Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Cigarette smoking is associated with airway inflammation and the development of chronic obstructive pulmonary disease (COPD). A variety of cells appear to be involved in this inflammatory process including neutrophils (1), eosinophils (2), macrophages (3), and T-lymphocytes (4). In animal models, exposure to cigarette smoke is associated with airway neutrophilia (5) and in studies in human volunteers there occurs an accumulation of neutrophils in bronchoalveolar lavage fluid of smokers when compared with nonsmoker control subjects (6). Neutrophils are a potential source of proteolytic enzymes thought to contribute to the destruction of the lung parenchyma that is characteristic of emphysema, whereas their capacity to generate oxidants has been linked to epithelial mucus production (7), a feature more closely identified with chronic bronchitis.

Interleukin-8 (IL-8) is a potent neutrophil chemoattractant and its levels are elevated in induced sputum of patients with COPD (8). It is produced by bronchial epithelial cells in response to oxidants present in cigarette smoke (9, 10) and this may play an important role in recruitment and activation of neutrophils in vivo. Bronchial epithelial cells also make several ligands for the epidermal growth factor (EGF) receptor (EGFR), including transforming growth factor (TGF)-, heparin-binding EGF-like growth factor (HB-EGF) and amphiregulin (AR) (11), which may contribute to epithelial maintenance and repair. These growth factors are produced as transmembrane precursor molecules whose processing and release (ectodomain shedding) is a highly regulated process involving metalloproteinases (12). Shedding of HB-EGF has been associated with transactivation of the EGFR by G-protein coupled receptors (13).

Because exogenous EGF can induce IL-8 production in primary bronchial epithelial cells (14) or the BEAS-2B epithelial cell line (15), we have investigated the role of autocrine ligands in the release of IL-8 from NCI-H292 bronchial epithelial cells in response to cigarette smoke (CS). Here we show that cigarette smoke extract (CSE) is able to induce expression and release of EGFR ligands and that synthesis and release of IL-8 in response to CSE is dependent on EGFR activation by these ligands. Moreover, the responses of nontransformed, primary epithelial cultures to CSE are similar to those of the H292 adenocarcinoma cells.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

NCI-H292 human pulmonary carcinoma cells were obtained from the American Type Culture Collection (Manassas, VA). The cells were maintained in RPMI-1640 containing 10% fetal bovine serum, 2 mM glutamine, 10 U/ml penicillin, and 10 µg/ml streptomycin at 37°C in a humified atmosphere of air containing 5% CO₂. The polyclonal sheep anti-EGFR antibody was raised against EGF affinity-purified receptors derived from A431 squamous carcinoma cell membranes (16) and was partially purified by (NH₄)₂SO₄ precipitation and diethylaminoethyl (DE-52; Whatman, Maidstone, Kent, UK) ion exchange chromatography (11). The antibody competed with ligand for binding to the EGFR. In control experiments the antibody blocked mitogenesis induced in murine fibroblasts or H292 cells by EGF, TGF-, and HB-EGF, but not by keratinocyte growth factor, or acidic or basic fibroblast growth factor. AG1478 was purchased from Biomol Research Laboratories Inc., Plymouth Meeting, PA.

Fiberoptic Bronchoscopy and Primary Bronchial Epithelial Cell Cultures

Brushed bronchial epithelial cells were obtained by fiberoptic bronchoscopy from seven subjects (mean age 45 yr, range 37-58 yr), all of whom were current cigarette smokers. All subjects were free from respiratory tract infections for a minimum of 4 wk before the study and abstained from smoking for 4 h before bronchoscopy. Written informed consent was obtained from all volunteers and ethical approval was obtained from the Joint Ethics Committee of Southampton University and General Hospital. Bronchoscopy was performed using a fiberoptic bronchoscope (FB-20D; Olympus, Tokyo, Japan) in accordance with standard published guidelines (17). Epithelial cells were obtained using a standard sterile single-sheathed nylon cytology brush. The cells were cultured in Bronchial Epithelial Growth Medium (BEGM) (Clonetics, San Diego, CA) in flasks coated with collagen using Vitrogen-100 (Nutacon, Leimuiden, The Netherlands). The epithelial cells were grown as monolayer cultures in BEGM and were used for assays at passage two.

CSE

The filters were removed from two Kentucky 1R4F research cigarettes (University of Kentucky, Lexington, KY) before attachment to a short length of tubing connected to a Buchner flask (Fisher Scientific, Loughborough, UK) containing 50 ml of RPMI-1640. The smoke was drawn into the medium under vacuum over a period of 1-2 min. The CSE was then filtered through a 0.22-µM Millex-GS (Millipore, Watford, UK) filter and used immediately.

Lactate Dehydrogenase Assay

H292 cells were seeded into 24-well culture trays (0.4 × 10⁵ cells/ cm²) and grown until confluent. The cells were rendered quiescent for 24 h in serum-free medium (Ultraculture; BioWhittaker, Wokingham, UK) and then exposed to Ultraculture ± 1 nM EGF or serial dilutions of CSE and for 24 or 48 h at 37°C, 5% CO₂. The culture supernatants were then analyzed using a lactate dehydrogenase (LDH) assay kit (Sigma Chemical, Poole, UK), according to the manufacturer's instructions.

Analysis of AR, HB-EGF, TGF-, and IL-8 Gene Expression

H292 cells were seeded into 25 cm² culture flasks at 0.4 × 10⁵ cells/cm² and prepared for assay as described for the LDH assay. The cells were exposed to Ultraculture ± 1 nM EGF or CSE in the presence or absence of 1 µM AG1478 for 6 or 24 h at 37°C, 5% CO₂. RNA was extracted from the cells using 750 µl/well of TRIzol reagent (Invitrogen, Paisley, UK) and 2 µg of total RNA reverse transcribed using random hexamer primers and AMV reverse transcriptase (RT) from Promega (Southampton, UK), following the manufacturer's protocol. The target primers and the probe, labeled with 5'-reporter dye FAM (6-carboxy-fluorescein) and 3'-quencher dye TAMRA (6-carboxy-N,N,N',N'-tetramethyl-rhodamine), were designed using Primer Express (Perkin-Elmer Biosystems, Warrington, UK). The primer and probe sequences are listed in Table 1. Primers directed against 18S rRNA were used as normalizing control and were obtained from Perkin-Elmer. cDNA standard curves were generated using serial dilutions of cDNA obtained from pooled samples. For each sample, measured in duplicate, the polymerase chain reaction (PCR) contained ~ 25 ng cDNA template, 250 nM fluorogenic probe, 900 nM of forward and reverse primers, 12.5 µl TAQMan universal PCR master mix (Oswell, Southampton, UK), 0.375 µl of primer, and probe mix for the 18S rRNA made up to 25 µl with water. "No template" and RT-negative samples were included as negative controls. The TAQMan PCR protocol was as follows: 50°C for 2 min, 95°C for 10 min, followed by 40 cycles of denaturation 95°C for 15 s, and annealing/extension 60°C for 1 min. Quantitation and real-time detection of the TAQMan PCR were followed on the a ABI Prism 7,700 sequence detection system and following completion of the PCR reaction, the thresholds for fluorescence emission baseline were set just above background levels on the FAM and VIC layers. The C_T values (cycle number at which the amplification resulted in a fluorescence reading above the threshold) for each of the standards was determined and a standard curve of log cDNA input versus C_T value was constructed; this exhibited linear amplification of target cDNA across the range of standards used. Each of the target gene C_T values were converted to cDNA concentrations by use of the appropriate standard curve; the values for the cytokine or growth factor genes were then normalized to their corresponding 18S rRNA values to control for cDNA input and to give relative values in arbitrary units. The relative primer efficiencies for each target gene, as indicated by the gradient of the standard curves, were within 2% of each other. The data was normalized by dividing the target gene value by the 18S rRNA value.

Analysis of Growth Factor and Cytokine Release

H292 cells were seeded into 24-well culture trays and prepared for experimentation as described for the LDH assay. The cells were exposed to Ultraculture ± 1 nM EGF or CSE in the presence or absence of 1 µM AG1478 or 500 µg/ml sheep anti-EGFR antibody for 48 h at 37°C, 5% CO₂. Primary bronchial epithelial cells were seeded into 24-well, Vitrogen-coated culture trays at 0.75 × 10⁵ cells/well in BEGM and cultured at 37°C, 5% CO₂. When confluent, the medium was replaced with basal medium (BEBM [Clonetics] containing 1% insulin/transferrin/sodium selenite media supplement [Sigma] and 1 mg/ml bovine serum albumin) and the cells were incubated for 24 h. The medium was then changed to basal medium ± dilutions of CSE and incubated for a further 36 h. The conditioned medium was then removed from the cells, clarified by centrifugation, and stored frozen until assay. Enzyme-linked immunosorbent assay (ELISA) kits for TGF- (CN Biosciences, Nottingham, UK.), AR (R&D Systems, Abingdon, UK), EGF, and IL-8 (Biosource International, Camarillo, CA) were used according to the manufacturer's instructions; the minimum detectable amount of each factor in these assays was 10, 15, 30, and 15 pg/ml, respectively. Neither AG1478 nor the anti-EGFR antibody interfered with the EGFR ligand ELISA methods.

Statistical Analysis

All H292 assays were performed at least twice and the data are the mean ± SD (n = 4). Data from the H292 cell assays and paired data from the primary epithelial cell cultures were compared using an unpaired Student's t test and the Wilcoxon test, respectively, using SPSS software (SPSS Inc., Chicago, IL).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effects of CSE on Cell Viability

The toxicity of CSE for H292 cells was assessed by release of LDH activity (Figure 1). The effect of CSE was dose-dependent: at concentrations of 8% and above, CSE was toxic to the cells; however, at lower concentrations (5-6%) cell survival was enhanced. EGF also promoted cell survival. The effects were similar at the 24 and 48 h time points.

View larger version (26K): [in this window] [in a new window]   Figure 1.   The effects of EGF and CSE on H292 cell viability. Cells were cultured for 24 h (black bars) or 48 h (gray bars) in serum-free medium ± 1 nM EGF or dilutions of CSE. Cell viability was determined by measurement of LDH released into the culture medium. Comparisons between control and treated cells were made using an unpaired Student's t test, *P < 0.05; **P < 0.02; ***P < 0.001.

Effects of CSE on IL-8 Expression and Release

Exposure of H292 cells to a nontoxic concentration of CSE (5%) caused an increase in IL-8 gene transcription. This was evident at 6 h and may have continued to increase 24 h after exposure, although the data for 24 h did not achieve statistical significance when compared with the control (Figure 2A). The EGFR- selective tyrosine kinase inhibitor, tyrphostin AG1478, completely blocked IL-8 gene transcription caused by CSE (Figure 2A; P < 0.01). Consistent with the increase in mRNA levels, there was a dose-dependent increase in IL-8 release into culture supernatants of cells treated with CSE and 5% CSE was comparable in potency with 1nM EGF (Figure 2B). AG1478 completely blocked both CSE- and EGF-induced IL-8 release (Figure 2B; P < 0.001), consistent with a role for the EGFR as a mediator of the CSE-induced effects on H292 cells. Furthermore, this appeared to be due, in part, to the activity of autocrine ligands, as application of neutralizing antibodies for the EGFR reduced CSE-induced IL-8 release by ~ 45% (P < 0.05) (Figure 2B). Under the same conditions, these antibodies also reduced EGF-induced IL-8 release by 65% (P < 0.005).

View larger version (21K): [in this window] [in a new window]   Figure 2.   Induction of mRNA expression and release of IL-8 from H292 cells. (A) H292 cells were exposed to 1 nM EGF, 5% CSE, or 5% CSE + 1 µM AG1478 for 6 h (black bars) or 24 h (gray bars). mRNA expression was then determined by TAQMan RT-PCR. Data were normalized relative to 18S rRNA and are expressed as arbitrary units. Comparisons between control and treated cells were made using an unpaired Student's t test, *P < 0.02. (B) Cells were cultured for 48 h in serum-free medium ± CSE or EGF in the absence or presence of AG1478 or anti-EGFR antibody. The cell medium was then assayed for IL-8 by ELISA. Comparisons between control ± antibody and treated cells ± antibody were made using an unpaired Student's t test, *P < 0.02; **P < 0.005.

Effects of CSE on EGFR Ligand Expression and Release

Although we have previously shown that bronchial epithelial cells synthesize several EGFR ligands, including TGF-, AR, and HB-EGF (11), the effects of CSE on EGFR ligand production have not been determined. To identify which EGFR ligands might be mediating the epithelial responses to CSE, we measured mRNA expression for TGF-, AR, and HB-EGF by quantitative PCR. As shown in Figure 3, the basal level of expression of these factors was relatively low but was enhanced by treatment with either 5% CSE or EGF. Furthermore, it was evident that each ligand had a distinctive time course for induction. TGF- expression was increased 10-fold 6 h after stimulation with EGF, but had returned to basal levels 18 h later (Figure 3A). A similar response was observed for CSE, although in this case it did not achieve statistical significance. AR expression was elevated in response to CSE at 6 h but was not significantly different from the control at 24 h (Figure 3B). By comparison, the rise in HB-EGF expression proceeded more slowly with a time-dependent increase occurring over 24 h (Figure 3C). As observed for IL-8 gene transcription, the EGFR-selective kinase inhibitor AG1478 blocked the effects of CSE on both TGF- and HB-EGF mRNA expression (significance comparisons for TGF- release at 6 h, 5% CSE versus 5% CSE + AG1478: P < 0.02; HB-EGF release at 6 and 24 h, 5% CSE versus 5% CSE + AG1478: P < 0.002 and P < 0.01, respectively). In contrast, AR mRNA expression at 6 h was 60% greater in the presence of CSE plus AG1478 compared with CSE alone (P < 0.05), suggesting that EGFR signaling might negatively regulate AR expression in response to CSE.

View larger version (15K): [in this window] [in a new window]   Figure 3.   Induction of mRNA expression for TGF- (A), AR (B), and HB-EGF (C). H292 cells were exposed to 1 nM EGF, 5% CSE or 5% CSE + 1 µM AG1478 for 6 h (black bars) or 24 h (gray bars). mRNA expression was then determined by TAQMan RT-PCR. Data were normalized relative to 18S rRNA. Comparisons between control and treated cells were made using an unpaired Student's t test, *P < 0.01; **P < 0.005.

We next investigated whether the CSE-induced increases in mRNA expression were accompanied by enhanced secretion of proteins into the culture medium. We assayed the conditioned medium for TGF- and AR by ELISA. HB-EGF could not be determined due to lack of suitable antibodies for ELISA; however, the relatively slow time course for induction of this growth factor suggested that it was not a primary mediator of the epithelial responses to CSE. We detected a small dose-dependent increase in TGF- release in response to CSE (Figure 4A), and this could be blocked by AG1478 (P < 0.005). However, in the presence of an EGFR-neutralizing antibody that prevented ligand binding and internalization of the receptor-bound ligand, there was a marked accumulation of TGF- compared with CSE treatment alone (P < 0.002), suggesting that, in the absence of the antibody, TGF- was rapidly utilized by the cells. CSE also enhanced the release of AR in a dose-dependent manner (Figure 4B). In this case, the concentration of AR was only slightly, but significantly, higher in the presence of anti-EGFR antibody (P < 0.02). In contrast with its effects on mRNA expression (Figure 3B), AG1478 totally blocked CSE-stimulated AR release (P < 0.001) (Figure 4B).

View larger version (25K): [in this window] [in a new window]   Figure 4.   Release of TGF- (A) and AR (B) from H292 cells. Cells were cultured for 48 h in serum-free medium ± CSE in the absence or presence of AG1478 or anti-EGFR antibody. The culture medium was then assayed for ligand by ELISA. Comparisons between control ± antibody and treated cells ± antibody were made using an unpaired Student's t test, *P < 0.02; **P < 0.005.

TGF-, AR, and IL-8 Release From Primary Bronchial Epithelial Cells

Although H292 cells provide a convenient model of the airway epithelium, they are a tumor-derived cell line. As overexpression of growth factor receptors and ligands is common in such transformed cell lines, we determined whether CSE stimulated the release of TGF- and AR from primary bronchial epithelial cells derived from seven different subjects. Significance comparisons between control cells and cells treated with 5% CSE were made using the Wilcoxon test. As seen for H292 cells, CSE significantly induced the production of TGF- and AR in a dose-dependent manner (P < 0.02), and this was associated with a corresponding increase in IL-8 release (P < 0.02) (Figure 5).

View larger version (19K): [in this window] [in a new window]   Figure 5.   Release of TGF- (A), AR (B), and IL-8 (C) from primary bronchial epithelial cells. Cells from seven different subjects, represented by the different symbols, were cultured for 36 h in serum-free medium ± CSE. The cell medium was then assayed for ligand and IL-8 by ELISA. Data are the mean of duplicate determinations.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Production of IL-8 by bronchial epithelial cells in response to oxidants present in cigarette smoke is well documented and has been linked to oxidant-mediated activation of nuclear factor (NF)-B (9, 10). EGF has also recently been demonstrated to initiate release of IL-8 from bronchial epithelial cells (15), consistent with its ability to activate NF-B (18). Although CSE is able to cause rapid, ligand-independent phosphorylation of the EGFR (19), no studies have previously explored the involvement of the EGFR as a mediator of CSE-induced changes in IL-8 gene transcription and protein release, nor have the effects of CSE on expression and release of autocrine EGFR ligands been investigated. Thus, our findings that CSE not only induces expression and release of several ligands for the EGFR, but that neutralizing antibodies partially block CSE-induced IL-8 release indicates a novel, causal relationship between autocrine EGFR ligand release and IL-8 production in response to CSE. As oxidants present in CSE are highly reactive, short-lived molecules whose effects are likely to be transient, the capacity of CSE to affect EGFR ligand processing and gene expression probably accounts for its ability to induce a sustained proinflammatory response.

Comparison of the ability of the anti-EGFR antibody to reduce EGF-induced and CSE-induced IL-8 release showed ~ 65% and 45% inhibition, respectively. The partial effectiveness of the antibody toward EGF presumably reflects a failure to totally block binding of the ligand to the EGFR. However, the lower level of inhibition in the case of CSE suggests that stimulation of EGFR phosphorylation occurs by oxidant-mediated, ligand-independent mechanisms as previously reported (19), as well as by direct autocrine ligand binding. This would explain the capacity of tyrphostin AG1478 to more efficiently block the CSE-induced responses. AG1478 is a highly selective inhibitor of the EGFR kinase activity. We have shown previously that it blocked EGF-induced mitogenesis in murine fibroblasts but not that induced by acidic or basic fibroblast growth factor or keratinocyte growth factor (20). Moreover, the effects of AG1478 on intracellular EGFR signaling are reported to be identical to those of a dominant-negative EGFR (21, 22).

Although CSE is eventually toxic for bronchial epithelial cells, in this study we found that sub-toxic concentrations of CSE promoted H292 cell survival. Although this response may appear paradoxical, it can be readily explained in the light of the fact that CSE caused increased expression and release of several ligands for the EGFR that are known to exert protective effects on epithelial cell survival. This may reflect a typical stress response by the cells that may be mediated by direct activation of EGFR ligand-processing enzymes by agents present in CSE, as well as indirectly by induction of ligand gene transcription. This release of autocrine ligands is consistent with our previous studies of scrape-wounded 16HBE 14o cells, where phosphorylation of the EGFR occurred rapidly in response to wounding, irrespective of the presence of exogenous ligand (20).

Our findings that EGF induces autocrine ligand expression in bronchial epithelial cells is consistent with previous studies in keratinocytes (23) and colonic epithelial cells (24) where induction of each member of the EGF ligand family showed a distinct temporal pattern of expression. EGF has also been demonstrated to promote shedding of TGF- from its membrane-bound precursor by a mechanism involving the EGFR and downstream signaling via the mitogen-activated protein kinase (MAPK) pathway (25), compatible with our findings that AG1478 blocked autocrine ligand release. In contrast, our studies are the first to report a direct effect of CSE on EGFR ligand expression. Based on the observation that the transcription of TGF- and AR mRNA preceded that of HB-EGF and IL-8, it is likely that one of these ligands is a primary mediator of the CSE-induced responses. Because blocking antibodies to the EGFR caused a preferential accumulation of TGF- in the culture medium, we conclude that this is the active ligand. In contrast, whereas AR release was increased by CSE, its consumption by the epithelial cells appeared to be limited, possibly due to its lower affinity for the EGFR relative to TGF- (26). Induction of TGF- and HB-EGF mRNA expression by CSE was blocked by AG1478, indicating that EGFR signaling was required for this response. In contrast, induction of AR mRNA by CSE was slightly enhanced by AG1478, suggesting that other unidentified pathways are activated by components of CSE that may contribute to AR gene expression. As AR release was selectively blocked by AG1478, this would be expected to lead to accumulation of the cell-associated growth factor precursor. Because the several members of the EGF ligand family have been shown to be proapoptotic in their unprocessed form (27), this dissociation between gene expression and release may be associated with the eventual toxic effects of CSE.

According to studies by Takeyama and coworkers (7, 19), reactive oxygen species produced by neutrophils cause ligand-independent transactivation of the EGFR, leading to goblet cell metaplasia and excess mucus secretion (30). However, antioxidants only partially inhibit epithelial cell mucin synthesis in response to CS (19), suggesting that other mechanisms also contribute to this response. As TGF- has been shown to be involved in the process of goblet cell differentiation (31), our studies suggest that the ability of CSE to induce expression and release of TGF- provides a direct mechanism whereby CS affects epithelial function and contributes to the mucus hypersecretory phenotype without the involvement of neutrophils.

In conclusion, we provide evidence that CSE utilizes the EGFR to stimulate IL-8 gene expression and release and that autocrine EGFR ligands are mediators of these responses. The ability of CSE to induce expression and release of several ligands for the EGFR suggests that it may have other direct effects on epithelial function linked to cell survival and differentiation. Furthermore, release of EGFR ligands may have other paracrine effects on underlying mesenchymal cells linked to airway remodeling.