Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Airway fibrosis is a feature of several environmentally related pulmonary diseases, including asthma and chronic bronchitis. During the progression of airway fibrosis in humans (1) and experimental animals (2), increased numbers of peribronchiolar fibroblasts and myofibroblasts deposit collagen that contributes to thickening of the airway wall. We previously described a model of airway fibrosis in rats induced by the intratracheal instillation of vanadium pentoxide (V₂O₅) (2), a transition metal that causes occupational asthma in workers in the petrochemical industry (3, 4). Moreover, trace amounts of vanadium compounds are present in air pollution particulate matter that cause airway hyperresponsiveness and inflammation (5, 6). Although metals associated with air pollution particles have been shown to induce the production of proinflammatory cytokines by human bronchial epithelial cells (HBECs) (7), the identity of cytokines and growth factors that mediate the proliferation of fibroblasts surrounding the airways is unknown.

Bronchial epithelial cell conditioned medium (CM), when fractionated by gel filtration chromatography, contains both growth-promoting and growth-suppressing peaks of activity for fibroblasts, yet the overall effect of the CM for fibroblasts is promitogenic (8, 9). Nakamura and coworkers reported that the growth-suppressive activity in the epithelial CM was due in part to transforming growth factor (TGF)- (8). TGF-1 is spontaneously produced by cultured bronchial epithelial cells (9), and is upregulated in the airway epithelium in patients with chronic obstructive pulmonary disease (10). Although TGF-1 appears to be a major growth inhibitory factor produced by bronchial epithelial cells for fibroblasts, the identity of the factor(s) that account for the growth-stimulatory activity in airway epithelial CM remains unclear. Insulin-like growth factor (IGF)-1, a cell cycle progression factor, has been reported to partially account for the fibroblast growth-promoting activity in human airway epithelial cell CM (11). The potential contribution of a variety of other growth factors in mediating airway epithelial-stimulated lung fibroblast growth has not been investigated. These growth factors include platelet-derived growth factor (PDGF) isoforms (12, 13), fibroblast growth factors (FGFs) (14), and members of the epidermal growth factor (EGF) family (15).

In this study we have investigated several growth factors as possible mediators of human lung fibroblast (HLF) mitogenesis stimulated by CM from normal HBECs (NHBECs). Fibroblast DNA synthesis stimulated by NHBEC-CM was blocked by a receptor tyrosine kinase inhibitor selective for the EGF receptor (EGF-R), but not by an inhibitor specific for PDGF receptor (PDGF-R). Experiments with neutralizing antibodies against several growth factors, including three members of the EGF family, identified heparin-binding EGF-like growth factor (HB-EGF) as a major NHBEC-derived mitogen for HLF. The mitogenic activity of HB-EGF was due in part to initiation of an FGF-2 autocrine loop. HB-EGF messenger RNA (mRNA) expression detected by reverse transcriptase/polymerase chain reaction (RT-PCR) in NHBECs was induced by the fibrogenic metal V₂O₅, but not by the inert metal titanium dioxide (TiO₂). Induction of HB-EGF mRNA by V₂O₅ was blocked by inhibitors of EGF-R tyrosine kinase (AG1478), p38 mitogen-activated protein (MAP) kinase (SB203580), and MAP kinase kinase (MEK) (PD98059). Together, these data suggest that HB-EGF and FGF-2 play a major role in fibroblast proliferation mediated by an injured airway epithelium.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents

V₂O₅ and TiO₂ were purchased from Aldrich Chemical (Milwaukee, WI). Tyrphostins AG1296 and AG1478 were purchased from Calbiochem (La Jolla, CA). Antihuman PDGF, FGF-2, IGF-1 EGF, HB-EGF, and TGF- neutralizing antibodies were purchased from R&D Systems (Minneapolis, MN). Recombinant human PDGF-BB, EGF, TGF-, IGF-1, and FGF-2 were purchased from Upstate Biotechnologies (Lake Placid, NY) or R&D Systems. Human recombinant HB-EGF was purchased from Santa Cruz Biotechnologies (Santa Cruz, CA) or R&D Systems. MEK inhibitor PD98059 was from New England Biolabs (Beverly, MA), and p38 MAP kinase inhibitor SB203580 was from Calbiochem Corp.

NHBEC Culture

Primary (passage 2) NHBECs were purchased from Clonetics Corp. (San Diego, CA). The cells were grown in Clonetics expansion medium consisting of BEBM medium and BEGM Singlequots supplemented with 10⁸ M retinoic acid, 25 ng/ml EGF, and bovine pituitary extract. The medium was changed every other day during the first week of culture and then every day until the cells were confluent. NHBECs were trypsin-liberated and cryopreserved for all further experiments. Submerged, undifferentiated cultures were established according to the method of Gray and coworkers (18) wherein passage 3 cells were seeded on plastic dishes at a density of 2,000 cells/cm² and fed with expansion medium until confluent. For collection of CM from submerged cultures that was used in mitogenesis assays, the cells were rinsed twice with serum-free defined medium (SFDM) (consisting of Ham's F-12 with CaCl₂ and N-2-hydroxyethylpiperazine-N'-ethane sulfonic acid [Hepes], supplemented with 0.25% bovine serum albumin [BSA] [Sigma, St. Louis, MO] and an insulin/transferrin/selenium mixture [Calbiochem]) and incubated in SFDM for 24 h. In experiments where NHBECs were treated with metals, the cells were rinsed twice with SFDM, then treated with the desired concentration of metal for 2 h, washed with SFDM, and incubated in fresh SFDM for 24 h. This strategy allowed for removal of the metal following NHBEC activation to prevent direct stimulation of fibroblasts in the mitogenesis assays described later. After collection of CM, cell debris was removed by centrifugation and aliquots of CM were stored at 80°C.

[³H]Thymidine Incorporation Assay using HLFs

HLFs (16 Lu) were purchased from American Type Culture Collection (Rockville, MD). HLFs (10⁶ cells) were seeded into 175-cm² plastic culture dishes and grown to confluence in 10% fetal bovine serum/Dulbecco's modified Eagle's medium, then trypsin-liberated and seeded into 24-well plates at a density of 2 × 10⁴ cells/cm². Once confluent, the cells were rendered quiescent for 24 h in SFDM. The cultures were rinsed and treated with recombinant growth factor (10 ng/ml unless otherwise indicated in the figure captions) or NHBEC-CM (1:1 dilution in SFDM) along with 5 µCi/ml [³H]thymidine (Amersham, Arlington Heights, IL) for 36 h. The cells were washed with Ham's F-12 at 25°C, placed on ice, and incubated with 0.5 ml/well 5% trichloroacetic acid (TCA) for 10 min. After washing three times with ice-cold distilled water, solubilization was performed with 0.5 ml/well 0.2 N NaOH containing 0.1% sodium dodecyl sulfate (SDS) for 30 min on an oscillating platform. A total of 100 µl of each sample was added to 1 ml of Ecolume (ICN, Costa Mesa, CA) and radioactivity was measured on a liquid scintillation counter. In experiments with tyrosine kinase inhibitors, the cells were treated with 100 µM AG1296 or AG1478 in vehicle (dimethyl sulfoxide) or vehicle alone for 1 h before the addition of recombinant growth factor or NHBEC-CM. In experiments with neutralizing antibodies, recombinant growth factors (10 ng/ml) or NHBEC-CM (1:1 dilution in SFDM) were incubated for 1 h at 37°C with 20 µg/ml neutralizing antibody (unless otherwise indicated in the figure captions).

RT-PCR

Total RNA from cultured NHBECs was extracted with TRI Reagent (Molecular Research Center, Cinnicinati, OH). To induce HB-EGF expression, NHBECs were exposed to metals for 3 h or pretreated with metabolic inhibitors or antioxidants for 1 h before metal exposure. RT-PCR was used to amplify a 750-base pair (bp) HB-EGF complementary DNA (cDNA) fragment from NHBECs that corresponded to bases 282-1035 of the published human HB-EGF cDNA sequence (19). Primer pairs were custom-designed by Life Technologies, Inc. (Gaithersburg, MD). The forward HB-EGF primer (24-mer) sequence was 5' GGT GCT GAA GCT CTT TCT GGC TGC 3'. The reverse HB-EGF primer (25-mer) was 5' ATT ATG GGA GGC CCA ATC CTA GAC G 3'. Oligonucleotide amplimers for -actin (which was used as the control gene for RT-PCR reactions) generated a 308-bp PCR cDNA fragment. The forward -actin primer (21-mer) sequence was 5' ATC GTG GGC CGC CCT AGG CAC 3'. The reverse -actin primer (22-mer) sequence was 5' TGG CCT TAG GGT TCA GAG GGG C 3'. Amplification was carried out using the GeneAmp RNA PCR Core kit from Perkin Elmer using a Perkin Elmer Cetus DNA thermal cycler according to the manufacturer's instructions (Perkin Elmer, Branchburg, NJ). Total RNA (1 µg per 20-µl reaction volume) was reverse transcribed into cDNA using random primers. A total of 40% for HB-EGF or 4% for -actin of the resulting cDNA was amplified using 0.2 µM of each primer. Denaturation was carried out at 95°C for 1 min. Annealing temperature was 55°C for 1 min, and extension was performed at 72°C for 1 min. PCR products were separated by electrophoresis in a 2% Seakem agarose gel (FMC, Rockland, ME) containing 50 ng/ml ethidium bromide and photographed with Polaroid type 55 film, the negative scanned on a Molecular Dynamics Densitometer (Molecular Dynamics, Sunnydale, CA), and the signal quantitated using the NIH 1.61 Image Program (NIH, Bethesda, MD). HB-EGF densitometric measurements were normalized against the corresponding -actin signal. The linear range for the PCR was established by plotting the intensity of signal versus PCR cycle number. The linear range for HB-EGF was 25 to 35 cycles. The 750-bp PCR product was identified in relation to a 250- to 3,500-bp ladder (Life Technologies). To verify that the amplified products were from mRNA and not genomic DNA contamination, negative controls were performed by omitting the RT from the RT reaction. In the absence of RT, no PCR products were observed.

Northern Blot Analysis

NHBECs were grown to ~ 70 to 80% confluency in 175-cm² plastic culture flasks, then rendered quiescent in SFDM for 24 h before treating for 3 h with 10 µg/cm² V₂O₅, 10 µg/cm² TiO₂, or fresh SFDM alone (control). Total RNA was isolated with TRI reagent (Molecular Research Center). A total of 20 µg of each sample was electrophoresed in 1% agarose/formaldehyde gels and capillary transferred onto BrightStar-Plus positively charged nylon membranes (Ambion, Inc., Austin, TX). A 2.36-kb human HB-EGF cDNA probe was kindly provided by Dr. Judith Abraham (Scios, Inc., Sunnyvale, CA). The probe was labeled with [-³²P]deoxycytidine triphosphate using a DECAprime II DNA labeling kit (Ambion). The hybridization and washing procedure for blotting was performed with a Northern Max-GLY Kit according to the supplied protocol (Ambion). The autoradiographic signal was visualized by exposing the film at 70°C for the appropriate time. Membranes were stripped and reprobed using a glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA as a housekeeping gene.

Western Blot Analysis

Confluent cell monolayers in 100-mm dishes were growth arrested in serum-free medium for 24 h before treatmtent with V₂O₅ or metabolic inhibitors. Cells were washed once with phosphate-buffered saline (PBS) on ice and 200 µl of lysis buffer (50 mM Hepes; 150 mM NaCl; 1% Triton X-100; 1 mM phenylmethylsulfonyl fluoride; and 20 µg/ml aprotinin, leupeptin, and pepstatin) was added to the monolayers. After 20 min at 0 to 4°C, the cell lysates were removed with no scraping. A total of 30 µg of protein per sample was separated by SDS-polyacrylamide gel electrophoresis, transferred to nitrocellulose membrane, and blocked for 2 h at 25°C with 0.5% nonfat milk in PBS buffer (20 mM Tris, 500 mM NaCl, and 0.01% Tween 20). The membrane was then incubated overnight at 4°C with a 1:1,000 dilution of monoclonal anti-phospho-EGF-R or anti-EGF-R polyclonal (Upstate Biotechnologies, Inc., Lake Placid, NY), followed by incubation for 2 h with a 1:2,000 dilution of appropriate horseradish peroxidase- conjugated secondary antibody. The immunoblot signal was visualized through enhanced chemiluminescence.

HB-EGF Enzyme-Linked Immunosorbent Assay

For detection of secreted HB-EGF protein, 175-cm² flasks of NHBECs were treated for 24 h with 30 ml of SFDM (containing low BSA, 0.01%) alone or supplemented with 10 µg/cm² V₂O₅ or TiO₂. A total of 30 ml of NHBEC-CM was collected, clarified by centrifugation, and concentrated to 300 µl using Centriplus-10 concentrators (Amicon, Beverly, MA). As a control, medium alone that was not incubated with cells was concentrated in the same manner. Serial dilutions of human recombinant HB-EGF (0.03 to 128 ng/ml) or NHBEC-CM were added 60 µl/well in 96-well immulon-4 plates (Dynatech, Chantilly, VA) and incubated overnight at 4°C. After washing the plate four times with PBS-Tween, 200 µl of blocking buffer (3% BSA in PBS-Tween) was added and the plate incubated for 1.5 h at 37°C. After washing the wells four times in PBS-Tween, 100 µl/well of 1 µg/ml goat antihuman HB-EGF (R&D Systems) was added and the plate incubated overnight at 4°C. The wells were washed again four times and 100 µl/well of 1:50,000 biotinylated rabbit antigoat immunoglobulin (Ig) G (Jackson ImmunoResearch, West Grove, PA) in blocking buffer was added for 1 h at room temperature. The wells were washed and 100 µl of 1:1,000 streptavidin-alkaline phosphatase (Jackson ImmunoResearch) in blocking buffer was added for 1 h at room temperature. After washing the wells six times, the enzyme-linked immunosorbent assay (ELISA) was developed with an alkaline-phosphatase substrate kit (Bio-Rad, Hercules, CA). The reaction was stopped by the addition of 50 µl/well 0.3 N NaOH and absorbance was read at 405 nm using a Dynatech MR 5000 microplate reader. The standard curve was linear between 0.03 and 32 ng/ml HB-EGF. NHBEC-CM absorbance values were converted to nanogram-per-milliliter values on the basis of the linear regression transformation of the standard curve.

FGF-2 ELISA

A commercially available FGF-2 ELISA kit (Quantikine high-sensitivity HS) was purchased from R&D Systems, and was used according to the manufacturer's instructions. CM was harvested from confluent, quiescent HLFs in 24-well dishes that had been treated with 1 ng/ml HB-EGF (1 ml SFDM/well) for 24 h. Unconcentrated CM was used in the FGF-2 ELISA.

Cytotoxicity Assay

The cytotoxic effects of V₂O₅ on NHBECs were measured using a CytoTox 96 nonradioactive cytotoxicity assay (Promega, Madison, WI). This kit detects the release of lactate dehydrogenase (LDH) from cells into the culture medium.

Statistical Analysis

The data presented are means ± standard error of the mean (SEM) for three to five experiments. Two sample t tests were performed to determine significant differences among control versus treatment groups (*P < 0.05, **P < 0.01).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

V₂O₅ but Not TiO₂ Increases the Fibroblast Mitogenic Activity in NHBEC-CM

To measure mitogenic activity in NHBEC-CM it was necessary to render NHBECs quiescent for 24 h in SFDM due to the fact that the basal growth medium used to maintain epithelial cells contained a variety of mitogens, including EGF (18). Therefore, NHBECs were rinsed with SFDM and incubated with fresh SFDM for 24 h, and then the NHBEC-CM was collected for [³H]thymidine uptake assays involving exposure of HLFs to a 1:1 dilution of NHBEC-CM in SFDM. A 1:1 dilution of NHBEC-CM caused a statistically significant (P < 0.05) ~ 3-fold increase in [³H]thymidine incorporation into HLF cultures as compared with treatment with SFDM alone (Figure 1). This increase in DNA synthesis was also reflected in an approximate 2-fold increase in fibroblast number as determined by hemocytometer counting of trypsin-liberated cultures 4 d after treatment with NHBEC-CM (data not shown). Exposure of NHBEC to V₂O₅ increased the release of fibroblast growth-promoting activity 2- to 3-fold in a concentration-dependent manner (Figure 1). In these experiments, NHBECs were treated with V₂O₅ for 2 h, then washed with SFDM and incubated with fresh SFDM for another 24 h before collection of NHBEC-CM. This strategy allowed for removal of the soluble metal from the NHBEC cultures to prevent direct metal-induced activation of the fibroblasts in the mitogenesis assay. Under these conditions, 10 µg/cm² V₂O₅ did not cause significant NHBEC cytotoxicity as measured by LDH release, whereas 50 and 100 µg/cm² V₂O₅ caused 10 to 20% and 70 to 80% cytotoxicity, respectively.

View larger version (19K): [in this window] [in a new window]   Figure 1.   Mitogenic effects of CM from V2O5- or TiO2-stimulated NHBECs on HLFs. Cultures of NHBECs were treated with increasing concentrations of V2O5 or TiO2 (1, 10, or 50 µg/ cm2) and the CM was harvested after 24 h and designated as CM (no metal treatment), CM 1 (1 µg/ cm2), CM 10 (10 µg/cm2), or CM 50 (50 µg/cm2). SFDM alone (no CM) served as the control. All samples were added to confluent, quiescent cultures of HLFs in the presence of 5 µCi/ml [3H]thymidine and incubated for 36 h before measuring uptake of radioactivity in TCA-precipitated DNA. Open bars: V2O5-treated cells; filled bars: TiO2-treated cells; hatched bars: no metal treatment. *P < 0.05, significant effect of CM compared with control. P < 0.05, significant effect of CM 10 or CM 50 as compared with CM treatment.

Fibroblast Growth-Promoting Activity in NHBEC-CM Is Due to HB-EGF

To determine whether members of either the EGF or PDGF families were mediating the NHBEC-CM-induced mitogenic response of HLFs, we pretreated HLFs with receptor tyrosine kinase inhibitors of the tyrphostin class that were selective for the PDGF-R (AG1296) or the EGF-R (AG1478) (20). Pretreatment of HLFs with AG1478 (100 µM) completely blocked the mitogenic activity of NHBEC-CM for HLFs, whereas AG1296 had no effect (Table 1). These data suggest that the mitogenic activity present in NHBEC-CM was mediated by a member of the EGF family. We then tested neutralizing antibodies against several EGF family members including EGF, TGF-, and HB-EGF. Anti-HB-EGF neutralizing antibody abolished the mitogenic potential of NHBEC-CM for HLFs (Figure 2). Dose-response experiments showed that a concentration of 20 µg/ml anti-HB-EGF was maximally effective in blocking HLF mitogenesis induced by HBEC-CM (data not shown). Antibodies to two other EGF family members, EGF and TGF-, had no blocking effect on HLF mitogenesis stimulated by NHBEC-CM (Table 2). However, the anti-EGF and anti-TGF- antibodies were effective in neutralizing [³H]thymidine uptake into HLFs stimulated by human recombinant EGF and TGF-, respectively (Table 3). We also observed that neutralizing antibodies to PDGF had no effect on blocking HLF mitogenesis stimulated by NHBEC-CM (Table 2), yet anti-PDGF antibody effectively neutralized human recombinant PDGF-BB (Table 3). Anti-FGF-2 partially inhibited the fibroblast growth-promoting activity in NHBEC-CM (Table 2).

View larger version (27K): [in this window] [in a new window]   Figure 2.   Inhibition of NHBEC-CM stimulated HLF mitogenesis by neutralizing anti-HB-EGF antibody. NHBEC-CM collected from cells treated with 50 µg/cm2 V2O5 (CM V50) was incubated with 20 µg/ml anti- HB-EGF or 20 µg/ml nonspecific IgG for 1 h, then added to confluent cultures of HLFs in the presence of [3H]thymidine as described in MATERIALS AND METHODS. **P < 0.01, significant effect of anti-HB-EGF treatment on CM V50-induced mitogenesis as compared with CM V50 alone.

Differential Effects of V₂O₅ and TiO₂ on HB-EGF mRNA and Protein Expression in NHBECs

RT-PCR was used to amplify a 753-bp HB-EGF cDNA fragment from cultured NHBECs that corresponded to bases 282-1035 of the published human HB-EGF cDNA sequence (19). After 30 amplification cycles, a single 753-bp band representing HB-EGF mRNA was observed (Figure 3). Increasing PCR cycle number demonstrated optical density saturation of the HB-EGF signal above 35 cycles (data not shown). Stimulation of NHBECs with V₂O₅ caused a concentration-dependent increase in HB-EGF mRNA that was nearly maximal at 10 µg/cm², whereas no increase was observed in cells treated with TiO₂ in the same dose range (Figure 3). No change in the expression of -actin was observed with either V₂O₅ or TiO₂. V₂O₅-induced upregulation of HB-EGF mRNA peaked at 3 h after stimulation and returned to a basal level of expression by 24 h after treatment (Figure 4). We also demonstrated induction of HB-EGF mRNA using Northern blot analysis. Treatment of NHBECs for 3 h with 10 µg/cm² of V₂O₅ caused an 8-fold increase in expression of the 2.5-kb HB-EGF transcript, whereas TiO₂ at the same concentration had no effect on HB-EGF gene expression (Figure 5). HB-EGF protein was detected in NHBEC-CM by ELISA (Table 4). No HB-EGF was detected in medium alone, whereas ~ 0.6 ng/ml HB-EGF was detected in CM from unstimulated NHBECs. Stimulation of NHBECs with 10 µg/cm² V₂O₅ caused a 3.5-fold increase in HB-EGF release into the medium, whereas TiO₂ did not significantly increase HB-EGF protein levels in the CM (Table 4). The relatively low concentrations of HB-EGF detected in the NHBEC-CM by ELISA (1 to 2 ng/ml HB-EGF) were biologically relevant, inasmuch as concentrations of recombinant HB-EGF induced an approximate 2-fold increase in [³H]thymidine uptake by human lung fibroblasts (Table 5).

View larger version (32K): [in this window] [in a new window]   Figure 3.   Concentration-dependent induction of HB-EGF mRNA by V2O5 in NHBECs. Cultures of NHBECs were treated with V2O5 or TiO2 at the indicated doses for 3 h before collecting total RNA for RT-PCR. (A) Representative RT-PCR results showing upregulation of HB-EGF mRNA stimulated by 10 and 50 µg/cm2 V2O5 (left, upper panel) but no induction by TiO2 (right, upper panel). RT-PCR results for -actin are shown in the lower right and left panels. (B) Densitometry of RT-PCR results showing HB-EGF signal normalized against the -actin signal. V2O5, open bars; TiO2, filled bars.

View larger version (30K): [in this window] [in a new window]   Figure 4.   Time course of HB-EGF mRNA upregulation by V2O5 in NHBECs. Cultures of NHBECs were treated with 50 µg/cm2 V2O5 or TiO2 at the indicated time points before collecting total RNA for RT-PCR. (A) Representative RT-PCR results showing upregulation of HB-EGF mRNA by 3 h after stimulation with V2O5 but no induction by TiO2. (B) Densitometry of RT-PCR results showing HB-EGF signal normalized against the -actin signal. V2O5, open bars; TiO2, filled bars.

View larger version (39K): [in this window] [in a new window]   Figure 5.   Northern blot analysis of HB-EGF in NHBECs treated with 10 µg/cm2 V2O5 or TiO2 for 3 h. (A) Representative Northern blot showing induction of the 2.36-kb HB-EGF mRNA by V2O5 (upper panel), constitutive expression of the 1.1-kb GAPDH mRNA (middle panel), and ethidium bromide-stained gel of 28s and 18s RNAs (lower panel). (B) Scanning densitometry was used to evaluate four separate experiments and normalize the HB-EGF signal to the GAPDH signal. **P < 0.01, V2O5 treatment compared with control (no addition).

Mechanism of V₂O₅-Induced HB-EGF mRNA Induction

The EGF-R-selective tyrosine kinase inhibitor AG1478 completely blocked the V₂O₅-induced increase in HB-EGF mRNA, whereas the PDGF-R-specific tyrosine kinase inhibitor AG1296 did not have a significant effect (Figure 6). V₂O₅ treatment of NHBECs caused phosphorylation of the EGF-R as determined by Western blot analysis using a monoclonal anti-phospho-EGF-R antibody, yet no change in the total amount of EGF-R was observed after V₂O₅ treatment (Figure 7). Moreover, V₂O₅-induced EGF-R phosphorylation and V₂O₅-induced HB-EGF secretion were inhibited in a concentration-dependent manner by AG1478 (Figure 7). Pretreatment of NHBECs with the MEK-1 inhibitor PD98059 abolished the V₂O₅-induced increase in HB-EGF mRNA expression, whereas pretreatment with the p38 MAP kinase inhibitor SB203580 blocked the V₂O₅-induced increase in HB-EGF mRNA by ~ 80% (Figure 8).

View larger version (27K): [in this window] [in a new window]   Figure 6.   Effect of EGF-R- and PDGF-R-specific tyrosine kinase inhibitors on V2O5-induced HB-EGF mRNA upregulation. Cultures of NHBECs were pretreated for 1 h with 100 µM AG1296 or AG1478 to inhibit phosphorylation of PDGF-R or EGF-R, respectively. NHBECs were then stimulated with 50 µg/cm2 V2O5 for 4 h before collecting total RNA for RT-PCR. (A) Representative result showing that AG1478, but not AG1296, significantly blocked the induction of HB-EGF mRNA by V2O5. (B) Densitometry of HB-EGF signal normalized against -actin.

View larger version (35K): [in this window] [in a new window]   Figure 7.   V2O5-induced phosphorylation of the EGF-R-induced and V2O5-induced secretion of HB-EGF by NHBECs is blocked in a concentration-dependent manner by AG1478. Cultures of NHBECs were pretreated for 1 h with an increasing concentration of AG1478 and then stimulated with 50 µg/cm2 V2O5 for 5 min before collecting cell lysates for phospho-EGF-R and EGF-R Western blotting (A) or 24 h before collecting conditioned medium for HB-EGF ELISA (B). Comparable concentrations of AG1478 reduced both V2O5-induced phosphorylation of EGF-R and V2O5-stimulated HB-EGF production by NHBECs.

View larger version (28K): [in this window] [in a new window]   Figure 8.   Effect of MAP kinase inhibitors on V2O5-induced HB-EGF mRNA upregulation in NHBECs. Cultures of NHBECs were pretreated for 1 h with 10 µM PD98059 or SB203580 to inhibit activation of MEK or p38 MAP kinase, respectively. NHBECs were then stimulated with 50 µg/cm2 V2O5 for 4 h before collecting total RNA for RT-PCR. (A) Representative result showing significant inhibition of V2O5-induced HB-EGF mRNA expression by PD98059 or SB203580. (B) Densitometry of HB-EGF signal normalized against -actin.

HB-EGF Mitogenic Activity Is Due in Part to FGF-2

Because an anti-FGF-2 neutralizing antibody partially blocked the mitogenic effect of NHBEC-CM on HLF (Table 2) whereas the anti-HB-EGF antibody caused complete inhibition of NHBEC-CM-induced HLF mitogenesis, we postulated that HB-EGF was acting in part by stimulating the production of FGF-2 by the HLFs. Confluent, quiescent cultures of HLFs were treated with 1 ng/ml HB-EGF for 24 h. This was approximately the same concentration of HB-EGF detected by ELISA in NHBEC-CM (see Table 4). FGF-2 was detected in unconcentrated HLF CM and was upregulated ~ 3-fold by HB-EGF treatment (Figure 9).

View larger version (33K): [in this window] [in a new window]   Figure 9.   Induction of FGF-2 production by HLFs after treatment with recombinant HB-EGF. Because anti-FGF-2 antibody caused significant inhibition of HLF mitogenesis stimulated by NHBEC-CM (Table 2), we postulated that HB-EGF produced by the NHBECs was causing the release of FGF-2 by HLF. Confluent HLFs in 75-cm2 flasks were rendered quiescent in SFDM and then treated with 1 ng/ml HB-EGF for 24 h. The HLF-CM was harvested and FGF-2 was measured using a commercially available ELISA kit. HB-EGF caused a ~ 3-fold increase in the amount of FGF-2 produced by HLFs. Data are means ± SEM of three separate experiments, each performed in triplicate. **P < 0.01, treatment versus control.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Human airway epithelial cells have been reported to produce factors that stimulate the mitogenesis of lung fibroblasts (8, 11). However, no clear connection has been made to establish the identity of the epithelial-derived growth factor(s) that mediate fibroblast growth. In this study we report that HB-EGF is a principal mitogen that is spontaneously produced by undifferentiated cultures of NHBECs and that stimulates the mitogenesis of HLFs in culture. This conclusion was reached on the basis of two observations: (1) An EGF-R tyrosine kinase inhibitor, AG1478, completely blocked NHBEC-mediated fibroblast growth (Table 1), and (2) a neutralizing anti-HB-EGF antibody, but not antibodies to other EGF family members (EGF, TGF-), abolished the mitogenic activity in NHBEC-CM (Figure 2 and Table 2). To our knowledge, this is the first report that establishes HB-EGF as a principal bronchial epithelial cell-derived mitogen for lung fibroblasts.

V₂O₅ caused a significant increase in HB-EGF mRNA and protein in NHBECs (Figures 3-5 and Table 4) and increased the HB-EGF-dependent mitogenic activity in NHBEC-CM (Figure 2). In contrast, the inert metal TiO₂ did not affect HB-EGF mRNA or protein levels. We further explored the mechanism through which V₂O₅ induced HB-EGF mRNA expression. The EGF-R-specific tyrosine kinase inhibitor AG1478, but not the PDGF-R-specific inhibitor AG1296, blocked the V₂O₅-induced increase in HB-EGF mRNA (Figure 6). We also demonstrated that V₂O₅ caused phosphorylation of the EGF-R (Figure 7). Moreover, both the phosphorylation of the EGF-R and secretion of HB-EGF were inhibited within the same concentration range of AG1478 (Figure 7). These data strongly suggest that the EGF-R is a central target of vanadium-induced stress that signals downstream pathways which culminate in HB-EGF gene expression.

We observed that the V₂O₅-induced increase in HB-EGF mRNA was inhibited by specific inhibitors of the p38 MAP kinase pathway or the extracellular signal-regulated kinase (ERK) pathway (Figure 8), suggesting that both of these MAP kinases are important signaling intermediates in causing elevated HB-EGF mRNA expression. The activation of the p38 MAP kinase has been linked to the production of inflammatory cytokines (21). Also, work by Samet and coworkers showed that vanadium and some other metals activate ERK, Jun amino-terminal kinase, and p38 MAP kinases in human bronchial epithelial cells (22). More recent findings by these same investigators showed that several metals (including vanadium, arsenic, copper, and zinc) activate the ERK pathway via phosphorylation of the EGF-R (23). In agreement with these findings, we recently reported that ERK phosphorylation in rat pulmonary myofibroblasts requires upstream activation of the EGF-R (24). The mechanisms through which vanadium activates p38 MAP kinase remain unclear, and further study should focus on upstream molecules targeted by vanadium that lead to p38 MAP kinase activation.

It is possible that V₂O₅-induced HB-EGF expression in NHBECs requires the generation of reactive oxygen species (ROS). Miyazaki and coworkers showed that exposure of cultured rat gastric epithelial cells to hydrogen peroxide (H₂O₂) increased HB-EGF gene expression (25). It is also known that vanadium compounds generate H₂O₂ and ·OH via redox cycling (26). Recently, we observed that N-acetyl-L-cysteine, a free-radical scavenger, blocked V₂O₅-induced gene expression of HB-EGF (L. Zhang and J. C. Bonner, unpublished observation). Therefore, V₂O₅ could induce HB-EGF expression via an oxidant-dependent mechanism. However, some vanadium compounds can act as competitive phosphatase inhibitors via an oxidant-independent mechanism (27). Therefore, the contribution of oxidant generation in mediating HB-EGF expression by V₂O₅ requires further study. It is possible that V₂O₅ could induce HB-EGF through the generation of ROS, by competitive inhibition of protein tyrosine phosphatases, or through a combination of both mechanisms.

We recently reported that intraperitoneal delivery of EGF-R tyrosine kinase inhibitor AG1478 or the PDGF-R tyrosine kinase inhibitor AG1296 reduced pulmonary fibrosis in rats after the intratracheal instillation of V₂O₅ (28). Further, Yi and coworkers showed that intratracheal delivery of recombinant PDGF caused obstruction of airways due to fibroblast proliferation (29). Although PDGF isoforms are likely to be important to fibroblast proliferation during the progression of fibroproliferative lung disease, our data with NHBECs suggest that the airway epithelium is not an important source of PDGF. However, we and others have shown that PDGF-BB contributes the majority of mitogenic activity in conditioned medium by rat or human alveolar macrophages that drives lung fibroblast mitogenesis (15, 30). Our previous finding that the EGF-R tyrosine kinase inhibitor AG1478 reduced pulmonary fibrosis (28) is consistent with the idea that HB-EGF is an important fibroblast mitogen. However, it is unknown whether the beneficial effect of AG1478 in vivo is due to inhibition of EGF-R phosphorylation induced by HB-EGF, EGF, or TGF-. Therefore, further investigation in vivo will be necessary to determine the overall contribution of HB-EGF to lung fibrogenesis.

A previous study by Cambrey and coworkers reported that IGF-1 was a major fibroblast mitogen produced by primary cultures of human airway epithelial cells (11). In that study, a neutralizing antiserum to IGF-1 inhibited fibroblast proliferation induced by epithelial cell-conditioned media by ~ 50%. In our hands, we did not observe any inhibitory effects of IGF-1 neutralizing antibody on HLF mitogenesis after stimulation with HBEC-CM. Several growth factors, including PDGF, act as competence factors that stimulate the G₀ to G₁ cell cycle transition, whereas IGF-1 is a progression factor that allows cells to progress through the cell cycle once they have reached the G₁ checkpoint. In our [³H]thymidine uptake experiments, we used an SFDM that contains insulin (a progression factor). Therefore, it is possible that insulin present in the SFDM substituted for HBEC-derived IGF-1 and thereby masked the possible mitogenic contribution of this growth factor.

Although anti-HB-EGF and AG1478 completely inhibited the HLF mitogenesis induced by CM from NHBECs, we observed that an anti-FGF-2 antibody also inhibited NHBEC-CM-induced HLF mitogenesis by ~ 50% (Table 2). The complete block of NHBEC-CM-stimulated mitogenesis by anti-HB-EGF together with the partial block by anti-FGF-2 is seemingly paradoxical. However, Peifley and coworkers reported that HB-EGF stimulates the production of FGF-2 by aortic smooth-muscle cells (31). Therefore, we pursued the hypothesis that the HB-EGF-induced mitogenic effect was due in part to the ability of HB-EGF to stimulate FGF-2 production by the HLFs. Indeed, we found that 1 ng/ml of recombinant HB-EGF (approximately the same concentration of HB-EGF measured in CM from V₂O₅-treated NHBECs; Table 4) caused a 3-fold increase in FGF-2 secreted by HLFs (Figure 9). This observation suggests that HB-EGF may exert its mitogenic effect on HLFs in part by initiating an FGF-2 autocrine loop.

HB-EGF produced by bronchial epithelial cells could play a role in airway remodeling and diseases such as asthma and chronic bronchitis. We have shown that HB-EGF is spontaneously produced by airway epithelial cells in culture. Moreover, HB-EGF mRNA expression is further increased by stimulation with V₂O₅. It is currently not known whether HB-EGF is upregulated in vivo after airway injury. However, HB-EGF is upregulated in rat kidney after acute injury (32) and during the progression of atherosclerosis in humans (33). Further, inflammatory cytokines such as tumor necrosis factor (TNF)- and oxidative stress increase gene expression of HB-EGF in vascular endothelial cells and gastric epithelial cells, respectively (25, 34). These same mediators (TNF- and oxidants) are potent activators of airway epithelial cells (35) and likely turn on HB-EGF expression during airway inflammation.

In summary, we report that cultured HBECs spontaneously produce HB-EGF, which is a major mitogen in the epithelial cell-conditioned medium that stimulates HLF mitogenesis. The transition metal V₂O₅ induces HB-EGF mRNA levels and further increases the release of mitogenic activity from NHBECs which is blocked by HB-EGF neutralizing antibody. The mitogenic effect of HB-EGF is due to direct activation of the EGF-R, but FGF-2 production stimulated by HB-EGF also appears to contribute to fibroblast mitogenesis after treatment with epithelial cell-CM. These data suggest that HB-EGF and FGF-2 contribute to the development of airway fibrosis after epithelial injury.