INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Interstitial lung diseases comprise a diverse group of diseases that cause disruption of alveolar structures, often resulting in pulmonary fibrosis with typical clinical, radiographic, and physiological consequences. Unfortunately, these disorders are progressive and refractory to the limited number of therapeutic options available to the clinician. A number of animal models, including the bleomycin-induced murine model of pulmonary fibrosis, have been developed to better understand the underlying pathophysiology and to test new therapeutic approaches to this disorder (1).

Injury to the alveolar epithelium appears to play a critical role in the pathogenesis of interstitial lung disease (2). In fact, there has been speculation that the extent of epithelial injury is a key determinant of the extent of fibrosis that develops in the lungs (3). The procoagulant/fibrinolytic balance in the bronchoalveolar space is important in the lung repair process (4, 5), and alveolar epithelial cells are key contributors to that balance. Alveolar epithelial cells express urokinase (u-PA), urokinase receptor, and plasminogen activator inhibitor type 1 (6, 7). Thus, their proliferative capacity and fibrinolytic potential, coupled with their strategic location at the site of initiation of fibrosis, make the alveolar epithelium an attractive therapeutic target for manipulating the repair process.

Recently, there has been interest in the role of hepatocyte growth factor in the lung repair process. Hepatocyte growth factor (HGF) is produced by mesenchyme and acts upon epithelial tissues via its receptor, c-met (8). In vitro studies have demonstrated that HGF is a potent mitogen for alveolar (9, 10) and bronchial epithelial cells (11) and also stimulates migration and morphogenesis (12). The expression of HGF is markedly increased in the rodent lung after injury with hydrochloric acid (13). HGF levels are also increased in the bronchoalveolar lavage fluid of patients with idiopathic pulmonary fibrosis, sarcoidosis, and the interstitial disease associated with rheumatoid arthritis (14).

Yaekashiwa and coworkers (15) recently demonstrated that continuous delivery of HGF to the systemic circulation suppressed pulmonary fibrosis induced by bleomycin in a murine model. HGF had a significant effect whether it was administered at the time of the injury (i.e., concomitant with bleomycin) or during the fibrotic phase (i.e., 2 wk after bleomycin). Taken together, these results suggest that HGF may have therapeutic potential for pulmonary fibrosis, and this spurred our interest in this area.

In the current study, we investigated the in vivo effects of HGF delivered directly to the mouse lung via the intratracheal route. Furthermore, we explored potential mechanisms of action in vitro using the A549 human epithelial cell line.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Reagents

Recombinant human HGF and sheep anti-serum to rhHGF were kindly provided by Drs. P. J. Godowski and R. H. Schwall (Genentech, Inc., South San Francisco, CA). Ham's F-12 culture medium, glycine hydrochloride, hydrogen peroxide, tranexamic acid, Triton X-100, casein, ethylenediaminetetraacetic acid (EDTA), trans-4-hydroxy-L-proline, citric acid, chloramine-T, normal goat serum, normal sheep serum, penicillin-streptomycin, and amphotericin B were obtained from Sigma Chemical Co. (St. Louis, MO). Bleomycin sulfate was a generous gift from Bristol-Myers-Squibb Pharmaceuticals (Princeton, NJ). Physiological saline and Americlear were from Baxter Healthcare Corp. (Deerfield, IL). Methoxyflurane was purchased from Pitman-Moore Inc. (Mundelein, IL). A549 cell line was obtained from American Type Culture Collection (Parklawn Drive, MD). The 75-cm² cell culture flasks and polyvinylpyrolidine-free polycarbonate filters with 8-µm pores (Nucleopore) were from Costar Corp. (Cambridge, MA). Paraformaldehyde (4%) in phosphate-buffered saline (PBS) was from Richard Allen Medical (Richland, MI). A monoclonal neutralizing antibody to human HGF was from R&D (Minneapolis, MN). Monoclonal antibody to proliferating cell nuclear antigen (PCNA) was purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Biotinylated goat anti-mouse immunoglobulin G (IgG) and rabbit anti-sheep IgG, Vectastain ABC kits for peroxidase and alkaline phosphatase, alkaline phosphatase substrate kit, diaminobenzidine, and methyl green were from Vector Laboratories, Inc. (Burlingame, CA). Ethylene glycol monomethyl ether, p-dimethyl aminobenzaldehyde, and Permount were from Fisher Scientific (Pittsburgh, PA). Methanol was from Mallinckrodt Specialty Chemicals Co. (Paris, Kentucky). Bovine albumin (fraction V) was obtained from ICN Immuno Biologicals (Costa Mesa, CA). Fetal calf serum was from GIBCO BRL (Grand Island, NJ). Automation buffer (10X) was from Biomeda Corp (Foster City, CA). Plasminogen and H-D-Val-Leu-Lys-p-nitroaniline dihydrochloride were from Chromogenix (Molandel, Sweden). Low-molecular-weight u-PA and human plasmin were obtained from American Diagnostica, Inc. (Greenwich, CT). Recombinant human keratinocyte growth factor (KGF) was obtained from Collaborative Biomedical, Becton Dickinson (Bedford, MA). An enzyme-linked immunosorbent assay (ELISA) system for rodent HGF was obtained from the Institute of Immunology, Tokyo, Japan.

Animal Experiments

We used 6-wk-old pathogen-free male C57BL/6 mice (Jackson Laboratory, Bar Harbor, ME) for these experiments. Intratracheal injection of reagents was carried out by the method of Esposito and coworkers (16). This method of administration was less invasive than tracheostomy and allowed repeated injection of reagents. Mice were anesthetized with methoxyflurane vapor and then mounted in a vertical position using a rubber band hooked over the upper incisor teeth. While holding the tongue anteriorly, the oropharynx was transilluminated, and the vocal cords directly visualized. Bleomycin (0.05625 U/ 20 gr body weight) in 30 µl of 0.9% NaCl was injected through the vocal cords into the trachea with a Hamilton syringe. A pilot dose-ranging study suggested that this was the optimal dosage to induce a moderate degree of fibrosis but cause minimal mortality. On Days 3 and 6 after the initial injection either 100 µg of rhHFG (n = 7 mice), heat-inactivated (100° C for 3 min) rhHGF (n = 3 mice), or buffer/saline mixture (n = 8 mice) was injected. The stock solution of rhHGF (Lot 21258-63, 4.6 µg/µl in 20 mM Tris-HCl, pH 7.5, 0.5 M NaCl) was diluted with 0.9% NaCl to a final concentration of 2.5 µg/µl, and 40 µl was administered with each injection. In control mice (n = 10 animals), 30 µl of saline was injected on Day 0 instead of bleomycin, and 40 µl of buffer/ saline mixture was injected on Days 3 and 6. The mice were sacrificed on Day 9 and their lungs analyzed by routine histology and hydroxyproline assays.

Histology and Immunohistochemistry

After perfusion of the lungs with PBS, lung tissue was fixed by instilling 4% paraformaldehyde in PBS through the trachea. The lungs were then cut out and submerged in the same fixative. The tissue was embedded in paraffin, and 4-µm sections of the lungs were placed on glass slides for further analyses. Sections were stained with hematoxylin-eosin and trichrome stain for routine histological examination. The extent of pulmonary fibrosis was evaluated according to the scoring system of Ashcroft and coworkers (17). The entire lung section was reviewed at a magnification of ×200. For each of the 30 to 35 microscopic fields needed to review the section, a score ranging from 0 (normal lung) to 8 (total fibrosis) was assigned. The mean score of all fields was taken as the fibrosis score of that lung section.

For immunohistochemical studies with anti-rhHGF sheep serum, the tissue was deparaffinized in Americlear and rehydrated with decreasing concentrations of ethyl alcohol. The slides were incubated at 94° C for 10 min and then treated with proteinase K (50 µg/ml in PBS) at 37° C for 10 min. The slides were rinsed with 0.1 M glycine-HCl (pH 2.2) for 20 min and then with automation buffer for an additional 20 min at room temperature. After blocking with 10% normal rabbit serum, the primary antibody was applied at a dilution of 1:200 and the slides incubated at 37° C for 30 min. After washing with PBS, biotinylated anti-sheep IgG antibody was applied and the slides incubated at 37° C for 30 min. After washing, avidin-biotin alkaline phosphatase complex was applied and the slides incubated at 37° C for 30 min, followed by the addition of the substrate solution. Color development was stopped by rinsing the slides in distilled water for 5 min. The slides were counterstained with neutral red. As a control for nonspecific staining, sheep serum (1:200) was applied instead of anti-HGF serum.

Determination of Lung HGF Content

Endogenous HGF content in the lung was measured by an ELISA (15). Intratracheal injection of bleomycin (n = 12) or saline (n = 6) was performed as described above. The mice were sacrificed on Day 14 and the lungs perfused to clear with saline. The lungs were then homogenized in four volumes of a buffer composed of 20 mM Tris-HCl (pH 7.5), 2 M NaCl, 0.1% Tween 80, 1 mM phenylmethylsulfonyl fluoride (PMSF), and 1 mM EDTA. The homogenate was centrifuged at 19,000 × g for 30 min at 4° C. The supernatant was used for quantitation of HGF, which was expressed as nanogram per gram wet lung tissue.

Evaluation of Epithelial Cell Proliferation in the Lung

For these experiments, a single dose of 100 µg of rhHGF or buffer/saline mixture was injected intratracheally into naive 6-wk-old mice. Mice were sacrificed 1, 2, 3, and 5 d after the injection and their lungs prepared for immunohistochemical analysis. Pulmonary epithelial cell proliferation was assessed by staining tissue sections for PCNA (18). After rehydration, slides were boiled in 0.05 M citric acid (pH 6.0) for 7 min. After cooling down to room temperature, endogenous peroxidase activity was blocked by incubating the slides in 3% H₂O₂ in methanol for 60 min at room temperature. The slides were next treated with blocking solution containing 5% normal goat serum, 2% casein, and 3% bovine albumin for 1 h, then the primary antibody (3.3 µg/ml) was applied to the tissue sections and incubated at 4° C overnight. Slides were washed with PBS and then biotinylated goat anti-mouse IgG antibody was applied and the sections incubated at 37° C for 30 min. After washing with PBS, avidin-biotin peroxidase complex was added and incubated at 37° C for 30 min. Diaminobenzidine solution was then applied. Color development was stopped by rinsing the slides with distilled water, and the slides were counterstained with methyl green. The reader was masked as to treatment group. Cells in which there was clear-cut nuclear staining were judged to be proliferating cells. For each lung, at least 1,000 bronchial epithelial cells and alveolar epithelial cells were counted. Proliferation rate was calculated by deriving a ratio of labeled cells to total cells counted (19).

Hydroxyproline Assay

For an evaluation of lung collagen content, hydroxyproline was measured by modifications of the previously described colorimetric method (20). After perfusion of the lungs with PBS, the left lung was cut out and hydrolyzed in 2 ml of 6 N HCl at 110° C for 14 h. The pH of the samples was then adjusted to between 6 and 7. A 2-ml aliquot from the total sample volume of 20 ml was added to 1 ml of 1.4% chloramine T and incubated at room temperature for 20 min, then 1 ml of 3.15 M perchloric acid was added and incubated for 5 min, followed by the addition of 1 ml of Erlich's solution (1 M p-dimethylaminobenzaldehyde in ethylene glycol monomethyl ether). After incubation at 65° C for 20 min, absorbance was measured at 562 nm with a Molecular Devices UV MAX plate reader. The assay was carried out in quadruplicate for each lung sample. The amount of hydroxyproline was determined by comparison with a standard curve prepared from known concentrations of reagent hydroxyproline.

Cell Surface Plasmin Generation and u-PA Activity

A549 human pulmonary epithelial cells were cultured in Ham's F-12 medium supplemented with 10% heat-inactivated fetal calf serum, L-glutamine (2 mM), penicillin (100 U/ml), steptomycin (100 µg/ml), and amphotericin B (250 µg/ml). For cell surface plasmin generation assays, confluent cells in 12-well plates were cultured in serum-free medium for 16 h and then treated with rhHGF (50 ng/ml), or no additive for 24 h. The medium was discarded, and then the cells were washed in PBS and incubated in serum-free Ham's F-12 medium containing plasminogen (20 µg/ml) for 3 h. The supernatants were collected and the cells again washed with PBS. Cell surface bound plasmin was then dissociated from the cell membrane by incubation of the cells in 250 µl of 1 mM tranexamic acid for 15 min at 37° C. The samples were assayed for plasmin activity with the chromogenic substrate, H-D-Val-Leu-Lys-p-nitroanilide (0.01 mM) in 0.1 M Tris, pH 8.0, as previously described (21). Briefly, the amount of p-nitroaniline released was measured at 410 nm with a Molecular Devices UV MAX plate reader and referenced to a plasmin standard curve. Data were expressed as mU of plasmin activity/well. Each experiment was performed in quadruplicate.

For the uPA activity assays, confluent cells in 12-well plates were incubated in serum-free medium for 16 h and then treated with rhHGF (10 and 100 ng/ml), rhKGF (10 and 100 ng/ml), or no additive for 24 h. Supernatants and lysates, obtained by the addition of 1% Triton X-100 in PBS on ice for 1 h, were then assayed for u-PA activity. Aliquots of samples were assayed in duplicate in 96-well plates with a total assay volume of 125 µl of 1% Triton X-100 in PBS containing 0.15 µM plasminogen and 1.5 mM H-D-Val-Lys-p-nitroanilide. The amount of p-nitroaniline released was measured at 410 nm with a Molecular Devices UV MAX plate reader and referenced to a u-PA standard curve prepared with low-molecular-weight human u-PA. Data were expressed as mU of u-PA activity/well. Each experiment was carried out in quadruplicate.

Cell Migration Assay

Cells were grown to confluence in complete medium, washed with PBS three times, and changed to serum-free medium. Fresh serum-free medium was added 12 h prior to the experiment. Trypsin-EDTA was used to prepare a cell suspension. After washing in serum-free medium, 45 µl of the cell suspension (1 × 10⁶ cells/ml in Ham's F-12) was placed in the upper wells of Boyden chambers and 25 µl of HGF (10, 30, and 100 ng/ml) or KGF (10, 30, and 100 ng/ml), diluted in Ham's F-12 medium, was placed in the lower wells. The upper and lower wells were separated by polyvinylpyrrolidone-free polycarbonate filters with 8-µm pores. After incubation for 6 h at 37° C in 95% room air/ 5% CO₂, the filters were removed and the top surface scraped to remove adherent cells. The filters were then stained in Diff-Quik, and the numbers of cells that migrated through the filters to the bottom surface were counted per high-power field (×400). Each assay was performed in triplicate.

Statistical Analysis

All values are expressed as the mean ± SEM. Groups were compared with a two sample t test or one-way ANOVA. p Values < 0.05 were considered to be significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In Vivo Expression of HGF after Bleomycin-induced Injury

To explore the potential role of HGF in the lung repair process, we first investigated the expression of endogenous HGF in the bleomycin-induced lung injury model using an immunohistochemical approach. We found significant and specific staining for HGF in areas of developing fibrosis (Figure 1). In contrast, uninvolved areas of the lung showed only slight staining of the interstitium, identical to our findings in control mice (data not shown). To confirm this increase in immunoreactivity, we performed an ELISA. We found that HGF levels were significantly higher in the lungs of bleomycin-treated mice (239.1 ± 20.1) as compared with saline-treated controls (107.4 ± 10.0), p < 0.002. 

View larger version (129K): [in this window] [in a new window]   View larger version (143K): [in this window] [in a new window]   Figure 1.   Expression of hepatocyte growth factor (HGF) in bleomycin-induced pulmonary fibrosis. Light microscopy of mouse lung on Day 14 after intratracheal bleomycin injection (0.05625 U). (A) Anti-HGF serum (1:200). (B) Normal sheep serum (1:200). Original magnification: ×200. Note the significant blue staining for HGF in A. The lack of staining in B confirms the specificity of this finding.

Exogenous Administration of HGF Induces Pulmonary Epithelial Cell Proliferation

We next examined the effect of an intratracheal injection of HGF on bronchial and alveolar epithelial cell proliferation in normal mouse lungs. At baseline, the proliferation rate of bronchial epithelial cells and alveolar epithelial cells was low, 0.7 ± 0.1% and 1.0 ± 0.2%, respectively, compatible with previous reports (19). A single intratracheal injection of HGF induced a significant increase in bronchial (Days 2 and 3) and alveolar (Days 1, 2, and 3) epithelial cell proliferation as compared with control mice (Figure 2). The maximal increase in PCNA labeling was observed on Day 2 after the HGF injection with a return to near baseline level by Day 5. 

View larger version (10K): [in this window] [in a new window]   View larger version (10K): [in this window] [in a new window]   Figure 2.   Effect of rhHGF on proliferation of epithelial cells in normal mouse lung. rhHGF (100 µg) or saline was injected into the lungs of the normal mice. The animals were sacrificed 1, 2, 3, or 5 d afterward and lung sections were prepared and stained with a monoclonal antibody to PCNA as described in METHODS. The proliferation rate was calculated by counting at least 1,000 cells for each sample (n = 4). (A) Bronchial epithelial cells. (B) Alveolar epithelial cells. HGF induced a significant increase in the proliferation of both bronchial and alveolar epithelial cells as compared with the saline-injected group (*p < 0.05).

Exogenous Administration of HGF Attenuates Collagen Accumulation in Bleomycin-induced Lung Injury

With the knowledge that HGF expression is markedly increased after lung injury and that exogenous administration of HGF stimulates epithelial cell proliferation, we postulated that HGF might favorably impact the lung repair process. To test that hypothesis, we explored the effect of HGF on bleomycin-induced pulmonary fibrosis using the experimental protocol described above. The body weights of the mice suggested a potential benefit of the HGF treatment. On Day 9, the saline/buffer-treated group weighed 22.3 ± 0.4 g, the bleomycin/buffer-treated group weighed 18.8 ± 0.3 g, and the bleomycin/HGF-treated group weighed 20.4 ± 0.7 g (p = 0.05 when comparing the bleomycin/buffer and bleomycin/HGF groups). The hydroxyproline contents of the left lung of the saline/buffer- and bleomycin/buffer-treated groups was 76.8 ± 16.0 µg, and 124.9 ± 11.9 µg, respectively (Figure 3). HGF administration resulted in a 53.4% inhibition of the bleomycin-induced increase in hydroxyproline content to 99.2 ± 10.5 µg. The difference between the bleomycin/buffer group and bleomycin/HGF treated groups was highly significant (p < 0.001). When heat-inactivated (100° C for the 3 min) rhHGF was administered in an identical regimen, there was no reduction in hydroxyproline content as compared with the bleomycin/ buffer group (n = 3, data not shown), indicating that the HGF effect was not simply due to a nonspecific protein effect. HGF treatment also resulted in a modest improvement in the histologic severity of fibrosis. The mean fibrosis scores for the bleomycin/buffer and bleomycin/HGF groups was 2.31 ± 0.07 and 1.95 ± 0.15, respectively (p = 0.037).

View larger version (12K): [in this window] [in a new window]   Figure 3.   Effect of rhHGF on lung hydroxyproline content. As expected, the hydroxyproline content of the bleomycin group was significantly higher than the control saline-treated group (*p < 0.001). This increase was significantly attenuated in the HGF-treated group (** p < 0.001 when compared with the bleomycin group).

HGF Increases Cell Surface Plasmin Generation and u-PA Activity Cells

To further explore the potential mechanisms of HGF's impact on the lung repair process, we examined its effects on the fibrinolytic system utilizing the A549 human pulmonary epithelial cell line. We found that rhHGF (60 ng/ml) increased cell surface plasmin generation as compared with untreated controls, (HGF = 2.6 ± 0.15 mU/well of plasmin activity versus control = 1.7 ± 0.2 mU/well, p < 0.05). This increase in plasmin generation was not due to an increase in cell number because there was no significant cell proliferation in the confluent culture conditions used for these experiments (data not shown). Increased u-PA expression is one potential explanation for the HGF-induced increase in cell surface plasmin. We found that HGF, but not KGF, stimulated a significant increase in u-PA activity (Figure 4).

View larger version (43K): [in this window] [in a new window]   Figure 4.   Effect of HGF and keratinocyte growth factor (KGF) on u-PA activity. Supernatants and lysates were collected 24 h after the addition of rhHGF (10, 100 ng/ml) or rhKGF (10, 100 ng/ml) to A549 cell cultures. u-PA activity was quantified as described in METHODS. HGF induced a significant increase in u-PA activity at 10 and 100 ng/ml (*p < 0.05, and **p < 0.01, respectively), whereas KGF decreased u-PA activity.

HGF Stimulates Cell Migration

HGF also stimulated A549 cell migration with a maximal effect at a concentration of 30 ng/ml (Figure 5). A checkerboard analysis demonstrated that this effect was chemotactic and not chemokinetic (data not shown). Preincubation with a neutralizing anti-HGF antibody blocked this effect. In contrast, KGF in concentrations ranging from 10 to 100 ng/ml did not stimulate cell migration.

View larger version (32K): [in this window] [in a new window]   Figure 5.   Effect of HGF on cell migration. A549 cells were incubated with HGF (10, 30, and 100 ng/ml) as described in METHODS. Fibronectin (FN) (10 ng/ml) was also included in the experiment as a positive control. Data shown are number of cells per 10 high-power fields (×400) (mean ± SEM). HGF showed chemotactic activity at 10 (*p < 0.05), 30 (***p < 0.001), and 100 ng/ml (**p < 0.01). Neutralizing antibody to HGF completely blocked the chemotactic activity. KGF at the same concentrations had no effect (data not shown).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Pulmonary fibrosis remains a devastating clinical disorder for which there are limited therapeutic options. A number of experimental approaches have been investigated in animal models including the inhibition of key cytokines and growth factors (22), but to date none of these approaches has come to fruition in the clinic. An alternative approach to modifying the cytokine network might be to stimulate the proliferative and repair capacity of the pulmonary epithelium. Such an approach has been considered with two heparin-binding growth factors, HGF and KGF.

HGF is a multipotent growth factor that has mitogenic, motogenic, and morphogenic effects on various epithelial cells including alveolar epithelial cells (8, 12) The alveolar epithelium plays a critical role in the complex repair process that follows lung injury. We found that intratracheal administration of exogenous rhHGF stimulates bronchial and alveolar cell proliferation in normal mouse lung and attenuates collagen accumulation in a bleomycin-induced model of pulmonary fibrosis. Furthermore, our in vitro experiments revealed that HGF stimulates fibrinolytic capacity and cell migration of A549 cells, potential mechanisms for the favorable impact of HGF on the repair process.

Our findings are generally in agreement with the report of Yaekashawa and coworkers (15), but the differences deserve to be highlighted. Using the same animal model, they delivered HGF continuously into the systemic circulation by an infusion pump. With this approach they were able to demonstrate a marked decrease in fibrosis at the histopathological level at Day 14 and Day 28 after bleomycin. In contrast, we found a significant attenuation of the bleomycin-induced increase in hydroxyproline content in the lungs at Day 9, but the improvement in the histopathological picture was much less dramatic. The difference in findings may simply relate to the earlier time point that we selected for study. Yanagita and coworkers (13) reported that c-met, the HGF receptor, was down-regulated for several days after lung injury induced by hydrochloric acid. If the same temporal expression pattern occurs following bleomycin-induced injury, then the lungs may be relatively refractory to the effects of HGF shortly after the injury and more responsive at later time points. Alternatively, the intratracheal delivery of HGF utilized in our experiments may have resulted in an uneven distribution throughout the lungs, thus minimizing its beneficial effects. In any event, both studies suggest that HGF may have therapeutic potential for the treatment of pulmonary fibrosis.

Despite their similarities, clear differences are emerging between HGF and KGF with respect to their effects on the lungs. KGF is critical in lung development (23). HGF may also play a role in development (24), but mice with targeted disruption of the HGF-like gene display normal embryogenesis and organogenesis (25). Both growth factors stimulate pulmonary epithelial cell proliferation in the normal rodent lung (26, 27). KGF protects against hyperoxic and bleomycin-induced injury when given prior to the injurious agent (28, 29), but it is not clear that postinjury treatment with KGF is effective. In contrast, we found that pretreatment with HGF had no effect on hydroxyproline content after bleomycin-induced injury (data not shown), whereas posttreatment on Days 3 and 6 attenuated the collagen accumulation in that model.

In addition to HGF's in vivo proliferative effect on the pulmonary epithelium, we found other potential mechanisms for its beneficial effect, namely the stimulation of fibrinolytic capacity and cell migration. Other investigators have demonstrated the critical role of the fibrinolytic system in determining the outcome of bleomycin-induced injury (30, 31). The alveolar epithelium, a primary target of HGF, clearly contributes to the fibrinolytic balance in the lungs (6, 7). HGF stimulates fibrinolytic activity in other epithelial cells (32), and likewise, we found that it upregulates u-PA activity and cell surface plasmin generation in A549 cells. In contrast, KGF did not stimulate fibrinolytic activity or cell migration. Of note, HGF is homologous to plasminogen (33), and urokinase has the capacity to cleave it, converting it from the inactive single-chain form to the active two-chain form (34) This provides a mechanism whereby the effect of HGF may be further amplified.

The in vivo effects of HGF may be critically dependent upon interactions with heparin and heparan sulfate proteoglycons. Heparin and heparan sulfate facilitate the interactions between HGF and its receptor (35). Matrix-bound stores of HGF exist in vivo as evidenced by the fact that the administration of heparin increases HGF levels (36). Indeed, one might speculate that the beneficial effects of heparin on lung repair (37) might be mediated in part by HGF. A better understanding of the interactions between HGF and heparin/heparan sulfate may provide insights into the therapeutic potential of this growth factor.

In summary, our results support a potential important role for HGF in the lung repair process. Further investigation is warranted to explore whether this growth factor will prove useful as a therapeutic agent for pulmonary fibrosis.