Introduction

Top Abstract Introduction Material and Methods Results Discussion References

Prolonged hypoxia causes proliferation and migration of smooth muscle cells in the arteriolar wall, a process known as vascular remodeling. Abnormalities in vascular tone and smooth muscle cell hyperplasia is common in a wide variety of pulmonary diseases, such as pulmonary fibrosis, pulmonary hypertension, and acute respiratory distress syndrome (1, 2). Hypoxia-induced hyperplasia and cell proliferation has also been observed in animal models (1). No data have been published on cell cultures obtained from human lung tissue.

Different growth factors including platelet-derived growth factor (PDGF), fibroblast growth factor (FGF), platelet-activating factor (PAF), tumor necrosis factor (TNF), interleukin (IL)-1, IL-6, and endothelin-1 are involved in cell proliferation and in hypoxia-mediated cell activation (3). PDGF and FGF are released from human mononuclear phagocytes cultured under hypoxic conditions and have been shown to enhance proliferation of bovine endothelial cells (4). Hypoxia induced the release of PAF from human umbilical vein-derived endothelial cells (HUVECs) (6, 7) and the release of PDGF and PAF in rat endothelial cells (3). Both factors have been shown to be involved in chemotactic activity and proliferation of rat pulmonary artery fibroblasts stimulated with cell culture medium obtained from endothelial cells cultured under hypoxic conditions (3). The presence of neutralizing anti-PDGF or anti-endothelin-1 antibodies significantly reduced hypoxia-induced proliferation of fibroblasts (3, 4).

Under normoxic conditions, both PDGF and PAF induce the synthesis of IL-6 and IL-8 in various cell types, including astrocytes (10), alveolar macrophages (11), endothelial cells (12), fibroblasts, vascular smooth muscle cells (VSMC), and mesenchymal cells (8, 11). We have shown previously that IL-6 plays an important role in the control of PDGF-induced cell proliferation of fibroblasts, VSMC, and mesangial cells (8). The mitogenic activity of PAF on human lung fibroblasts may also involve the action of IL-6 (9).

IL-6 concentration is increased in the serum of patients with primary pulmonary hypertension, a disease associated with enhanced proliferation of VSMC in lung arteries (13). Therefore, an important role of PDGF and interleukins in the pathogenesis of lung diseases associated with hypoxia has been suggested (14). Because hypoxia enhances the production of growth factors and cytokines in animal models and HUVECs, we hypothesized that PAF and PDGF are important mediators for hypoxia-induced cytokine synthesis and cell proliferation. These cellular responses to hypoxia may be followed by tissue structural changes leading to narrowing of blood vessels and fibrosis.

In this study, we demonstrated an enhanced production of IL-6 and IL-8 under hypoxic conditions in human pulmonary fibroblasts and pulmonary VSMC, and investigated the role of PDGF and PAF on hypoxia-induced cytokine synthesis and on cell proliferation.

	Material and Methods

Top Abstract Introduction Material and Methods Results Discussion References

Cell Cultures

Ten primary cell lines of VSMC were established from pulmonary arteries obtained from patients undergoing lobectomy or pneumonectomy for peripheral lung cancer. Arteries were kept overnight in Hanks' buffered salt solution (HBSS; Seromed, Fakola, Basel, Switzerland) at 4°C before stripping off the intimal cell layer and residual adventitial tissue by forceps. The resting media of the vessels was cut into small pieces (3 to 5 mm) and transferred into prewetted cell culture flasks (Falcon; Inotech, Basel, Switzerland). VSMC were allowed to grow out by incubating the vessel tissue for 1 wk in minimal essential medium (Seromed) supplemented with 5% fetal calf serum (FCS; Seromed), and 20 mM Hepes buffer (Seromed). Cells were subcultivated in the same medium after trypsin treatment, and experiments were performed between passages 2 and 6.

VSMC were characterized by immunohistochemical staining using monoclonal antibodies specific for either smooth muscle cell actin, keratin, fibronectin, laminin, or von Willebrand factor (all antibodies; Boehringer Mannheim, Mannheim, Germany). In brief, cells were grown in Lab-Tek tissue culture chamberslides (Miles, Scientific Division, Naperville, IL) until confluence and fixed in 4% paraformaldehyde. Nonspecific protein binding was blocked by incubating the cells in phosphate-buffered saline (PBS; Seromed) supplemented with 0.5% (wt/vol) bovine serum albumin (BSA; Fluka Chemie, Buchs, Switzerland) for 20 min. The slides were then incubated with one of the foregoing antibodies (60 min), washed with PBS (3×), and further incubated with a fluorescein-linked antirabbit or antimouse immunoglobin G (IgG) (Boehringer Mannheim). Preparations were washed (3×) with PBS, mounted with Fluorosave reagent (Calbiochem-Novabiochem, San Diego, CA), and analyzed using a microscope equipped with epilumination and specific filters (560 nm; Axiophot, Carl Zeiss Inc., Oberkochem, Germany) (17). Nonspecific binding of the fluorescein-linked antibody was excluded using the second antibody alone.

Five primary cell lines of pulmonary fibroblasts were grown out from sterile peripheral lung tissue biopsies and cultured in RPMI 1640 (Seromed), supplemented with 10% FCS, and 8 mM L-glutamine (Seromed) as described previously (15, 16). No antibiotics or antimycotics were added to the culture medium at any time.

The phenotype of the fibroblasts was determined by immunohistochemistry. Similar to VSMC, primary cells were grown in Lab-Tek tissue culture chamber/slides and fixed in 4% paraformaldehyde. Nonspecific protein binding was blocked by PBS containing 0.5% BSA (20 min). Cells were incubated for 30 min with one of the foregoing monoclonal antibodies. After three washes with blocking buffer, slides were further incubated for 30 min with either fluorescein-coupled antirabbit IgG or antimouse IgG (Boehringer Mannheim). After washing, the preparations were mounted with Fluorosave reagent and observed on a microscope as described previously (17).

Prior to stimulation subconfluent (80% density) cultures of cells in passages 2 to 6 were serum-deprived for 48 h with low serum medium (Dulbecco's modified Eagle's medium or RPMI 1640, respectively, supplemented with 0.1% FCS and 20 mM Hepes). Low serum medium was exchanged every 12 h to avoid autostimulation of the cells (15).

Cell Culture Conditions

Normoxic culture conditions were defined as follows: 21% O₂, 77% N₂, and 5% CO₂. For hypoxic culture conditions, the concentration of O₂ was reduced to 3% by replacement with N₂, keeping CO₂ constant at 5%. Subconfluent, quiescent cells were incubated either under hypoxic or normoxic conditions for various time points (0, 2, 4, 6, 8, 12, 18, 24, 36, or 48 h).

Isolation of RNA and Northern Blotting

Expression of transcripts encoding for IL-6, IL-8, and the constitute HLA- gene was determined using Northern blot analysis. Total RNA was extracted as described (15- 17), and 10 µg of total RNA were size-fractionated in a standard 1% agarose/7% formaldehyde gel. RNAs were transferred onto Hybond nylon membranes (Amersham Corp., Buckinghamshire, UK) by capillary blotting overnight in 10× saline sodium citrate (SSC) (1× SSC = 0.15 M NaCl, 1.5 M Na₃-citrate, pH 7.0) and crosslinked by ultraviolet irradiation. Membranes were prehybridized for 1 h at 45°C with QuickHyb (Stratagene GmbH, Zürich, Switzerland) supplemented with 1 mg/ml heat-denatured salmon sperm DNA. Blots were either hybridized for 1 h with 20 ng of a 3'-end-labeled ([³²P]dATP) IL-6 or IL-8-antisense oligonucleotides, respectively (R&D, Abington, UK), or hybridized overnight with 20 ng of a random prime labeled ([³²P]dATP) HLA- cDNA probe (American Type Culture Collection, No. 57474, Rockville, MD) used as a constitutive control. Blots were washed twice at room temperature with 5× SSC/0.1% sodium dodecyl sulfate (SDS) (15 min each) followed by a final wash with 0.1× SSC/0.1% SDS at 55°C for IL-6 or IL-8, respectively, or at 68°C for the HLA- probe. Blots were then exposed to Kodak (Eastman Kodak, Rochester, NY) X-omat films overnight or up to 48 h at 70°C using an intensifying screen. Hybridization signals were analyzed densitometrically using an automated scanner system supported with the National Institutes of Health image program (15).

Transcription Rate of Interleukins

Nuclei were prepared from subconfluent cell cultures (2 × 10⁶ cells) kept under hypoxic or normoxic culture conditions at various time points (0, 1, 2, 4, 8, and 12 h); transcription rates of IL-6 and IL-8 genes were determined as described earlier (18). In brief, cells were washed (three times) with ice-cold PBS, scraped, and resuspended in 1 ml of PBS containing 0.5% Nonidet P-40. Cells were kept on ice (5 min) before nuclei were pelleted by centrifugation (1,000 × g, 3 min, 4°C). Isolated nuclei were washed with elongation reaction buffer (20 mM Tris, pH 7.9; 20% glycerol, 140 mM KCl, 10 mM MgCl₂, 1 mM dithiothreitol), resuspended in 100 µl of buffer, counted using a Neugebaur chamber slide, and adjusted to 1 × 10⁵ nuclei/µl. To elongate IL-6 and IL-8 mRNA, 50 µl of nuclei were incubated with [-³²P]UTP (3,000 Ci/mmol; Amersham), 10 mM NTP mix (ATP, CTP, and GTP; Amersham) and creatine kinase (50 ng/µl; Amersham). The mixture was incubated at 37°C (10 min), and RNA was isolated by guanidinium-isothiocyanate as described previously and ethanol precipitated at 20°C. RNA was resuspended in 10 µl 0.2 M Tris (pH 7.4) and incubated at 37°C (20 min) in the presence of 1 mM MgCl₂ and DNase (50 ng/µl), and was purified again. To determine the amount of labeled specific RNAs, linearized single-stranded cDNA-probes for IL-6 (ATCC, No. 67153) or a cDNA probe for IL-8 (R&D; No. BPR 100) were used, respectively. Equal amounts of the cDNAs (10 µg) were spotted onto nylon membranes, fixed by UV crosslinking, and prehybridized in QuickHyb (Stratagene) buffer at 65°C, for 1 h. Each nuclear RNA extract was incubated with a strip of IL-6 or IL-8 prelabeled nylon membranes for 24 h at 65°C and washed with high stringency conditions (1× SSC, 60°C, 2 × 30 min) before nylon membranes were exposed to BioMax-film for 24 to 72 h. Afterward, bands were excised and counts per minute were in a liquid scintillation counter. In parallel, the intensities of bands were determined using a computer-supported image analysis system (X-ray, Germany).

Quantitation of Secreted Interleukins

Concentration of the secreted interleukins was analyzed using enzyme-linked immunosorbent assays (EIA; R&D). In brief, cells (1 × 10⁴/ml) were seeded onto a 24-well cell culture plate (Falcon) and cultivated until they achieved 80% confluence. The cells were kept in low serum medium for 48 h prior to hypoxic culture conditions. Culture medium aliquots (500 µl) were collected at different time points (0, 12, 24, 36, and 48 h). The concentration of secreted IL-6 and IL-8 was determined for each sample in duplicate.

Mitogenicity

The mitogenic effect of hypoxic culture conditions was determined by incorporation of [³H]thymidine (1 µCi/ml) following standard protocol as described by Lindl and colleagues (18). Percentage of de novo synthesis of DNA was calculated comparing the ratio of the incorporation of [³H]thymidine under hypoxic conditions with that achieved under normoxic conditions at the same time point. All experiments were performed at least in triplicate for each cell line.

PAF-Receptor Antagonist and Neutralizing PDGF Antibodies

The PAF-receptor antagonist WEB2170 was provided by Boehringer Ingelheim (Ingelheim, Germany). According to earlier studies (6), WEB2170 was used at concentrations ranging from 2 to 20 µM. Cells were preincubated in the presence of WEB2170 for 30 min prior to exposure to hypoxia. Neutralizing anti-PDGF-AB antibodies were used at a concentration range of 1 to 100 µg/ml, according to the instructions of the distributor (R&D). Cells were preincubated for 30 min in the presence and absence of the anti-PDGF antibodies prior to culture under hypoxia.

Statistical Analysis

Data were analyzed using a computer-supported statistical program (Statview; Apple Computer, Cupertino, CA). Homogeneity of two groups was analyzed by two-tailed Student's t test for each time point or concentration, respectively.

	Results

Top Abstract Introduction Material and Methods Results Discussion References

Hypoxia Induced Expression of Interleukin-6 and Interleukin-8 Genes

Under normoxia, the transcription of both IL-6 and IL-8 remained consitutively low for up to 24 h (Figure 1). The signal for the mRNA encoding for IL-6 (1.3 ± 0.1), or IL-8 (1.2 ± 0.05), respectively, was normalized to that of the constitutive control gene (HLA-) and was only detected when filters were exposed for more than 48 h. No significant change of the low constitutive transcription of both genes was observed during the investigated time period in any of the two cell types (Figure 1).

View larger version (22K): [in this window] [in a new window]   View larger version (28K): [in this window] [in a new window]   Figure 1.   Northern blot analysis of IL-6 and IL-8 mRNA under normoxia and hypoxia, in human pulmonary fibroblasts (a) and VSMC (b). Each data point represents the mean ± SEM of four independent experiments in both cell types. Transcription rates (c, d) were determined using nuclear run-on assays for both interleukins at the time point of maximal transcription. Data are presented as results of photodensitometrical analysis. Each data point represents mean ± SE estimated in three independent experiments.

In contrast, when cells were exposed to hypoxia (3% O₂) transcription of IL-6 or IL-8 rapidly increased in both cell types within 24 h (Figure 1). Transcription of the IL-6 gene was induced as early as 2 h after exposure of cells to hypoxia and reached a maximum at 4 h (5.3×) in fibroblasts (Figure 1a). In VSMC exposed to hypoxia, the transcription of IL-6 was maximal after 8 h (4.8×; Figure 1b). The transcription of the IL-8 gene started to increase after 4 h in fibroblasts and peaked at 12 h (5.8×; Figure 1a). In VSMC, the transcription of IL-8 started to increase after 2 h of hypoxia and reached a maximum at 12 h (4.8×; Figure 1b). As shown in Figures 1c and 1d, a nuclear run on experiments revealed that the hypoxia-dependent increase of mRNA signals for both interleukins was due to enhanced de novo synthesis of mRNAs rather than to accumulation of mRNAs. These data were confirmed by experiments using actinomycin D (10 µg/ml) (data not shown).

On the protein level, spontaneous secretion of IL-6 or IL-8 under normoxic conditions was low and resulted in a slow increase of the concentration of both interleukins in the cell culture medium (IL-6: 4.7 ng/10,000 cells/h; IL-8: 2.9 ng/10,000 cells/h). Fibroblasts and VSMC showed a similar pattern of spontaneous secretion of IL-6 and IL-8 under normoxic conditions (Figure 2).

View larger version (21K): [in this window] [in a new window]   Figure 2.   Kinetics of IL-6 and IL-8 secretion under normoxia and hypoxia in human pulmonary fibroblasts (a) and VSMC (b). Each data point represents the mean ± SEM of four independent experiments in both cell types.

Hypoxia-dependent transcription of both interleukin genes was followed by secretion of copious amounts of the respective interleukin proteins into the culture medium (Figure 2). The secretion of IL-6 into the fibroblast cell culture medium started to increase 8 to 10 h after cells were exposed to hypoxia, reaching a maximum after 48 h (Figure 2a), whereas in VSMC the secretion of IL-6 started at 18 h, reaching a maximum at 36 h (Figure 2b). Similarly, hypoxia-dependent secretion of IL-8 protein started 18 to 24 h after cells were exposed to 3% O₂ and continued to increase over the observed time period (48 h) in both cell types (Figure 2).

Role of PAF and PDGF in Hypoxia-Mediated Expression of Interleukin-6 and Interleukin-8

As shown in Figures 3a and 3b, PAF (1 × 10⁹ M) induced the secretion of both interleukins IL-6 and IL-8. In the presence of the PAF antagonist WEB2170, this increase of interleukin synthesis was inhibited by 72 ± 6.5% (Figures 3c and 3d). As described previously, hypoxic culture conditions also induced the expression of IL-6 and IL-8 in both cell types. In the presence of the PAF antagonist WEB 2170, the hypoxia-induced transcription of IL-6 and IL-8 genes was reduced. Reduction of transcription of both interleukins by WEB2170 was similar in fibroblasts and VSMC. In the presence of WEB2170, the hypoxia- induced transcription of IL-6 was reduced by 47 ± 6.1% in fibroblasts and by 49 ± 7.3% in VSMC (Figure 3d). Hypoxia-induced transcription of IL-8 was reduced by 49 ± 4.7% in fibroblasts and by 48 ± 4.2% in VSMC (Figure 3d). The inhibitory effect of WEB2170 on transcription of both interleukins was dose dependent (data not shown).

View larger version (26K): [in this window] [in a new window]   View larger version (26K): [in this window] [in a new window]   Figure 3.   Effect of PAF antagonist WEB2170 and neutralizing anti-PDGF antibodies on hypoxia-induced IL-6 and IL-8 transcription in human pulmonary fibroblasts and VSMC. The stimulatory effect of PAF and PDGF on IL-6 and IL-8 synthesis was assessed in both fibroblasts (a) and VSMC (b). The inductive effect of hypoxia on interleukin synthesis is shown for both fibroblasts (c) and VSMC (d). The inhibitory effect of the PAF antagonist WEB2170 and neutralizing anti-PDGF antibodies was investigated at time point of maximal transcription for both cell types. Each data point represents the mean ± SEM of four independent experiments in both cell types.

PDGF-AB (10 ng/ml) induced the expression of IL-6 and IL-8 in both fibroblasts and VSMC (Figures 3a and 3b). This enhanced expression of the two interleukin genes was inhibited in the presence of anti-PDGF-AB-antibodies by 85 ± 9.4%. In the presence of neutralizing anti-PDGF antibodies, the hypoxia-induced transcription of IL-6 and IL-8 was partly inhibited in both cell types. Transcription of IL-6 was significantly reduced by 56 ± 3.9% in fibroblasts (Student's t test: P < 0.01) and by 62 ± 5.7% in VSMC (Student's t test: P < 0.005; Figures 3c and 3d). Hypoxia-dependent de novo synthesis of IL-8 mRNA was inhibited by 50 ± 4.7% in fibroblasts (Student's t test: P < 0.01) and by 68 ± 7.7% in VSMC (Student's t test: P < 0.001; Figures 3c and 3d).

Similar to transcription, hypoxia-induced secretion of IL-6 and IL-8 was significantly reduced when cells were preincubated with WEB2170. The maximal inhibitory effect of WEB2170 on hypoxia-induced secretion was achieved at a concentration of 10 µg/ml. Inhibition of IL-6 secretion was reduced by 60 ± 2.3 % in fibroblasts (Student's t test: P < 0.001) and by 48 ± 5.7% in VSMC (Student's t test: P < 0.01) in the presence of WEB2170 (Figures 4a and 4b). Hypoxia-stimulated secretion of IL-8 was reduced by 59 ± 3.8% in fibroblasts (Student's t test: P < 0.001) and by 53 ± 5.7% in VSMC (Student's t test: P < 0.005; Figures 4a and 4b).

View larger version (29K): [in this window] [in a new window]   View larger version (31K): [in this window] [in a new window]   Figure 4.   Effect of PAF antagonist WEB2170 and neutralizing anti-PDGF antibodies on hypoxia-induced IL-6 and IL-8 secretion by human pulmonary fibroblasts and VSMC. The inhibitory effect of both PAF antagonist WEB2170 and neutralizing anti-PDGF-antibodies on hypoxia-induced cytokine secretion was investigated 48 h after addition of the stimulus (fibroblasts, a; VSMC, b). The specific inhibitory effect of WEB2170 and neutralizing anti-PDGF antibodies is shown for fibroblast (c) and VSMC (d). Each data point represents the mean ± SEM of four independent experiments in both cell types.

Similar to their inhibitory effects on transcription, neutralizing anti-PDGF antibodies diminished the secretion of IL-6 by 58 ± 5.7% in fibroblasts (Student's t test: P < 0.005) and by 64 ± 5.7% in VSMC (Student's t test: P < 0.001; Figures 4a and 4b). Hypoxia-mediated secretion of IL-8 was reduced by 51 ± 5.7% in fibroblasts (Student's t test: P < 0.01) and by 51 ± 5.7% in VSMC (Student's t test: P < 0.01) in the presence of neutralizing anti-PDGF antibodies; Figures 4a and 4b).

Hypoxia Induced Cell Proliferation

Hypoxia induced marked cell proliferation in primary human pulmonary fibroblasts and pulmonary VSMC. Under normoxic conditions, both cell types exerted a low spontaneous incorporation of tritium thymidine over a time period of 48 h. Following exposure to 3% O₂ thymidine, incorporation increased significantly within 24 h. Compared with cell proliferation under normoxic conditions, hypoxia enhanced cell proliferation of primary fibroblasts by 335 ± 26% (Figure 5a) and proliferation of VSMC by 234 ± 15% (Figure 5b) at 48 h.

View larger version (27K): [in this window] [in a new window]   View larger version (22K): [in this window] [in a new window]   Figure 5.   Effect of PAF antagonist WEB2170 and neutralizing anti-PDGF antibodies on hypoxia-induced cell proliferation; (a) human pulmonary fibroblasts and (b) VSMC. The inhibitory effect was investigated at 48 h. Analysis of mitogen and antagonist specificity effects are presented for fibroblasts (c) and VSMC (d). Each data point represents the mean ± SEM of four independent experiments in both cell types.

Role of PAF and PDGF in Hypoxia-Induced Cell Proliferation

Both PAF and PDGF enhanced incorporation of tritium thymidine within 48 h in both cell types (Figures 5c and 5d). The mitogen-induced cell proliferation was inhibited with their respective antagonists by 85 ± 2.9% (fibroblasts, Figure 5c), or 91 ± 7.3% (VSMC; Figure 5d). In the presence of the PAF-receptor antagonist WEB2170 at a dose of 10 µg/ml, hypoxia-mediated cell proliferation was reduced by 40 ± 3.7% in fibroblasts (Student's t test: P < 0.001; Figure 5a) and by 42 ± 2.9% in VSMC (Student's t test: P < 0.001; Figure 5b). The addition of neutralizing anti-PDGF antibodies (10 µg/ml) reduced hypoxia-dependent cell proliferation by 41 ± 8.3% in fibroblasts (Student's t test: P < 0.001; Figure 5a) and by 39 ± 3.1% in VSMC (Student's t test: P < 0.005; Figure 5b).

	Discussion

Top Abstract Introduction Material and Methods Results Discussion References

In this study, hypoxia could be shown to induce markedly cell proliferation and de novo synthesis of IL-6 and IL-8 in human pulmonary fibroblasts and VSMC. We demonstrated that PAF and PDGF are both involved in hypoxia-induced interleukin release and cell proliferation.

Increased production of cytokines, enhanced synthesis of extracellular matrix components, and proliferation of fibroblasts and VSMC play important roles in the pathogenesis of different lung diseases that are associated with hypoxia. In pulmonary hypertension, proliferation of VSMC and enhanced deposition of collagens could be demonstrated in small vessels (14, 20). The location of pulmonary fibroblasts allows communication between the vascular compartment and the alveolar space, and therefore facilitates the formation of interleukins and other cell activators under pathophysiologic conditions (21, 22). Morphometric studies suggested that pulmonary vascular reactivity to vasodilators is lost when concentric media hypertrophy and intimal fibrosis occur, leading to disease progression in primary pulmonary hypertension (23). Prolonged hypoxia causes proliferation and migration of VSMC and accumulation of extracellular matrix in the arterial wall (20). Fibroblast proliferation with enhanced production of extracellular matrix and marked fibrosis are also important features of acute respiratory distress syndrome and interstitial lung diseases.

The question whether hypoxia contributes to fibroblast and VSMC activation with subsequent fibrosis has not been investigated in human lung cells. On the basis of results obtained in animal studies, we expected that hypoxia induces cell activation in the human lung. We therefore established cultures of primary human pulmonary fibroblasts and VSMC grown from surgically removed lung tissue and used these cell cultures for all experiments described. Because PAF, PDGF, IL-6, and IL-8 are essentially involved in cell activation of human pulmonary fibroblasts and VSMC under normoxic conditions (8, 9, 11, 13, 17, 21), we analyzed the network of these cytokines and growth factors under hypoxic conditions.

PDGF is thought to play an important role in the pathophysiology of tissue structural changes in severe pulmonary hypertension (14) and in pulmonary fibrosis (24). Using a bleomycin-induced lung injury in a rat model, PDGF has been shown to be essentially involved in the tissue repair processes (25). PDGF production by macrophages has been shown in rats, leading to fibroblast proliferation and chemotaxis (26, 27). The presence of PDGF in lung biopsies obtained from patients with pulmonary fibrosis was demonstrated by in situ immunostaining (28). Although PDGF was observed in both healthy and fibrotic tissue, the percentage of PDGF-positive alveolar macrophages and alveolar epithelial cells was increased in patients with pulmonary fibrosis (28). Recent data suggest that hypoxia induces the expression of PDGF in human umbilical vein endothelial cells (3) and bovine endothelial cells (4). Enhanced release of PDGF has also been described in hypoxic mononuclear phagocytes, where it contributes to proliferation of endothelial cells under hypoxic conditions (4). The possible role of PDGF in hypoxia- induced cell proliferation is supported by the findings of Dawes and coworkers (3), who reported that neutralizing anti-PDGF antibodies reduce hypoxia-induced fibroblast proliferation by 55% in a rat model. We observed a similar inhibitory potency of neutralizing anti-PDGF antibodies on hypoxia-induced cell proliferation in human pulmonary fibroblasts and VSMC. In parallel to the reduced cell proliferation, hypoxia-induced expression of IL-6 and IL-8 synthesis was markedly decreased in the presence of anti-PDGF antibodies in both cell types. These data support a role of PDGF in hypoxia-mediated cell activation and fibrosis in the human lung.

PAF is a potent proinflammatory phospholipid mediator that modulates vascular function in the human lung (29). We have shown previously that PAF is a potent mitogen for human pulmonary fibroblasts, and stimulates the release of IL-6 and IL-8 by these cells under normoxic conditions (9). It has been reported that PAF is a potent stimulator for de novo synthesis of cytokines in mouse fibroblasts (30), alveolar macrophages (11), and human neutrophils (31). The role of PAF on concerning activation of pulmonary VSMC has not yet been discussed in the literature. It has been shown that hypoxia is associated with increased secretion of PAF in animal endothelial cell cultures (5, 6). Enhanced PAF release was also shown in HUVECs cultivated under hypoxic conditions (6). Our data show that hypoxia-induced synthesis of IL-6 and IL-8 was reduced by 45% to 52%, respectively, in the presence of the PAF-receptor antagonist WEB2170. Similarly to its inhibitory effect on cytokine release, WEB2170 also downregulated the hypoxia-mediated cell proliferation of human pulmonary fibroblasts and VSMC.

Our data demonstrate that hypoxia-induced activation of pulmonary fibroblasts and VSMC involves the action of PAF and PDGF. However, inhibition of a single proinflammatory mediator was not sufficient to counteract hypoxia-mediated cytokine release and cell proliferation.

There is increasing evidence in the literature that IL-6 is more than an immuno-modulating cytokine. An intracellular precursor of IL-6 is essentially involved in the regulation of cell proliferation (8, 9, 32). In this context, a possible role of IL-6 in the pathogenesis of different lung diseases associated with proliferation of fibroblasts and VSMC is supported by clinical data. Increased serum concentration of IL-6 has been reported in patients suffering from severe pulmonary hypertension (13). In patients with acute respiratory distress syndrome, enhanced concentrations of IL-6 and IL-8 were described in bronchoalveolar lavage fluid and in serum (2).

In vitro, hypoxic cell culture conditions resulted in a considerable increase of de novo synthesis of IL-6 and IL-8 within 12 h. Hypoxia also induced marked cell proliferation. However, because the synthesis rate of the two interleukins occurs shortly after exposure of the cells to hypoxia-enhanced cytokine, release cannot be explained by the hypoxia-associated increase of cell numbers. Regarding the markedly enhanced expression of both IL-6 and IL-8 in two different human pulmonary cell types under hypoxic conditions, our data support the relevance of increased cytokine synthesis for development of fibrosis in acute respiratory distress syndrome (2). Interestingly, IL-6 has recently been shown to induce the expression of the vascular-endothelial growth factor (34, 35). The vascular-endothelial growth factor is a main regulator for vascular tone, vascular permeability, and angiogenesis (33), which contributes to hypoxia-initiated vascular remodeling (34, 35).

In conclusion, our data demonstrate that hypoxia induces the expression of IL-6 and IL-8 in human pulmonary fibroblasts and VSMC, and stimulates proliferation of both cell types. These cellular responses to hypoxia involve formation and release of PAF and PDGF. Our data support the importance of proinflammatory factors and cytokines in the pathogenesis of hypoxia-associated changes of lung tissue structure of the human lung.