Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Interleukin (IL)-5 is a pleiotropic cytokine initially described as a B-cell growth factor for its ability to promote the proliferation and production of immunoglobulin M (IgM) by the murine B-cell tumor line BCL-1 (1). Subsequently, IL-5 has been shown to exhibit biological activity on cells of diverse hemopoieitic lineages. IL-5 plays a pivotal role in the proliferation, differentiation, survival, activation, and chemoattraction of eosinophils (2), enhances histamine release and leukotriene C₄ generation from human basophils, and promotes IL-2-induced proliferation and differentiation of human and murine cytotoxic T lymphocytes (3). Because it is coordinately expressed along with a cluster of cytokines encoded on human chromosome 5q_31-33, secretion of IL-5 has been closely linked with eosinophil recruitment and priming that occurs in allergic diseases such as bronchial asthma, both at rest and following allergen exposure (4). It is generally thought that the Th2- and Tc2-like T cells, mast cells, basophils, and eosinophils are the principal cellular sources of this cytokine (2).

Using a sensitive reverse transcription-polymerase chain reaction (RT-PCR) enzyme-linked immunosorbent assay (ELISA) system we demonstrate here that bronchial tissue obtained from healthy human volunteers expresses IL-5 mRNA constitutively. Although T lymphocytes, mast cells, basophils, and eosinophils are the known cellular sources of IL-5 messenger RNA (mRNA) in the airway tissue, these cells normally do not express IL-5 mRNA unless stimulated with antigen or other factors such as IgE in the presence of stem cell factor (5, 6). Using in situ hybridization Ying and coworkers (2) have shown that up to 5% of IL-5 mRNA-positive cells in the bronchoalveolar lavage fluid (BALF) obtained from asthmatic subjects do not colocalize to T cells, mast cells, eosinophils, or macrophages, suggesting that cells other than these have the capacity to produce IL-5 at least at the mRNA level. We show here that exposure to 0.12 ppm ozone for 2 h in healthy human subjects leads to a decrease in the amounts of IL-1, IL-5, and granulocyte macrophage colony-stimulating factor (GM-CSF) mRNA expression in the bronchial tissue obtained 1.5 h after exposure ceases. It is known that the oxidative effects of ozone are largely confined to the epithelial surface and that ozone induces epithelial cell damage and shedding (7, 8). IL-1 and GM-CSF mRNA are known to be produced mainly by epithelial cells in the healthy airway tissue (9), and damage to the airway epithelium following exposure to ozone may explain the decrease in IL-1 and GM-CSF observed in the bronchial tissue. The constitutive expression of IL-5 mRNA in the healthy bronchial tissue and its concomitant decrease with IL-1 and GM-CSF following ozone inhalation led us to hypothesize that epithelial cells might be a source of IL-5 mRNA.

Using transformed human airway epithelial cell lines A549 and 16HBE14o, and primary human bronchial and nasal epithelial cells obtained from healthy human airways and grown in culture, we demonstrate here that these cells constitutively express IL-5 mRNA. Stimulation of primary human nasal epithelial cells with tumor necrosis factor (TNF)- in vitro induced an upregulation of IL-5 mRNA expression. Immunohistochemical staining of bronchial tissue obtained from healthy human volunteers demonstrated immunoreactivity for IL-5 localized to the epithelium, and A549 cells when stimulated with TNF- and interferon (IFN)- demonstrated low, but detectable, levels of IL-5 protein in culture supernatants. To our knowledge, this is the first report demonstrating constitutive expression of IL-5 mRNA and the production of IL-5 protein by human airway epithelial cells.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Ozone Exposure Protocol

Twelve healthy, nonatopic, nonsmoking volunteers (three female, nine male; mean age 27.9 yr) who had no history of asthma or other respiratory illness and had normal lung function measurements were exposed to filtered air and 0.12 ppm ozone randomly in a single-blinded crossover design. At least 3 wk separated the two exposure events. During the 2-h exposure, mild exercise (VE = 20 liters/ min/m²) on a bicycle ergometer was alternated with rest in 15-min periods. The exposure chamber setup and the ozone-generating system have been described previously (10). Endobronchial biopsies from the main carina and the subcarinae of the middle lobe bronchus and lingula were obtained under local anesthesia 1.5 h after the end of each exposure, with a fiberoptic bronchoscope using a well- established protocol (10). Two biopsies from each subject were placed in ice-cooled acetone containing protease inhibitors and were processed into glycol methacrylate (GMA) resin for immunohistochemical staining. A further two biopsies were stored in liquid nitrogen at 70°C until further analysis for cytokine mRNA expression by the RT-PCR ELISA method. The exposures were performed at the Medical Division of National Institute for Working Life at Umeå, Sweden. An informed, written consent was obtained from all subjects, and the study was approved by the local ethics committee.

Cell Cultures

The alveolar adenocarcinoma cell line A549 was obtained from the American Type Culture Collection, Rockville, MD. The virally tranformed human bronchial epithelial cell line 16HBE14o was a kind gift from Dr. D. Gruenert (University of California at San Francisco). Both cell lines were maintained in Eagle's minimal essential medium (Biowhittaker, Walkersville, MD) containing 4 mM L-glutamine, 80 mg/ml gentamicin, and 10% heat-inactivated fetal bovine serum (Sigma, Dorset, UK). The 16HBE14o retains differentiated morphology and function of normal human epithelia. Primary bronchial epithelial cells were obtained as outgrowths from explants of airway epithelium. The epithelium was microdissected from the underlying connective tissue of bronchial airway obtained at surgical resection. Two- to 3-mm portions of the explants were plated onto Primaria plastic dishes (Falcon, Los Angeles, CA) and cultured for 2 to 3 wk in M199 medium (GIBCO, Middlesex, UK) supplemented with 2% Ultrosor G lyophilized (GIBCO). During this time, epithelial cells grew to form a confluent monolayer that was 3 to 4 cm in diameter around each portion of explant tissue. These cells retain epithelial characteristics in vitro as demonstrated by the expression of cytokeratin staining. All cells were grown at 37°C in 5% CO₂, 95% air, and saturated humidity to near confluence before being harvested for PCR analysis. Nasal epithelial cells were isolated from the nasal mucosa, obtained from patients undergoing surgery for therapeutic reasons. Briefly, the mucosa was dissected into small strips and then incubated at 37°C for 90 min in keratinocyte medium containing 0.1% trypsin. After filtering the resultant cell suspension, epithelial cells were plated directly into six-well culture plates (Nunc, Paisley, Scotland) and grown to confluence in keratinocyte medium before being harvested for PCR analysis.

RT-PCR ELISA

RNA isolation and reverse transcription. Total RNA was extracted from bronchial biopsies (approx 2 mm²) and A549, 16HBE14o, primary bronchial cells, and primary nasal epithelial cells (5 × 10⁶ cells each) using Trizol (Life Technologies, Paisley, Scotland) according to the manufacturer's instructions, and then was precipitated with isopropranol and washed in 80% ethanol. With poly (dT)₁₈ as primer and 15 U of AMV Reverse Transcriptase (Promega, Southampton, UK), polyadenylic acid mRNA was reverse transcribed for 1 h at 42°C in the presence of 1 mM dNTPs, 5 mM MgCl₂, 40 U RNAsin (Promega), and reverse transcriptase buffer (10 mM Tris-HCl, pH 8.8; 50 mM KCl; 0.1% Triton X-100) (Promega) to produce complementary DNA (cDNA).

PCR and digoxigenin incorporation. PCR amplification of the cDNA was carried out using primer pairs specific for the constitutively expressed gene adenine phosphoribosyl transferase (APRT) and the cytokines IL-1, IL-5, IL-8, TNF-, IFN-, and GM-CSF (primer sequences, see Table 1). Two and one-half microliters of cDNA (1/10th of the RT product) was amplified using 1 U Taq DNA polymerase (Promega) in the presence of 15 pmol of both primers, 2.5 µl PCR digoxigenin (DIG) labeling mix (0.2 mM deoxyadenosine triphosphate, deoxycytidine triphosphate, and deoxyguanosine triphosphate; 0.19 mM deoxythymidine triphosphate; 0.01 mM DIG-11-deoxyuridine triphosphate) (Boehringer Mannheim, Mannheim, Germany), magnesium-free PCR buffer (50 mM KCl; 10 mM Tris-HCl, pH 9.0; 0.1% Triton X-100) (Promega), and 1 mM MgCl₂. Target cDNA was amplified using a three-temperature PCR, with primer annealing temperatures optimized for each cytokine (56°C for APRT; 50°C for IL-1, IL-5, and IL-8; 54°C for TNF- and IFN-; and 60°C for GM-CSF).

DIG detection. Ten microliters of the PCR product was denatured with 40 µl denaturation solution from the PCR ELISA kit (Boehringer Mannheim) for 15 min. Biotinylated, labeled "capture probes" (Table 1) specifically designed to hybridize with each cytokine PCR product were then hybridized to the complementary DIG-labeled PCR product and immobilized on duplicate wells of streptavidin-coated microtiter plates at 37°C for 3 h. After thorough washing to remove free antibodies, bound PCR products were detected by incubation for 30 min at 37°C with an anti-DIG antibody conjugated to horseradish peroxidase, followed by reaction with the substrate 2,2'-Azino-di-(3-ethyl benzthiazoline sulfonate). During green color development, the absorbance (at 405 nm) was measured with an ELISA plate reader. Absorbance readings were taken after 10 min and at 5-min intervals thereafter, for up to 30 min. For some of the PCR products, the upper limit of the colorimetric detection system used in this PCR ELISA was reached before the full 30-min color development had elapsed. Therefore, a time point of 20 min, at which the absorbance values were still within the dynamic range of the ELISA, was used throughout. Mean absorbance values (at 20 min) were calculated for the duplicate samples. Before being compared between treatments (exposure to air and exposure to ozone), the levels of cytokine transcripts were normalized to APRT and expressed as a percentage (level of cytokine products/level of APRT product × 100). The RT-PCR ELISA technique for cytokine mRNA detection has been well established in our laboratory (11).

Optimizing the PCR for semiquantitative analysis. In preliminary experiments, cDNA samples were subjected to 25, 30, 35, and 40 PCR cycles in the presence of DIG labeling mix and levels of PCR products measured using the RT-PCR ELISA. Absorbance readings were plotted against the PCR cycle number to identify the range over which PCR products were accumulating exponentially. For the housekeeping gene APRT and the cytokines listed previously, PCR products were still accumulating exponentially after 30 PCR cycles (data not shown). Measurement of levels of PCR products for APRT, IL-1, IL-5, IL-8, TNF-, IFN-, and GM-CSF in bronchial biopsies were therefore made after 30 PCR cycles.

DNA sequencing to confirm IL-5 RT-PCR product. To confirm the identity of the IL-5 PCR product, 50 µl of PCR product from 10 × 10⁶ 16HBE14o cells was purified using QIAquick spin columns (QIAGEN, Hilden, Germany) and 40 ng was subjected to dye terminator cycle sequencing using an ABI prism Big Dye Terminator Cycle sequencing Ready Reaction Kit (PE Applied Biosystems, Foster City, CA) using the 5' PCR primer. The sequencing reaction was run on an ABI 3 and DNA sequencer, and the resulting sequence was compared with that of IL-5 (Genebank accession no. X04688) and was found to match.

IL-5 Protein Measurements in Culture Supernatants

Culture supernatants were obtained from 5 × 10⁶ A549 and 16HBE14o epithelial cells grown in culture flasks at baseline, and from 5 × 10⁶ A549 cells stimulated for 18 h with TNF- (10 ng/ml) alone and TNF- with IFN- (both at 30 ng/ml) and then concentrated using Centriprep centrifugal concentrators (Amicon, Beverly, MA) from 15 to 0.8 ml. The IL-5 protein was measured on 200 µl of the supernatant using a standard IL-5 ELISA kit (Quantikine human IL-5 ELISA kit; R&D Systems, Minneapolis, MN).

Immunohistochemical Staining of Bronchial Tissue

Processing the bronchial biopsy in GMA and immunostaining were performed using a standard protocol as described before (12). In brief, bronchial biopsies obtained from healthy human volunteers were processed in GMA resin. Two-micrometer-thick sections were cut and floated onto ammonia water (1:500), picked onto 0.01% poly-L-lysine-coated glass slides, and allowed to dry at room temperature for 1 h. The sections were treated to block endogenous peroxide. Nonspecific antibody binding was blocked with undiluted culture supernatant for 30 min, followed by the primary IL-5 monoclonal antibody (mAb) (kind gift from P. H. Hissey, Glaxo Wellcome, East Sussex, UK) (13) at a dilution of 1:800, which was applied and incubated at room temperature overnight. After rinsing in tris-buffered saline (TBS), biotinylated rabbit antimouse IgG Fab (Dako Ltd., High Wycombe, UK) was applied for 2 h, followed by streptavidin-biotin horseradish peroxide complex (Dako) for another 2 h. After being rinsed in TBS, amino ethyl carbamazole in acetate buffer (pH 5.2) and hydrogen peroxide were used as substrate to develop a peroxide-dependent red color reaction. The sections were then counterstained with Mayer's hematoxylin.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Cytokine mRNA Changes in the Bronchial Tissue after Acute Exposure to Ozone

There was constitutive expression of IL-1, IL-5, IL-8, TNF-, IFN-, and GM-CSF mRNA in the healthy bronchial tissue obtained following exposure to air (Table 2). One and one-half hours after exposure to 0.12 ppm ozone for 2 h, there was a significant decrease in the relative amounts of IL-1, IL-5, and GM-CSF mRNA expression in the bronchial tissue compared with air (Figure 1, Table 2). No significant changes were noted in the relative amounts of IL-8, TNF-, and IFN- mRNA (Table 2).

View larger version (15K): [in this window] [in a new window]   Figure 1.   Changes in relative expression of IL-5 mRNA in bronchial biopsies obtained from healthy human airways 1.5 h after exposure to 0.12 ppm ozone or filtered air for 2 h.

Constitutive Expression of IL-5 mRNA in Epithelial Cells and Its Upregulation with TNF-

RT-PCR ELISA on A549, 16HBE14o, and primary human bronchial and primary human nasal epithelial cells grown in culture at baseline demonstrated constitutive expression of IL-5 mRNA in three repeat experiments (Figure 2). The relative amount of IL-5 mRNA standardized and expressed as a percentage of the housekeeping gene APRT was as follows: A549, 61.29%; 16HBE, 55.62%; primary bronchial, 54.49%; and primary nasal epithelial, 14.9% (Figure 3). Stimulation of primary nasal epithelial cells with 30 ng/ml TNF- for 6 h induced a 2.6-fold relative increase in IL-5 mRNA as detected by RT-PCR ELISA (Figure 3).

View larger version (45K): [in this window] [in a new window]   Figure 2.   Detection of constitutive IL-5 mRNA in A549 cells and primary nasal and primary bronchial epithelial cells grown in tissue culture using RT-PCR ELISA.

View larger version (22K): [in this window] [in a new window]   Figure 3.   Relative levels of IL-5 mRNA expression in unstimulated A549 cells, 16HBE cells, primary nasal and bronchial epithelial cell cultures (mean ± SEM, n = 3), and in primary nasal epithelial cell cultures exposed to 30 ng/ml TNF- (n = 2).

IL-5 Immunoreactivity in the Bronchial Epithelium

Immunohistochemical staining of bronchial tissue obtained from healthy human volunteers and immunostained using mAbs against IL-5 revealed immunoreactivity for IL-5 in the bronchial epithelium (Figure 4A). Control sections stained with an irrelevant mouse IgG₁ mAb did not show any immunoreactivity (Figure 4B).

View larger version (83K): [in this window] [in a new window]   View larger version (72K): [in this window] [in a new window]   Figure 4.   (a) Identification of IL-5 immunoreactivity in epithelium of bronchial biopsy obtained from a normal healthy subject. IL-5 stained by mAb and immunoperoxidase technique. Original magnification: ×40. (b) Isotype control for IL-5 mAb staining in sequential section of bronchial biopsy.

IL-5 Protein Measurement in Cell Supernatants

Using a standard ELISA detection technique, IL-5 protein was not detectable in any of the epithelial cell A549 and 16HBE14o culture supernatants at baseline. A549 cells (5 × 10⁶ cells) when stimulated with TNF- and IFN- (both at 30 ng/ml for 18 h), demonstrated low, but detectable, levels of IL-5 protein (4.54 pg/ml), whereas no IL-5 protein was detected from A549 cells stimulated with TNF- (10 ng/ml) or IFN- (10 ng/ml) alone for 18 h.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Using an RT-PCR ELISA to study the effects of ozone on cytokine mRNA expression in the healthy human bronchial mucosa, we found that ozone exposure led to a decrease in the relative amounts of IL-1, IL-5, and GM-CSF mRNA that could be identified in bronchial biopsies obtained 1.5 h after exposure. This finding was unexpected and led us to wonder whether epithelial cells might be a source of IL-5 mRNA, because (1) it is well recognized that following inhalation, ozone cannot penetrate more than a few microns into the air/lung interface before it reacts (14), thus making the epithelial cells lining the respiratory airways the logical initial target site for ozone- induced lung injury; indeed it is known that the oxidative effects of ozone are mainly confined to the epithelial surface (7); (2) ozone can induce epithelial damage and shedding (7, 8); and (3) the other two cytokines that showed reductions following ozone inhalation (IL-1 and GM-CSF) are known to be produced constitutively by epithelial cells (9). We therefore hypothesized that the epithelial cells could be a source of IL-5 mRNA. Exploiting the sensitivity of the RT-PCR ELISA system, we then demonstrated that in culture the epithelial cell lines A549 and 16HBE14o and primary human nasal and bronchial epithelial cells all constitutively express IL-5 mRNA. Culture supernatants derived from A549 cells stimulated with TNF- and IFN- demonstrated low, but detectable, levels of IL-5 protein, and bronchial biopsies obtained from healthy human airways demonstrated immunoreactivity for IL-5 in the bronchial epithelium. To our knowledge this is the first report demonstrating constitutive gene transcription of IL-5 and the production of IL-5 protein by human airway epithelial cells.

We have used a sensitive RT-PCR ELISA system to semiquantify relative changes in cytokine mRNA following exposure to air and ozone. To achieve meaningful quantitation of the PCR product, it is important to limit the number of PCR cycles so that the accumulation of cytokine and APRT products is still in the linear exponential phase, and to ensure that the ELISA development times are within the dynamic range of the colorimetric detection system. Both of these conditions were met in this study. This technique therefore not only has the unique advantage of being more sensitive than the standard RT-PCR gel electrophoresis method, but also offers numerical figures that have been used to compare relative changes in gene transcript products.

We have used two human epithelial cell lines (A549 and 16HBE14o) and primary cultures of bronchial and nasal epithelial cells to investigate further the possibility that airway epithelial cells can produce IL-5. A549 and 16HBE14o are well-established human airway epithelial cell lines. The purity of the primary epithelial cultures used in this study was more than 95%, as assessed by light microscopic examination of cells stained with mAbs specific for the epithelial cell marker cytokeratin. Using the described RT-PCR ELISA, we have consistently (n = 3) demonstrated mRNA for IL-5 in both the epithelial cell lines and the primary epithelial cultures. Furthermore, expression of IL-5 mRNA was upregulated 2.6-fold in primary nasal epithelial cells stimulated with TNF-. We have confirmed the identity of the IL-5 PCR products using fluorescence-based automated DNA sequencing. Also, the use of a specific biotin-labeled internal capture probe complementary to a region of the IL-5 PCR product during the RT-PCR ELISA protocol provides verification of the specificity of the PCR reaction product.

Our ability to detect IL-5 mRNA in contrast with some previous studies that have failed to demonstrate the presence of IL-5 mRNA in human airway epithelial cells either by in situ hybridization (15) or by RT-PCR with product detection by ethidium bromide staining of agarose gels (16) may be due to differences in the sensitivity of the techniques used. In situ hybridization is not as sensitive as RT-PCR, and failure to detect IL-5 mRNA using agarose gel electrophoresis of RT-PCR products could be due to the presence of IL-5 transcripts at a level below the sensitivity of ethidium bromide visualization. The sensitivity of RT-PCR product detection by ELISA is at least 100-fold greater than that of fluorescent staining on agarose gels. Using activated peripheral blood mononuclear cells, we have previously demonstrated that the RT-PCR ELISA detection method is able to detect as little as 20 pg of cytokine PCR products, levels that are undetectable by ethidium bromide staining (11).

RT-PCR ELISA carried out on bronchial tissue obtained from 12 healthy human subjects in this study demonstrated constitutive expression of IL-5 mRNA in all subjects, the cellular source of which cannot be determined, as RNA was extracted from the whole bronchial tissue. T cells, mast cells, basophils, and eosinophils are the known sources of IL-5 in the airway tissue. Eosinophils are not present in healthy airway tissue and are therefore unlikely to be the source of IL-5 mRNA. T cells and mast cells are found in healthy airway tissue but express IL-5 mRNA normally only in an activated state (5, 6). Using in situ hybridization, Ying and colleagues have previously demonstrated that up to 5% of IL-5 mRNA-positive BALF cells obtained from asthmatic subjects did not have the phenotype of T lymphocytes, macrophages, mast cells, or eosinophils, suggesting that other cell types have the capacity to produce IL-5 at least at the mRNA level (2). Perhaps coincidentally, in asthmatics, epithelial cells constitute between 3 and 5% of the BALF cell population (17). The ability to detect IL-5 mRNA in cultured epithelial cells in this study suggests that at least some of the IL-5 mRNA detected by RT-PCR ELISA in the healthy bronchial biopsies in our study originates from airway epithelial cells. In support of this conclusion, it has recently been demonstrated, using in situ hybridization with ³⁵S labeled riboprobes, that human bronchial epithelial cells from healthy and nocturnal asthmatic subjects can express IL-5 mRNA (18, 19), although these data have not been formally published.

The presence of IL-5 mRNA in epithelial cells does not necessarily mean that the protein is produced and released. We therefore looked for the presence of IL-5 protein in epithelial cell culture supernatants using standard ELISA techniques and immunohistochemical staining in bronchial tissue. Measurable levels of IL-5 protein were not detected in culture supernatants from primary or secondary transformed epithelial cell lines. However, when the A549 epithelial cells were stimulated with high concentrations of TNF- and IFN-, we were able to demonstrate low, but detectable, levels of IL-5 protein. The amount of IL-5 protein generated and released into culture supernatants by unstimulated primary and transformed epithelial cells may be below the sensitivity threshold of the standard ELISA technique. It may also be that the stimuli that we have used in vitro are inadequate to stimulate the epithelial cells to release IL-5 protein and therefore do not correspond to the complex dynamic processes occurring in vivo. Using immunohistochemical staining on bronchial tissue obtained from healthy human subjects, we have been able to demonstrate immunoreactivity for IL-5 protein in the bronchial epithelium. It is difficult to ascertain that the cellular source of IL-5 immunoreactivity in the epithelium are only the epithelial cells, because there are several other cell types present in the bronchial epithelium that could be the source of IL-5 immunostaining; however, the pattern of the staining observed and the fact that T cells and mast cells commonly found in the airway epithelium produce IL-5 only in an activated state suggest that the cell source of constitutive IL-5 protein in the healthy bronchial epithelium is more likely to be the epithelial cell.

Airway epithelial cells exist at the interface with the external environment and are therefore the first cells exposed to inhaled irritants, allergens, and noxious stimuli. These cells not only act as a physical barrier but also play an important role as a metabolically active physicochemical layer, capable of synthesizing and releasing a large number of inflammatory mediators, including eicosanoids (prostaglandin [PG]E₂, PGF₂, thromboxane A₂, 15-hydroxy-eicosatetraenoic acid) (20), and are also capable of expressing a multitude of signaling cytokines, including chemoattractants (IL-8, regulated on activation, normal T cells expressed and secreted, eotaxin, monocytic chemotactic protein-1, Gro-) and proinflammatory (IL-1, IL-6, TNF-), anti-inflammatory (IL-10), and hemopoieitic growth factors (GM-CSF, granulocyte CSF, CSF-1) (9, 21). All of these are thought to contribute to the proliferation, differentiation, activation, and chemoattraction of various inflammatory cells in the airway mucosa. Recently, human airway epithelial cells have also been shown to be a source of IL-2, which is likely to promote local proliferation of T cells in the epithelial microenvironment (24). The epithelial cells have been known to support eosinophil recruitment, activation, and survival, which is thought to be mediated through the secretion of eosinophil chemoattractants (e.g., IL-8, eotaxin) and survival enhancing factors (e.g., GM-CSF) (25). The present study is the first report to show that human airway epithelial cells constitutively produce IL-5 mRNA and protein and that IL-5 mRNA expression is induced by stimulation with TNF-. IL-5 produced by epithelial cells could further support eosinophil recruitment and activation in the airway mucosa. We propose that IL-5 produced by airway epithelial cells could serve a critical role in the proliferation, differentiation, activation, and survival of eosinophils in the airway mucosa, and is therefore likely to contribute to the eosinophilic inflammation in the bronchial tissue seen in conditions such as bronchial asthma.