Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The urokinase/plasmin system is an important enzyme system mediating fibrin degradation both within the vascular bed and at numerous extravascular sites, including alveolar surfaces (1). The fibrinolytic potential of urokinase is augmented by its capacity to bind to its high-affinity cell-surface receptor, uPAR (CD87). The glycosylphosphatidylinositol-anchored uPAR concentrates urokinase at cellular surfaces and promotes activation of cell-associated plasminogen important to activation of metalloproteases and cellular migration through fibrinous barriers (4). uPAR is remarkably widely expressed, being present constitutively on endothelial and epithelial lining cells of the lung and other tubular structures (5, 7). uPAR is also inducible, being expressed on most migrating cells where it localizes to their leading edge (8, 9), suggesting a role for uPAR in promoting focused extracellular proteolysis important to migration through extracellular matrices, especially fibrin-rich matrices that commonly accumulate during injury to the lung (2, 3). The capacity of uPAR to localize proteolysis at the leading edge of migrating cells has implicated this receptor in biologic processes as disparate as wound repair, inflammation, angiogenesis, and tumor development (1, 10).

In addition to its role in promoting pericellular proteolysis, in vitro studies indicate that urokinase initiates at least two other biologic processes independently of its protease activity. First, occupancy of uPAR by urokinase or a nonproteolytic amino-terminal receptor-binding fragment thereof consisting of residues 1-48, the growth factor domain (GFD), promotes cellular adhesion of myelomonocytic cells to the matrixlike form of vitronectin (13). The observation that soluble cell-free uPAR binds vitronectin provides evidence of a direct interaction between uPAR and vitronectin (14). The strength of interaction between uPAR and vitronectin is proportional to concurrent uPAR occupancy with urokinase. This phenomenon is not dependent on proteolytic activity as GFD, larger nonproteolytic amino-terminal fragments of urokinase (ATF), relatively inactive pro-urokinase, and two-chain active urokinase all bind uPAR with high affinity (5) and promote uPAR/vitronectin binding (13, 14). Second, binding of urokinase and, more specifically, ATF to uPAR is chemotactic for myeloid cells and operates at least in part through initiation of cellular signals leading to reorganization of the cytoskeleton (15, 16). The capacity of urokinase to regulate adhesion and migration through both its catalytic domain and its receptor-binding domain underscores its central involvement in cell movement.

Depite extensive in vitro study, a specific role for the uPAR-binding region of urokinase in vivo remains undefined. Using a model of metastasis of human prostate cells in immunodeficient mice, Crowley and colleagues (17) reported reduced pulmonary metastases in the presence of circulating ATF. Persistent levels of circulating ATF were achieved by injection of a fusion protein composed of the amino-terminal fragment of human urokinase and the Fc portion of immunoglobulin (Ig)G, the latter of which greatly prolongs the circulating half-life of the fusion protein. Reduction of metastasis was ascribed to decreased pericellular proteolysis resulting from the displacement of active urokinase from uPAR expressed by the prostate cells. However, we have recently demonstrated that occupancy of uPAR by ATF retards tumor cell migration across vitronectin in vitro by increasing cellular adhesion to vitronectin (18). Thus uPAR could affect metastasis by regulating nonproteolytic as well as proteolytic events at the cell surface.

Mice with targeted deletion of a functional uPAR gene were recently reported to have delayed peritoneal neutrophil accumulation in response to inflammatory stimuli (19). However, whether the capacity of uPAR to bind urokinase is important to this functional defect is again unknown. The aim of the present study was to explore the effect of uPAR occupancy on cellular migration in vivo and to characterize further the importance of proteolysis to this migration. Using a murine model, we now report that the presence of a circulating amino-terminal fragment of urokinase in the form of GFD markedly reduces chemotaxin-induced migration of neutrophils into the lung. Neutrophil emigration in response to the murine homologue of interleukin (IL)-8, KC, remains robust in mice lacking functional urokinase, whereas the presence of GFD reduces cellular migration in these mice. In contrast, GFD does not diminish cellular migration induced by KC in mice lacking functional uPAR. These observations suggest that the inhibitory effects of GFD on uPAR-dependent cellular migration in this model do not reflect disruption of pericellular proteolysis but rather changes in cellular adhesiveness and/or signaling.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Materials

Recombinant murine KC was prepared as described (20) and its purity was verified by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Murine GFD- human IgG fusion protein (GFD-Fc) was prepared as described (21), resuspended in phosphate-buffered saline (PBS), and sterile filtered. Human IgG, obtained from Sigma Chemical Company (St. Louis, MO), was similarly resuspended in PBS and sterile filtered. Calf lung surfactant extract was the kind gift of Dr. Edward Ingenito (Brigham and Women's Hospital, Boston, MA).

C57Bl/6 wild-type mice were obtained from Charles River Laboratories (Wilmington, MA). Mice with targeted disruptions of the urokinase (22) and uPAR (23) genes were obtained from Peter Carmeliet (Center for Transgene Technology, Flanders Interuniversity, Leuven, Belgium).

Preparation of Pegylated Murine GFD

A synthetic gene encoding the GFD of human urokinase (residues 1-48) was modified by site-directed mutagenesis to change four residues in the receptor-binding loop to those in murine urokinase (N23Y, N27R, H29R, and W30R). This mutant protein was expressed as for the human protein using a yeast system (24). The protein was purified from the conditioned medium by Fractogel-S ion exchange chromatography. Murine uPAR binding assays were performed as described (21), and the concentration that produces 50% inhibition (IC₅₀) for the mutant protein is < 1 nM. The modification of this protein with a single polyethylene glycol moiety was performed using the methods described by Gaertner and Offord (25). Briefly, a solution of purified protein (1-1.5 mg/ml) was incubated with a 1.5-fold molar excess of sodium meta periodate for 30 min in the dark at room temperature. The protein aldehyde was then separated from remaining periodate by diafiltration against 30 mM sodium acetate, pH 4.5. The purified protein aldehyde was then reacted with a 2-fold molar excess of 20 kD polyethylene glycol hydrazide (Shearwater Polymers, Huntsville, AL) for 24 h at 37°C. The desired product was separated from the remaining polyethylene glycol hydrazide by cation exchange chromatography and characterized by mass spectrometry, amino-acid analysis, endotoxin measurement, and receptor-binding assays. The final product has a size of 26 kD, < 1 endotoxin units (e.u.)/mg endotoxin, and has an IC₅₀ of 1 nM in the murine uPAR-binding assay.

In Vivo Neutrophil Chemotaxis Assay

All animal experiments were performed in accordance with the National Institutes of Health guidelines for the humane care and treatment of laboratory animals and approved by the Harvard Medical Area Standing Committee on Animals. Mice were lightly anesthetized via halothane inhalation. Recombinant KC, 1.5 µg in 75 µl sterile PBS containing 5 mg/ml calf lung surfactant extract, was instilled in 5-10 µl aliquots into alternate nares via pipet allowing for full insufflation between aliquots. Sham treatment consisted of nasal insufflation of PBS containing calf lung surfactant extract alone. The addition of calf lung surfactant extract to the KC mixture was noted to promote an even distribution of chemotactic response throughout the lungs (D. Waltz, unpublished observations). In some experiments, mice underwent intraperitoneal injection of GFD-Fc, pegylated murine GFD (GFD-PEG), or human IgG 18 h before nasal insufflation. After nasal insufflation, the mice were allowed to recover fully. Euthanasia via intraperitoneal injection of pentobarbital was carried out after 4 h. The trachea was intubated under direct visualization, and whole lung lavage was carried out in serial aliquots with a total of 4 ml of 0.6 mM ethylenediaminetetraacetic acid (EDTA) in PBS. The volume of recovered lavage fluid was recorded, and total white blood cell counts obtained via hemocytometer. Differential white blood cell counts were performed on cytospin specimens of lavage fluid containing 50,000 cells per slide.

To determine the effects of GFD-Fc and human IgG on circulating neutrophil counts, whole blood was obtained 18 h after intraperitoneal injection by intracardiac puncture in euthanized mice. Red blood cells were lysed by resuspension in 0.15 M NH₄Cl/1 M KHCO₃/0.01 M Na₂EDTA for 2 min on ice and total and differential white blood counts performed.

Alveolar Permeability Assay

Alveolar permeability was assessed by injection of 100 µl Evans blue dye (6.25 mg/ml in 0.9% NaCl) into a tail vein at the time of KC nasal insufflation. Evans blue dye binds to serum proteins and the appearance of dye in lung lavage fluid reflects alveolar permeability (26). After euthanasia at 4 h, whole lung lavage was performed as described previously, and blood obtained by cardiac puncture was subjected to centrifugation and 1:50 dilution in PBS. The ratio of the absorbances of whole lung lavage fluid and diluted serum at 620 nm was taken as the permeability index.

Statistical Methods

Two-tailed unpaired Student's t test was performed for comparison of mean experimental values (27). Where depicted, error bars represent the standard deviation (SD).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Nasal Insufflation of KC Increases Pulmonary Neutrophil Influx and Protein Leakage, and This Increase Is Attenuated by Prior Administration of GFD-Fc

Baseline measurements in untreated and sham-treated, wild-type mice revealed the recovery of ~ 0.7 × 10⁶ leukocytes by whole lung lavage, of which 99% were macrophages. Nasal insufflation of recombinant murine KC, a potent neutrophil chemoattractant (20), resulted in a marked increase in recovered leukocytes at 4 h, which were 85- 97% neutrophils (Figure 1). We chose this model to explore the influence of uPAR binding by GFD because it is known that neutrophils stimulated by chemotaxins express cell-surface uPAR (28). Moreover, mice deficient in uPAR were recently reported to have a reduced peritoneal accumulation of neutrophils in response to local inflammatory stimuli (19). Intraperitoneal administration of 200 µg human IgG 18 h before KC nasal insufflation had little effect on the number or type of leukocytes recovered (Figure 1). In contrast, prior administration of 200 µg GFD-Fc markedly attenuated this neutrophil influx.

View larger version (11K): [in this window] [in a new window]   Figure 1.   GFD-Fc reduces KC-induced neutrophil influx in wild-type mice. The potent murine neutrophil chemoattractant KC, 1.5 µg, was instilled via nasal insufflation 18 h after intraperitoneal administration of saline or 200 µg of either human IgG or a fusion protein consisting of the Fc portion of human IgG and a nonproteolytic amino-terminal uPAR-binding fragment of murine urokinase (GFD-Fc). Whole lung lavage was performed at 4 h and the number of leukocytes recovered was determined. Each circle represents the value from a single animal and the horizontal lines indicate the mean values of each group. *P < 0.05 versus KC alone and the IgG control. IP = intraperitoneal.

Histologic sections of lungs recovered from similarly treated animals revealed a marked neutrophil infiltrate in mice treated with KC (Figure 2). The neutrophils were distributed throughout the interstitial and alveolar compartments. Prior administration of GFD-Fc attenuated this neutrophil infiltrate. Of note, after administration of GFD-Fc and KC insufflation, no neutrophil accumulation was observed within blood vessels or within alveolar walls. These observations indicate that the decrease in neutrophils recovered by lung lavage after GFD-Fc administration is not due to increased intrapulmonary neutrophil adhesion or margination.

View larger version (139K): [in this window] [in a new window]   Figure 2.   Representative photomicrographs. The lungs from IgG- (a, c) and GFD-Fc- (b, d) treated mice were recovered 4 h after KC instillation and subjected to fixation followed by staining with hematoxylin and eosin. Bars represent 40 µm (a, b) and 10 µm (c, d).

The marked inflammatory infiltrate noted on histologic examination and whole lung lavage in response to KC administration was accompanied by increased alveolar permeability to protein. Alveolar permeability to protein can be assessed using Evans blue dye as a marker (26). When Evans blue dye was administered intravenously at the time of KC nasal insufflation the amount of dye present in lung lavage fluid at 4 h more than doubled (Figure 3). Consistent with the effects of GFD-Fc on accumulation of alveolar neutrophils, prior administration of GFD-Fc (but not IgG) significantly attenuated this increased permeability.

View larger version (11K): [in this window] [in a new window]   Figure 3.   GFD-Fc reduces KC-induced alveolar permeability in wild-type mice. Intraperitoneal administration of IgG or GFD-Fc was carried out as in Figure 1. Evans blue dye was administered intravenously at the time of KC instillation, and the ratio of dye in lung lavage fluid to a 1:50 dilution of serum as determined by spectrophotometry is expressed as the permeability index. The permeability index of animals not receiving KC is shown for comparison. The mean ± SD of six separate animals in each group is shown. *P < 0.05 versus IgG control. IP = intraperitoneal.

GFD-Fc consists of the Fc portion of human IgG coupled to murine GFD (21). The Fc fragment prolongs the circulating half-life of the fusion protein in the peripheral circulation. We considered the possibility that this protein might decrease the number of uPAR-expressing cells in the peripheral circulation by binding to uPAR and "marking" these cells for Fc-dependent clearance. Peripheral leukocyte counts were obtained from wild-type mice 18 h after intraperitoneal injection of saline, human IgG, or GFD-Fc (n = 8 in each group). Total leukocyte counts were 0.87 ± 0.43, 0.92 ± 0.39, and 0.83 ± 0.49 × 10⁶ cells/ml (mean ± SD) in the three groups, respectively, with a similar percentage of neutrophils (P > 0.05). These data indicate that GFD-Fc administration did not result in neutrophil clearance from the peripheral circulation.

To explore further the possibility that the Fc portion of the GFD-Fc was important to its anti-inflammatory effect, we repeated the experiments with a murine GFD covalently coupled to polyethyleneglycol (GFD-PEG). This protein has a molecular mass of ~ 26 kD and, like GFD-Fc, has a relatively long circulating half-life. Intraperitoneal injection of 400 µg GFD-PEG 18 h before KC insufflation reduced KC-induced neutrophil influx ~ 50% compared to the effect of 400 µg IgG (Figure 4), consistent with the effects of GFD-Fc (Figure 1). The differences in the inflammatory response after administration of GFD-Fc and IgG (Figures 1-3) were thus not due to differences in intrapulmonary margination or adhesion of neutrophils, differences in the circulating numbers of neutrophils, or effects mediated by binding of human Fc to mouse Fc receptors.

View larger version (9K): [in this window] [in a new window]   Figure 4.   GFD-PEG reduces KC-induced neutrophil influx in wild-type mice. Intraperitoneal administration of 400 µg of either IgG or GFD-PEG was followed by instillation of KC and lung lavage as in Figure 1. The mean ± SD of 8-10 animals in each group is shown. *P < 0.05 versus IgG control.

GFD-Fc Attenuates KC-Induced Neutrophil Influx in Urokinase / but Not uPAR / Mice

The consistent inhibitory effect of GFD-Fc on KC-induced neutrophil influx into the lung (Figure 1) could be due either to displacement of proteolytically active urokinase from uPAR or to effects of uPAR occupancy on cellular adhesion and/or migration independently of urokinase activity. To distinguish between these possibilities, the responses of wild-type and urokinase-deficient mice to KC in the absence and presence of GFD-Fc were compared. Nasal insufflation of KC in urokinase-deficient mice resulted in an increase in leukocytes recovered by lung lavage (Figure 5), of which 90-95% were neutrophils. The numbers of leukocytes recovered from urokinase-deficient mice in response to KC were no different from that of normal C57Bl/6 mice (Figure 1). Similar to the results seen in wild-type mice, prior administration of GFD-Fc, but not IgG, markedly attenuated this neutrophil influx.

View larger version (13K): [in this window] [in a new window]   Figure 5.   GFD-Fc reduces KC-induced neutrophil influx in urokinase-deficient mice. Intraperitoneal administration of IgG or GFD-Fc was followed by instillation of KC and lung lavage as in Figure 1. Each circle represents a single animal and the horizontal lines indicate the mean values of each group. *P < 0.05 versus KC alone and the IgG control. IP = intraperitoneal.

The effects of the GFD fusion protein on KC-induced neutrophil influx were next assessed in uPAR-deficient mice. As seen in Figure 6, the neutrophil influx induced by KC nasal insufflation was similar in uPAR +/+ and / mice. The magnitude of the response to KC was greater than that seen in the wild-type and urokinase-deficient mice (Figures 1 and 5), which were of a different genetic background. In a manner similar to that seen in wild-type and urokinase-deficient mice, prior administration of GFD-Fc, but not IgG, attenuated the neutrophil influx in uPAR +/+ mice. In contrast, GFD-Fc did not abrogate the influx of neutrophils in mice lacking functional uPAR, verifying that the effect of the GFD fusion protein is dependent upon the presence of uPAR.

View larger version (26K): [in this window] [in a new window]   Figure 6.   GFD-Fc does not attenuate KC-induced neutrophil influx in uPAR-deficient mice. Administration of IgG, GFD-Fc, and KC were followed by lung lavage as in Figure 1. The bars represent the mean ± SD of five uPAR-deficient or wild-type littermate control animals in each group. *P < 0.05 versus KC alone and the IgG control. IP = intraperitoneal.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

This study is the first to directly demonstrate a nonproteolytic role for uPAR in vivo. Whereas neutrophil emigration into the lung in response to the chemotactic agent KC was similar in wild-type mice and mice deficient in either urokinase or uPAR, the presence of a circulating nonproteolytic, receptor-binding fragment of urokinase attenuated neutrophil influx in a uPAR-dependent manner. This result supports and extends recent observations indicating that uPAR expression facilitates cell migration in vivo. May and colleagues (19) reported that uPAR-deficient mice had impaired peritoneal accumulation of neutrophils in response to inflammatory signals. Similarly, several groups of investigators have reported that chimeric fusion proteins similar to that used in the current study interfere with tumor cell metastasis and, in some cases, tumor-associated angiogenesis (17, 21). None of these data, however, address the possibility that uPAR can function in vivo independently of urokinase activity. Our data extend these studies by demonstrating that engagement of uPAR by GFD impairs the in vivo response to a potent murine chemotaxin, KC, in urokinase-deficient mice to a similar degree as that of wild-type mice. Thus, whereas displacement of active urokinase from its surface receptor may impair cell migration through physical barriers, engagement of uPAR per se by a nonproteolytic agonist exerts functional consequences on neutrophil migration in vivo.

The mechanism by which uPAR occupancy attenuated KC-induced neutrophil emigration into the lung is uncertain. Turnover of uPAR from the cell surface is mediated by the low-density lipoprotein/₂-macroglobulin receptor (LRP) and is important in cellular detachment and migration (29). This turnover is promoted by binding of plasminogen activator inhibitor type 1 (PAI-1) to uPAR-bound urokinase (12). As GFD lacks a binding site for PAI-1, uPAR/urokinase complexes would be expected to be cleared less rapidly from the cell surface, thus promoting an adhesive cellular phenotype. Indeed, we previously demonstrated that the presence of inactive urokinase, incapable of binding PAI-1, promoted stable cellular adhesion when compared with active urokinase (32). The decreased recovery of neutrophils in lung lavage fluid could then be related to increased adhesion in the interstitial and/or alveolar spaces of the lung. Careful scrutiny of multiple histologic sections of lung, however, as seen in the photomicrographs (Figure 2), indicates that there are not increased numbers of neutrophils in these lung compartments. While it is possible that neutrophils have been sequestered elsewhere in the body, this seems an unlikely explanation of our findings.

We believe it more likely that occupancy of uPAR by an agonist that is inefficiently cleared from the cell surface interferes in some way with integrin function. Prior studies of urokinase-mediated chemotaxis in vitro showed that high levels of urokinase or ATF impaired rather than promoted myelomonocytic cell migration across collagen-coated porous membranes (33). Similarly, we previously showed that the integrin Mac-1 (CD11b/CD18) and uPAR are functionally linked on monocytes (34). Saturation of uPAR with ATF inhibited postreceptor binding events attributable to Mac-1, including adhesion to fibrinogen. uPAR has been shown to physically associate with Mac-1, a ₂ integrin, as well as ₁ integrins (35). The molecular basis for these in vitro observations is unclear but could stem from the presence of uPAR in one or more multiprotein membrane complexes capable of emitting signals in response to matrix attachment, integrin clustering, or uPAR occupancy by urokinase (36, 37). Integrin function, especially ₂ integrin function, is known to be important to neutrophil accumulation at inflammatory sites (19), although isolated Mac-1-deficient mice do not show impaired neutrophil migration in vivo (38). Elucidation of the exact mechanism by which ATF impairs uPAR function in vivo will require further study.

It should be noted that whereas engagement of uPAR by GFD impairs neutrophil emigration in response to KC, the complete absence of uPAR has no apparent effect on this emigration (Figure 6). There are several possible explanations for this observation. Administration of GFD may result in uPAR-mediated alterations in cellular signaling and/or adhesiveness, accounting for the observed results, whereas uPAR itself, in the absence of ligand binding, has no role in cellular migration. Alternatively, it is possible that redundant and/or adaptive mechanisms compensate for the absence of uPAR. Finally, and more likely, is the possibility that uPAR has a nonessential, yet integral, role in neutrophil emigration. Occupancy of uPAR with GFD may then interfere with signaling events important to cellular migration in the presence of uPAR as suggested above. There is precedent for this hypothesis in observations from Mac-1-deficient mice (38), and P-selectin- and ICAM-1-deficient mice (39).

Cellular migration requires a delicately orchestrated series of events, including cellular signaling, mobilization of the cytoskeleton, and attachment to and detachment from the substrate (40). Observations that uPAR is localized to the leading edge of migrating cells (8, 9), associates with integrins (28, 34, 35, 41), and mediates ligand-induced signaling important to migration (37, 42, 43) implicate this receptor in migration and identify this receptor as a potential target for therapeutic intervention in inflammatory states characterized by excessive neutrophil accumulation and in tumor progression (12). Our data indicate that uPAR may promote cellular migration by both protease-dependent and protease-independent mechanisms, consistent with its role as a multifunctional receptor. This conclusion has implications for future attempts to regulate uPAR function in vivo. Prior attempts to modulate cellular migration by affecting uPAR function have been based on the presumption that inhibition of uPAR-dependent cellular urokinase activity by reagents that disrupt binding of urokinase to uPAR would be the most efficacious approach. The demonstration that uPAR regulates cellular migration independently of urokinase activity suggests that the use of uPAR-binding ATF may not be the best strategy for blocking uPAR function in vivo. In vitro, ATF has a lesser effect on uPAR-dependent cell migration of neutrophils and smooth muscle cells than do reagents (peptides) that interfere with uPAR-integrin associations (H. Chapman, unpublished observations). Antibodies against uPAR also appear to have a more profound affect on cell migration in vitro than does simple uPAR occupancy by urokinase or ATF (28). Our in vivo data raise the possibility that future attempts to attenuate inflammation or tumor cell migration by interfering with uPAR function may be more successful if reagents that not only block urokinase binding but also inhibit uPAR-dependent integrin function are used.