Introduction
Bone metastasis is a significant complication in tumor progression and a contributing cause of patient mortality. Contributing factors to such skeletal "homing" include the ability of the tumor to stimulate remodeling of the bone matrix thereby providing a niche supportive of its survival and growth [reviewed in [
1]]. Fibronectin is a component of the bone matrix and is necessary for osteoblast differentiation and survival [
2‐
4]. Recent in vivo studies have demonstrated a role for specific matrix-remodeling proteases (e.g. plasminogen activators) in the maintenance of bone mass and tissue composition [
5], suggesting both fibronectin and the plasminogen activator system are important regulators of bone formation.
Formation of fibronectin matrix requires a cell-driven mechanical stretching of fibronectin, which is progressively incorporated into a dense detergent-insoluble fibrillar network via interactions with other cell-associated fibronectin dimers [reviewed in [
6]]. The assembly of the fibronectin matrix requires activated α5β1 integrins [
7,
8]. Regulation of α
5β
1 integrin activation is thought to involve changes in integrin conformation which affect its affinity for fibronectin [
9]. Downstream of fibronectin ligation, additional integrin dependent steps in the regulation of matrix assembly may occur in response to changes in key intracellular signaling pathways [
10] or through the formation of complexes with either cytoskeletal proteins [
11] or cell surface molecules including the glycosylphosphatidylinositol (GPI)-anchored urokinase-type plasminogen activator receptor (uPAR) [
12‐
14]. High levels of uPAR expression have been shown to lead to formation of uPAR/β1 complexes which modulate integrin signaling and adhesive function [
15,
16]. uPAR effects on α5β1 function are complex and may result in either gain or loss of function depending on cellular context [
15,
17]. Several peptides have been identified which can modulate the functional association of integrins with uPAR [
12,
18]. Among them, peptide P25, isolated from a peptide library screening, is able to directly bind to uPAR and modulate integrin activity [
12,
15,
19,
20].
uPAR binds directly to the somatomedin B (SMB) domain of vitronectin and supports adhesion of a number of different cell types [
21,
22]. Plasminogen activator inhibitor Type I (PAI-1), the major physiological regulator of uPA activity, binds in close proximity to the uPAR recognition site within the SMB domain and can competitively inhibit or displace uPAR-vitronectin interactions [
22,
23]. The PAI-1 binding site on vitronectin is close to the only RGD motif in vitronectin [
21] and the binding of PAI-1 to vitronectin can sterically inhibit integrin dependent adhesion to this motif [
22,
24]. Thus, PAI-1 can modulate the association of both uPAR- and/or integrins with vitronectin.
Treatment of human dermal fibroblasts with the uPAR ligand, P25, results in a marked increase in the polymerization of the fibronectin matrix and in the level of β1 integrin activation [
13,
14]. In the present study, we now show that overexpression of uPAR in MG-63 cells increases the effects of P25 on both fibronectin matrix assembly as well as integrin activation. In addition, P25 has no effect on matrix assembly in uPAR null cells. Our results also show that uPAR and PAI-1 synergistically regulate the amount of fibronectin deposition into the matrix. The positive regulation of PAI-1 on fibronectin assembly requires the presence of vitronectin suggesting that PAI-1 and vitronectin work cooperatively to regulate the amount of fibronectin matrix in MG-63 cells.
Discussion
We have identified a role for the plasminogen activator system and vitronectin in the regulation of fibronectin matrix assembly in the human osteoblast-like cell line, MG-63. uPAR acts in concert with PAI-1 to stimulate fibronectin matrix deposition by increasing the level of α5β1 integrin activation. In addition, the effects of P25 on assembly of fibronectin matrix were modulated by the level of uPAR expression. Cells overexpressing uPAR were more sensitive to the effects of P25, while uPAR -/- cells were unable to respond to P25. Reexpression of uPAR in uPAR -/- cells restored the P25 response confirming that the effect of P25 on matrix assembly was specific to uPAR. This finding is in agreement with earlier studies which have shown that matrix assembly can be regulated through uPAR [
14]. Other studies have shown that in some tumor cells, high levels of uPAR expression correlated with an increase in the formation of fibronectin matrix [
43]. In our studies, neither increasing the levels of uPAR in MG-63 nor restoring uPAR to uPAR -/- cells increased the incorporation of fibronectin into matrix, unless the cells were stimulated with P25. This suggests that in MG-63 cells as well as MEFs, basal levels of matrix assembly are not under uPAR regulation.
P25 induced a several fold increase in the rate of assembly of exogenous fibronectin into previously established matrix. Furthermore, when suspended MG-63 cells were allowed to adhere to fibronectin, P25 increased the number and size of ECM contacts as well as the level of activated β1 integrin present in the ECM contacts. Several earlier studies have shown that polymerization of fibronectin matrix originates in cell adhesion sites, requires the presence of activated β1 integrins and can be modulated by changes in the level of integrin activation [
7,
8,
37]. Taken together, these results suggest that in MG-63 cells ligation of uPAR with P25 results in an increase in both assembly of exogenous fibronectin into preformed matrix as well as an increase in the initiation of matrix assembly sites in newly adherent cells. The effects of P25 on matrix assembly were significantly attenuated in the presence of a blocking antibody to the β1 integrin, suggesting that the P25 induced increase in matrix assembly was the direct result of the increase in β1 integrin activation. This observation is in agreement with previous studies showing that addition of P25 to cells stimulates the adhesion function of the α5β1 integrin [
15,
19]. P25 induced matrix assembly was not completely inhibited by β1 blocking antibodies suggesting that P25 may stimulate alternate mechanisms of matrix assembly by promoting the binding of fibronectin to β3 integrins or to unfolded matrix fibronectin molecules [
44,
45].
P25 directly interacts with uPAR and shares some sequence homology with integrin α subunits [
12,
15,
18]. uPAR has been found to both physically and functionally associate with several different integrin receptors [
46,
47]. The direct interaction of P25 with uPAR may well affect uPAR modulation of integrin function [
12,
48], however its mechanism of action remains unclear. Dimerization/oligomerization of uPAR can affect its subcellular localization by driving uPAR into lipid rafts [
49]. As lipid rafts are sites of integrin activation [
50], P25-mediated changes in uPAR localization to lipid rafts could promote the activation of raft-associated integrins.
PAI-1 may play a role in the regulation of the uPAR-dependent formation of fibronectin matrix. Up-regulation of matrix assembly by PAI-1 was observed under conditions of uPAR stimulation and was accompanied by an increase in the activation of the β1 integrin. The PAI-1 mediated increase in integrin activation and fibronectin matrix assembly required the binding of PAI-1 to vitronectin, but was independent of the binding of PAI-1 to uPA. The mechanism whereby PAI-1 binding to vitronectin might further increase β1 integrin activation is not known. Binding of PAI-1 to the SMB domain of vitronectin blocks the binding of both uPAR and vitronectin-integrin receptors to vitronectin [
22,
24], suggesting that loss of either uPAR-vitronectin and/or integrin-vitronectin binding may be part of the mechanism by which PAI-1 modulates fibronectin matrix formation. Vitronectin negatively regulates matrix assembly [
51‐
53] therefore, "disengagement" of either uPAR or vitronectin integrins from vitronectin by PAI-1 may release the inhibitory effect of vitronectin on matrix assembly. Alternatively, PAI-1 mediated release of uPAR from vitronectin may free uPAR to form complexes with other cell surface molecules. Complexes of uPAR with α5β1 integrins would then promote further integrin activation [
13,
17]. Alternatively, uPAR may complex with other receptors including G-protein linked receptors or growth factor receptors which affect signaling molecules important in integrin function [
12,
48,
54]. A recent study has reported that PAI-1 initiates cellular detachment from both vitronectin and fibronectin substrates, suggesting a PAI-dependent negative effect on the adhesive function of both β3 and β1 integrins [
55]. In these studies, the effects of PAI-1 on cell adhesion were independent of vitronectin, but required the binding of PAI-1 to uPA. The reason for the discrepancy between the two studies may reflect the choice of cell type. Our experiments were performed on MG-63 cells which assemble a robust extracellular matrix, while the earlier studies used HT-1080 fibrosarcoma cells which do not assemble an extracellular matrix.
Bone destruction during metastasis is marked by loss of collagenous matrix [
56]. Deposition and organization of collagenous matrix is intrinsically linked to the assembly of the fibronectin matrix which serves as the core scaffolding on which the collagenous matrix is built [
57,
58]. Our results in MG-63 cells highlight a new mechanism by which PAI-1 and uPAR work synergistically to increase the formation of fibronectin matrix and suggests that therapeutic antagonists of this pathway may attenuate the inappropriate deposition of excess matrix often seen in osteoblastic lesions. Alternatively, modalities directed at increasing levels of fibronectin in bone matrix (i.e., uPAR ligands) might prove useful in the control of osteolytic metastasis, particularly under conditions where inhibitors of bone resorption are not effective [
59]. Increased levels of fibronectin, moreover, would be expected to promote the differentiation and survival of osteoblasts, which are often rendered inactive in sites of bone metastasis. Future studies are needed to determine whether the vitronectin/plasminogen activator system are therapeutic targets for the control of bone metastasis.
Methods
Reagents and antibodies
Unless otherwise stated, all chemicals were purchased from Sigma (St. Louis, MO). The uPAR ligand, peptide P25, sequence AESTYHHLSLGYMYTLN, and the scrambled peptide, S25, sequence NYHYLESSMTALYTLGH, were synthesized by Cell Essentials (Boston, MA). The anti-β
1 antibodies, clones P5D2, 6S6, and HUTS-4 were purchased from Chemicon (Temecula, CA). Control mouse IgG and antibody to β-actin were obtained from Sigma (St. Louis, MO). Secondary antibodies goat anti-mouse and goat anti-rabbit HRP were purchased from Bio-Rad (Hercules, CA). AlexaFluor
594-labeled goat anti-rat antibody was obtained from Molecular Probes (Eugene, OR). 9EG7 antibody was obtained from Pharmingen (San Diego, CA) and IST-9 antibody from Abcam (Cambridge, England). The pCEP4 plasmid was from Invitrogen (Carlsbad, CA) and the pCEP4 plasmid containing the full length cDNA for human uPAR was described previously [
15] and was the gift of Dr. H. Chapman (University of California-San Francisco). Recombinant active PAI-1, active uPA, and uPAR polyclonal antibody (399R) were from American Diagnostica (Greenwich, CT). The non-vitronectin binding PAI-1 mutant, Q123K, was from Molecular Innovations (Southfield, MI) and the dual PAI-1R (T333→R and A335→R) which resulted in the loss of the protease inhibitory activity of PAI-1 but did not affect its binding affinity for vitronectin, was kindly provided by Dr. D. Lawrence (Holland Lab, American Red Cross, Rockville, MD) [
25]. The fragment of vitronectin 40–459, which lacks the somatomedin B domain, was the gift of Dr R.P. Czekay (Albany Medical College, Albany, NY), and was expressed and purified as described previously [
21].
Cell culture
The MG-63 osteosarcoma cells were grown in Dulbecco's modified Eagle's medium (DMEM, Invitrogen), containing antibiotics (penicillin-streptomycin) and 10% fetal bovine serum (FBS, Hyclone Laboratories, Logan, UT). The uPAR+/+ and uPAR-/- mouse embryo fibroblasts (MEF) cell lines have been described previously [
26] and were provided by Dr. Steve Gonias (University of California-San Diego). MEF were cultured in DMEM containing 10% FBS, plus antibiotics.
Isolation of the uPAR overexpressing MG-63 cell lines
Two μg of the expression vector (pCEP4) or the pCEP4-uPAR plasmid were transfected into MG-63 cells (3 × 105) using Lipofectamine 2000 (Invitrogen). The following day, cells were trypsinized and transferred into two 100 mm diameter tissue culture plates. Selection (100 μg/ml hygromycin) was started 48 hours after the transfection. Clones were identified, isolated with cloning rings, and expanded for testing. All transfected cell lines were screened for uPAR expression by Western blotting. Three clones overexpressing uPAR and three clones containing only vector were selected for expansion and further characterization.
Transient transfection of uPAR in MEF
uPAR-/- MEFs were transfected with 4 μg of pcDNA 3.1+ plasmid containing the full length cDNA for human uPAR (pcDNA-uPAR) and 12 μl LipofectAMINETM, according to the manufacturer instructions (Invitrogen, Carlsbad, CA). Control cells were transfected with empty vector. Transfection efficiencies were approximately 60%, as determined by counting fluorescent cells co-transfected with pEGFP plasmid (Clontech). MEF were transfected for 24 hours before matrix incorporation assay. Cells transfected with empty vector (pLDNA3.1+) served as controls.
Purification of fibronectin and vitronectin
Human plasma fibronectin was purified from a fibronectin- and fibrinogen-rich by-product of Factor VIII production by ion exchange chromatography on DEAE-cellulose (Amersham Biosciences) as described previously [
27]. Vitronectin was purified from fibronectin- and fibrinogen-depleted human plasma by heparin-Sepharose affinity chromatography according to the method of Yatohgo et al. [
28]. Purified plasma fibronectin (400 μg) was iodinated with 1 mCi of Na
125I (Perkin Elmer Life Sciences) as described previously [
29]. Iodinated fibronectin was mixed with bovine albumin, 1 mg/ml, dialyzed against phosphate-buffered saline, and frozen at -80°C until used. Fibronectin (2 mg/ml) was derivatized with AlexaFluor
488 according to the manufacturer's protocol (Molecular Probes).
Matrix incorporation assay
MG-63 cells were seeded onto 6 well plates (2.5 × 10
5 cells/well) in complete medium. The following day, cultures were incubated for 6 hours with
125I-fibronectin (2 μg/ml; 1 × 10
6 cpm/ml) in DMEM plus 0.02% BSA in the presence of the uPAR ligand, P25, or the control peptide, S25. For isolation of detergent insoluble matrix, cells were rinsed in PBS, extracted in 1% deoxycholate (DOC) (in 20 mM Tris (pH 8.8) buffer containing 2 mM phenylmethylsulfonyl fluoride, 2 mM EDTA, 2 mM N-ethylmaleimide, and 2 mM iodoacetic acid). Detergent insoluble matrix was obtained by centrifugation at 18,000 rpm for 40 min and associated radioactivity present in the pellet measured using gamma scintillation. Preincubation with 20 nM PAI-1 (active or mutant) or uPA/PAI-1 complex and
125I-fibronectin (2 μg/ml; 1 × 10
6 cpm/ml) was for 20 min in DMEM+BSA 0.02% before peptide addition. The uPA/PAI-1 complex was prepared by reacting active two chain uPA with an equimolar concentration of PAI-1 for 10 min at 37°C as described previously [
30]. In some experiments, cells were pre-incubated with blocking antibody to the β1 integrin prior to incubation with
125I-fibronectin. To determine the total cell layer-associated fibronectin, cell cultures were pretreated for 20 min with different concentrations of PAI-1 or uPA with
125I-fibronectin (2 μg/ml; 1 × 10
6 cpm/ml) in DMEM at 37 °C and then stimulated with either P25 or S25 for 6 hours. After incubation, cells were rinsed three times in PBS, then scraped directly into 1 ml of 1% deoxycholate and the total cell bound
125I-fibronectin was counted. In some experiments, tissue culture wells were precoated with either 10 μg/ml fibronectin, vitronectin or vitronectin fragment 40–459 in PBS for 2 hours at 37°C.
Fluorescent microscopy
To localize the P25-dependent plasma fibronectin deposited into matrix with endogenous cell-derived fibronectin, MG-63 cells were cultured in complete medium for 24 hours. Cell monolayers were incubated for an additional 16 hours in serum-free medium containing 50 μM S25 or P25 and plasma fibronectin (2 μg/ml) derivatized with AlexaFluor488. Cell layers were then rinsed, fixed with 3% paraformaldehyde, permeabilized with 0.2% Triton for 5 min, and blocked for 1 hour with 1% BSA in PBS. Endogenous cell-derived fibronectin was visualized by incubating cells with IST-9 monoclonal antibody (2.5 μg/ml) against cellular fibronectin. This antibody recognizes fibronectin containing the EDA domain which is found exclusively in cell-derived fibronectin. Cells were rinsed and incubated with goat anti-mouse AlexaFluor594-labeled secondary antibody. To examine the formation of ECM contacts, coverslips were coated with fibronectin (10 μg/ml), and blocked in 1% BSA. MG-63 cells were allowed to attach for 6 hours on fibronectin-coated coverslips in DMEM containing soluble AlexaFluor488-fibronectin (20 μg/ml) and peptide (P25 or S25: 50 μM). Cells were washed, fixed in 3% paraformaldehyde, permeabilized with 0.2% Triton for 5 min and blocked with 3% milk. Cells were then incubated for 1 hour with 5 μg/ml 9EG7 antibody against active β1 integrin and bound antibody visualized using AlexaFluor594-conjugated goat anti-rat antibody. Fluorophores were visualized using an Olympus BMX-60 microscope equipped with a cooled LCD sensi-camera. Images were acquired using Slidebook Software (Intelligent Imaging Innovation, Inc., Denver, CO) and processed using Adobe Photoshop.
Immunoblotting
Cell layers (3 × 105/well) were washed with PBS and lysed with hot sample buffer [75 mM Tris-HCl (pH 6.8), 2% SDS, 10% glycerol]. The cell lysates were boiled for 2 min and the proteins were separated by SDS-PAGE (Bio-Rad, Hercules, CA), transferred to nitrocellulose membranes and blotted with the indicated antibodies.
Integrin activation assay
MG-63 cell lines were grown overnight (6 × 10
4 cells/well) in 24-well plates in complete medium. The following day, cells were treated for 2 hours in DMEM containing either P25 or S25 and subsequently incubated for 1 hour at 37°C with either 100 ng/ml of HUTS-4 antibody, which recognizes the activated conformation of the β
1 integrin [
31‐
33], or 100 ng/ml P5D2, an antibody that recognizes all forms of β
1 integrin. Cells were rinsed with PBS, fixed with 3% paraformaldehyde, blocked in 3% BSA, and incubated for 1 hour with HRP-conjugated goat anti-mouse antibody. Freshly prepared substrate (0.1 M citrate buffer, 0.5 mg/ml
o-phenyl-enediamine, 1 μl/ml 30% hydrogen peroxide pH 5.0) was added to each well, and the color was allowed to develop. The reaction was stopped with the addition of 2 N sulfuric acid, and the OD was measured at
A490. Measurements were corrected for light scattering by subtracting the OD obtained at
A650. In some experiments, cells were incubated overnight on either fibronectin or vitronectin-coated plates in DMEM containing 0.02% BSA and preincubated for 20 min with PAI-1 (1 μg/ml) before treatment with peptides.
Acknowledgements
These studies were supported by CA-58626 from the National Institutes of Health. Elizabeth Monaghan-Benson is supported by an NIH Training Grant, T32-HL-07194, from the National Heart, Lung & Blood Institute. The authors thank Drs. Paul Higgins and Ralf-Peter Czekay for critical review of the manuscript and Drs. Chapman, Lawrence, and Czekay for their generous gift of reagents.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
DV designed and carried out most of the experiments and drafted the manuscript. EM-B participated in performing and designing several experiments. PM-L conceived the study, coordinated the experimental questions, and prepared the final version of the manuscript. All authors read and approved the final manuscript.