Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The respiratory epithelium represents the first line of lung defense and is frequently exposed to different, potentially damaging agents, such as infectious bacteria and viruses, pollutants, toxic materials, and mechanical or inflammatory insult. Although the existence of epithelial damage in asthma, bronchitis, and bronchiolitis is widely accepted, the behavior and properties of epithelial cells during restitution and the involvement of soluble mediators and their cellular receptors in these processes are poorly understood.

The epidermal growth factor receptor (EGFR) plays a prominent role in the maintenance and repair of epithelial tissues. This receptor tyrosine kinase can be activated by one of several structurally related ligands including epidermal growth factor (EGF) (1), transforming growth factor- (TGF-) (2), heparin-binding EGF-like growth factor (HB-EGF) (3), amphiregulin (AR) (4), betacellulin (BTC) (5), and epiregulin (6). The effects of these growth factors on target cells are pleiotropic, ranging from induction of DNA synthesis and alterations in cell adhesion and motility to stimulation of differentiated cell function (7). This ability of growth factors to regulate several facets of cell behavior is probably an important factor in controlling the individual phases of tissue restitution. Indeed, a direct role for EGF, TGF-, and to a lesser extent HB-EGF in cutaneous wound healing is already well established (8, 9). The involvement of EGF-like growth factors in human lung repair has been suggested by the observation that TGF- is present in edema fluid of patients with acute lung injury (10). Moreover, increased EGFR and EGF immunoreactivities have recently been reported in asthmatic airways (11). In rats, the concentration of TGF- is reported to increase after bleomycin-induced lung injury (12), and HB-EGF is increased in experimentally induced pulmonary hypertension (13).

EGFR is the prototype member of the c-erbB receptor-coupled tyrosine kinase family, which comprises EGFR (c-erbB1), human EGF receptor-1 (HER1), c-erbB2 (HER2), c-erbB3 (HER3), and c-erbB4 (HER4) (14). Binding of cognate ligand appears to stabilize erbB receptors in an activated dimeric form (15). In recent years it has been recognized that in addition to the formation of homodimers, the repertoire of activated c-erbB receptors can be expanded through formation of heterodimers comprising two different members of the family (15, 16). The importance of this heterologous association is that a specific family member can be activated in the absence of its cognate ligand. For example, EGF induces tyrosine phosphorylation of c-erbB3 through formation of EGFR/c-erbB3 heterodimers (17); similarly, EGFR can become activated by heregulin- (a ligand for c-erbB3 and c-erbB4) through the formation of EGFR/c-erbB4 heterodimers (18). Heterodimerization has important consequences on the affinity of the receptor for ligand (19, 20), intracellular signaling (19, 21), and cellular responses (22). It is therefore likely that the pleiotropic effects of the EGFR ligands are mediated at least in part by heterodimeric receptors. This complex network of interactions between c-erbB receptors is further expanded by the broader receptor specificity of BTC (23) and HB-EGF (24), which recognize c-erbB4 as well as EGFR. Thus, there exists a complex combinatorial relationship within the c-erbB receptor and ligand families that offers the potential to finely regulate the molecular and cellular processes that need to be coordinated to effect tissue repair.

Although the presence of EGF and EGFR in human lung tissues has been demonstrated by radioimmunoassay (25) and immunohistochemistry (26), much of the work on the c-erbB family has been done in the context of cancer (27), in which scant attention has been paid to the fine detail of receptor and ligand expression within the bronchial mucosa. Furthermore, no studies have determined whether bronchial mucosal cells coexpress more than one member of the c-erbB family. In the present study, immunohistochemistry was used to examine the distribution of the c-erbB-family receptors and their ligands in human bronchial mucosa and in human bronchial epithelial cells grown to confluence on coverslips, whereas the identity of these substances' mRNAs was confirmed in cultures derived from bronchial epithelial cells by means of reverse transcription-polymerase chain reaction (RT-PCR).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Tissue Specimens

Samples of human bronchial epithelium were obtained from material removed from six subjects undergoing surgical resection procedures. In all cases specimens were taken from the affected lung distant from the lesion requiring surgery. All samples were studied by a pathologist to confirm the absence of abnormality. Specimens were processed into glycolmethacrylate (GMA) resin (Park Scientific, Northampton, UK) as previously described (31).

Cell Culture

The H292 human epithelial lung cancer cell line (32) was obtained from the American Type Culture Collection and grown in RPMI 1640 medium supplemented with 10% (vol/vol) heat-inactivated fetal bovine serum (FBS). The virally transformed human bronchial epithelial cell line 16HBE 14o- (33) was a gift from Dr. D. C. Gruenert of the Cardiovascular Research Institute of the University of California, San Francisco, and was grown in Eagle's modified essential medium (EMEM), supplemented with 10% (vol/vol) FBS. Primary cultures were grown from explants of mucosa that had been microdissected away from underlying connective tissue of bronchial airway specimens obtained from surgical resection. Small, 2-3-mm portions of the explants were plated onto Primaria culture dishes (Becton Dickinson, Oxford, UK) and cultured for 2-3 wk in modified M199 medium (GIBCO BRL, Paisley, Scotland). During this time, epithelial cells grew to form a confluent monolayer that was 3-4 cm in diameter around each portion of tissue. These cells have been shown to retain epithelial characteristics in vitro, including expression of cytokeratins and desmosomes (34). All cells were grown to around 90% confluence, unless otherwise indicated, before being harvested for RT-PCR analysis, flow cytometry, or immunocytochemistry.

Detection of c-erbB Receptors and Ligands by RT-PCR

Total RNA was extracted from cultured cells through the acid guanidinium-thiocyanate-phenol-chloroform method (35). RT and nested PCR for detection of EGF, TGF-, HB-EGF, and AR mRNA was done as previously described (36). Detection of mRNA for the four c-erbB receptors was done essentially according to the same protocol as that for the EGFR ligands, except that amplification was done with the sequences and annealing temperatures shown in Table 1. Nested primers were not used for detection of the c-erbB receptors. To ensure RNA quality, all preparations were subjected to analysis of -actin expression (36).

Antibodies

Two anti-EGFR antibodies were used: a sheep anti-EGFR polyclonal antibody (an IgG fraction obtained from immune serum raised against EGFRs that had been purified from A431-cell plasma membranes by EGF-affinity chromatography), and a mouse monoclonal anti-EGFR1 antibody (37). Both the mouse monoclonal antibody against c-erbB2 and the rabbit polyclonal antibody against c-erbB3 were obtained from Transduction Laboratories (supplied by Affiniti Research Products, Ltd., Exeter, UK). Immunoprecipitation experiments confirmed that each anti-c-erbB receptor antibody was specific for an individual c-erbB receptor and did not cross-react with other members of the c-erbB family (data not shown). The mouse monoclonal anti-EGF antibody (clone 3D3, [38]) was a gift from Prof. K. Nishikawa (Kanazawa Medical University, Kanazawa, Japan), and that against TGF- (clone Ab-20) was purchased from Cambridge BioScience (Cambridge, UK). Rabbit polyclonal antibodies to AR were as previously described (39); chicken anti-HB-EGF raised against the intracellular domain of HB-EGF was a gift from Dr. R. Adam (Childrens Hospital, Boston, MA), and affinity-purified goat anti-BTC was from R&D Systems (Abingdon, UK). None of the antibodies showed cross- reactivity toward other members of the EGF ligand family by Western blot analysis or enzyme-linked immunosorbent assay (ELISA) (data not shown). Optimal dilutions of antibodies for staining tissues (and cells) were determined by titration, and were as follows: sheep anti-EGFR: 25 µg/ml (50 µg/ml); mouse anti-EGFR: 1:40; mouse anti- c-erbB2: 1:50 (1:25); rabbit anti-c-erbB3: 1:20 (1:20); mouse anti-EGF: 1:20 (1:20); mouse anti-TGF-: 1:20 (1:20); chicken anti-HB-EGF: 1:100 (1:50); rabbit anti-AR: 1:100; goat anti-BTC: 1:20. For fluorescence-activated cell sorting (FACS) analysis, antibodies were routinely used at twice the concentration required for immunocytochemistry. Peroxidase-conjugated rabbit antisheep/goat immunoglobulins (Dako, Wycombe, UK), and rabbit antichicken immunoglobulins (Jackson ImmunoResearch Laboratories Inc., West Grove, PA) were used at 1:100 or 1:500, respectively. Biotinylated IgG Fab fragments (Dako) were used as follows: swine antirabbit at 1:300, rabbit antigoat at 1:200, and rabbit antimouse at 1:300. Fluorescein isothiocyanate (FITC)-conjugated secondary antibodies were used at 1:80, 1:150, and 1:50 for sheep, rabbit, and mouse immunoglobulins, respectively.

Immunohistochemistry

The technique described by Britten and colleagues (31) was applied to 2-µm sections of GMA-embedded tissue. Endogenous peroxidase activity was inhibited before blocking with DMEM containing 10% (vol/vol) FBS and 1% (wt/vol) bovine serum albumin (BSA). The sections were then incubated with a panel of primary antibodies for either 1 h (for polyclonal antibodies) or overnight (for monoclonal antibodies) at room temperature. After washing, relevant species-specific, biotinylated IgG Fab fragments were applied to the sections for either 1 h (for polyclonal antibodies) or 2 h (for monoclonal antibodies). This was followed by incubation with streptavidin-biotin- horseradish peroxidase complex at 1:200 (Dako) and visualization with 0.02% aminoethyl carbazole in acetate buffer (pH 5.2) containing 0.3% H₂O₂. Sections were counterstained with Mayer's haematoxylin and mounted in p-xylene-bis-pyridium bromide (DPX).

For immunocytochemical staining of cultured bronchial epithelial cells, cells were grown to 90% confluence on sterile coverslips before fixation in cold methanol for 10 min. After blocking for 20 min with DMEM containing 10% FBS and 1% BSA, cells were first incubated with a panel of primary antibodies and then with relevant peroxidase-conjugated secondary antibodies before being visualized with 0.5 mg/ml diaminobenzidine (DAB) in PBS containing 0.01% H₂O₂. The coverslips were counterstained with Harris's haematoxylin and mounted in DPX.

All immunostaining experiments included control slides unexposed to primary antibody, with substitution of an unrelated antibody of the same isotype or preincubation of the antibody with a 10-fold molar excess of immunizing peptide (EGF, TGF-, AR, HB-EGF, BTC). In the case of the anti-EGFR antibody, control experiments were performed by preincubating the antibody with detergent-solubilized plasma membrane vesicles (1 mg/ml) prepared from the EGFR-overexpressing A431 vulval carcinoma cell line.

Flow Cytometry

Cultures of bronchial epithelial cells detached from dishes through the use of cell dissociation solution (Sigma Company LTD, Poole, Dorset, UK), washed, counted, and suspended in incubation medium comprising Hanks' balanced salt solution containing 2% (vol/vol) FBS and 0.1% (wt/vol) sodium azide. Aliquots containing 5 × 10⁴ cells were incubated for 1 h on ice in the absence or presence of primary antibody (see the preceding discussion), washed twice, and then incubated with FITC-labeled secondary antibodies for 1 h at 4° C. After two further washes, cells were resuspended in 0.5 ml of incubation medium and surface expression of EGFR, c-erbB2, and c-erbB3 was analyzed with a Becton Dickinson FACScan with Lysis II software. For each tube, 10,000 events were collected.

In order to determine whether cell density affected cell-surface c-erbB receptor levels, H292 and 16HBE 14o cells were seeded into 90-mm diameter Petri dishes over a range of cell densities, from 4 × 10² to 1 × 10⁵ cells/cm². In order to reduce effects of nutrient depletion, high- and low-density cultures were seeded into the same culture vessel, which had been subdivided into two with a plastic spacer sealed with sterile petroleum jelly. After the cells were allowed to adhere for 12 h, the spacer was removed and the cells were cultured in the same medium until the highest density culture had been confluent for 3 d, after which FACS analysis was done on all cultures.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Expression of the c-erbB Receptor Family and Ligands by Human Bronchial Epithelial Cells Cultured In Vitro

RT-PCR of mRNA extracted from primary cultures of human bronchial epithelial cells or from two well-established bronchial epithelial cell lines (H292 and 16HBE 14o-) demonstrated the presence of mRNA transcripts encoding c-erbB1, c-erbB2, and c-erbB3. No c-erbB4 mRNA was detected (Figure 1, upper panel, and Table 2); this was not due to any failure of the RT-PCR methodology, since a positive signal was detected in a control cell line known to express c-erbB4 (CB4 cells [40], a gift from Professor Y. Yarden of the Weizmann Institute, Rehovot, Israel) (data not shown). Transcripts encoding EGF, TGF-, HB-EGF, and AR were also detected in the primary cell cultures and cell lines (Figure 1, lower panel, and Table 2).

View larger version (54K): [in this window] [in a new window]   Figure 1.   RT-PCR analysis of the expression of c-erbB receptors (upper panel ) and ligands (lower panel ) in human bronchial epithelial cells. RNA was isolated from H292, 16HBE 14o, or primary human bronchial epithelial cells (HBEC) and was reverse transcribed, and the cDNA was amplified with use of the primers listed in Table 1. In upper panel, L = 123bp ladder; 1 = EGFR/c-erbB1; 2 = c-erbB2; 3 = c-erbB3; and 4 = c-erbB4. In lower panel, H = HB-EGF, A = amphiregulin, E = EGF, and T = TGF-. Blank = negative control.

In accordance with the demonstration of mRNA transcripts by RT-PCR, immunocytochemical staining of the cultured cells confirmed the presence of EGFR, c-erbB2, and c-erbB3. Thus, 16HBE 14o, H292, and primary human bronchial epithelial cells all exhibited membrane staining for EGFR (Figure 2a and Table 2), c-erbB2 (Figure 2c and Table 2), and c-erbB3 (Figure 2d and Table 2); we also observed that H292 cells consistently exhibited additional cytoplasmic staining for EGFR (not shown). Preadsorption of the anti-EGFR antibody with detergent-solubilized A431-cell plasma membrane vesicles abolished EGFR immunostaining (Figure 2b).

View larger version (129K): [in this window] [in a new window]   Figure 2.   Immunocytochemical staining of 16HBE 14o cultures of human bronchial epithelial cells. Cultures were stained for the presence of EGFR (a and b), c-erbB2 (c), or c-erbB3 (d) as described in the MATERIALS AND METHODS section. In (b), the anti-EGFR antibody was preadsorbed with detergent-solubilized, EGFR-rich, A431-cell plasma membrane vesicles. Scale bar = 20 µm.

Surface expression of EGFR in near-confluent, unstimulated H292, 16HBE 14o-, and early-passage primary bronchial epithelial cells was evaluated through flow cytometry. As found by immunocytochemistry, all the cultured airway epithelial cells expressed EGFR (Figure 3), with H292 cells showing slightly lower expression than either 16HBE 14o or the primary cultures. In order to determine whether cell density influenced surface c-erbB receptor levels, we performed a single experiment using 16HBE 14o and H292 cells, each cultured at six different cell densities; no difference in EGFR, c-erbB2, or c-erbB3 immunofluorescence was observed through FACS analysis under the conditions used (data not shown).

View larger version (14K): [in this window] [in a new window]   Figure 3.   FACS analysis of surface EGFR expression on cultured human bronchial epithelial cells. Live, unfixed H292 cells (A), 16HBE 14o cells (B), and primary HBEC (C ) were stained for the presence of EGFR as described in the MATERIALS AND METHODS section. Representative histograms of cells stained either with sheep anti-EGFR antibodies (black peaks) or no primary antibody (white peaks) are shown.

The presence of EGF, TGF-, and HB-EGF was also confirmed in each cell line (Table 2), in which staining was found to be diffuse and cytoplasmic. In none of these experiments was labeling observed when the primary antibodies were omitted from the immunocytochemical protocol or when an irrelevant antibody from the same species was used in place of the primary antibody (data not shown).

Detection of EGFR and Other c-erbB Receptors in Normal Human Bronchial Mucosa

Immunohistochemical staining of GMA-embedded sections of surgical specimens of bronchial mucosa showed strong staining for EGFR in the submucosal glands and bronchial epithelium, as seen in Figure 4a. The endothelium was also consistently stained by anti-EGFR antibody, whereas the submucosal connective tissue did not stain. Positive EGFR immunostaining in the epithelium (Figure 4b) and endothelium was abolished by preadsorption of the antibody with EGFR-rich A431-cell plasma membrane vesicles, as shown for the bronchial epithelium in Figure 4c; staining in the submucosal glands was also diminished, but some residual staining was attributed to nonspecific binding of the antibody to mucus. A similar distribution of staining within the bronchial mucosa was obtained with the anti- c-erbB2 and anti-c-erbB3 antibodies.

View larger version (156K): [in this window] [in a new window]   Figure 4.   Immunohistochemical staining of c-erbB receptors in GMA-embedded sections of human bronchial epithelium. Sections were stained as described in the MATERIALS AND METHODS section, using sheep anti-EGFR (a, b, and c); mouse anti-EGFR (EGFR1) (d), mouse anti-c-erbB2 (e), and rabbit anti-c-erbB3 antibodies ( f ). In c, the sheep anti-EGFR antibody was preadsorbed with detergent-solubilized, EGFR-rich, A431-cell plasma membrane vesicles. In a, the arrows indicate the bronchial epithelium, e = endothelium, and g = glands. Some staining in the submucosal glands was attributed to nonspecific binding to mucus. Scale bars = 100 µm (a) and 30 µm (b-f ). For all high-power plates, the sections are oriented such that the lumen is at the top and the mucosa at the bottom.

Within the bronchial epithelium, EGFR was detected in all specimens, with strong immunostaining associated with the cell membrane of epithelial cells (Figure 4b). This was particularly evident between basal cells and the basal aspect of columnar cells, although weak reactivity was seen on the brush border of the epithelium. A similar pattern of staining was observed with a monoclonal antibody directed against EGFR (clone EGFR1) (Figure 4d). The patterns of staining in the bronchial epithelium were similar for c-erbB2 (Figure 4e) and c-erbB3 (Figure 4f), although staining for c-erbB2 was considerably weaker.

Detection of the EGFR Ligand Family in Normal Human Bronchial Mucosa

Intense immunostaining for EGF, TGF-, HB-EGF, AR, and BTC was observed in the submucosal glands. Bronchial epithelial immunostaining for EGF, AR, TGF-, and HB-EGF was particularly evident between basal cells and the basal aspect of columnar cells (Figures 5a to 5d), in a pattern similar to that seen for c-erbB receptors. In each case, no labeling was observed when the primary antibodies were preadsorbed with immunizing peptide, as shown in Figure 5e for EGF; similarly, no staining was observed when the primary antibody was omitted or when an irrelevant antibody of the same isotype was used in place of the primary antibodies (not shown), thus confirming the specificity of the immunolocalization method. In the case of BTC, no epithelial staining was evident (Figure 5f), even though strong and specific immunostaining was demonstrated in the submucosal glands of the bronchial mucosa (Figure 5g versus Figure 5h).

View larger version (125K): [in this window] [in a new window]   Figure 5.   Immunohistochemical staining of EGFR ligands in GMA-embedded sections of human bronchial epithelium and submucosal glands. Sections were stained for the presence of EGF (a, e), AR (b), TGF- (c), HB-EGF (d), and BTC ( f-h) as described in the MATERIALS AND METHODS section. In e and h, the antibodies were preadsorbed with EGF and BTC, respectively. Scale bar = 30 µm. Tissue orientation is as in Figure 4.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Numerous immunohistochemical studies have demonstrated widespread expression of EGFR and c-erbB2 in neoplastic human epithelial tissues, including lung carcinomas (27- 30). In these studies, areas of histologically normal bronchial epithelium were also examined, and EGFR was found to be localized to basal cells, whereas weak immunostaining for c-erbB2 was found in all layers of the pseudostratified epithelium. More recently, precise localization of EGFR in the adult human lung has been described with immunoelectron microscopy (26). In this latter study, EGFR immunoreactivity was found on basal cells of the bronchial epithelium, and was limited to the intercellular lateral cell membrane, whereas the basal surface attached to the basement membrane was negative. Thus, our data for immunolocalization of EGFR and c-erbB2 in GMA-embedded sections of bronchial epithelium are consistent with previously published data.

Gullick and colleagues (41) have reported moderate levels of expression of c-erbB3 in bronchial epithelium; however, the precise localization of this c-erbB3 was not described. Our finding that c-erbB3 expression paralleled that of EGFR and c-erbB2 suggests the possibility that these receptors are coexpressed within the same cell, and are able to form heterodimers that can regulate intracellular signaling within the basal cells of the bronchial epithelium. Perturbation of the level of expression of even just one of these receptors, as may occur after epithelial damage, could dramatically alter the proportions of heterodimeric receptor combinations (42), leading to activation of a different subset of signaling intermediates.

Information on the expression of c-erbB4 in the airways is severely limited, presumably because few antibodies to this protein are available. Indeed, the lack of a specific anti-c-erbB4 antibody precluded our determination of the expression of this receptor in the bronchial mucosa. However, our studies with RT-PCR and primary bronchial epithelial cells suggest that c-erbB4 is not expressed by this cell type. Although low levels of c-erbB4 mRNA have been detected with Northern blot analysis of mRNA extracted from whole lung (43), it is possible that these transcripts were derived from neuronal or muscle cells, since these types of cells are known to express high levels of c-erbB4.

In order to determine whether EGFR could be functionally active within the bronchial epithelium, we also examined the occurence of five related EGF-like peptides in human bronchial epithelium. Each of these growth factors is synthesized as a membrane bound precursor and is proteolytically cleaved to generate the mature growth factor (44). In the case of EGF and TGF-, the growth factors are freely diffusible, and detection of immunoreactivity associated with a particular cell does not necessarily identify that cell as the source of the growth factor. Furthermore, AR and HB-EGF have heparin-binding domains (3, 4) that facilitate their interaction, and hence localization, with heparan sulfate proteoglycans on the cell surface or in the extracellular matrix. Because many inflammatory cells including macrophages, platelets, eosinophils, and T lymphocytes are known to synthesize TGF-, EGF, or HB-EGF (3, 45), the contribution of these cells to provision of EGFR ligands cannot be ignored. However, the ability of bronchial epithelial cells to synthesize EGFR ligands was evident in the present study, and also in a previous study (50), in which AR, TGF-, HB-EGF, and EGF mRNAs were detected in lung tissue with RT-PCR or by Northern blotting. Because bronchial epithelial cells require EGF for growth in vitro, it is possible that this growth factor is responsible for induction of autocrine ligand expression, as has been observed in cultured keratinocytes (51) and colonic epithelial cells (52). The ability of EGF to induce autocrine ligand expression in vitro may reflect an important mechanism for sustaining tissue repair after injury.

In accordance with a previous immunohistochemical study of EGF expression in human lung, we observed that serous acinar cells are a major site of EGF immunoreactivity. However, we extended these observations by demonstrating glandular expression of TGF-, HB-EGF, AR, and BTC in human lung. It is likely that all of these growth factors contribute to the EGF-like growth factor activity that is secreted into the fluid that bathes the apical surface of the bronchial epithelium (53). In our study we also observed immunostaining for EGF, TGF-, HB-EGF and AR, but not BTC, between the basal and columnar epithelial cells. This pattern of staining was similar to that observed for EGFR, suggesting that it may represent ligand bound to receptor. However, in the case of TGF-, HB- EGF, and AR, staining was also particularly evident within the cytoplasm of the columnar epithelial cells, where it was confined to the perinuclear and basal regions of the cell, suggesting polarized export of the ligand to the basal aspect of the columnar cell. Basolateral release of TGF- (54) and AR (55) has previously been observed in polarized cultures of colonic epithelial cells, suggesting that there may be a common mechanism for directed release of these ligands by epithelial cells. The significance of the basolateral localization of EGF-like peptides in columnar epithelial cells is evident when viewed in the context of EGFR expression. Our staining data suggest that a juxtacrine mechanism may exist in which ligands are presented by the columnar epithelial cells to the EGFRs present on basal cells. Juxtacrine ligand synthesis appears to be an important regulator of c-erbB4 and c-erbB2 activity in neuronal and cardiac development (56), and a similar mechanism acting on the EGFR may control basal cell function in bronchial epithelium. Further studies, using in situ hybridization and immunoelectron microscopy, will be required to study the cells responsible for the synthesis of EGFR ligands and the exact cellular localization of these ligands.

Although autocrine ligand synthesis was originally proposed to lead to malignancy through uncontrolled stimulation of the EGFR, it is clear that malignant transformation occurs only when the EGFR is expressed at extremely high levels (57). Under conditions of normal receptor expression, autocrine, as well as juxtacrine and paracrine, stimulatory mechanisms for the EGFR can contribute to normal cell behavior. For example, parallel expression of EGFR, EGF, and TGF- has been observed in developing and postnatal human lung, and it has been suggested that these growth factors regulate lung development and maturation through an autocrine mechanism (58). Furthermore, IFN- has been found to induce prostaglandin G/H synthase-2 (PGHS-2) by an indirect mechanism involving upregulation of TGF-, HB-EGF, and AR, which in turn activate an autocrine EGFR-mediated signaling pathway leading to induction of PGHS-2 expression (59). Although little is known about EGF-mediated regulation of bronchial epithelial repair in humans, EGF is known to be involved in bronchoalveolar repair in experimental animals through effects on cell migration (60) and proliferation (61).

In summary, the present study indicates that bronchial epithelial cells coexpress several members of the c-erbB family of receptor tyrosine kinases, suggesting the potential for productive c-erbB receptor interactions in bronchial epithelium. Further study of these interactions may help to define their role in activation of the bronchial epithelium in response to toxic insult, infection, and inflammation, as well as their role in maintenance and repair of this crucial barrier, whose function it is to protect the airway microenvironment from external stimuli.