INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

THE PLASMIN (Pm) system is involved in normal bone remodeling through a variety of effects, including initiating bone resorption through the activation of latent collagenases (30) and through the activation of osteoblastic regulatory peptides transforming growth factor- (TGF-) (21, 27) and insulin-like growth factors I (IGF-I) and II (11). The effects of the Pm system on osteoblast function are highly regulated and are localized to the cell surface and underlying extracellular matrix (ECM) (8, 12).

Pm activation of IGFs comprises the proteolysis of specific IGF-binding proteins (IGFBPs), thus affecting the bioactivity of IGFs (11). In vivo, IGFs are complexed to IGFBPs in the circulation (15) and within the osteoblastic pericellular environment (1, 2, 18). Pm preferentially cleaves IGFBPs, dissociating IGFs and thus allowing IGFs to interact with their specific cell surface receptors (11). In addition, IGFBPs are known to both stimulate and inhibit IGF function in bone (2, 9, 23). For example, the major bone IGFBP, IGFBP-5, augments IGF function in osteoblasts (2, 23) and targets IGFs to the osteoblastic cell surface (1, 8), where an active Pm system mediates IGF access to the IGF-I receptor (8, 12).

Specific regions within the IGFBP-5 molecule are likely to mediate these various processes. The heparin-binding domain of IGFBP-5, IGFBP-5-(201218), modulates the binding of IGFBP-5 to cellular and extracellular pericellular domains (7). The IGFBP-5-(1169) fragment, which contains only minor heparin-binding domain regions, modifies osteoblast function independently of IGFs by directly stimulating mitogenesis (3). This effect is most likely mediated through its binding to a specific membrane protein (1). However, in either case, endogenous glycosaminoglycan moities do not mediate pericellular binding of IGFBP-5 (1, 7).

The same region of IGFBP-5, IGFBP-5-(201218), has also been reported to inhibit the proteolysis of intact IGFBP-5 in fibroblast-conditioned media (4, 25) and to inhibit metalloproteinase-dependent proteolysis of IGFBP-4 in osteoblast-conditioned media (14). Moreover, heparin itself is also known to inhibit IGFBP-5 proteolysis (4, 25). Whether heparin or other glycosaminoglycans are involved in stabilizing Pm-associated IGFBP-5 proteolytic events in the osteoblastic pericellular environment is unknown. The role of the IGFBP-5-(201218) domain in the proteolysis of IGFBP-5 by Pm remains to be determined.

The purpose of the present study was to evaluate the potential effect of the heparin-binding synthetic peptide IGFBP-5-(201218) on Pm-mediated IGFBP-5 proteolysis, including direct inhibition of Pm action and modification of the activation of the zymogen plasminogen (Pg) to active Pm.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Materials. Recombinant human IGF-I was purchased from GroPep, (Adelaide, Australia). Recombinant intact human IGFBP-5 (IGFBP-5_I) and IGFBP-5-(1169) were produced in a baculovirus expression system and purified by affinity chromatography and reverse-phase high-performance liquid chromatography (RP-HPLC) (3). Glucose Pm was purified from fresh frozen plasma using lysine-Sepharose in the presence of 1 mM benzamidine, 50 µg/ml trypsin inhibitor, and 20 kallikrein-inhibitor units/ml aprotinin (12). Human Pm and tissue-type Pg activator (tPA) was purchased from American Diagnostica (Greenwich, CT). Human single-chain urokinase PA (scuPA) was the generous gift of Dr. A. Mazar (Abbott Laboratories, Abbott Park, IL). scuPA was activated by Pm at a ratio of 20:1 (scuPA/Pm) for 30 min at 23°C. Pm was removed from activated urokinase PA (uPA) by aprotinin-Sepharose. The IGFBP-5 peptides IGFBP-5-(201218), IGFBP-5-(138152), and IGFBP-5-(130142) were synthesized and purified by RP-HPLC. IGF-I, IGFBP-5_I, and IGFBP-5-(201218) were iodinated by the chloramine T method to a specific activity of ~150 µCi/µg protein (11).

Charcoal IGFBP assay. Pm, at 50 ng/ml, was incubated in 50 mM tris(hydroxymethyl)aminomethane (Tris) and 2.5% bovine serum albumin (BSA), pH 7.4, with 6.25 ng/ml IGFBP-5 in the presence of IGFBP-5 peptides [IGFBP-5-(201218), IGFBP-5-(130142), or IGFBP-5-(138152)], as indicated in Figs. 1-10. After a 1-h 37°C incubation in a total reaction volume of 145 µl, proteolytic reactions were terminated by the addition of 5 µg/5 µl aprotinin. ¹²⁵I-labeled IGF-I was then added, and incubations were continued in a total volume of 200 µl for 1 h at 37°C. Complexes of ¹²⁵I-IGF-I and IGFBP-5 were determined by the addition of 400 µl 1% activated charcoal, 0.2 mg/ml protamine sulfate, and 1% BSA in phosphate-buffered saline (PBS), pH 7.4. Reactants (600 µl total volume) were vortexed and incubated in a 4°C icebath for 15 min. Incubations were terminated by centrifugation for 5 min at 14,000 g, and 500-µl supernatants, containing ¹²⁵I-IGF-I/IGFBP-5 complexes, were counted for radioactivity. Binding of ¹²⁵I-IGF-I to IGFBP-5_I in the absence of Pm represented control binding. Nonspecific binding was determined by the inclusion of 100 ng of unlabeled IGF-I to non-Pm-treated IGFBP-5_I and ¹²⁵I-IGF-I, and all binding results were corrected for this value. Results are presented as percentages of nonplasmin control.

Heparin bead and heparin-Sepharose IGFBP assays. Stock slurry solutions of 10% hydroxyapatite (HA) (Calbiochem, La Jolla, CA) and 1:5 heparin-Sepharose beads (Pharmacia Bioteck, Piscataway, NJ) were maintained at 4°C in 30 mM Tris-acetate, 10 mM sodium phosphate, 0.05% Tween 20, and 0.2% sodium azide, pH 7.4 (assay buffer). For an assay, beads were suspended by gently swirling, and 50 µl of the 10% HA bead slurry or 10 µl of the 1:5 heparin-Sepharose were placed in 1.5-ml screw-cap plastic Eppendorf tubes and equilibrated with ¹²⁵I-IGFBP-5_I [~100,000 counts/min (cpm)] in a total volume of 200 µl for 30 min at 23°C, with occasional vortexing. Unbound ¹²⁵I-IGFBP-5_I was removed by washing the bead pellet twice with 1 ml assay buffer. Pm (250 ng/ml) and IGFBP-5-(201218) reactants (10 µg/ml) were incubated with immobilized ¹²⁵I-IGFBP-5_I in a total volume of 200 µl at 37°C with occasional vortexing. At 30- and 60-min intervals, tubes were centrifuged (10,000 g for 30 s), and 25-µl aliquots of supernatants were used to determine released radioactivity. Radioactivity released from beads was corrected for total incubation volume (200 µl) and is presented as the percentage of total IGFBP-5 originally bound to beads (%total IGFBP bound) or percentage of Pm-mediated release (%Pm control).

Peptide protease fingerprinting assay. Protease fingerprinting assays were carried out as previously described (10), with the primary exception that the reaction buffer was changed to 30 mM Tris acetate, 10 mM sodium phosphate, 0.1% Tween 20, and 0.2% sodium azide, pH 7.4 (assay buffer). For the assays, ¹²⁵I-IGFBP-5_I (100,000 cpm), Pm, IGFBP-5 peptides, and other reactants were incubated (as described in Figs. 1-10) in 50 µl assay buffer for 1 h at 37°C. Reactions were terminated by the addition of 1 µg aprotinin, followed by 25 µl 3× nonreducing sample buffer and boiling for 3 min. Reactants were eluted through a 17.5% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and the gel was dried and autoradiographed.

Assessment of Pm proteolysis by ligand blot assay. Experiments to assess Pm treatment of IGFBP-5 and the effect on IGFBP-5 function as determined by ligand blot assay were performed as previously described (11). IGFBP-5_I (100 ng/ml) was reacted with increasing concentrations of Pm in the presence or absence of 10 µg/ml IGFBP-5-(201218) in 1.5-ml Eppendorf plastic screw-cap tubes in 50 µl of 30 mM Tris acetate, 10 mM sodium phosphate, 0.1% Tween 20, and 0.2% sodium azide, pH 7.4. Incubations were terminated after 1 h at 37°C with the addition of 10 µg aprotinin. Twenty-five microliters of 3× nonreducing SDS-PAGE sample buffer were added and samples were boiled for 3 min. Reactants were eluted through an 11% SDS-PAGE and electroblotted onto Immobilon-P (Millipore, Bedford, MA). Blots were blocked with 3% BSA in Tris-buffered saline and equilibrated with 5 × 10⁶ cpm ¹²⁵I-IGF-I overnight at 4°C. Unbound ¹²⁵I-IGF-I was removed by extensive washing, and the blot was air dried and autoradiographed.

Plate binding assays. Binding of IGFBP-5 peptides to Pg was characterized using an immobilized Pg-based assay system. Ninety-six-well immunological plates (NUNC, Fisher Scientific, Pittsburgh, PA) were coated with 5 µg/ml of Pg in 0.1 M Na₂CO₃, pH 9.8, overnight at 4°C. The plates were rinsed with 200 µl PBS and blocked with 200 µl 10 mM Tris · HCl, 150 mM NaCl, 0.05% Tween 80, 1% BSA, and 0.02% sodium azide, pH 7.5, for 1 h at 37°C. Plates were rinsed twice with 200 µl PBS and once with 200 µl of 30 mM Tris acetate, 10 mM sodium phosphate, 0.1% Tween 20, and 0.2% sodium azide, pH 7.4 (assay buffer). ¹²⁵I-IGFBP-5_I (100,000 cpm) or ¹²⁵I-IGFBP-5-(201218) (20,000 cpm) was incubated with increasing concentrations of IGFBP-5-(201218), IGFBP-5-(130142), IGFBP-5-(138152), or heparin in 200 µl assay buffer for 4 h at 4°C. Wells were rinsed twice with 200 µl of ice-cold assay buffer. Bound radioactivity was solubilized with 200 µl 1 N NaOH, transferred to 12 × 75-mm glass test tubes, and counted for radioactivity.

Amidolytic assays. The effect of IGFBP-5 peptides on Pg activation to Pm was determined using Pm chromogenic substrate S-2551 (Pharmacia). IGFBP-5-(201218) (5-40 µg/ml), IGFBP-5-(130142) (20 µg/ml), IGFBP-5-(138152) (20 µg/ml), or heparin (500 nM) was preincubated with 25-400 nM Pg (final concentration) for 3 min at 37°C in the presence of 0.2 mM S-2551 (final concentration). Assay buffer was 30 mM Tris acetate, 10 mM sodium phosphate, 0.1% Tween 20, and 0.2% sodium azide, pH 7.4. After preincubation, 5 nM of uPA was added to initiate Pg activation, and the absorbance change at 405 nm was monitored every 10 s for 5 min at 37°C by use of a Molecular Devices ThermoMax microplate reader (Menlo Park, CA). The initial reaction velocites were determined and, where noted (see Figs. 7 and 8), Lineweaver-Burk plots were calculated.

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

Pm degrades IGFBP-5 into several smaller fragments (Fig. 1A, lane 2); IGF-I bound to IGFBP-5 does not inhibit the Pm effect (lane 3). Heparin prevented Pm degradation of IGFBP-5 (Fig. 1B), whereas IGFBP-5-(1169) did not. This suggested that heparin may inhibit proteolysis by its interaction with the carboxy-terminal portion of IGFBP-5. Because heparin binds predominantly to the 201-218 amino acid region of IGFBP-5 (5, 16), we coincubated this peptide with Pm and found that it was capable of inhibiting Pm-induced proteolysis. This response was specific because other basic peptides within the IGFBP-5 molecule [IGFBP-5-(130132) and IGFBP-5-(138152)] failed to alter the Pm response.

View larger version (73K): [in this window] [in a new window]   Fig. 1.   Plasmin degradation of 125I-labeled intact human insulin-like growth factor-binding protein-5 (125I-IGFBP-5I) and effect of insulin-like growth factor I (IGF-I), heparin, and IGFBP-5 peptides. A: plasmin degradation of complexed and uncomplexed 125I-IGFBP-5I. 125I-IGFBP-5I [100,000 counts/min (cpm)] was incubated with or without 10 ng of IGF-I for 1 h at 37°C to form IGF-I-complexed IGFBP-5 and uncomplexed IGFBP-5. Plasmin (Pm, 2 µg/ml) was added, and reactants were incubated for 1 h at 37°C in a total volume of 50 µl. Incubations were terminated by the addition of 25 µl 3× SDS-polyacrylamide gel electrophoresis (PAGE) nonreducing sample buffer, boiled 3 min, and eluted over 17.5% SDS-PAGE. Gel was dried and autoradiographed. Lane 1, 125IGFBP-5I alone; lane 2, 125I-IGFBP-5I + Pm; lane 3, 125I-IGFBP-5I/IGF-I complex + Pm. B: protection of 125I-IGFBP-5I from Pm degradation by heparin and IGFBP-5 peptides. 125I-IGFBP-5I (100,000 cpm) was reacted with 50 ng/ml Pm, as indicated, for 1 h at 37°C. Pm proteolysis was terminated by the addition of 1 µg aprotinin. Reactants were separated by 17.5% SDS-PAGE. Various factors were coreacted with Pm, as indicated: heparin, 1 mg/ml; IGFBP-5-(130142), 10 µg/ml; IGFBP-5-(138142), 10 µg/ml; IGFBP-5-(201218), 10 µg/ml; or IGFBP-5-(1169), 1 µg/ml.

To quantitate the inhibitory effect of IGFBP-5-(201218), we utilized an IGF-I charcoal-binding assay and found that increasing concentrations of Pm resulted in a concentration-dependent loss of IGF-I-binding activity, with a half-maximally effective concentration (EC₅₀) of 20 ng/ml (Fig. 2A). The presence of 10 µg/ml IGFBP-5-(201218) resulted in a shift of the EC₅₀ to 100 ng/ml. The protective effect of the peptide was concentration dependent, with maximal protection achieved at 10 µg/ml IGFBP-5-(201218) (Fig. 2B). The protective nature of IGFBP-5-(201218) was specific, because two other basic IGFBP-5 fragments, IGFBP-5-(138152) and IGFBP-5-(130142), failed to protect IGFBP-5 from Pm proteolysis (Fig. 3). Similar inhibitory results were observed with ¹²⁵I-IGFBP-5 peptide mapping assays (Fig. 4). Although the peptide was protective under all conditions tested, the inhibition of proteolysis could be overcome with higher concentrations of Pm, suggesting that the peptide was acting as a competitive inhibitor of Pm action (Fig. 4). Additionally, ligand blot analysis confirmed that the protective effects of IGFBP-5-(201218) were specifically affecting IGFBP-5 and its ability to bind IGF-I (Fig. 5).

View larger version (11K): [in this window] [in a new window]   Fig. 2.   IGFBP-5-(201218) protects the IGF-I-binding ability of IGFBP-5I from Pm degradation. A: increasing concentrations of Pm were incubated with 6.25 ng/ml IGFBP-5I for 1 h at 37°C in the presence or absence of 10 µg/ml IGFBP-5-(201218). 125I-IGF-I/IGFBP-5I complexes were determined by charcoal IGFBP assay. Values are means ± SE of 3 separate experiments performed in duplicate. B: Pm at 50 ng/ml was incubated with IGFBP-5 with increasing concentrations of IGFBP-5-(201218) for 1 h at 37°C. 125I-IGF-I/IGFBP-5I complexes were formed and determined as described above.

View larger version (12K): [in this window] [in a new window]   Fig. 3.   Protection of IGFBP-5I from Pm proteolysis is specific for IGFBP-5-(201218). Pm, at 50 ng/ml, was incubated with 6.25 ng/ml IGFBP-5I in the absence or presence of IGFBP-5-(201218), IGFBP-5-(130142), or IGFBP-5-(138152). Incubations were for 1 h at 37°C and were terminated by addition of 5 µg aprotinin. 125I-IGF-I was then added, and incubations were continued for an additional 1 h at 37°C. 125I-IGF-I/IGFBP-5I complexes were determined by charcoal IGFBP assay. Bars, means ± SE of triplicate determinations.

View larger version (106K): [in this window] [in a new window]   Fig. 4.   IGFBP-5-(201218) protection of 125I-IGFBP-5I from Pm: effect of increasing Pm concentration. 125I-IGFBP-5I was incubated with increasing concentrations of Pm, with and without 10 µg/ml IGFBP-5-(201218), for 1 h at 37°C in a total volume of 50 µl. Incubations were terminated by addition of 1 µg aprotinin. Proteolysis was assessed by protease fingerprinting assay, as described in EXPERIMENTAL PROCEDURES.

View larger version (70K): [in this window] [in a new window]   Fig. 5.   Ligand blot analysis of IGFBP-5-(201218) protection of IGFBP-5I proteolysis. IGFBP-5I (100 ng/ml) was reacted with increasing concentrations of Pm in the presence or absence of 10 µg/ml IGFBP-5-(201218). Incubations were for 1 h at 37°C. Pm proteolysis was terminated by addition of 1 µg aprotinin. Reactants were run over 11% SDS-PAGE under nonreducing conditions. Ligand blot was performed with 125I-IGF-I followed by autoradiography.

We next examined whether a protective effect on Pm proteolysis could be seen when intact IGFBP-5 was immobilized to a solid support. This is physiologically relevant because IGFBP-5 is stored in bone matrix (6, 26). As shown in Table 1, ¹²⁵I-IGFBP-5 that was bound to HA (paradigm for the inorganic bone matrix) or to heparin-Sepharose beads (paradigm for the organic bone matrix) could be proteolyzed by Pm, and the degradation was partially inhibited by IGFBP-5-(201218). Within a given bead type (HA or heparin-Sepharose), IGFBP-5-(201218) inhibited Pm-induced release of IGFBP-5 radioactivity at each time point (P < 0.002; Student's t-test).

View this table: [in this window] [in a new window]   Table 1.   IGFBP-5-(201218) inhibits plasmin-mediated proteolysis of 125I-IGFBP-5I bound to HA beads and heparin-Sepharose

To further assess the mechanism of the IGFBP-5-(201218) inhibitory effect, we considered whether IGFBP-5-(201218) interfered with the binding to IGFBP-5 substrate to Pm. IGFBP-5-(201218) effectively inhibited the binding of ¹²⁵I-IGFBP-5 to Pg with a half-maximally inhibiting concentration (IC₅₀) of 100 ng/ml (Fig. 6). In contrast, IGFBP-5-(130142) exhibited minimal competition, and IGFBP-5-(138152) was slightly more inhibitory with an IC₅₀ of 20 µg/ml. Heparin inhibited the binding of ¹²⁵I-IGFBP-5_I to Pg at an IC₅₀ of 1 µg/ml (data not shown). The heparin concentration required to inhibit IGFBP-5 binding may reflect the presence of two low-affinity heparin-binding sites on IGFBP-5 in addition to the high-affinity heparin-binding site at residues 205-212 (16). The question remained, Did IGFBP-5-(201218) directly bind to Pg, thus directly competing for IGFBP-5 binding to Pg? To access this we used an assay that quantitates the extent of ¹²⁵I-IGFBP-5-(201218) binding to Pg. We found that saturable binding of the IGFBP-5-(201218) to Pg occurred in a specific manner with an IC₅₀ of 50 ng/ml (data not shown). IGFBP-5-(130142) did not compete for ¹²⁵I-IGFBP-5-(201218) binding to Pg, and IGFBP-5-(138152) was minimally effective as a competitor, exhibiting only 40% inhibition of binding at 20 µg/ml. Heparin exhibited a biphasic response, with <500 ng/ml heparin increasing ¹²⁵I-IGFBP-5-(201218) binding to Pg (maximum of 155% of control) and >500 ng/ml heparin decreasing ¹²⁵I-IGFBP-5-(201218) binding with an IC₅₀ of 60 µg/ml (data not shown). On the basis of these data, one possible mechanism whereby IGFBP-5-(201218) inhibits the Pm-induced proteolysis of intact IGFBP-5 is by directly competing for binding to Pm, thus interfering with the binding of Pm to substrate IGFBP-5.

View larger version (16K): [in this window] [in a new window]   Fig. 6.   IGFBP-5-(201218) inhibits binding of 125I-IGFBP-5I to plasminogen (Pg). Plates were coated with 5 µg/ml Pg overnight at 4°C. 125I-IGFBP-5I (100,000 cpm/well) were incubated in the presence of increasing concentrations of IGFBP-5-(201218), IGFBP-5-(130142), and IGFBP-5-(138152). Bovine serum albumin (BSA)-coated wells were used to determine nonspecific binding. Incubations were in 200 µl assay buffer for 4 h at 4°C. Unbound 125I-IGFBP-5I was removed by rinsing the wells twice with ice-cold assay buffer. Values are means of duplicate determinations expressed as a ratio of bound radioactivity in the presence of competitor over the bound radioactivity in the absence of any competitor (B/Bo) 125I-IGFBP-5I.

Another possible mechanism whereby fragments of IGFBP-5 can affect Pm proteolysis is through the activation of Pg to proteolytic active Pm by PA. To evaluate this possibility, IGFBP-5-(201218), IGBP-5-(130142), and IGFBP-5-(138142) were tested for their ability to influence Pg activation by uPA by use of an amidolytic assay that measures Pm proteolysis. Only IGFBP-5-(201218) reduced the activation of Pg by uPA (Fig. 7). To evaluate activation efficiency, Lineweaver-Burk plots were determined for IGFBP-5-(201218) and compared with those generated by heparin, a known stimulator of Pg activation. As shown in Fig. 8, IGFBP-5-(201218) decreased Pg activation efficiency by more than twofold, whereas heparin increased efficiency by more than threefold. IGFBP-5-(201218) was also found to inhibit tPA activation of Pg, similarly to uPA (data not shown). IGFBP-5-(201218) did not affect the specific amidolytic activities of Pm, uPA, or tPA (data not shown), confirming specificity of IGFBP-5-(201218) to actual activation of Pg. Natural Pm fragments of IGFBP-5 were also tested to confirm that Pm degradation products of IGFBP-5 were capable of affecting the activation of Pg to Pm. Pm-induced fragments of IGFBP-5 were bound to heparin-Sepharose, and heparin-binding fragments eluting >300 mM NaCl were collected, precipitated to remove excess NaCl, and found to reduce the activation efficiency of Pg by uPA (Fig. 9). To exclude the possibility that substances other than IGFBP-5 fragments might be affecting Pg activation, mock protease digestions without IGFBP-5 were performed and found not to change control activation (data not shown).

View larger version (10K): [in this window] [in a new window]   Fig. 7.   IGFBP-5-(201218) inhibits urokinase Pg activator (uPA) activation of Pg. Increasing concentrations of IGFBP-5-(201218) (), 20 µg/ml IGFBP-5-(130142) (), or 20 µg/ml IGFBP-5-(138152) () were preincubated with 200 nM Pg and 0.2 mM S-2551 for 3 min at 37°C. Amidolytic activities were determined on addition of 5 nM uPA for 5 min at 37°C. Initial reaction velocities (V) were determined as described in EXPERIMENTAL PROCEDURES.

View larger version (11K): [in this window] [in a new window]   Fig. 8.   Activation efficiency of Pg by uPA: contrasting effects of IGFBP-5-(201218) and heparin. IGFBP-5-(201218) (40 µg/ml, A) or heparin (500 nM, B) was preincubated with increasing concentrations of Pg (25-400 nM) and S-2551 for 3 min at 37°C. uPA (5 nM) was added to initiate catalytic reactions. Initial reaction V values were determined, and results are presented as Lineweaver-Burk plots. Symbols are means of 2 separate experiments.

View larger version (14K): [in this window] [in a new window]   Fig. 9.   Pm-generated IGFBP-5 heparin-binding fragment(s) inhibit activation of Pg by uPA. Five micrograms of IGFBP-5I were digested with 100 nM Pm for 30 min at 37°C. Pm was removed by ultrafiltration (30,000 molecular-weight cutoff filter), and heparin-binding fragments of IGFBP-5 were collected by heparin-Sepharose chromatography and trichloroacetic acid (TCA) precipitation, as described in EXPERIMENTAL PROCEDURES. TCA pellet was solubilized in 100 µl of assay buffer, and 20-µl aliquots were reacted with 100-400 nM Pg and 5 nM uPA to determine amidolytic activity.

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

Pm activation at the osteoblast surface has the potential for regulating growth factor responses by an action that involves its separation from a carrier protein. Both TGF- and the IGFs, which are dependent on carrier protein interactions, become activated after Pm proteolysis (11, 21, 27). In the case of the IGF system, IGFBP-5/IGF-I complexes appear to be susceptible to the serine protease action of Pm in a predictable fashion. For example, we found that IGF-I, when bound to IGFBP-5_I, does not prevent Pm proteolysis of IGFBP-5 and that proteolytic fragments of IGFBP-5 do not bind to IGF-I. Thus IGF-I should be free to interact with surface receptors without interference from IGFBP-5 degradation products. We have previously demonstrated that cell surface generation of Pm on osteoblastic cells results in proteolysis of endogenous IGFBP species as well as specific IGFBP species IGFBP-1 (11), IGFBP-3 (10), and IGFBP-4 (10). As is demonstrated by Pm-proteolyzed IGFBP-1 (11), proteolytically released IGF-I retains its full biological potential. Using U₂OS osteoblastic cells as a source of IGFBP-5, we demonstrated that cell surface-generated Pm activity will reduce IGFBP-5 in the conditioned media by 78.6% (data not shown). These data support a role for Pm in the regulation of IGFBP-5 function in the osteoblastic pericellular environment.

The interest that specific fragments of IGFBP-5 may possess biological activity relates to previous work showing that truncated forms of NH₂-terminal (1, 3) and COOH-terminal (25) peptides stimulate osteoblast activity (3) and inhibit generalized IGFBP-5 proteolysis (25), respectively. Peptides responsible for the latter effect have been shown to bind heparin while lacking IGF-binding capability (1). In the current report, we have extended the study of heparin-binding peptides of IGFBP-5 to evaluate whether they affect the Pm system itself, assuming that if an effect were present it would likely be inhibitory in nature. Our results, which confirm this hypothesis, show that the heparin-binding synthetic peptide IGFBP-5-(201218) modifies Pm proteolysis of intact IGFBP-5 by at least two independent mechanisms: 1) blocking the interaction of Pm with IGFBP-5 and 2) inhibiting Pg activation to Pm.

The dichotomy of the heparin effects in regulating Pm proteolysis of IGFBP-5 illustrates the complexity of the regulatory mechanisms. The ability of heparin and IGFBP-5-(201218), but not IGFBP-5-(1169), IGFBP- 5-(130142), or IGFBP-5-(138152), to block Pm proteolysis of IGFBP-5 suggests that at least one important cleavage site may reside within the carboxy-terminal region of the IGFBP-5 molecule. Whether the heparin inhibitory action occurs by binding to and blocking the IGFBP-5-(201218) region or whether heparin induces a conformational change in the tertiary structure of IGFBP-5 is not apparent from these studies. However, because solid-phase heparin does not prevent Pm proteolysis of IGFBP-5, it is likely that the proteolytic inhibitory effect of heparin results from inhibition of Pm binding to IGFBP-5, as the ability of heparin to inhibit IGFBP-5 binding to Pg suggests. This concept is further supported by our findings and those of others (20, 28) demonstrating that heparin stimulates the activation of Pg to Pm by PAs (Fig. 8B).

View larger version (15K): [in this window] [in a new window]   Fig. 10.   Model summarizing immediate effects of heparin and IGFBP-5-(201218) on Pm proteolysis of IGFBP-5. Arrows from heparin and IGFBP-5-(201218) to Pg or Pm denote direct binding interaction with the respective zymogen or protease. +, Stimulation of activity; , inhibition of activity.

The synthetic peptide IGFBP-5-(201218) contains binding sites for heparin and Pg, localizing the regulation of Pm proteolysis to the carboxyl portion of IGFBP-5. Consistent with this notion, we found that IGFBP-5-(201218) binds directly to Pg and blocks the binding of IGFBP-5 to Pg. This suggests that IGFBP-(201218) inhibits Pm proteolysis of IGFBP-5 by blocking the interaction of Pm with IGFBP-5. This may occur through IGFBP-5-(201218) binding to Pm and directly blocking the Pm binding site and/or through IGFBP-5-(201218) binding to IGFBP-5 and blocking the cleavage site for Pm. Additional new evidence now indicates that IGFBP-5-(201218) does bind to intact IGFBP-5 (8a). Whether the binding sequences within IGFBP-5-(201218) are the same for heparin, Pg, and IGFBP-5 is not clear at this time. A recent report by Arai et al. (5) indicated that Arg-201, Lys-202, Lys-206, and Arg-214 are involved in heparin binding to IGFBP-5, suggesting the likelihood of a definite binding site overlap among heparin, Pg, and IGFBP-5 within the IGFBP-5-(201218) sequence. This would be similar to shared binding domains for Pg, tPA, heparin, and collagen/gelatin, which are also present in two other Pm-labile proteins, laminin and fibronectin (17, 22).

Although IGFBP-5-(201218) clearly interferes with the Pm interaction with IGFBP-5, it also inhibits Pg activation to Pm. The mechanism appears to be mediated through decreased Pg activation efficiency. This is in contrast to the effects of heparin, which increased Pg activation efficiency. Peptide fragments of various Pm substrate proteins are known to modify the activation of Pg to Pm (19, 20, 24, 28, 31). Whereas fragments of fibrin (20, 31) and laminin (19) stimulate the activation of Pg, vitronectin (19) inhibits Pg activation. Similar to the effects of IGFBP-5-(201218) reported here, the heparin-binding domain of vitronectin (29) is the responsible element inhibiting Pg activation. Interestingly, peptide sequences within the heparin-binding domain of fibronectin can exhibit both inhibitory and stimulatory effects on Pg activation (24, 28). Because such bioactive fragments can be included in proteolytic products, including IGFBP-5 proteolytic products, these ECM-derived peptide fragments have the capacity to modify subsequent proteolysis, both as an autocrine loop for a specific protein substrate and as a general proteolytic capacity of the Pm system.

The effects of heparin and IGFBP-5-(201218) on Pg activation and Pm degradation of IGFBP-5 are summarized in Fig. 10. The activation of Pg to Pm by PA is differentially regulated by heparin and IGFBP-5-(201218), whereas both heparin and IGFBP-5-(201218) inhibit Pm proteolysis of IGFBP-5. The actions of heparin and IGFBP-5-(201218) appear to require direct binding to Pg and Pm. This direct binding results in a conformational change in Pg that alters its activation efficiency and competitively blocks the binding of IGFBP-5 to Pm, which inhibits IGFBP-5 proteolysis.

In the osteoblastic pericellular environment, Pm regulation of IGF-I interaction with IGFBP-5 is likely localized to the cell surface interface with the fluid and ECM solid phases (1, 8, 10, 12), where Pm is protected from abundant fluid phase protease inhibitors. IGF-I inhibits uPA synthesis by the osteoblast (13) and reduces the binding affinity of the Pg/Pm receptor (12). IGFBP-5-(201218) downregulates Pm-associated proteolysis by reducing the activation efficiency of Pg and by blocking the interaction of Pm with IGFBP-5. The extent to which feedback control of Pm activation occurs with natural IGFBP-5 peptides that are produced by Pm degradation is currently being investigated.