understanding the mode of action of IGFBP-3, as either a growth inhibitor or stimulator, and the structural determinants of such action remain important goals.
The concentrations of plasma-derived IGFBP-3 required for stimulation of EGF activity in MCF-10A cells (10 -100 ng/ml) are similar to those required for its inhibitory activity in this cell line in the absence of EGF (20), up to an order of magnitude lower than those used in the demonstration of IGFBP-3 bioactivity in some other cell systems (7,41), and similar to the levels expressed by many cell types (1). In view of the fact that MCF-10A cells secrete ϳ30 ng/ml IGFBP-3 under the conditions used in these experiments (20), it is somewhat surprising that they remain sensitive to exogenous IGFBP-3 at similar concentrations, and may reflect differences between the cellderived IGFBP-3 and exogenous IGFBP-3 used in this study. It is noteworthy that, when compared within the same experiment, recombinant IGFBP-3 appeared to show increased potency compared with plasma-derived IGFBP-3, with significant enhancement of EGF activity with 1 ng/ml adenoviral IGF-BP-3, compared with 10 ng/ml plasma-derived IGFBP-3. Adenoviral IGFBP-3 is essentially unphosphorylated, whereas plasma-derived IGFBP-3 has ϳ1 mol/mol serine phosphorylation (data not shown), raising the possibility that differences in the potency of IGFBP-3 from alternative sources may be caused by its differential phosphorylation.
Our observation of activation of the p44/42 and p38 MAPK signaling pathways, but not Akt/PKB, is in contrast with a study that showed, in fibroblasts, the PI 3-kinase pathway appears to be involved in IGFBP-3 potentiation of IGF action, with IGFBP-3 increasing the sensitivity of Akt/PKB to phosphorylation by IGF-I (47). This implies that IGFBP-3 may impact on numerous growth factor-regulated pathways, perhaps in a cell-or growth factor-specific manner, to enhance cell proliferation.
Although attenuation of either the p44/42 or p38 MAP kinase pathways was sufficient to block the potentiating effect of IGF-BP-3, inhibition of the p44/42 MAP kinase pathway alone had the additional effect of reinstating the growth-inhibitory activity of IGFBP-3 in MCF-10A cells. We have shown previously a similar reversal of refractoriness to growth inhibition by IGF-BP-3 by blocking p44/42 MAP kinase activation in MCF-10A ras cells, in which chronic activation of this pathway occurs as a result of transfection with oncogenic ras (20). The results of the present study confirm that activation of Ras-MAPK signaling ablates the growth-inhibitory activity of IGFBP-3, and, importantly, extend these findings to show that interactions between IGFBP-3 and Ras-dependent signaling pathways may result in enhanced growth-stimulatory signaling in breast cells. Inhibition of this pathway has the potential both to block stimulatory activity arising from the interaction of IGFBP-3 with other growth factor systems and to restore its inhibitory activity. Clearly, the identification of the factors involved in the potentiation of IGFBP-3 of EGF action will be the next important step in delineating its growth-stimulatory role in breast cancer cells, and explaining its association with highly malignant cancers with poor prognosis.