Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Although asthma is an inflammatory disorder of the conducting airways, high-resolution computer tomography and postmortem and biopsy studies have revealed airway wall thickening comprising changes in the epithelium and underlying mesenchyme. This involves epithelial damage thickening of the lamina reticularis, smooth-muscle hyperplasia, microvascular congestion, edema, and neuronal proliferation. Although mild to moderate asthma can be treated effectively with inhaled corticosteroids, there are many asthmatics with chronic symptoms who have a greatly impaired quality of life (1), exhibit a component of fixed airflow obstruction, and have clear evidence of airway wall remodeling (2). By affecting the airway structure and its mechanical and functional properties, remodeling provides an explanation for the incomplete resolution of bronchial hyperresponsiveness (BHR) with inhaled corticosteroids (3) and the accelerated decline in lung function that has been observed over time (4).

The deposition of interstitial collagens in the lamina reticularis is a unique feature of asthma. This appears to be due to the activity of subepithelial mesenchymal cells with features of myofibroblasts whose numbers are increased in asthma in proportion to the thickness of the lamina reticularis (5). Myofibroblasts have a phenotype intermediate between that of a fibroblast and a smooth-muscle cell (SMC), and ultrastructural analyses of these cells in bronchial tissue has shown characteristic intracellular bundles of filaments, abundant rough endoplasmic reticulum, and irregularly shaped nuclei (6). Transforming growth factor (TGF)-, whose levels are elevated in bronchoalveolar lavage fluid (BALF) from asthmatic subjects (7), is a potent inducer of myofibroblast differentiation (8); and it has been reported that it can promote survival of rat lung myofibroblasts by blocking interleukin (IL)-1-induced apoptosis (9).

Most forms of asthma are characterized by mast cell and eosinophil infiltration resulting from T-lymphocyte polarization toward a T helper (Th)-2 phenotype leading to coordinate secretion of IL-3, -4, -5, -9, and -13 and granulocyte macrophage colony-stimulating factor encoded in a cluster on chromosome 5q_31-33 (10). IL-4 and IL-13 are key cytokines in this repertoire on account of their roles in T-cell differentiation toward a Th-2 phenotype and isotype switching of B cells to immunoglobulin (Ig) E production (11). However, more recently, attention has focused on their role in airway remodeling and BHR due to their effects in transgenic animal models (12). In particular, overexpression of IL-13 in the bronchial epithelium of mice causes goblet-cell metaplasia, subepithelial fibrosis, and smooth-muscle proliferation associated with marked BHR, in addition to lymphocyte and eosinophil infiltration (13). Similarly, in a murine model of allergic asthma, blockade of IL-13 using a soluble fusion protein prevented allergen-induced asthma including increases in mucus cell numbers in the airways (14). By comparison, mice expressing an IL-4 transgene had goblet-cell metaplasia and high levels of mononuclear cells in the lungs but an absence of airway wall fibrosis or BHR (15).

In view of the differential effects of IL-4 and IL-13 on subepithelial fibrosis in these transgenic mouse models, we compared their ability to act as profibrogenic factors in asthma. We established bronchial fibroblast cultures from atopic asthmatic subjects and compared the ability of IL-4 and IL-13 to promote myofibroblast transformation with that of TGF-. Because the bronchial epithelium and underlying attenuated fibroblast sheath function as a trophic unitthe epithelial-mesenchymal trophic unit (EMTU) (16, 17)through bidirectional provision of growth and survival factors, we also examined the ability of IL-4 and IL-13 to stimulate epithelial production of TGF-. Our results suggest that whereas the Th-2 cytokines cause direct activation of fibroblasts with a proinflammatory outcome, their effects on subepithelial fibrosis and airway wall remodeling in asthma are indirect and are a result of modulation of TGF- release from bronchial epithelial cells. Further, unlike findings in experimental animals, we could not distinguish any differences between IL-4 and IL-13.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Subjects

Subjects with mild to moderately severe asthma were characterized according to symptoms, pulmonary function, and medication. Assessment of asthma severity was in accordance with the GINA guidelines on the diagnosis and management of asthma (18). Bronchial biopsies were obtained from four individuals with a mean age of 35 yr, two of whom were receiving inhaled steroids and two who were receiving Ventolin only. The mean forced expiratory volume in 1 s (FEV₁) for this group was 84.5 (percent of predicted FEV₁; range 53 to 114). Brushed epithelial cells were obtained from eight individuals with a mean age of 28 and mean FEV₁ of 83 (percent of predicted FEV₁; range 69 to 103). All subjects were nonsmokers and were free from respiratory tract infections for a minimum of 4 wk. Written informed consent was obtained from all volunteers and ethical approval was obtained from the Joint Ethics Committee of Southampton University and General Hospital. All subjects were tested for atopy using a panel of common aeroallergens, including house dust mite (HDM) extract (Dermatophagoides pteronissinus), grass pollen, tree pollen, cat dander, dog dander, candida, aspergillus, as well as negative (saline) and positive (histamine) controls. Tests were considered positive if a wheal response of 3 mm greater than the negative control was observed. Subjects with positive responses to HDM were selected for inclusion in the study.

Fiberoptic Bronchoscopy

Bronchial biopsies and bronchial brushings were obtained by bronchoscopy using a fiberoptic bronchoscope (Olympus FB-20D; Olympus, Tokyo, Japan) in accordance with standard published guidelines (19). Those moderately asthmatic subjects treated with inhaled corticosteroids withheld this medication for a minimum of 1 wk before bronchoscopy. After an overnight fast, subjects received premedication with nebulized salbutamol (2.5 mg) and intravenous atropine (0.6 mg). Light sedation was achieved using midazolam (0 to 5 mg intravenously). Local anesthesia was achieved by applying topical 10% lignocaine spray to the oropharynx and 1% lignocaine solution to the lower airways via the bronchoscope. Bronchial biopsies were obtained using alligator forceps, whereas epithelial cells were obtained using a standard sterile single-sheathed nylon cytology brush. This was passed by direct vision via the bronchoscope channel into the lower airways, and five to six consecutive brushings were sampled from the bronchial mucosa of the second- and third-generation bronchi. Cells were harvested into 5 ml sterile phosphate-buffered saline (PBS) after each brushing. At the completion of the procedure, 5 ml of 20% fetal bovine serum (FBS)/ RPMI was added and the sample was centrifuged at 150 × g for 5 min to pellet the cell suspension.

Primary Bronchial Fibroblast Cultures

Six submucosal biopsies from each subject were placed in a petri dish with 10% FBS/Dulbecco's modified Eagle's medium (DMEM) containing 50 IU/ml penicillin, 50 µg/ml streptomycin, and 2 mM L-glutamine, and were cut into pieces using sterile scalpel blades. The tissues were incubated in a humified incubator at 37°C, 5% CO₂, for approximately 1 wk, during which time fibroblasts migrated from the tissue and proliferated on the base of the culture dish. The fibroblast cultures were then passaged weekly. Cultures were used for assays between passages 2 and 8.

Primary Bronchial Epithelial Cell Cultures

The characterization and growth properties of brushed bronchial epithelial cells have been described elsewhere (20). The cells were cultured at 37°C, 5% CO₂, in Bronchial Epithelial Growth Medium (BEGM; Clonetics, San Diego, CA). For assays, cells (passage 2 or 3) were seeded into 24-well trays (1 ml BEGM/well containing 5 × 10⁴ cells/ml) and cultured until approximately 80% confluent. The medium was then replaced by 1 ml/well of Bronchial Epithelial Basal Medium (BEBM; Clonetics) containing 1% of insulin, transferrin, and sodium selenite (ITS) media supplement (Sigma, Poole, Dorset, UK) for 24 h to render the cells quiescent. The medium was then replaced with 1 ml/well of BEBM/ITS in the absence or presence of 20 ng/ml IL-4, IL-13 (Peprotech EC, London, UK), and/or 1 µM dexamethasone. The cells were incubated for a further 48 h. After this period the medium was removed from the cells and assayed for TGF-2 by enzyme-linked immunosorbent assay (ELISA) (see later section).

Fibroblast Proliferation Assays

Cells were seeded into 24-well trays at 2.5 × 10⁴ cells/well, 500 µl/ well 10% FBS/DMEM, and incubated at 37°C for 6 h. The medium was then changed to serum-free medium (SFM) (Ultraculture; Biowhittaker, Wokingham, Berks, UK) in the absence or presence of TGF-1 (Sigma), TGF-2 (Sigma), IL-4, or IL-13. At specified time periods the medium was removed and the cells were fixed with 500 µl/well of formol saline (4% formaldehyde and 0.15 M NaCl) at room temperature for a minimum period of 1 h. Methylene blue dye was used to assess the growth of fibroblast and epithelial cell cultures (21). The fixed cells were stained with 250 µl/well of 1% methylene blue in 10 mM disodium tetraborate, pH 8.5, for 30 min. Excess dye was washed from the trays with 10 mM disodium tetraborate, pH 8.5, and the trays were then blotted. The dye was extracted from the cells by addition of 500 µl/well of 1:1 0.1 M HCl/ethanol for 30 min at room temperature. The absorbance at 630 nm (A₆₃₀) of individual wells was determined using a microplate spectrophotometer (MultiScan Ascent; Affinity Sensors, Cambridge, UK). A₆₃₀ was directly proportional to cell number over the range of observed cell densities.

ELISA for -Smooth Muscle Actin

Cells were plated into 96-well trays (1 × 10⁴ cells/well in 100 µl 10% FBS/DMEM) and incubated at 37°C for 6 h. The medium was then changed to Ultraculture without or with TGF-2, IL-4, or IL-13 at the doses stated in RESULTS. After 3 d the medium was removed and the cell monolayers were air-dried. The cells were then fixed and permeabilized with 100 µl/well of methanol for 30 min and air-dried again. The quantity of 100 µl of antibody buffer (PBS containing 1% bovine serum albumin, 1% ovalbumin, and 0.1% Tween-20) was added to each well and incubated at 37°C for 30 min to block nonspecific binding sites. The cells were then incubated with anti--smooth muscle actin (-SMA) antibody (1/2,000 in antibody buffer) at 37°C for 2 h. After washing with PBS, bound antibody was detected using a secondary horseradish peroxidase-conjugated antimouse IgG antibody (1/1,000 in antibody buffer) for 1 h at 37°C. After a final wash in PBS, peroxidase activity was detected by addition of 150 µl/well of 15 mM TMB in 0.1 M sodium acetate, pH 6.0. The color reaction was stopped by the addition of 50 µl/well 2 M H₂SO₄. The trays were then read on a microplate spectrophotometer at 450 nm with a 630-nm reference filter. The dynamic range of the ELISA was 0.45 absorbance units and showed sigmoidal characteristics. The expression of -SMA was confirmed by Western blotting, which showed a qualitatively similar upregulation of the protein in response to increasing doses of TGF-.

Collagen Gene Expression

BF1 fibroblasts were seeded into six-well culture trays in 10% FBS/DMEM and grown to 90% confluence, after which the medium was replaced by Ultraculture for 24 h. The serum-starved cells were treated with Ultraculture in the absence or presence of dilutions of TGF-2, IL-4, or IL-13 for 18 h. For analysis of collagen gene expression, the RNA was extracted using TRIzol reagent (Life Technologies, Paisley, UK) and contaminating DNA was removed by deoxyribonuclease digestion on RNeasy Mini Kits (Qiagen, Crawley, West Sussex, UK) in accordance with manufacturer's instructions. Total RNA (2 µg) was reverse transcribed using oligo (dT)₁₅ primers and avian myeloblastosis virus transcriptase from the Reverse Transcription System (Promega, Southampton, UK), following the manufacturer's protocol. The primers for pro-1 collagen I and fluorogenic probe, labeled with 5'-reporter dye 6-carboxy-fluorescein (FAM) and 3'-quencher dye 6-carboxy- N,N,N',N'-tetramethyl-rhodamine (TAMRA), were designed using Primer Express (Perkin-Elmer Biosystems, Warrington, UK). The sequences were: sense primer, 5'-CCCTGGAAAGAATG GAGATGAT-3'; and antisense primer, 5'-AAACCACTGA AACCTCTGTGTCC-3'; and probe, FAM-5'-CGGGCAAT CCTCGAGCACCCT-3'-TAMRA. Housekeeping gene primers and probe for the endogenous control glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were obtained from Perkin-Elmer. Complementary DNA (cDNA) standard curves were generated using serial dilutions of cDNA (0.1 to 100 ng) obtained from untreated fibroblast cultures. For the samples, each 25-µl polymerase chain reaction (PCR) reaction contained 25 ng cDNA, 100 nM fluorogenic probe, 200 nM primers, and 12.5 µl TAQMan universal PCR master mix (Perkin-Elmer). No-template controls and reverse transcription-negative samples were also included as 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 anealing/extension at 60°C for 1 min. Quantitation and real-time detection of the TAQMan PCR were followed on the on ABI Prism 7700 sequence detection system, and after completion of the PCR, the thresholds for fluorescence emission baseline were set just above background levels on the FAM and VIC layers (~ 15 to 20 cycles). Standard curves were constructed for target genes and the GAPDH endogenous control, and the amount of target and endogenous control were calculated. The data were normalized by using the ratio of the amount of target gene relative to endogenous control.

Enzyme-Linked Immunoassays

Fibroblasts were plated into six-well culture trays (1 × 10⁵ cells/well in 2 ml 10% FBS/DMEM) and incubated at 37°C for 6 h. The medium was then changed to Ultraculture with or without TGF-2 (150 pg/ml), IL-4, or IL-13 (20 ng/ml). After 3 d the medium was removed from the trays and assayed for cytokines by ELISA using the following kits and according to manufacturer's instructions. Human endothelin (ET)-1: Quantikine ELISA, R&D Systems (Barton Lane, Abingdon, Oxfordshire, UK) (minimum detectable dose < 1 pg/ml); human vascular endothelial growth factor (VEGF): Quantikine ELISA, R&D Systems (minimum detectable dose < 5 pg/ml); human eotaxin: Cytoscreen ELISA, Biosource International (Camarillo, CA) (minimum detectable dose 2 pg/ml); human TGF-2: E_max ELISA, Promega (Madison, WI) (minimum detectable dose 32 pg/ml).

Statistical Analysis

Unless otherwise stated, paired data were compared by the Wilcoxon test using SPSS software (SPSS, Inc., Chicago, IL).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Establishment of Primary Fibroblast Cultures

Primary bronchial fibroblast cultures were derived from biopsies taken from four patients with mild to moderate atopic asthma. The cells were characterized by immunohistochemistry using antibodies against the mesenchymal cell marker vimentin, the SMC marker myosin heavy chain, and the myofibroblast and SMC marker -SMA. All four cell lines stained homogenously positive for vimentin but were negative for myosin heavy chain and -SMA, indicating that the cultures were of fibroblast origin and were not contaminated with SMC or myofibroblasts (data not shown). The fibroblasts grew rapidly for at least the first ten passages and assays were performed with cells between passages 3 and 8.

Stimulation of Cell Proliferation

Each fibroblast culture was found to proliferate in SFM and had no requirement for exogenous growth or attachment factors (Figure 1), suggesting that they were capable of secreting autocrine growth factors and extracellular matrix proteins. Addition of TGF-1 or -2 to the SFM significantly increased the proliferation rate of the fibroblasts. This effect was most apparent at the later time points where the cells achieved higher densities compared with cells grown in SFM alone (Figure 1). In these assays, the two TGF- isoforms were equipotent. When we compared the effects of IL-4 or IL-13 (20 ng/ml), both of the cytokines significantly stimulated proliferation but were found to be approximately 50% less potent than TGF- (Figure 1, inset).

View larger version (40K): [in this window] [in a new window]   Figure 1.   Effects of TGF-, IL-4, and IL-13 on fibroblast proliferation. BF1 cells were seeded at a low density and then grown for 6 d in (a) SFM alone or with 150 pg/ml of (b) TGF-1 or (c) TGF-2. Cell density on indicated days was determined by the uptake of methylene blue dye as described in MATERIALS AND METHODS. Data are means ± standard deviation (SD) of triplicate determinations. Inset shows a comparison of the growth-stimulatory effects of TGF-2, IL-4, and IL-13. BF1-4 cells were grown for 5 d in (I) SFM alone or plus (II) 150 pg/ml TGF-2, (III) 20 ng/ml IL-4, or (IV) 20 ng/ml IL-13. Cell density on Day 5 was then determined by the uptake of methylene blue dye. Data are means ± SD of triplicate determinations. Significance comparisons between control and treated cells were made using an unpaired t test; *P = 0.002.

Induction of Myofibroblast Differentiation

The ability of TGF- to promote proliferation of the fibroblasts appeared to be due, in part, to suppression of contact inhibition inasmuch as a greater degree of cell-cell overlap was observed in TGF--treated cultures. The appearance of the cells indicated that they had acquired a contractile phenotype and that they had differentiated into myofibroblasts (data not shown). To confirm that TGF- had induced myofibroblast differentiation, we developed an ELISA procedure to quantitate -SMA expression by cells. In this method the cells were cultured in a 96-well tray in the absence or presence of each cytokine. The cells were then fixed and permeabilized in the culture tray and assayed for -SMA using a standard ELISA protocol. Figure 2 shows induction of -SMA expression in BF3 and BF4 cells. Although transformation into myofibroblasts was marked and dose-dependent in the presence of TGF-2, there was negligible upregulation of -SMA by IL-4 and IL-13 and there was no dose-dependency. Similar results were obtained for BF1 and BF2 cells. Induction of -SMA and assembly into filaments following TGF-1 or TGF-2 treatment of cells was confirmed using both confocal immunofluorescence microscopy and electron microscopy (data not shown).

View larger version (28K): [in this window] [in a new window]   Figure 2.   The effects of IL-4 and IL-13 on -SMA expression. (A) BF3 or (B) BF4 fibroblasts were cultured for 3 d in SFM with or without serial dilutions of IL-4, IL-13, or TGF-2. The -SMA content of the cells was then measured using an ELISA technique as described in MATERIALS AND METHODS. The A450 of control cells in SFM has been subtracted from the data. Data are means ± SD of triplicate determinations.

Induction of Collagen-I Gene Expression

Myofibroblast differentiation is associated with an increase in extracellular matrix production and TGF-1 and TGF-3 have previously been demonstrated to stimulate collagen secretion by primary human lung fibroblasts (22). Whereas TGF-1 is present in the airways (7), the 2 isoform has been shown to be produced by bronchial epithelial cells in response to injury (23). Therefore, we further characterized the effects of TGF-2 on the fibroblasts. We compared the effects of TGF-2, IL-4, and IL-13 on procollagen-1 gene expression in BF1 and BF4 cells using TAQMan PCR. As shown in Figure 3 for BF1 cells, TGF-2 caused a dose-dependent increase in collagen-I gene expression. The fibroblasts were exquisitely sensitive to TGF-2, with the lowest dose tested (150 pg/ml) significantly stimulating gene expression. In contrast, none of the doses of IL-4 or IL-13 tested significantly increased collagen-I messenger RNA (mRNA) levels. Comparable results were obtained for BF4 cells.

View larger version (33K): [in this window] [in a new window]   Figure 3.   Induction of procollagen 1 gene expression in BF1 fibroblasts. BF1 fibroblasts were cultured for 18 h in SFM with or without serial dilutions of IL-4, IL-13, or TGF-2. Procollagen 1 gene expression was then determined by quantitative PCR as described in MATERIALS AND METHODS. Procollagen gene expression is expressed as a ratio of GAPDH gene expression. Data are means ± SD of a minimum of six determinations. Significance comparisons between control and treated cells were made using an unpaired t test; *P < 0.05; **P < 0.0005.

Stimulation of Mediator Release

To determine whether TGF- and IL-4 or IL-13 differentially affected the secretory activity of the fibroblast cultures, conditioned medium was taken from cells treated with TGF-2 (150 pg/ml), IL-4, or IL-13 (20 ng/ml) and assayed for ET-1, VEGF, and eotaxin by ELISA. Whereas TGF-2 had a major stimulatory effect on secretion of ET-1 and VEGF (Figures 4A and 4B), IL-4 and IL-13 were comparatively poor stimuli with only a minor, albeit statistically significant, effect on production of the cytokines. These data suggest that myofibroblast transformation favors release of mitogens that can drive smooth muscle and vascular proliferation. In contrast, eotaxin release was markedly increased by both IL-4 and IL-13, whereas the effect of TGF-2 was not significant (Figure 4C). Thus, Th-2 cytokines would appear to favor a proinflammatory rather than remodeling phenotype in the fibroblasts.

View larger version (31K): [in this window] [in a new window]   Figure 4.   Measurement of mediators released by stimulated cells. BF1-4 cells were grown for 3 d in (a) SFM alone or plus (b) 150 pg/ml TGF-2, (c) 20 ng/ml IL-4, or (d) 20 ng/ml IL-13. Conditioned medium was removed from the cells and tested for (A) ET-1, (B) VEGF, or (C) eotaxin by ELISA. Data are means of duplicate determinations. Inset plots show the combined data for (I) SFM, (II) TGF-2, (III) IL-4, and (IV) IL-13. Significance comparisons between control and treated cells were made using an unpaired t test. ns, not significant; *P < 0.02.

TGF-2 Release from Primary Epithelial Cell Cultures

To explain the inability of IL-13 to directly induce a remodeling phenotype by the bronchial fibroblasts, we postulated that its in vivo effects in transgenic mice were an indirect consequence of cellular activation at the site of transgene expression, i.e., the bronchial epithelium. Inasmuch as we have demonstrated previously that bronchial epithelial cells release TGF-2 (23, 24), we determined the effects of IL-4 and IL-13 on release of this growth factor from primary cultures of epithelial cells which were established from bronchial brushings taken from patients with mild to moderate atopic asthma. After culture of confluent monolayers in SFM alone or in the presence of 20 ng/ml IL-4 or IL-13 for 48 h, there was a 2- to 3-fold increase in TGF-2 in the culture medium of the treated cells (Figure 5). The enhanced release of TGF-2 was not inhibited by the presence of 1 µM dexamethasone, indicating that the response was steroid-unresponsive. In contrast, no eotaxin was detectable (< 2 pg/ml) in these epithelial supernatants and no eotaxin mRNA could be detected by reverse transcriptase-PCR of RNA extracted from the cells (data not shown).

View larger version (19K): [in this window] [in a new window]   Figure 5.   Release of TGF-2 from primary bronchial epithelial cells. Cell monolayers were cultured for 2 d in SFM alone or plus 20 ng/ml IL-4 or IL-13 and with or without 1 µM dexamethasone. Conditioned medium was removed from the cells and tested for TGF-2 by ELISA. All samples were tested in duplicate; n = 8.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Asthma is a complex genetic disease with multiple genes interacting with the environment to modify both susceptibility and severity of disease. IL-4 and IL-13 appear to play key roles in this genetic susceptibility with promoter and/or functional polymorphisms in their own genes, those of their composite receptors, and their intracellular mediator, signal transducer and activator of transcription-6, many of which have been linked to asthma (12). Whereas the effects of IL-4 and IL-13 on T- and B-cell switching can explain their proallergic effects in asthma, observations derived from experimental animals suggest that IL-13 may play a more important role than IL-4 in airway remodeling through its effects on subepithelial fibrosis and BHR (13). Because these findings have significant implications for the utility of anti-IL-4 versus anti-IL-13 targeted therapies, we considered it important to establish whether these differences might also occur in human airways.

In asthma, there is an increase in subepithelial mesenchymal cells with features of myofibroblasts whose numbers have been found to increase in proportion to the thickness of the sub-basement membrane collagen layer (5). These cells correspond to the attenuated fibroblast sheath described by Evans and colleagues (16) as lying adjacent to the lamina reticularis, and form a network similar to hepatic stellate cells which, when activated by liver damage, are the effector cells responsible for fibrosis (25). Therefore, to study the profibrogenic contribution of IL-4 and IL-13 in asthma, we used fibroblasts grown from bronchial biopsies obtained from asthmatic subjects. The ability of these cells to proliferate in the absence of exogenous growth factors is similar to that observed for fibroblasts from interstitial lung fibrosis which behave more like neonatal fibroblasts than normal adult lung fibroblasts (26). However, a report by Dubé and associates (27) described primary bronchial fibroblasts from asthmatic subjects as having lower basal DNA synthesis than corresponding fibroblasts from normal individuals. These authors (27) did not report a proliferative effect of TGF-1 on lung fibroblasts under reduced serum conditions. This contrasts with the clear stimulation of cell proliferation by both TGF-1 and TGF-2 observed in the present study, where the treatments were carried out in SFM.

Using primary bronchial fibroblasts, it was possible to demonstrate myofibroblast differentiation in response to the potent profibrogenic effects of TGF-1 and TGF-2 at both ultrastructural and biosynthetic levels by confirmation of induction of -SMA expression and collagen gene expression. In contrast, neither IL-4 nor IL-13 was able to promote differentiation, suggesting that these cytokines are not directly responsible for induction of remodeling responses in the subepithelial fibroblast layer. However, it remains to be determined whether these cytokines may have other "remodeling" effects on the fibroblasts such as an effect on matrix turnover.

During fetal lung development, epithelial and mesenchymal cells function as a "trophic unit" through the release of soluble mediators, including TGF-s, epidermal growth factor receptor ligands, and fibroblast growth factors that mediate bidirectional communication to direct airways growth and branching (28). In view of the extent of epithelial damage in asthma and the observation that injury causes bronchial epithelial cells to release growth factors that drive the myofibroblast proliferation and induce expression of interstitial collagens, we have proposed that the EMTU becomes reactivated in asthma (17). This concept led us to consider whether IL-4 and IL-13 might functionally interact with the EMTU via the epithelium rather than by a direct effect on the submucosal fibroblasts. Significantly, both IL-4 and IL-13 were found to stimulate release of TGF-2 from bronchial epithelial cells, suggesting that both cytokines have the potential to use the epithelium to translate remodeling responses to the underlying mesenchyme. Because expression of IL-4 in the bronchial epithelium in transgenic mice failed to cause subepithelial fibrosis (15), it remains to be determined whether this reflects an intrinsic difference in the responsiveness of murine epithelial cells to IL-4 or whether our in vitro system failed to reflect cell-cytokine interactions as they occur in vivo. Therefore, our human tissue- derived system may prove to be an important tool that complements and validates studies using animal models. It was also notable that the enhanced release of TGF- induced by IL-4 and IL-13 was corticosteroid-unresponsive, a finding that suggests that anti-IL-4 or anti-IL-13 targeted therapies may have the potential to interrupt some of those remodeling responses that persist at the severe end of the disease spectrum even in the face of corticosteroid treatment.

Mice in which transgenes for IL-6, IL-11, or IL-10 have been separately expressed in the bronchial epithelium using the CC-10 promoter all develop subepithelial matrix deposition in proportion to smooth-muscle hyperplasia and greatly enhanced BHR (29, 30). Although the relationship between myofibroblast activation and the underlying smooth-muscle mass in asthma has not yet been studied, there is an accumulation of migratory, contractile cells in the lamina reticularis (6) and an increase in ET-1 in BALF after allergen exposure (31). Analysis of the effect of TGF- on the secretory activity of the asthmatic fibroblasts demonstrated that, unlike the Th-2 cytokines, it was a strong inducer of growth factors (ET-1 and VEGF) that can cause increases in the smooth muscle and microvasculature. Thus, we propose that the EMTU functions as an integrated unit in which the bronchial epithelium coordinates responses to injury (23, 24) or to Th-2 cytokines through release of TGF-. This then acts on the underlying fibroblast sheath, causing transformation into activated myofibroblasts that become the key effector cells that drive airway remodeling both through their ability to synthesize and secrete matrix proteins and through their production of growth factors that drive autocrine proliferation and paracrine growth of the smooth muscle, nerves, and vessels.

Although neither IL-4 nor IL-13 was able to induce significant profibrogenic responses from the asthmatic fibroblasts per se, they were both potent inducers of the eosinophil chemoattractant eotaxin, as has been reported previously using fibroblasts from lung resections (32). Thus, it seems likely that the main effect of IL-4 and IL-13 on the submucosal fibroblast population is to drive a proinflammatory response through recruitment of eosinophils into the airways. Inasmuch as eosinophils are also a potent source of TGF-, this may represent another indirect contribution of IL-4 and IL-13 toward airway remodeling. In contrast with IL-4 and IL-13, TGF- failed to induce expression of eotaxin, consistent with its known anti-inflammatory activity. Indeed, our data and those of the transgenic mice suggest that anti-inflammatory mediators tend toward fibrotic responses, whereas proinflammatory mediators are less potent profibrogenic agents. Although we were able to detect eotaxin secretion from the fibroblast cultures, no eotaxin was detectable from the undifferentiated primary epithelial cell cultures. Because eotaxin expression is observed in the bronchial epithelium of the peripheral airways of asthmatic subjects (33), our failure to detect this chemokine may reflect our use of epithelial cells from the large airways or a requirement for epithelial differentiation.

In conclusion, we have shown that both IL-4 and IL-13 modulate epithelial cell function and, through these cells, their influence on the underlying mesenchymal cells is amplified. Therefore, we propose that IL-4 and IL-13 are key cytokines that contribute to asthma severity and chronicity by augmenting responses within the EMTU. The availability of in vitro systems of human asthmatic bronchial epithelial cells and fibroblasts will enable detailed characterization of these interactions as well as early evaluation of novel, targeted interventions directed toward the aberrant responses of airway structural cells.