Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Fibrotic pulmonary disease is characterized by accumulation of extracellular matrix collagen that is localized to the alveolar interstitium as a consequence of increased proliferation of lung fibroblasts. A feature of this chronic inflammatory process is an infiltration of leukocytes, with eosinophils (Eos) playing a prominent role (1). In animal models of bleomycin-induced pulmonary fibrosis it has been shown that infiltration of Eos precedes and parallels the development of lung fibrosis (2), whereas in patients with fibrotic lung disease, the presence of Eos has been found to correlate with a worse prognosis of the lung fibrotic process (3, 4). Similarly, ultrastructural studies of asthmatic biopsies have demonstrated the presence of collagen deposition in the abnormally thickened and dense lamina reticularis beneath the bronchial epithelium that is associated with the presence of both Eos and lung fibroblasts (5). Eos have the ability to produce vasoactive mediators that may contribute to edema formation; and via cytokines, such as transforming growth factor (TGF)-, TGF-, granulocyte macrophage colony-stimulating factor (GM-CSF), interleukin (IL)-3, IL-5, and IL-6, they may promote epithelial proliferation, metaplasia, angiogenesis, fibroblast proliferation, matrix generation, and tissue remodeling (6).

On the basis of their ability to produce different cytokines, murine CD4⁺ T lymphocytes have been classified as Th1 or Th2 (9). Th1 cells produce IL-2, interferon (IFN)-, and high levels of tumor necrosis factor (TNF)-, but not IL-4 or -5; whereas Th2 cells secrete IL-4, IL-5, IL-6, IL-9, and IL-10, and low levels of TNF-, but not IL-2 or IFN- (9, 10). Both Th cells produce IL-3 and GM-CSF. Th1 cells mediate delayed-type hypersensitivity, cell-mediated immunity, and immunoglobulin (Ig) class switching to IgG2a, whereas Th2 cells induce humoral immunity by activating B cells, promoting antibody production, and inducing class switching to IgG1 and IgE. In humans, most CD4⁺ T cells express a mixed cytokine profile. However, upon antigen or allergen stimulation they can be polarized into a Th1 or Th2 cytokine pattern (11, 12). In allergic inflammation the Th2-type cytokines, including IL-4 and IL-5, promote the differentiation and recruitment of Eos, whereas the Th1 type downregulates Eo differentiation by suppressing the development of Th2 cells (13). In relation to the fibrotic pulmonary process, animal studies have shown that IL-4 promotes collagen deposition as well as Eo tissue recruitment (14, 15). Moreover, a type-2 cytokine pattern has been found to predominate in the pulmonary interstitium of patients with cryptogenic fibrosing alveolitis (16). On the other hand, it has been demonstrated that IL-4 increases both fibroblast growth and collagen production, whereas IFN- suppresses these fibroblast activities (17). IL-4 has also been found to induce the production of the Eo attractant eotaxin by dermal fibroblasts (18). Because lung fibroblasts play a central role in both lung fibrotic disease and bronchial asthma, we have hypothesized that these cells may be a source of a number of Eo attractants. Therefore, we have stimulated lung fibroblasts with either Th1 or Th2 cytokines and investigated the presence of Eo attractants in the culture supernatants. Because some members of the chemokine family are chemotactic for Eos, we have also studied the expression of regulated on activation, normal T cells expressed and secreted (RANTES), monocyte chemotactic protein (MCP)-3, MCP-4, eotaxin, and eotaxin-2 (19), in both cytokine-stimulated and nonstimulated lung fibroblasts.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Culture of Lung Fibroblast

Human lung tissue was obtained from subjects undergoing lung surgery for either bronchiectasis or bronchial carcinoma removal. Specimens obtained from the noninvolved lung tissue were cut into small fragments and cultured in 75-cm² flasks containing minimum essential medium (Life Technologies-AM2MED, Allschwil, Switzerland) supplemented with penicillin (10 U/ml), streptomycin (100 µg/ml), and glutamin (2 mM), in 5% CO₂ in air at 37°C. After a period of 3 to 4 wk, when the growth of fibroblasts was established, lung fragments were removed and the first passage was performed. The cells were washed in phosphate-buffered saline (PBS)and then covered in a minimal quantitiy of 0.05% trypsin and 0.02% ethylenediaminetetraacetic acid. After 5 min incubation, the resultant cell suspension from each flask was plated directly into six 75-cm² flasks (NUNC, Wiesbaden, Germany) and grown to confluence. Subsequently, cultures were either frozen in liquid nitrogen or passaged into 500-cm² flasks and maintained for up to five passages.

Cytokine Stimulation of Lung Fibroblasts

Fibroblasts from the fifth passage were stimulated with either Th1 or Th2 cytokines including TNF-, IFN-, or IL-4 (Pepro Tech, London, UK) at a concentration of 20 ng/ml. Combination of either TNF-/IFN- or TNF-/IL-4 was also used to stimulate lung fibroblasts. After 24, 48, or 72 h of cytokine stimulation, culture supernatants were collected and assayed for Eo chemotaxis. Lung fibroblasts were restimulated at Days 6 and 12, and culture supernatants were collected at Days 6, 12, and 25 over the 25-d period. Eo chemotactic activity (ECA) was studied as described below.

Purification of Chemoattractants by High-Performance Liquid Chromatography (HPLC)

Chemoattractants were purified by sequential heparin- affinity chromatography, reverse-phase (RP-8) chromatography, microbore cation exchange chromatography, and reverse-phase (RP-18) chromatography.

Initially, supernatants from cytokine-stimulated lung fibroblasts (500 ml per experiment, derived from approximately 5 × 10⁸ fibroblasts) were concentrated using Amicon YM3 filters (Amicon Corp., Danvers, MA) and then diluted in 10 ml of 10 mM Tris/10 mM citrate buffer at pH 8.0 (equilibration buffer) before application to a heparin- sepharose affinity column (HiTrap; Pharmacia, Uppsala, Sweden). Unbound material was washed from the column using equilibration buffer, and bound material was then eluted using 1 M NaCl in the same buffer. This material was concentrated using Amicon filters as described previously and resuspended in 0.1% trifluoroacetic acid (TFA), pH 3, before application onto a preparative wide-pore RP-8 column (300 × 7 µm, C₈ Nucleosil, 250 × 12.6 mm; Macherey and Nagel, Düren, Germany). Proteins were eluted via a linear gradient of acetonitrile in water containing 0.1% TFA (flow 3 ml/min), and absorbance was measured at 215 nm. Fractions obtained from the column were lyophilized before assay for Eo chemotaxis.

Microbore HPLC Purification of Attractants

Fractions obtained from preparative RP-8 HPLC, containing ECA, were concentrated as described previously and resuspended in 20 mM ammonium formate, pH 4.0, before application onto a micro Mono S-HPLC column (attached to a Smart-Micro-HPLC System; Pharmacia) previously equilibrated with 50 mmol/liter ammonium formate, pH 4.0, containing 25% (vol/vol) acetonitrile. Proteins were eluted with a linear gradient of NaCl (maximum 1 mol/litter) with a flow rate of 100 µl/min, and absorbance was measured at 215, 254, and 280 nm. Fractions containing ECA were further purified by reverse-phase chromatography using a microbore RP-18 HPLC column. Integration values were calculated from known amounts of recombinant eotaxin (Pepro Tech) that were used for calibration.

Isolation of Eos

Blood obtained from healthy donors or subjects with mild eosinophilia (5 to 10% of peripheral blood leukocytes) was collected into citric acid/dextran, and Eos were isolated with the use of discontinuous Percoll density gradient as previously described (20). Eo preparations were > 90% pure.

Chemotaxis Assay

Lung fibroblast supernatant HPLC fractions (10 to 20 µl) were lyophilized, dissolved in PBS containing 0.1% bovine serum albumin, and then assayed for Eo chemotaxis. Eo chemotaxis was performed using the Boyden chamber technique and indirect counting of migrated Eos by means of an established endogenous component chemotaxis assay using -glucuronidase as marker enzyme (20). Chemotactic activity is expressed as a chemotactic index, calculated as stimulated migration divided by random migration.

To investigate whether IL-5 synergizes with eotaxin, Eos were primed with IL-5 (either 125 pg/ml or 5, 10, 100, or 250 ng/ml) for 30 min before the chemotaxis assays. Lung-derived eotaxin was used in a concentration range of 0.01 to 1,000 ng/ml.

Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis (SDS-PAGE) and Western Blot

SDS-PAGE was performed using PhastGel high-density gels (Pharmacia) according to the manufacturer's protocol. Briefly, electrophoresis was performed at pH 6.4 with PhastGel SDS buffer strips that contained 0.112 M Tris/ 0.112 M acetate buffer. Proteins were fixed with 0.5% (vol/vol) glutaraldehyde containing 0.1% (wt/vol) sodium thiosulfate in 30% aqueous (vol/vol) ethanol containing 0.4 M/liter sodium acetate at pH 6.0, and were visualized by the use of a silver staining kit (Sigma, St. Louis, MO).

Western blot analysis was performed to detect RANTES using a specific monoclonal antibody (mAb) raised against RANTES (mAb 2DS). Full details of this procedure have been published previously (21).

Amino Acid Sequence Determinations

Amino acid sequences were determined by Edman degradation in an Applied Biosystems 476 A pulsed liquid protein sequencer using reverse-phase HPLC for detection of phenylthiohydantoin-amino acid derivatives.

Cysteine residues were confirmed on filter reduction with tributylphosphine and alkylation with 4-vinylpyridine.

Chemokine Measurements

Solid phase enzyme-linked immunosorbent assays (ELISAs). Measurements of the immunoreactive chemokines RANTES, MCP-3, and MCP-4 in HPLC fractions were performed by solid-phase ELISA (22), using the appropriate antibodies (anti-RANTES mAb ID2 was a gift from Dr. Michael Sticherling [University of Kiel, Germany]; mouse anti-MCP-3 mAb and rabbit anti-MCP-4 polyclonal antibodies were purchased from Pepro Tech).

Sandwich ELISAs. Concentrations of eotaxin and RANTES in the culture supernatants of cytokine-stimulated lung fibroblasts were measured via sandwich ELISAs. For the eotaxin ELISA measurements, ELISA plates (CovaLink; Nunc Roskilde, Denmark) were coated with the 10 µg/ml of the 164.4 monoclonal mouse antihuman eotaxin monoclonal antibody (University of Tokyo, Tokyo, Japan). The antibody was allowed to bind overnight at 4°C, and the plates were then washed three times with PBS containing 0.05% Tween 20 before wells were blocked with blocking reagent (Boehringer Mannheim, Mannheim, Germany) for 1 h. After the plates were washed as previously described, either eotaxin standards (purified from the supernatant of dermal fibroblast) or samples were added and plates were incubated overnight at 4°C. The plates were washed again and 0.025 µg of the 174.44 biotinylated monoclonal mouse anti-eotaxin antibody was added to each well. After 3 h incubation, plates were washed and streptavidin-horseradish peroxidase (Amdex, Copenhagen, Denmark) was added. After 1 h incubation, plates were developed by o-phenylendiamine dihydrochloride (Sigma). Concentrations of eotaxin in samples were calculated from the standard curve. The lower limit of detection was 40 pg/ml eotaxin. Details of the RANTES ELISA have been described previously (22).

Reverse Transcriptase-Polymerase Chain Reaction (RT-PCR)

Total RNA from lung fibroblasts was isolated by acidic guanidinium thiocynate-phenol-chloroform extraction. One microgram total RNA was reverse transcribed using an oligo T18 primer and standard reagents (GIBCO BRL, Eggenstein, Germany). Intron spanning sets of primers specific for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and eotaxin and appropriate chemokines were used to differentiate between genomic and complementary DNA (cDNA) templates. Primer sequences for chemokines were RANTES: forward (F) primer 5'CATCCTCATTGCTACTGCCCTCTG-3', reverse (R) 5'CGGGTTCACGCCATTCTCCT-3'; MCP-3: (F) 5'-ACCAAAC-CAGAAACCTCCAATTC-3', (R) 5'AGGTAGAGAA-GGGAGGAGCAT-3'; MCP-4: (F) 5'-CTTGCAGAGGCTGAAGAGCTATG-3', (R) 5'-CTCAACCCCTGGGAACCGA-3'; eotaxin (23), eotaxin-2: (F) 5'-CTCACGGGCTCTGTGGTC-3', (R) 5'-GGTTTGGTTGCCAG-GATA-3'. cDNA corresponding to 50 ng RNA served as template in a duplex-PCR reaction containing 0.8 µm of primer specific for RANTES, MCP-3 and MCP-4, or eotaxin and eotaxin-2. Amplification and analysis of the PCR products were performed as previously described (23).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

TNF- Induces the Production of Two Eo Attractants

TNF- is a cytokine that promotes differentiation and development of Th1 cells (24). To investigate whether lung fibroblasts produce Eo chemoattractants we stimulated these cells, which were grown from normal lung tissue, with 20 ng/ml TNF-. Analysis of the crude fibroblast culture supernatant showed the presence of ECA, whereas supernatants from nonstimulated fibroblasts contained negligible ECA (data not shown). To test the hypothesis that chemokineswhich have the ability to bind heparincould contribute greatly to ECA, fibroblast culture supernatants were initially separated by heparin-affinity chromatography, followed by reversed HPLC chromatography (RP-8 column). Eo chemotaxis assays showed that the nonbinding proteins from the heparin column, as well as the lipids containing effluent, contained negligible ECA. In contrast, heparin-binding fractions were found to be active and revealed two peaks of ECA when separated by preparative reverse-phase HPLC (Figure 1A).

View larger version (17K): [in this window] [in a new window]   Figure 1.   Eo chemotactic activities released in the culture supernatant of either TNF-- (A), IFN-- (B), or IL-4-stimulated (C ) lung fibroblasts. Material eluted from the heparin affinity column was applied to a reverse-phase chromatography (RP-8), and bound material was washed from the column using a two-step linear gradient of acetonitrile over 40 min. Eo chemotaxis to individual fractions is given as a chemotactic index (open and filled bars). Whereas TNF- and IL-4 led to strong ECA peaks (chemotactic index > 2), IFN- led only to very weak activity. Filled bar represents ECA containing RANTES immunoreactivity. Negative control (buffer), per definition, gave a chemotactic index of 1.

The early peak of ECA was further purified by microcation-exchange chromatography followed by micropore reverse-phase (RP-18) HPLC, giving a single protein peak that contained the whole ECA (Figure 2A). SDS-PAGE analysis showed that this molecule migrated as 13-kD protein. The late "peak" of Eo activity was also purified to homogeneity using the same chromatography protocol described previously and upon SDS-PAGE analysis this activity migrated as 8-kD protein. Both peaks of ECA were detected in the supernatant of cultured fibroblasts when stimulated for 24 h to 25 d with TNF-.

View larger version (18K): [in this window] [in a new window]   Figure 2.   Purification of the RP-8 chromatography early (A) and late (B) peaks of ECA released by TNF--stimulated lung fibroblasts (see Figure 1A). The two ECAs further purified by micro- Mono S HPLC (not shown) were applied separately to a micro-reverse-phase column (RP-18) and bound material was washed from the column using a two-step linear gradient of acetonitrile over 40 min. Upon SDS-PAGE the early peak of Eo activity migrated as a 13-kD protein (A, lane 2), whereas the late peak of ECA gave an 8-kD molecular mass (B, lane 2). Western blot was performed with the Phast System, and the late peak of ECA (B, lane 4) was stained using a mAb ID2 against RANTES. Lane 3 shows platelet-derived RANTES used as a positive control.

IFN- Induces a Weak Peak of Eo Activity

To investigate whether IFN- induces lung fibroblasts to release Eo attractants, these cells were stimulated with IFN-. Analyses of the cell culture supernatant by heparin-sepharose chromatography followed by reverse-phase chromatography revealed the presence of a weak peak of Eo activity (chemotactic index 1.3 [Figure 1B]). Further purification of this chemoattractant by microcation-exchange chromatography followed by reverse-phase (RP-18) chromatography was not successful because the amount of this ECA was too low.

IL-4 Induces a Single Peak of Eo Activity

To test the hypothesis that a Th2 cytokine can induce a different pattern of release of Eo attractants than that observed with TNF- or IFN-, lung fibroblasts were stimulated with IL-4. HPLC analysis of supernatants obtained from these stimulated fibroblastsusing the same methods described previouslyshowed the presence of a single peak of ECA (Figure 1C). Interestingly, this Eo attractant also exhibited an identical elution profile and the same biochemical behavior (mobility upon SDS-PAGE analysis and retention time upon HPLC) as the early peak of activity induced by either TNF- or IFN-.

IL-4, As Well As IFN-, Potentiates the Ability of TNF- to Induce Eo-Attractant Production

The effect of a combination of TNF- with either IFN- or IL-4 on the ability of lung fibroblasts to release Eo attractants was also studied. Like TNF- (Figure 1A), either TNF- and IFN- or TNF- and IL-4 stimulate lung fibroblasts to release two peaks of activity, with RP-8 chromatography elution profiles that were identical to those peaks induced by TNF- alone. Interestingly, the major peak of Eo activity induced by combination of TNF- and IFN- was the late peak, whereas the major peak induced by TNF- and IL-4 was the early peak. Purification to homogeneity of the early peak of Eo activity induced by TNF- and IL-4 finally resolved into three biochemically distinct peaks of Eo activity (Figure 3) that, upon SDS- PAGE analysis, were found to give single bands at 12.5, 13, and 12.8 kD, respectively (Figure 3). In contrast, when purified to homogeneity, the late peak of Eo activity (eluting from the RP-8 HPLC column) induced by either TNF-/ IFN- or TNF-/IL-4 migrated as an 8-kD molecule.

View larger version (14K): [in this window] [in a new window]   Figure 3.   TNF- and IL-4 stimulate the production of eotaxin variants. ECA off RP-8 HPLC resolved into two peaks of ECA (A) by microcation-exchange chromatography (Mono S column). Further purification of the first peak of ECA off micro-Mono S HPLC by micro-RP-18 HPLC (B) resulted in two peaks (12.5 kD [lane 2] and 13 kD [lane 3]), whereas separation of the second peak (C ) resolved as a single peak of ECA (12.8 kD [lane 5]). Lanes 1 and 4 contain recombinant Escherichia coli-derived eotaxin.

Characterization of Eo Attractants

To obtain structural information of the 12.5 to 13-kD Eo activities, N-terminal amino acid sequence analyses were performed. These revealed the sequences GPASVPTTXXFNL (50%) and ASVPTTXXF (40%) as predominant forms, which have previously been identified as eotaxin. Interestingly, all three peaks of Eo activity were found to be composed of both eotaxin forms and another N-terminally truncated form starting with Ser-Val (10%).

To test whether the late peak of Eo activity originates from RANTES, Western blot experiments were performed using a specific anti-RANTES mAb. As a result we found that Peak 2 of Eo activity induced by either TNF-, TNF-/ IL-4, or TNF-/IFN- corresponded to RANTES (Figure 2B and data not shown).

Biologic Activity of Eotaxin Variants

To investigate whether the lung fibroblast-derived eotaxin variants exert different biologic activity on Eos, chemotaxis assays were performed. These in vitro experiments demonstrated that eotaxins Peak 1 and Peak 3 exert more potent ECA (Figure 4) than does eotaxin Peak 2. 

View larger version (12K): [in this window] [in a new window]   Figure 4.   Chemotactic activity of eotaxin variants. Eotaxin variants were purified from the supernatant of TNF-- and IL-4-stimulated fibroblasts as described in Figure 3 and MATERIALS AND METHODS. ECA of the eotaxin variants is given as a chemotactic index.

IL-5 has previously been shown to enhance Eo migration in response to recombinant eotaxin in vivo (25); however, the priming effect in vitro is a subject of controversy (26). We have primed Eos with different concentrations of IL-5 (125 pg/ml to 250 ng/ml) and studied their chemotactic response to natural eotaxin. We found that IL-5 increased mobility of Eos but did not enhance the Eo chemotactic response to any of the eotaxin variants (data not shown).

Chemokine Immunoreactivity Measurements

Eotaxin and RANTES measurements in culture supernatant. To quantify the amount of eotaxin and RANTES released by cytokine-stimulated lung fibroblasts, the concentrations of these chemoattractants were measured by ELISA. TNF-, IL-4, and to a much greater extent, a combination of TNF- and IL-4 were the major stimuli for eotaxin release by lung fibroblasts (Figure 5). Interestingly, a combination of TNF- and IFN- was a less potent stimulus than TNF- alone (Figure 5), suggesting that IFN- suppresses TNF--induced eotaxin production in lung fibroblasts.

View larger version (84K): [in this window] [in a new window]   Figure 5.   Determination of eotaxin and RANTES in the culture supernatants of cytokine-stimulated lung fibroblasts 24 (open bars) and 48 h (closed bars) after stimulation. Results shown are representative experiments performed in triplicate estimated for the release of the chemokines per 106 lung fibroblasts.

Chemokine immunoreactivity in HPLC fractions. To obtain further insight into the relative production of Eo- active chemokines by lung fibroblasts, we investigated chemokine immunoreactivity in the HPLC fractions by solid-phase ELISA. As shown in Figure 1C, RANTES immunoreactivity correlated with peaks of Eo activity previously identified as RANTES. Similarly, very weak immunoreactivity MCP-3 was detected in a few HPLC fractions; however, they did not correlate with peaks of ECA (data not shown). Using a polyclonal MCP-4 antibody, attempts were made to detect MCP-4 immunoreactivity. However, this antibody gave high background and no MCP-4 immunoreactivity was detected.

Messenger RNA (mRNA) Expression of Eo-Activating Chemokines

After 6 h of stimulation, TNF- induced mRNA expression for both eotaxin and RANTES in lung fibroblasts. IL-4 was found to be a potent inducer of eotaxin mRNA expression but a weak stimulus for RANTES mRNA expression. Dose-concentration experiments showed that IL-4, at concentrations as low as 0.1 ng/ml, was a potent inducer of eotaxin mRNA expression (Figure 6A). In contrast, IFN- was a poor stimulus for both eotaxin and RANTES mRNA expression (Figures 7A and 7B). TNF- was found to synergize with both IL-4 and IFN- to enhance eotaxin mRNA expression. Interestingly, the enhanced eotaxin gene expression induced by TNF-/IFN- contrasts with the low eotaxin protein release by lung fibroblasts, which suggests that the eotaxin production induced by TNF-/IFN- may be regulated at the post-transcriptional level. On the contrary, TNF-/IFN- is a potent inducer of both RANTES mRNA and RANTES protein. Other Th1 and Th2 cytokines, such as IL-2, IL-5, IL-6, and IL-10, did not induce eotaxin or RANTES mRNA expression in fibroblast (data not shown).

View larger version (39K): [in this window] [in a new window]   Figure 6.   (A) Eotaxin mRNA expression in lung fibroblasts in response to a concentration range of IL-4 (0.1 to 100 ng/ml). (B) Chemokine gene expression in normal lung tissue. Amplified chemokine cDNA is marked by ">". See legend to Figure 7.

View larger version (46K): [in this window] [in a new window]   Figure 7.   Gene expression of eotaxin (A), RANTES (B), MCP-3 (C ), and MCP-4 (D) relative to GAPDH in lung fibroblasts. After 6 h of stimulation with TNF-, IFN-, IL-4, or TNF-/IFN-, TNF-/IL-4 total RNA was isolated from lung fibroblasts. One microgram of total RNA was reverse transcribed, and appropriate sets of specific primers were used to detect each of the above chemokines by semiquantitative RT-PCR (MATERIALS AND METHODS). Chemokine mRNA expression relative to GAPDH expression was calculated from the densitometrically determined amounts of the corresponding amplification products.

Because we did not detect biologically active MCP-3 or MCP-4, we investigated whether lung fibroblasts express mRNA for these chemokines. After 6 h of stimulation, TNF-, IFN-, or IL-4 induced the accumulation of MCP-3 mRNA in lung fibroblasts (Figures 7C and 7D). Moreover, TNF- was synergetic with either IFN- or IL-4 to induce MCP-4 mRNA expression. Lung fibroblasts did not express eotaxin-2 mRNA under any conditions (data not shown).

To investigate whether the Eo-activating CC-chemokines are normally implicated in leukocyte trafficking, we also studied gene expression in normal lung tissue using RT-PCR. This demonstrated that, with the exception of the MCP-3 mRNA, mRNAs for eotaxin, RANTES, and MCP-4 are constitutively expressed in normal lung tissue (Figure 6B).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In this study we have characterized Eo attractants released by lung fibroblasts stimulated with either Th1- or Th2-type cytokines. IL-4 induced the release of a single peak of ECA that was found to correspond to eotaxin. In contrast, IFN- had negligible effects on the ability of these cells to release this attractant. The combination of TNF- and IL-4 greatly increased the production of eotaxin, whereas that of TNF- and IFN- predominantly induced RANTES production.

Previous independent reports have shown that IL-4 and IFN- regulate eotaxin mRNA expression in several cell types, including endothelial cells, epithelial cells, and dermal fibroblasts (22, 29), and demonstrated that either of these cytokines synergizes with TNF- to induce eotaxin mRNA expression further. Using multiple steps of chromatography, we have recently reported that IL-4 induces biologically active eotaxin from dermal fibroblasts (18), whereas Lilly and colleagues (31) demonstrated that A549 epithelial cells stimulated with IL-1 released very small amounts of eotaxin immunoreactivity (45 pg/10⁶ cells). To date, however, a systematic examination of the production of other Eo-activating chemokines (such as RANTES, MCP-3, MCP-4, and eotaxin-2) by lung fibroblasts upon Th1- or Th2-type cytokine stimulation has not been carried out. The present study demonstrates that IL-4 is a potent stimulus for the release of eotaxin from lung fibroblasts (140 ng eotaxin/2.5 × 10⁵ fibroblasts), whereas IFN- has negligible effects on these cells to release eotaxin. Interestingly, the amount of immunoreactive eotaxin induced by IL-4 and a combination of IL-4 and TNF- from lung fibroblasts is 3,800 and 20,000 times (respectively) greater than that induced by cytokine-stimulated A549 epithelial cells, as reported by Lilly and associates (31). This observation, together with the finding that very low concentrations of IL-4 (0.1 ng/ml) are required to induce eotaxin mRNA expression, places the fibroblasts as the major cellular source of eotaxin. In a separate study, we recently investigated eotaxin production by other cell types, including lymphocytes and monocytes, and demonstrated that these cells produce minute amounts of eotaxin compared with lung fibroblasts (unpublished data). An unexpected finding in the present study was that whereas the combination of TNF- and IFN- strongly upregulates the expression of eotaxin mRNA in lung fibroblasts, these cytokines did not enhance the eotaxin protein production, which suggests that the eotaxin regulation by combination of these cytokines may take place at post-transcriptional level.

We have recently shown that dermal fibroblasts release several glycosylated forms of eotaxin with a molecular weight of 13 kD. Consistent with this observation, the present study shows that TNF- also stimulates lung fibroblasts to release a 13-kD eotaxin. Interestingly, a combination of TNF- and IL-4 induces several forms of eotaxin with molecular weights of 12.5, 12.8, and 13 kD, which suggests some heterogeneity between lung and dermal fibroblasts. Interestingly, the 12.5- and 12.8-kD eotaxins isolated from lung fibroblasts showed a more potent in vitro chemotactic activity than did the 13-kD one. Further studies need to define whether these eotaxin variants differentially stimulate Eo recruitment in vivo. In this study, however, none of the eotaxin variants synergized with IL-5 to enhance Eo migration. This finding is consistent with the study conducted by Heath and coworkers, who showed similar results using recombinant eotaxin (27), and disagrees with a recent study showing that IL-5 enhances the Eo chemotactic response to eotaxin (28). The finding that IL-5 primed bone-marrow guinea pig (BM-GP) Eos to respond to eotaxin suggests that BM-GP Eos exhibit different characteristics from circulating human Eos (26).

The present study shows that neither IL-4 nor IFN- induced RANTES production by lung fibroblasts. In contrast, TNF- not only induced RANTES production but also stimulated eotaxin release, both of which were found to be biologically active on Eos. This finding supports a previous study showing that injection of TNF- into rat skin induces Eo recruitment at the injection site (32) and suggests that this cytokine recruits Eos via the production of both RANTES and eotaxin. We have also demonstrated that the combination of TNF- and IFN- preferentially stimulates RANTES mRNA expression and biologically active RANTES protein. This finding is in agreement with the study conducted by Marfaing-Koka and colleagues, which showed that a combination of IFN- and TNF- stimulates RANTES release by endothelial cells (33). However, these authors found that IL-4 suppressed the ability of TNF- to induce RANTES production from endothelial cells, whereas we found that IL-4 did not affect the ability of TNF- to release RANTES. In the present study the effects of other Th1 and Th2 cytokines on eotaxin and RANTES expression were also investigated. However, neither IL-2, -5, -6, nor -10 stimulated chemokine expression in lung fibroblasts.

Although lung fibroblasts were found to express mRNA for MCP-3 and MCP-4 upon appropriate cytokine stimulation, these cells did not secrete biologically active MCP-3 or MCP-4. We detected sparse immunoreactive MCP-3 in few HPLC fractions, but the finding that this immunoreactivity did not induce Eo chemotaxis suggests that MCP-3, when released by lung fibroblasts, could be cleaved into an inactivated form, or that the anti-MCP-3 mAb used in the solid-phase ELISA could react with another chemokine. Because manufacturers have demonstrated that this antibody is specific against MCP-3 (it does not cross-react with MCP-1, MCP-2, or other structurally related chemokines), this last observation would seem unreasonable. However, it must be taken into account that members of the chemokine family are rapidly growing in number, and the possibility of cross-reactivity with an unknown chemokine cannot be ruled out.

We studied gene expression of Eo-activating chemokines in normal lung tissue using RT-PCR and demonstrated that mRNAs for eotaxin, RANTES, and MCP-4, but not MCP-3, are constitutively expressed in the normal lung tissue. This finding suggests that, with the exception of MCP-3, all of the above chemokines may regulate leukocyte trafficking under normal conditions. To study the role of chemokines in Eo recruitment in lung disease, Chensue and coworkers developed mice models of lung granuloma formation characterized by either Th1 or Th2 cytokine profile (34). This allowed them to demonstrate that Eo recruitment into the lung granuloma formation induced by PPD (Th1 profile characterized by increased IFN- and IL-2 concentrations in cultured lung lymph nodes) is associated with RANTES mRNA expression, whereas that induced by N. brasiliensis (Th2 profile characterized by increased IL-4 and IL-5 concentrations) correlated with eotaxin mRNA expression (34). A type-2 cytokine pattern has been found to predominate in both the pulmonary interstitium of patients with cryptogenic fibrosing alveolitis (16) and in bronchial biopsies of asthmatic patients (35). We have demonstrated that endobronchial allergen challenge leads to the release of RANTES into the airway epithelial lining fluid (bronchoalveolar lavage [BAL]) of asthmatic patients (21, 36), and Lamkhioued and associates detected eotaxin immunoreactivity in steady-state asthma (28). Similarly, two separate studies have reported increased mRNA expression for either RANTES or eotaxin in bronchial biopsies derived from asthmatic patients (28, 37). On the other hand, patients with systemic sclerosis who developed fibrotic lung disease have been reported to release RANTES immunoreactivity in the BAL fluid (38). We hypothesize that the release of IL-4 into the bronchial tissue may polarize fibroblasts to produce eotaxin, whereas TNF-/IFN- may stimulate these cells to predominantly release RANTES. Indeed, the demonstration that asthmatics released increased concentrations of Th2 cytokines into their airways (35), together with the finding that there is increased collagen deposition in the lamina reticularis beneath the bronchial epithelium (5), suggests that these cytokines regulate chemokine production by lung fibroblasts.

In summary, the present study has demonstrated that lung fibroblasts produce two biologically Eo-activating chemokines including eotaxin and RANTES upon appropriate cytokine stimulation. IL-4 specifically stimulates the production of eotaxin, whereas IFN- had a negligible effect on the release of this chemokine. IL-4 synergized with TNF- to release eotaxin; and IFN-, despite enhanced mRNA expression, suppressed the TNF--induced release of this chemokine by lung fibroblasts. These observations, together with the finding that TNF- and IFN- preferentially stimulate RANTES production, suggest that Th1 cytokines may polarize lung fibroblasts to produce RANTES, whereas IL-4, a Th2-type cytokine, polarizes these cells to release eotaxin. Because these two chemokines activate Eos through the CCR3 receptor, the development of an antagonist for this receptor may have a major impact in the treatment of lung eosinophilic diseases, including bronchial asthma and lung fibrotic disease.