Regulation of Vascular Smooth Muscle Cell Responses to Insulin-like Growth Factor (IGF)-I by Local IGF-binding Proteins*

Insulin-like growth factor (IGF)-I is a pleiotropic hormone that regulates vascular smooth muscle cell (VSMC) migration, proliferation, apoptosis, and differentiation. These actions are mediated by the IGF-I receptor. How activation of the same receptor by the same ligand leads to these diverse cellular responses is not well understood. Here we describe a novel mechanism specifying VSMC responses to IGF-I stimulation, distinctive for the pivotal roles of local IGF-binding proteins (IGFBPs). The role of local IGFBPs was indicated by comparing the activities of IGF-I and des-1–3-IGF-I, an IGF-I analog with reduced binding affinity to IGFBPs. Compared with IGF-I, des-1–3-IGF-I was more potent in stimulating DNA synthesis but much less potent in inducing directed migration of VSMCs. When the effects of individual IGFBPs were tested, IGFBP-2 and IGFBP-4 were found to inhibit IGF-I-stimulated DNA synthesis and migration. IGFBP-5 had an inhibitory effect on IGF-I-stimulated DNA synthesis, but it strongly potentiated IGF-I-induced VSMC migration. By using a non-IGF-binding IGFBP-5 mutant and an IGF-I-neutralizing antibody, it was demonstrated that IGFBP-5 also stimulates VSMC migration in an IGF-independent manner. This effect of IGFBP-5 was inhibited by soluble heparin and by treating cells with heparinase. Mutation of the heparin-binding motif of IGFBP-5 reduced its migration promoting activity. These findings suggest that local IGFBPs are important determinants of cellular responses to IGF-I stimulation, and a key player in this paradigm is IGFBP-5. IGFBP-5 not only modulates IGF-I actions, but it also stimulates cell migration by interacting with cell-surface heparan sulfate proteoglycans.

It is now understood that most, if not all, IGFs in the extracellular environment are bound to specific, high affinity IGFbinding proteins (IGFBPs). IGFBPs are a family of secreted proteins that specifically bind IGF-I and IGF-II with affinities that are equal to or greater than those of the IGF-IR. Six distinct IGFBPs, designated as IGFBP-1-6, have been isolated and characterized from a variety of vertebrate species, ranging from humans to teleost fish (14 -17). The circulating IGF⅐IGFBP complexes prolong the half-lives of IGFs and inhibit the potential hypoglycemic effects of high concentrations of circulating IGFs. Locally expressed IGFBPs are thought to provide a means of localizing IGFs to specific cells and regulating their local availability and biological activity (14,15). Previous studies in porcine, bovine, rat, and mouse indicate that mammalian VSMCs secrete IGFBP-2, IGFBP-3, IGFBP-4, and IGFBP-5 (6, 9, 18 -23). These IGFBPs, when added exogenously to cultured VSMCs in combination with IGF-I, can either inhibit or potentiate IGF-I-induced DNA synthesis and migration (6, 19, 21, 24 -26). Recent studies have also revealed that the abundance of these local IGFBPs vary significantly under different pathophysiological conditions and different cellular contexts. For example, higher levels of IGFBP-5 were found in atherosclerotic plaques compared with tissue from the normal vessel wall (27). In vitro, synthetic VSMCs maintained on fibronectin synthesize and release more IGFBP-5 compared with differentiated contractile cells grown on laminin and type IV collagen (27). Likewise, the biosynthesis and secretion of IGFBP-5 by porcine VSMCs decreased by 18-fold when cell density increases from 30 to 100% confluence, whereas the concentrations of IGFBP-4 increased up to 15-fold (6).
Because different IGFBPs have different biological effects (14,15), and because the availability of IGFBPs is differentially regulated by the cellular context, we hypothesized that locally produced IGFBPs may be important in determining specific responses to the IGF signal. The goals of this study were to determine whether IGFBPs regulate the chemotactic and mitogenic responses of VSMCs to IGF stimulation, and to elucidate the individual role(s) and underlying mechanisms of local IGFBPs in regulating VSMC migration and proliferation.
Cell Culture and Transfection-Porcine VSMCs were isolated from thoracic aorta of 3-week-old piglets (22). The cells were grown in 10-cm dishes (Falcon, BD Biosciences) in DMEM supplemented with glutamine (4 mM), penicillin (100 units/ml), streptomycin (100 g/ml), and 10% FBS. The medium was changed every 4th day until the cells became confluent. Before stimulation experiments, medium was changed to serum-free DMEM (SFM) for 20 -24 h. This SFM was then replaced with fresh SFM plus the indicated growth factors for various times. Cultured porcine VSMCs were transfected with the pMEP4/ IGFBP-5 construct or empty pMEP4 following a procedure published previously (30). The transfected cells were trypsinized, plated on three 10-cm plates, and grown in DMEM plus 10% FBS with 120 g/ml hygromycin B. Medium was analyzed for secretion of IGFBP-5 as described below.
Western Immunoblot, Immunoprecipitation, and Ligand Blot Analysis-In order to identify the forms of IGFBPs secreted by VSMCs, conditioned media were concentrated 20 times through a Centricon-10 microconcentrator (Amicon) by centrifugation (22). The proteins were separated by 12.5% SDS-PAGE under nonreducing conditions, and they were transferred to filters (Immobilon P, 0.45-m pore size, Millipore Corp., Bedford, MA) and subjected to immunoblot and ligand blot analysis following standard procedures (6).
Biological Assays-The effects of IGF-I and/or IGFBPs on VSMC DNA synthesis and growth was examined by direct cell counting, [ 3 H]thymidine incorporation, and 5-bromo-2-deoxyuridine (BrdUrd) labeling assays as reported previously (8). The effects of IGF-I, IGFBPs, and selected soluble glycosaminoglycans (GAGs) on VSMC migration were investigated using the trans-well migration assay and gold particle cell motility assay. The trans-well migration assay measures the directional movements of VSMCs across a porous membrane, as described previously (8). The gold particle cell motility assay measures cell motility in the absence of a concentration gradient (31,32). Briefly, a uniform carpet of gold particles was prepared on BSA-coated glass coverslips placed in 6-well plates. Colloidal gold-coated coverslips were rinsed and then seeded with 1 ϫ 10 4 VSMCs in 2 ml of DMEM or DMEM ϩ 100 ng/ml IGF-I. After 6 h, cells were fixed with 3.5% glutaraldehyde, air-dried, and mounted on glass slides. The gold particle-free clear zone surrounding the cells was photographed, and VSMC motility was represented as the average area of at least 60 pericellular clear zones, quantitated using Scion Image software (Frederick, MD).
Removal of Cell-surface GAGs by GAG Lyase Treatment-Enzymatic treatment with GAG lyases was performed following published methods (33). Briefly, VSMCs cells were incubated with the GAG lyases in 1ϫ phosphate-buffered saline containing 0.1% BSA, 0.2% gelatin, and 0.1% glucose for 40 min at 37°C in a CO 2 incubator. Cells were then washed 3 times with 1ϫ phosphate-buffered saline, trypsinized, washed again with SFM, and used for migration assays.
Statistical Analysis-Values are means Ϯ S.E. Differences among groups were analyzed by one-way analysis of variance followed by Fisher's protected least significance difference test using StatView (Abacus Concept, Berkeley, CA). Significance was accepted at p Ͻ 0.05.

The Local Repertoire of IGFBPs Is Critical in Determining
Whether VSMCs Migrate or Proliferate in Response to IGF-I Stimulation-To test the hypothesis that the local repertoire of IGFBPs may affect the cellular responses to IGF-I stimulation, we examined the effects of IGF-I and des-1-3-IGF-I on VSMC growth and migration. Des-1-3-IGF-I is an IGF-I analog that binds to IGF-IR with normal affinity but has greatly reduced affinity for IGFBPs (34). As shown in Fig. 1A, both IGF-I and des-1-3-IGF-I stimulated [ 3 H]thymidine incorporation in a dosedependent fashion. Compared with IGF-I, however, des-1-3-IGF-I was more potent at concentrations ranging from 1 to 20 ng/ml. At 10 ng/ml, des-1-3-IGF-I and IGF-I induced a 362.6 Ϯ 10 and a 221.3 Ϯ 15.7% increase over the control, respectively. The difference between these two groups was statistically significant (p Ͻ 0.05, n ϭ 4). Similarly, 20 ng/ml of des-1-3-IGF-I caused a 444.2 Ϯ 37.7% increase which is significantly higher than the 304.7 Ϯ 31.4% increase induced by IGF-I at the same concentration (p Ͻ 0.05, n ϭ 4). When the peptide concentration was increased to 50 ng/ml and higher, the difference between the two peptides disappeared, indicating saturation of endogenous IGFBPs by excess IGF-I added.
When the chemotactic activities of IGF-I and des-1-3-IGF-I were examined, an opposite relationship was observed, i.e. IGF-I was a much stronger chemoattractant (Fig. 1B). At 20 ng/ml, IGF-I caused 833 Ϯ 182% increase over the control group (p Ͻ 0.01, n ϭ 3), whereas des-1-3-IGF-I only caused a 450 Ϯ 113% increase over the control group (p Ͻ 0.05, n ϭ 3). The difference between the IGF-I and des-1-3-IGF-I groups at this concentration was also statistically significant (p Ͻ 0.05). At 50 ng/ml, des-1-3-IGF-I induced migration was 441.7 Ϯ 71.2%. This value was significantly lower than that of the IGF-I group (1183 Ϯ 176.4%, p Ͻ 0.01). Similarly, IGF-I at a higher concentration (100 ng/ml) caused a significantly greater increase in VSMC migration than that of des-1-3-IGF-I (1216.7 Ϯ 154.3% versus 208.3 Ϯ 109.3%, p Ͻ 0.05). These data suggest that the endogenous IGFBPs strongly promote VSMC migration toward IGF-I, but they inhibit IGF-I-induced DNA synthesis. The distinct dose-response curves of IGF-I and des-1-3-IGF-I on cell migration are also suggestive of different mechanisms underlying their actions.

Different IGFBPs Exhibit Different Biological Effects in Regulating Mitogenic and Chemotactic Actions of IGF-I-Cultured
porcine VSMCs synthesize and secrete IGFBP-2, IGFBP-4, and IGFBP-5 (22). To determine the individual effects of these IGFBPs, we tested the ability of each IGFBP to modulate the mitogenic and chemotactic actions of IGF-I. IGFBP-2 and IGFBP-4, when added together with IGF-I at an equal molar concentration, significantly inhibited IGF-I-stimulated BrdUrd incorporation ( Fig. 2A). Neither IGFBP-2 nor IGFBP-4 alone caused any significant change. In migration assays, IGF-I (50 ng/ml) alone induced a 642 Ϯ 181% increase (p Ͻ 0.05) (Fig.  2B). The addition of IGFBP-2, at an equal molar concentration (200 ng/ml), reduced IGF-I-induced migration to 133 Ϯ 28% of the control; this was significantly less than that observed in the IGF-I alone group (p Ͻ 0.01). IGFBP-4 (150 ng/ml) had a similar, but more modest, inhibitory effect; it caused a 52% reduction in IGF-I-induced migration (p Ͻ 0.05). When either IGFBP-2 or IGFBP-4 was added in the absence of IGF-I, no significant change was observed (Fig. 2B).
When added with IGF-I, IGFBP-5 caused a statistically insignificant 21% increase in IGF-I-induced DNA synthesis ( Fig.  2A). In contrast, addition of IGFBP-5 together with IGF-I at the same concentrations resulted in a 1393 Ϯ 372% increase in directed migration toward IGF-I (Fig. 2B). This value was significantly higher than that of the IGF-I alone group (642 Ϯ 181%, p Ͻ 0.05). Intriguingly, addition of IGFBP-5 alone resulted in a statistically significant increase as compared with the control group (216 Ϯ 32%, p Ͻ 0.05). IGFBP-5 alone had no effect on DNA synthesis in these cells. To examine further the effect of IGFBP-5, cells were subjected to different concentrations of IGFBP-5 in the presence and absence of IGF-I (20 ng/ml). As shown in Fig. 2C, addition of IGFBP-5 stimulated VSMC migration in a dose-dependent fashion with or without IGF-I. These results suggest that IGFBP-2 and IGFBP-4 act as inhibitors of IGF-I actions, whereas IGFBP-5 potentiates IGFinduced VSMC migration. The data also suggest that IGFBP-5 may alter the VSMC motility by itself.

Expression of IGFBP-5 in VSMCs Increases the Basal and the IGF-I Induced Cell Migration but Decreases Cell Growth-
Earlier studies showed that porcine VSMCs secrete a protease that specifically degrades IGFBP-5, but this protease does not cleave IGFBP-2 or IGFBP-4 (22,35). The presence of this potent IGFBP-5 protease complicates the interpretation of the data obtained with the use of exogenous IGFBP-5. To define further the role of IGFBP-5, porcine VSMCs were transfected with an IGFBP-5 expression construct, and three independent groups of cells that stably expressed the transgene (ϩBP5) were obtained. The three ϩBP5 clones secreted considerably more IGFBP-5 (250, 345, and 180% compared with the wildtype control (WT), respectively). The empty expression vector (Mock)-transfected cells had low levels of IGFBP-5, comparable with those of WT. No significant variation in the IGFBP-2 levels was detected among ϩBP5, Mock, and WT cells (data not shown).
The growth curves of three independent clones of Mock cells were similar to that of WT (Fig. 3A). In contrast, all three ϩBP5 clones had significantly reduced growth rate when compared with WT and Mock throughout the period. The difference was highly significant at 288 h (p Ͻ 0.01). Next, the IGF-Iinduced DNA synthesis in these cells was examined by [ 3 H]thymidine incorporation assays. As shown in Fig. 3B, IGF-I stimulated DNA synthesis in a dose-dependent manner in all groups. However, the basal levels of [ 3 H]thymidine incorporation in the ϩBP5 clones were only 26 Ϯ 7% of those observed in WT and Mock cells, and the level remained significantly lower even in the presence of increasing concentrations of IGF-I. Together, these data indicate that overexpression of IGFBP-5 inhibits IGF-I-stimulated porcine VSMC growth. The effect of IGFBP-5 expression on cell migration is shown in Fig. 4. In the absence of IGF-I, Mock-transfected cells were similar to the WT group in the basal migration rate (4.8 Ϯ 0.2 versus 3.7 Ϯ 1.2 cells/well). IGF-I caused a similar degree of chemotactic response in the WT and Mock cells (Fig. 4A). In contrast, the ϩBP-5 cells showed significantly higher levels of migration at all the concentrations of IGF-I tested (p Ͻ 0.01). Even in the absence of IGF-I, these cells showed an 8-fold increase over the WT control (29.6 versus 3.7 cells/well, p Ͻ 0.01, n ϭ 3). Closer observation of the response curves revealed that IGFBP-5 overexpression caused a left shift in the doseresponse curve to IGF-I. The maximum response was detected at 5 ng/ml of IGF-I in the ϩBP5 cells, which is much lower than the maximal response concentrations seen in WT and Mock cells (50 ng/ml), suggesting that IGFBP-5 enhances the cellular responsiveness to the IGF-I stimulation. To examine further the effect of IGFBP-5, cells from ϩBP5, Mock, and WT groups were analyzed using the gold particle cell motility assay. As shown in Fig. 4B, WT and Mock cells showed a similar degree of motility. The ϩBP5 clones showed a mean increase of 264 Ϯ 52% compared with WT cells. To determine whether the effect of IGFBP-5 on basal cell motility was IGF-dependent, an IGF-I-neutralizing antibody (Sm1.2) was added. This antibody has been shown to neutralize IGF-I actions in porcine VSMCs (6). Addition of Sm1.2 did not cause any significant change in basal or IGFBP-5-induced motility. These results suggest that overexpression of IGFBP-5 in VSMCs not only potentiates IGF-Iinduced chemotaxis but also stimulates basal cell motility rate through a ligand-independent mechanism(s).

IGFBP-5 Stimulates VSMC Migration via an IGF-independ-
ent Mechanism-Further evidence for IGF-independent action of IGFBP-5 came from the check board analysis. IGF-I is a chemoattractant for VSMCs. It stimulated VSMC migration when added to the lower chamber (Fig. 5A). When the IGF-I concentration gradient was eliminated by adding the same concentration of IGF-I to both lower and upper chambers, it did not stimulate VSMCs migration. In contrast, addition of IGFBP-5 to both chambers resulted in an increase that was comparable with that in the lower chamber alone (Fig. 5A), suggesting that the effect of IGFBP-5 on VSMC migration is not chemotactic.
Because it is known that cultured porcine VSMCs synthesize and secrete IGF-I (6), it is possible that the exogenously added IGFBP-5 influences VSMC migration by interacting with the endogenous IGF-I. To clarify this issue, migration assays were performed in the presence and absence of the IGF-I-neutralizing antibody (Sm1.2). As shown in Fig. 5B, Sm1.2 completely abolished the chemotactic action of IGF-I (50 ng/ml) without affecting the level of basal migration. The addition of Sm1.2, however, did not change the higher basal migration rate of the IGFBP-5 group.
To ascertain that the action of IGFBP-5 is independent from its IGF binding, we next examined the activity of a non-IGF binding IGFBP-5 mutant. This mutant has a 1000-fold reduced affinity for IGF-I (28) and has been used to determine IGF-Iindependent effects of IGFBP-5 on cellular physiologic processes in cell types that secrete IGF-I (36). As shown in Fig. 5C pared with the BSA control). This value was even higher than that of the wild-type IGFBP-5 group (21.8 Ϯ 3.9 cells/well, a 376% increase over the BSA control) although the difference was not statistically significant. Taken together, these data indicate that IGFBP-5 stimulates VSMC migration through an IGF-I-independent mechanism.
IGFBP-5 Stimulates VSM Migration by Interacting with Cell-surface GAGs-Because IGFBP-5 binds to heparin and GAG-containing proteoglycans located on VSMC surface (29,37) and because a peptide derived from the heparin-binding motif of IGFBP-5 has been shown to stimulate mesangial cell migration (38), we postulated that IGFBP-5 may regulate VSMC motility by interacting with GAGs containing proteoglycans on the cell surface. If this were correct, then one or more of the four groups of GAGs found on membrane-associated proteoglycans should act as competitive inhibitors of IGFBP-5 actions. We therefore tested the impact of the 4 major GAGs found on membranes, i.e. heparin/heparan sulfate (HS), chondroitin A, chondroitin B, and chondroitin C. Of the four GAGs, heparin at the concentration of 100 g/ml resulted in a complete inhibition of IGFBP-5 activity (Fig. 6A). Chondroitin sulfate B, which shares the most structural similarity to heparin, exhibited partial inhibition (47%). In contrast, chondroitin sulfate A and chondroitin sulfate C at the same concentration had no inhibitory effect (Fig. 6A). To prove that cell-surface GAGs are indeed involved in the ligand-independent action of IGF-BP-5, different cell-surface GAGs were removed by digestion with specific lyases. As shown in Fig. 6B, treatment of cells with chondroitinase AC, which specifically digests chondroitin sulfates A and C, was ineffective in inhibiting IGFBP-5 action. In contrast, treatment with heparinase III (heparitinase), an enzyme mostly active on heparan sulfate and heparin (39), abolished the cellular response to IGFBP-5 (Fig. 6B). To provide more definitive evidence that the binding of IGFBP-5 to cell-surface HS proteoglycans is required for its IGF-independent action on VSMC migration, we analyzed the effect of an IGFBP-5 mutant, K202A/K206A/K207A. This mutant has reduced heparin binding but has normal binding affinity to IGF-I in the absence of heparin (29). As shown in Fig. 6C, the K202A/ K206A/K207A mutant caused a 180 Ϯ 32% increase, which is significantly lower (p Ͻ 0.05) than that of the wild-type IGFBP-5 (380 Ϯ 68%). These data indicate that the IGF-independent action of IGFBP-5 requires its interaction with cellsurface HS proteoglycans. DISCUSSION In this study, we have unraveled a novel mechanism regulating specific types of biological responses of VSMCs to IGF-I. Our data suggest that the local repertoire of IGFBPs is critical in determining whether VSMCs migrate or proliferate in response to IGF-I stimulation. To our knowledge, this is the first demonstration that locally produced IGFBPs play a role in specifying cellular responses to IGF-I. Among the three locally produced IGFBPs in porcine VSMCs, a key player in this new paradigm, is IGFBP-5. This IGFBP can not only inhibit the cell growth-promoting action of IGF-I in VSMC in similar fashion to IGFBP-2 and IGFBP-4, but it also has the unique ability to stimulate VSMC motility through an IGF-independent mechanism. We further show that the IGF-independent action of IGFBP-5 requires its interaction with cell-surface HS proteoglycans.
The critical role of IGFBPs in specifying differential growth and chemotactic responses of VSMCs to the IGF signal is supported by the different activities of IGF-I and des-1-3-IGF-I. Compared with IGF-I, des-1-3-IGF-I was more potent in stimulating DNA synthesis but was much less potent than IGF-I in stimulating VSMC migration. Previous studies have shown that des-1-3-IGF-I has greatly reduced affinity for IGFBPs but binds to IGF-IR with normal affinity (34) and activates mitogen-activated protein kinase and phosphatidylinositol 3-kinase cascades (13). These data indicate that the endogenous IGFBPs produced by VSMCs strongly promote VSMC migration toward IGF-I, but they in total have an inhibitory effect on IGF-I-induced DNA synthesis. The distinct dose-response curves of IGF-I and des-1-3-IGF-I on cell migration are also suggestive of different signaling mechanisms underlying their actions.
When the individual roles of the three local IGFBPs were examined more closely, IGFBP-2 and IGFBP-4 were found to inhibit IGF-I-stimulated DNA synthesis as well as directed cell migration. These observations are in agreement with previous studies (6,19,24,25). Because the IGF-IR mediates the biological actions of the IGFs and because soluble IGFBPs bind to IGFs with higher affinity than that of the IGF-IR, the inhibitory effects of IGFBP-2 and IGFBP-4 can be rationally attributed to their competition for the ligand with cell-surface receptors. IGFBP-5, on the other hand, had limited effect in modulating IGF-I-stimulated BrdUrd incorporation when added exogenously. The lack of effect by IGFBP-5 is likely due to the presence of a potent IGFBP-5 protease in these cells and the lengthy incubation time of this particular assay. In a previous study, we found that the endogenously secreted IGFBP-5 by these cells was completely degraded into non-IGF-binding fragments under serum-free conditions, whereas IGFBP-2 and IGFBP-4 were present as intact proteins under the same condition (22). This is also consistent with previous observations that g concentrations of exogenous IGFBP-5 could have inhibitory or potentiating effect on IGF-induced thymidine incorporation, but this effect was very modest and dependent on the incubation time. For example, when co-incubated with IGF-I for a relatively short time, IGFBP-5 inhibited IGF-I-induced DNA synthesis; when co-incubated for longer periods, IGFBP-5 potentiated IGF-I-induced DNA synthesis (6). In this study, we have clarified the role of IGFBP-5 in VSMC growth using VSMCs stably transfected with IGFBP-5. The ϩBP5 cells grew slower than Mock and WT cells. In addition, they had lower basal, and IGF-I-induced, DNA synthesis rates. These data indicate that endogenously overexpressed IGFBP-5 has an inhibitory effect on IGF-I-induced cell growth. When the effect of IGFBP-5 in regulating IGF-I-stimulated directed migration was examined, a different picture emerged. In this case, IG-FBP-5 caused a further increase in migration when co-incubated with IGF-I. The maximal response concentration of IGF-I shifted from 50 to 5 ng/ml in the ϩBP5 cells. This suggests that IGFBP-5 inhibits the mitogenic effect of IGF-I but enhances its chemotactic effect. Our finding is in agreement with recent observations made in mouse osteosarcoma cells and skeletal muscle cells, where IGFBP-5 inhibited the mitogenic effects of IGF-I, but enhanced differentiation of osteosarcoma cells and stimulated the myogenic effect of IGF-I in skeletal muscle cells (40,41).
A novel finding made in this study is that in addition to its ability to modulate IGF actions, IGFBP-5 also regulates VSMC motility through a ligand-independent mechanism. This conclusion is supported by several lines of evidence. First, treating porcine VSMCs with exogenous IGFBP-5 alone resulted in a significant increase in the number of migrating cells. This effect of IGFBP-5 is not due to its possible interaction with the endogenously secreted IGF-I, because the IGF-I-neutralizing antibody did not affect its ability to stimulate VSMC migration, but potently abolished the IGF-I-induced migration at the same concentration. Likewise, the IGFBP-5-transfected cells had an 8-fold increase in basal migration in the absence of exogenous IGF-I. This was not caused by the transfection and/or clone selection, because all three ϩBP5 cones showed similar elevation in motility. There was no such change in basal migration rate in mock-transfected cells. In addition, expression of IGFBP-5 in VSMCs resulted in a significant increase in cell motility (2.6-fold) measured by the gold particle motility assay. This increase could not be suppressed by adding the IGF-I-neutralizing antibody. Furthermore, the effect of IGFBP-5 on VSMC migration is not chemotactic because adding IGFBP-5 to both upper and lower chambers caused a similar level of increase in migration compared with adding IGFBP-5 to lower chamber alone. This is different from the IGF-I action, which is a chemoattractant for VSMCs (4,8). Finally, the non-IGF-binding mutant IGFBP-5 is equally or even more potent in stimulating VSMC migration. Therefore, IGFBP-5 has a direct effect on VSMC motility. This is agreement with an increasing number of reports in the literature indicating that IGFBP-5 may exert intrinsic biological actions independent from IGFs. For instance, IGFBP-5 has been shown to stimulate bone growth through an IGF-independent mechanism in vitro and in vivo (42,43).
The results of this study indicate an involvement of cellsurface HS proteoglycans in the migration-promoting effect of IGFBP-5. Early studies have shown that soluble GAGs interfere with binding of IGFBP-3 and IGFBP-5 to cell surfaces and extracellular matrix (44 -46). A peptide containing amino acids 201-218 of IGFBP-5 competitively inhibited IGFBP-5 binding to proteoglycans (25,37,46). Further delineation of specific amino acids indicated that the heparin-binding motif, Lys 206 -Arg 207 -Lys 208 -Gln 209 -Cys 210 -Lys 211 -(BBBXXB), contained the primary binding site (29). The functional significance of the IGFBP-5 and cell-surface HS proteoglycan interaction and the involvement of this GAG-binding motif in the ligand-independent action of IGFBP-5 were not known. In this study, we show that heparin, but not chondroitin sulfate A-C, inhibited IGF-BP-5-stimulated VSMC migration. Furthermore, enzymatic re-moval of HS-containing, but not chondroitin-containing proteoglycans, abolishes the migratory responses to IGFBP-5. Another line of evidence came from our analysis of the K202A/ K206A/K207A IGFBP-5 mutant. This mutant has reduced GAG binding but has normal binding affinity to IGF-I (29). It has significantly weaker activity compared with the wild-type IGFBP-5 in stimulating VMCS migration. Together, these findings strongly suggest that the interaction of IGFBP-5 with cell-surface-associated HS GAGs is important for the IGF-independent actions of IGFBP-5 on cell migration.
The signaling mechanism(s) underlying the cell-surface HS proteoglycan(s)-mediated, IGF-I-independent action of IGF-BP-5 in cell migration is not well understood at present. Andress (47) has reported that a carboxyl-truncated IGFBP-5 fragment binds to a 420-kDa membrane protein in human bone cells and that heparin competitively inhibited this binding interaction. Further study by the same author indicated that IGFBP-5 treatment led to an increase in Ser and Thr phosphorylation of this protein, suggesting that it could be an IGFBP-5 signaling receptor (48). More recent studies (49 -51) have shown that exogenously added IGFBP-5 could be translocated into the nuclei of cultured human T47D breast cancer cells and osteoblast cells through a yet unidentified pathway. Proteoglycans are proteins that have a diversity of cellular function, including roles in cellular adhesion, motility, differentiation, and growth (see Ref. 52). Cell-surface proteoglycans are known to act as cellular receptors for several viral proteins, including human immunodeficiency virus Tat protein, adeno-associated virus type 2 virions, foot-and-mouth disease type O virus, herpes simplex virus, types 1 and 2, and dengue virus (see Ref. 53). They are also involved in the internalization of a number of growth factors and cytokines, including fibroblast growth factor-2 (54, 55), hepatic growth factor (56), chemokine platelet factor 4 (57), and others. Future studies are needed to determine which HS proteoglycan(s) is involved in the cell migration stimulatory activity of IGFBP-5 and to elucidate the underlying signaling mechanism(s).