Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Long-term treatment with low-dose erythromycin (EM), a macrolide antibiotic, has been accepted as an effective treatment for diffuse panbronchiolitis (DPB) (1). This finding has been extended to include other airway inflammatory diseases such as bronchiectasis associated with chronic sinusitis, chronic bronchitis, and bronchial asthma (2, 4). Several newly synthesized macrolides, including clarithromycin (CAM) and roxithromycin, have been confirmed to provide the same beneficial effects as EM for patients with chronic airway inflammation (9). The mechanisms of effectiveness of macrolides for treating airway inflammatory diseases have been explained by their antiinflammatory effects (1). The antiinflammatory effects of macrolides include inhibition of cytokine production in neutrophils (4), monocytes (16), and bronchial epithelial cells (9, 17, 18); suppression of T-cell proliferation by EM (19); and in vivo inhibition of vascular smooth-muscle cell migration by rapamycin (20).

Interleukin (IL)-8 is a potent neutrophil chemotactic and activating factor produced by various cell types including bronchial epithelial cells (21). IL-8 proteins are abundant in sputum from patients with chronic bronchitis, bronchiectasis, and cystic fibrosis (28), and in bronchoalveolar lavage fluid (BALF) from patients with DPB (4, 29), in relation to neutrophil chemoattractant activity in these diseases. EM was found to significantly reduce the number of neutrophils or levels of IL-8 protein in BALF from patients with DPB and chronic bronchitis (4). Further, EM repressed neutrophil influx into the alveoli in response to Proteus mirabilis (13) or challenge with recombinant human IL-8 (5). These lines of evidence suggest that one of the antiinflammatory mechanisms of macrolides relates to the inhibition of IL-8 production. The goal of the present study was to elucidate the molecular mechanisms of IL-8 inhibition by macrolides through use of cells of the human bronchial epithelial cell line BET-1A.

Bronchial epithelium is now recognized as not merely a boundary of ambient air along the bronchial tree, but also as an active biomodulator via production of a variety of proinflammatory mediators (25, 30, 31). Among the bioactive substances produced by bronchial epithelial cells, IL-8 has been found to be one of the most important mediators of airway inflammation (9, 17, 18, 25).

In the present study we found, by using the luciferase reporter gene assay and electrophoretic mobility shift assay (EMSA), that CAM repressed tumor necrosis factor (TNF)--induced IL-8 gene transcription in BET-1A cells, and that the activator protein (AP-1) site on the IL-8 gene was one of the elements responsible for a CAM-related repression of TNF--induced promoter activation.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Cell Cultures

BET-1A, a human bronchial epithelial cell line transformed by the SV40 virus, was a generous gift of Dr. C. C. Harris of the National Cancer Institute, Bethesda, MD (32). BET-1A cells were cultured in 6-cm culture dishes filled with serum-free light harvesting complex (LHC)-9 medium (Biofluids, Inc., Rockville, MD) containing 25 µg/ ml amphotericin B, 25 U/ml penicillin, and 25 µg/ml streptomycin (GIBCO BRL/Life Technologies, Inc., Gaithersburg, MD) at 37°C in a humidified incubator containing 5% CO₂ in air (33). The surfaces of the culture dishes were coated with coating solution consisting of a mixture of 10 µg/ml fibronectin (Collaborative Research, Inc., Bedford, MA), 30 µg/ml collagen (Vitrogen 100; Collagen Corp., Palo Alto, CA), and 10 µg/ml crystallized bovine serum albumin (BSA; Biofluids) dissolved in LHC basal medium (Biofluids). All experiments were done after the cells had reached 70-80% confluence.

Macrolides and Stimulants

EM (Dainippon Pharmaceutical Co., Osaka, Japan) and CAM (Taisho Pharmaceutical Co., Tokyo, Japan) were dissolved in methanol at a final concentration of 2 mg/ml, and the solution was stored at 20°C. Cephazolin (CEZ; Fujisawa Pharmaceutical Co., Osaka, Japan) and ampicillin (ABPC; GIBCO BRL) were dissolved in distilled water. We also used human recombinant TNF- (100 U/ml; Genzyme Corp., Cambridge, MA) as a positive stimulant for IL-8 gene expression in bronchial epithelial cells (25- 27). Before the introduction of macrolides, CEZ, or ABPC, the culture medium was replaced with fresh, antibiotic-free LHC-9. The cells were then incubated for 24 h. The culture media containing macrolides or other antibiotics such as CEZ and ABPC were changed every 24 h. Dosages of CAM used in this study produced the same levels as those of this drug that are achievable in patients who are normally prescribed CAM at 400 to 800 mg/d (34).

Cell Growth

Effects of macrolides on cell proliferation and viability were evaluated with the alamarBlue assay (BioSource International, Camarillo, CA) (35). BET-1A cells were cultured at 37°C for 0-72 h in 96-well plastic plates at a density of 4,000 cells/well in 20 µl of alamarBlue dye solution in 180 µl LHC-9 medium in the presence or absence of EM or CAM (0, 3, and 30 µg/ml). The optical density (OD) of the culture medium was measured at 595 nm with a spectrophotometric microtiter plate reader (Model 450; Bio-Rad Laboratories, Inc., Hercules, CA) at 0, 1, 4, 6, 8, 24, 36, 48, 60, and 72 h. In these analyses, cell proliferation and viability correlated well with the OD of the medium in each well. Morphology of BET-1A cells was evaluated visually throughout the course of the experiments.

IL-8 Messenger RNA Transcript Levels

The levels of IL-8, type I TNF receptor (TNF-RI), and control glyceraldehyde-3-phosphate dehydrogenase (GAPDH) messenger RNA (mRNA) transcripts were evaluated by Northern blot analysis (25, 27, 36). Total cellular RNA was isolated through the RNAsol method (Cinna/Biotex Laboratories, Inc., Houston, TX) (37). RNA (15 µg) was subjected to formaldehyde-agarose gel electrophoresis, transferred to a nylon membrane (Nytran; Schleicher & Schuell, Inc., Keene, NH), hybridized with a ³²P-labeled IL-8, TNF-RI, or GAPDH complementary DNA (cDNA) probe generated through a random priming method (38, 39), and evaluated autoradiographically. Autoradiographic signals were analyzed with a laser densitometer (Atto Densitograph; Atto Corp., Tokyo, Japan). The IL-8 cDNA used as a probe (pPB248) was a 750-bp cDNA segment that included the sequence from the PstI site of exon 1 to the BamHI site of exon 4 (25). The probe for TNF-RI was a 304-bp cDNA segment including the sequence from the HindIII site to the EcoRI site of human TNF-RI cDNA (40). The human GAPDH cDNA probe was a 1,007-bp cDNA segment constructed by means of the polymerase chain reaction (PCR) as described previously (27). Autographic signals were analyzed with an Ultroscan densitometer (LKB 2222-020; Pharmacia P-L Biochemicals Inc., Milwaukee, WI).

In the experiments for dose dependency, BET-1A cells were incubated with 0-10 µg/ml CAM for 72 h, followed by stimulation with TNF- for 4 h. In the experiments for the time course of IL-8 gene expression, cells were preincubated for 0-96 h with 10 µg/ml CAM and then stimulated with TNF- for 4 h. Following incubation, total RNA was isolated and the levels of IL-8 and GAPDH mRNA transcripts were evaluated by Northern blot analysis as described earlier.

To estimate the effects of macrolides on the stability of TNF--induced IL-8 mRNA transcripts, BET-1A cells were exposed to actinomycin D (10 µg/ml; Sigma Chemical Co., St. Louis, MO) for 1-6 h after 96 h of preincubation with or without CAM (20 µg/ml) before being stimulated with TNF- for 4 h. Total cellular RNA was extracted at each time point, and IL-8 mRNA levels were evaluated through Northern blot analysis as described earlier and were quantified through laser densitometry.

IL-8 Protein Levels in Culture Supernatants

To evaluate IL-8 protein levels in culture supernatants, BET-1A cells were cultured for 96 h with CAM and were stimulated with TNF- for an additional 24 h. IL-8 protein levels in the culture supernatants were analyzed with a specific sandwich enzyme-linked immunosorbent assay (ELISA; Quantikine; R&D Systems, Inc., Minneapolis, MN) (27). Briefly, a microtiter plate was coated with a murine monoclonal antibody specific for IL-8. Experimental samples or standard samples of recombinant IL-8 were added to individual wells. After three washes to remove unbound protein, a polyclonal antibody specific for IL-8 and conjugated to horseradish peroxidase was added to the wells. After three additional washes, a substrate solution containing hydrogen peroxide and tetramethylbenzidine was added. The reaction was stopped by adding 2 N sulfuric acid. The color generated was determined by measuring the OD at 450 nm, using a spectrophotometric microtiter plate reader (Model 450; Bio-Rad). The standard curve produced in this manner was linearized on a log/log scale and was subjected to regression analysis. The results were reported as means ± SEM in pg/ml culture supernatant.

Effects of CAM on Promoter Activity of IL-8 Gene 5'-Flanking Sequences

Transfection vectors containing fusion genes of 5'-flanking sequences of the IL-8 gene and a luciferase reporter gene were constructed from a pUC8-derived vector (pCMV- luciferase) (25). The IL-8 promoter 5'-flanking region, a 1,525-bp EcoRI-HindIII fragment spanning from bp 1481 to +44 (numbering based on the sequence of the IL-8 gene reported by Mukaida and colleagues [41]), and its sequentially deleted fragments (starting from 391, 335, 130, 112, and 78 to +44 bp), were prepared by PCR and cloned into luciferase expression vectors by replacing the cytomegalovirus (CMV) promoter in pCMV-luciferase expression plasmids between unique XhoI and HindIII sites, as previously described (25). A promoterless luciferase plasmid (pLuc0) was used as a negative control. Site-directed mutagenesis of the IL-8, AP-1, nuclear factor (NF)-IL-6, and NF-B binding sites was done by using pN130, as previously described (42).

BET-1A cells were transfected through the lipofectin method (GIBCO BRL) (43). Before transfection, the cells were preincubated for 48 h in the presence or absence of CAM (20 µg/ml). Each luciferase expression plasmid vector (10 µg) and a CMV promoter--galactosidase expression plasmid (3 µg; Promega Corp., Madison, WI) were mixed in deionized water to a final volume of 100 µl. Lipofectin reagent (100 µg) was then added and the preparation was allowed to stand at room temperature (RT) for 15 min. The lipofectin-plasmid complexes were added to the culture medium, covering the cells in a 6-cm culture dish, and were incubated at 37°C for 24 h. The plasmid-containing medium was then replaced with 2 ml of fresh culture medium, and the cells were incubated in the presence or absence of CAM for another 24 h (finally incubated with CAM for 96 h). To evaluate the effects of TNF- on reporter gene expression, we added TNF- (100 U/ml) to the culture medium 6 h before harvesting of the transfected cells.

To measure reporter gene expression, we washed cells with Mg²⁺/ and Ca²⁺-free phosphate buffered saline (PBS) and lysed them in 400 µl of reporter lysis buffer (Promega) by incubating at RT for 15 min. The cell lysates were retrieved by scraping, vortexing, and centrifuging them at 15,000 rpm in a microcentrifuge for 2 min. Twenty microliters of the supernatant was then evaluated for luciferase activity, using a luminometer (Atto Luminescenser JNR; Atto Corp., Tokyo, Japan). The protein concentration of extracts was measured according to the Bradford method (44) (Bio-Rad), with BSA as a standard. -Galactosidase activity was analyzed with the -Galactosidase Enzyme Assay System (Promega).

Extraction of Nuclear Proteins and EMSA

BET-1A cells were preincubated with or without CAM (20 µg/ml) for 96 h in 10-cm culture dishes and then stimulated with TNF- (100 U/ml) for 1 h. Nuclear extracts were prepared according to the method described by Wong and coworkers, with some modifications (45). Treated cells were washed in PBS, retrieved by scraping, and pelleted by centrifugation. Cell pellets were suspended in 400 µl of Buffer A (10 mM 4-(2-hydroxyethyl)-1-piperazine-N'-2-ethanesulfonic acid [Hepes], pH 7.8; 10 mM KCl; 0.1 mM ethylenediaminetetraacetic acid [EDTA]; 1.5 mM MgCl₂; 0.2% Nonidet P-40; 1 mM dithiothreitol [DTT]; 0.1 mM phenylmethylsulfonyl fluoride [PMSF]) and incubated for 10 min with vortexing every 2 min. Cell lysates were centrifuged at 6,000 rpm in a microcentrifuge for 5 min, after which the pellets were resuspended and incubated in 50 µl of Buffer C (20 mM Hepes, 420 mM NaCl, 0.1 mM EDTA, 1.5 mM MgCl₂, 25% glycerol, 1 mM DTT, 0.5 mM PMSF) on ice for 20 min. After centrifugation of this preparation at 15,000 rpm in a microcentrifuge for 15 min, nuclear proteins contained in the supernatant were stored at 80°C. Protein concentration was determined by the Bradford method.

EMSA was done with the Gel Shift Assay System (Promega). Briefly, AP-1 oligonucleotide (5'-CTAGTGATGAGTCAGCCGGATC-3'), and NF-B oligonucleotide (5'-GATCCAGGGGACTTTCCCTAGC-3') (41) were end-labeled with [-³²P] adenosine triphosphate (3,000 Ci/mmol; Amersham Pharmacia Biotech, Buckinghamshire, UK), using T4 polynucleotide kinase, and were then purified in Microspin G-25 columns (Amersham Pharmacia Biotech, Uppsala, Sweden) and used as probes for EMSA. To complete the specific binding reactions, each of the 100-fold molar excesses of unlabeled AP-1, NF-B, and SP-1 (5'-ATTCGATCGGGGCGGGGCGAGC-3') oligonucleotide was added to the binding mixture before addition of the labeled probe. Nuclear extract proteins (10 µg) were preincubated with the binding buffer (10 mM Tris-HCl, pH 7.5; 50 mM NaCl; 0.5 mM EDTA; 1 mM MgCl₂; 0.5 mM DTT; 4% glycerol; 0.05 mg/ ml poly[deoxyinosine-deoxycytosine]) for 5 min and then incubated with the labeled probe at RT for 20 min. Each sample was electrophoresed in a 4% nondenaturing polyacrylamide gel in 0.5× Tris borate-EDTA (TBE) buffer at 100 V for 1 h. The gel was dried and subjected to autoradiography.

Statistics

To determine the cell growth curve and results of IL-8 protein analyses and luciferase activity analyses, statistical significance was analyzed, using analysis of variance with Bonferroni's correction for multiple comparisons after testing. All results are expressed as mean ± SEM, and results were considered significant at P < 0.05.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Inhibition of IL-8 Gene Expression by Macrolides in BET-1A Cells

Northern blot analyses showed that resting BET-1A cells expressed 1.8-kb IL-8 mRNA transcripts at a very low level. TNF- markedly increased the IL-8 mRNA transcript levels. Although macrolides, including EM and CAM, did not themselves change IL-8 gene expression in resting BET-1A cells, pretreatment with EM or CAM reduced TNF--induced IL-8 gene expression (Figure 1). In contrast, EM and CAM, alone or in combination with TNF-, did not modulate TNF-RI gene expression or control GAPDH gene expression in BET-1A cells. Neither CEZ nor ABPC modulated TNF--induced IL-8 gene expression (data not shown).

View larger version (44K): [in this window] [in a new window]   Figure 1.   Effects of macrolides on IL-8 and TNF-RI gene expression in the human bronchial epithelial cell line BET-1A. Results shown are the IL-8 mRNA transcript levels (top), TNF-RI mRNA transcript levels (middle), and control GAPDH mRNA transcript levels (bottom) of BET-1A cells treated as subsequently described. RNA was extracted from resting cells (lane 1), cells incubated for 72 h with EM (0.1 and 10 µg/ml, lanes 2 and 3) or CAM (0.1 and 10 µg/ml, lanes 4 and 5), and cells stimulated with TNF- alone (100 U/ml for 4 h, lane 6) or incubated with EM (0.1 and 10 µg/ml, lanes 7 and 8) or CAM (0.1 and 10 µg/ml, lanes 9 and 10) 72 h before TNF- stimulation. Northern blotting was performed as described in MATERIALS AND METHODS. The sizes of IL-8 mRNA transcripts are indicated. Bars represent densitometric units obtained from an autoradiogram, which are expressed as the ratios of IL-8 mRNA to GAPDH mRNA transcripts obtained from Northern blot analyses.

Effects of CAM on BET-1A Cell Growth

The alamarBlue assay demonstrated that neither higher (30 µg/ml) nor lower doses (3 µg/ml) of CAM affected the proliferation or viability of BET-1A cells (Figure 2). The same results were obtained in the case of EM (data not shown). In microscopic observations, although BET-1A cells grew to up to 90-100% confluence in a 96-well culture plate after incubation for 72 h, no morphologic changes were observed in either resting BET-1A cells or CAM-treated BET-1A cells.

View larger version (16K): [in this window] [in a new window]   Figure 2.   Effects of CAM on BET-1A cell growth. Cells were cultured with alamarBlue dye solution in the absence or presence of a low dose (3 µg/ml) and a high dose (30 µg/ml) of CAM for 72 h, and the OD at 595 nm was measured at the indicated time points. Proliferation and viability of BET-1A cells were determined in relation to the growth curve, based on the measurement of OD at each time point. Data are mean ± SEM of four independent experiments.

Dose- and Incubation Time-Dependent Inhibition by CAM of TNF--Induced IL-8 Gene Expression in BET-1A Cells

We investigated the dose-dependent effects of CAM on the IL-8 mRNA transcript levels upregulated by TNF- in BET-1A cells (Figure 3). Preincubation with CAM inhibited IL-8 gene expression in a dose-dependent manner. In addition, inhibition of TNF--induced IL-8 gene expression was dependent on the preincubation time with CAM (Figure 4). In contrast, control 1.4-kb GAPDH mRNA transcripts were consistently expressed in both resting BET-1A cells and cells exposed to TNF- and CAM.

View larger version (46K): [in this window] [in a new window]   Figure 3.   Dose-dependent inhibition of IL-8 gene expression by CAM in BET-1A cells. Results shown are IL-8 mRNA transcript levels (top) and control GAPDH mRNA transcript levels (bottom) of BET-1A cells treated as subsequently described. RNA was extracted from resting cells (lane 1) and cells stimulated with TNF- alone (100 U/ml for 4 h, lane 2) or incubated with CAM at the concentrations of 0.01, 0.1, 1, 10 µg/ml (lanes 3, 4, 5, and 6, respectively) for 72 h before TNF- stimulation. Northern blotting was performed as described in MATERIALS AND METHODS. Bars represent densitometric units obtained from an autoradiogram, which are expressed as the ratios of IL-8 mRNA to GAPDH mRNA transcripts obtained from Northern blot analyses.

View larger version (44K): [in this window] [in a new window]   Figure 4.   Effects of incubation time with CAM on IL-8 gene expression in BET-1A cells. Results of Northern blot analyses show IL-8 mRNA transcript levels (top) and control GAPDH mRNA transcript levels (bottom) of BET-1A cells treated as subsequently described. RNA was extracted from resting cells (lane 1) and cells stimulated with TNF- alone (100 U/ml for 4 h, lane 2) or incubated with CAM (10 µg/ml) for 24, 48, 72, and 96 h (lanes 3, 4, 5, and 6, respectively) before TNF- stimulation. Northern blotting was performed as described in MATERIALS AND METHODS. Bars represent densitometric units obtained from an autoradiogram, which are expressed as the ratios of IL-8 mRNA to GAPDH mRNA transcripts obtained from Northern blot analyses.

Effects of CAM on TNF--Induced IL-8 Protein Release in BET-1A Cells

Although resting BET-1A cells released IL-8 protein into the culture supernatants at low levels (347 ± 175 pg/ml/ 24 h), TNF- markedly increased IL-8 secretion by BET-1A cells (2,408 ± 212 pg/ml/24 h; Figure 5). In the presence of 10 µg/ml CAM, IL-8 protein levels in culture supernatants were significantly reduced to 1,625 ± 161 pg/ ml/24 h (P < 0.05 versus the levels with TNF-). EM also decreased TNF--induced IL-8 protein release (data not shown). These data were very consistent with the data obtained in Northern blot analyses, shown in Figure 1.

View larger version (12K): [in this window] [in a new window]   Figure 5.   Effects of CAM on IL-8 protein release from BET-1A cells in culture supernatants with mediation by TNF-. Supernatants of BET-1A cells incubated for 24 h were evaluated for IL-8 protein levels using an ELISA. Results shown were obtained from supernatants of resting cells (Rest), cells treated with TNF- (100 units/ml) alone (TNF), and cells treated with TNF- in the presence of CAM (CAM+TNF; 10 µg/ml, 96-h preincubation). Data are mean ± SEM of four independent experiments. * P < 0.05, compared with TNF--treated cells.

CAM Did Not Alter the Stability of IL-8 mRNA Transcripts in BET-1A Cells

To evaluate whether or not macrolides could change the stability of IL-8 mRNA transcripts, we examined the effects of CAM on the stability of these transcripts by inhibiting RNA synthesis with actinomycin D. IL-8 mRNA transcript levels in TNF--stimulated BET-1A cells fell, with a half-life of 3 h (Figure 6). The addition of CAM to BET-1A cells did not alter the half-life of IL-8 mRNA transcripts. This suggests that downregulation of TNF-- induced IL-8 mRNA transcript levels following CAM pretreatment is not modulated by changes in the stability of IL-8 mRNA transcripts.

View larger version (12K): [in this window] [in a new window]   Figure 6.   Effects of CAM on stability of IL-8 mRNA transcripts in BET-1A cells. Data shown are the effects of inhibition of RNA synthesis on IL-8 mRNA levels in TNF--stimulated (100 U/ml for 4 h) BET-1A cells preincubated with or without CAM (20 µg/ ml for 96 h). Cells were harvested at the indicated times after the addition of actinomycin D (10 µg/ml), and the extracted RNA was evaluated by Northern blot analysis. Autoradiographic signals were quantified by laser densitometry.

CAM Repressed IL-8 Gene Promoter Activity in BET-1A Cells

To examine the effects of macrolides on the transcriptional activity of the IL-8 gene, we transfected the fusion genes of 5'-flanking sequences of the IL-8 gene and a luciferase reporter gene into BET-1A cells in the presence or absence of CAM. To determine the efficiency of each transfection, a CMV promoter--galactosidase plasmid was cotransfected with a luciferase reporter gene into BET-1A cells, and -galactosidase activity was analyzed. CAM did not affect the activity of -galactosidase in BET-1A cells, confirming that CAM did not change the transfection efficiency of the plasmids. In resting BET-1A cells, the construct pNAF (sequence from 1,481 to +44 bp of the IL-8 promoter 5' flanking region) showed potent promoter activity (Figure 7). The deletion constructs pN391, pN335, pN130, and pN112, in which 5'-flanking sequences were variously deleted from 1,481 to 113 bp, also showed relatively potent promoter activity in resting BET-1A cells. Promoter activity declined to lower levels as the sequence was further deleted from 112 to 79 bp.

View larger version (20K): [in this window] [in a new window]   Figure 7.   Effect of CAM on promoter activity of the 5'-flanking region of the IL-8 gene in BET-1A cells. IL-8 promoter activity was evaluated with fusion genes consisting of sequentially deleted 5'-flanking regions of the IL-8 gene and a luciferase reporter gene. The constructs of fusion genes are shown to the left. Levels of luciferase expression by the fusion gene constructs are shown relative to expression of the construct pNAF in the resting state. Relative values of luciferase activity were adjusted for transfection efficiency as determined from -galactosidase expression through cotransfection with a CMV promoter--galactosidase plasmid. For evaluation of the effect of CAM on IL-8 promoter activity in BET-1A cells, the cells were preincubated in the presence or absence of CAM (20 µg/ml for 96 h), transfected with the fusion genes, and then stimulated with TNF- (100 U/ml) for 6 h. Data are mean ± SEM of five independent experiments. * P < 0.05 compared with the luciferase activity in TNF--treated cells.

Stimulation with TNF- markedly increased the luciferase activity in constructs containing sequences from 1,481 to 79 bp, which included AP-1 and NF-B binding sites (P < 0.05 versus the resting state). Neither the constructs pN78, from which the sequence from 1,481 to 79 bp was deleted, nor the negative control pLuc0, produced any response to TNF- stimulation. The construct pN130, which lacked the sequence from 1,481 to 131 bp, produced the strongest luciferase activity in TNF-- stimulated cells.

CAM significantly repressed TNF--induced IL-8 promoter activity in the construct pNAF (P < 0.05 versus TNF-). Although deletion of the sequence below 336 bp (constructs pN391 and pN335) did not result in significant repression of luciferase activity by CAM, the construct pN130, containing the AP-1 binding site, produced a significant decrease in TNF--induced IL-8 promoter activity (P < 0.05 versus TNF-). Importantly, the construct pN112, including the NF-B-like binding site but not the AP-1 site, did not produce inhibition of TNF--induced IL-8 promoter activity (P > 0.19 versus TNF-). The same was noted for resting or TNF--stimulated cells. The construct pN78 (78 to +44 bp) produced only weak promoter activity in CAM-treated cells.

Because the promoter region of the IL-8 gene contains three cis elements of gene transcription, the AP-1, NF-IL-6, and NF-B binding sites, we also examined the relative contribution of each element to the repression by CAM of IL-8 gene transcription. Mutation of the AP-1 binding site failed to repress the TNF--induced enhancement of luciferase activity (Figure 8). Mutation of the NF-B binding site abrogated the TNF--induced increase in luciferase activity. In contrast, mutation of the NF-IL-6 binding site did not change the repression by CAM of the TNF-- induced enhancement of luciferase activity.

View larger version (14K): [in this window] [in a new window]   Figure 8.   Effects of mutation of AP-1, NF-IL-6, and NF-B binding sites on promoter activity of the IL-8 gene. BET-1A cells were transfected with a pN130 luciferase expression vector containing mutations of the AP-1, NF-IL-6, or NF-B binding sites. The cells were then stimulated with TNF- (6 h) in the presence or absence of CAM (20 µg/ml for 96 h). Relative values were normalized to -galactosidase activity as an internal control, and were expressed as the percent of luciferase activity in resting cells. Data shown are mean ± SEM in quadruplicate assays from one representative experiment. Similar experiments were repeated two times; mut = mutant.

CAM Inhibited AP-1 Binding Activity in BET-1A Cells

To investigate the effects of CAM on protein-DNA complex formation with nuclear extracts from BET-1A cells, we performed EMSA, using specific oligonucleotide probes for the AP-1 and NF-B binding-site regions of the IL-8 gene. As shown in Figure 9, the nuclear extracts from resting cells showed weak AP-1 binding activity. Nuclear proteins extracted from TNF--stimulated BET-1A cells showed strong binding activity for the AP-1 oligonucleotide probe. Importantly, CAM downregulated the nuclear protein-AP-1 oligonucleotide complex formation induced by TNF-. This binding was specifically inhibited by an excess of unlabeled AP-1 oligonucleotide, but not by unlabeled SP-1 oligonucleotide. This suggests that the binding was specific to AP-1.

View larger version (46K): [in this window] [in a new window]   Figure 9.   CAM-inhibited AP-1 binding in BET-1A cells. A free, labeled AP-1 consensus oligonucleotide was used as a negative control (lane 1). Nuclear extracts from resting BET-1A cells (lane 2), cells with 1-h stimulation with TNF- alone (100 U/ml, lane 3), and cells preincubated with CAM (20 µg/ml for 96 h, lane 4) before TNF- stimulation were incubated with the labeled AP-1 oligonucleotide. For binding competition, nuclear extracts from TNF--treated BET-1A cells were incubated with the labeled AP-1 probe in the presence of the unlabeled AP-1 oligonucleotide (lane 5) or unlabeled SP-1 oligonucleotide (lane 6).

In addition, although the nuclear extracts from resting cells did not show any binding to the NF-B oligonucleotide probe, TNF- dramatically induced binding activity for the NF-B probe (Figure 10). Importantly, CAM had no effect on nuclear protein-NF-B oligonucleotide complex formation induced by TNF-. An excess of unlabeled NF-B oligonucleotide blocked this binding, but unlabeled SP-1 oligonucleotide did not.

View larger version (66K): [in this window] [in a new window]   Figure 10.   Lack of effect of CAM on NF-B binding in BET-1A cells. A free, labeled NF-B consensus oligonucleotide (lane 1) was used as a negative control. BET-1A cells were incubated with or without CAM (20 µg/ml) for 96 h and stimulated with TNF- (100 U/ml) for 1 h. Nuclear extracts from resting BET-1A cells (lane 2), TNF--treated cells (lane 3), and TNF--treated cells preincubated with CAM (lane 4) were incubated with the labeled NF-B oligonucleotide. Lanes 5 and 6 indicate results of coincubation with the unlabeled NF-B oligonucleotide or with unlabeled SP-1 oligonucleotide.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The main goal of our study was to clarify the molecular mechanism of IL-8 gene repression by macrolides. We showed that CAM, a synthetic macrolide, repressed TNF-- induced IL-8 gene expression and protein secretion in cells of the human bronchial epithelial cell line BET-1A. We confirmed these results in human primary bronchial epithelial cells (unpublished data). We also confirmed that CAM inhibited TNF--induced IL-8 gene expression in both a dose- and incubation time-dependent manner. Our reporter gene analyses and EMSA showed that the AP-1 binding site was one of the elements on which CAM acted suppresively in a 5'-flanking region of the IL-8 gene.

Repression of TNF--induced IL-8 gene expression by CAM occurs through distinct mechanisms for the following reasons. First, since CAM did not affect either the proliferation or viability of BET-1A cells, the inhibition of IL-8 gene expression by CAM was not caused by the cytotoxic effect of this drug. Second, TNF-RI mRNA transcript levels did not change with exposure to macrolides, suggesting that macrolides do not modulate TNF--induced IL-8 mRNA transcripts, at least at the TNF-RI gene expression level. Third, the addition of CAM also repressed IL-8 gene expression induced by other stimulants, such as phorbol 12-myristate 13-acetate or supernatant from the culture of Pseudomonas aeruginosa (unpublished data). Additionally, EM has been reported to repress H. influenzae endotoxin-induced release of IL-6 and IL-8 (17), and IL-1- induced IL-8 gene expression in human bronchial epithelial cells (9). Thus, the repression of IL-8 gene expression by macrolides is not restricted by TNF- stimulation.

Although IL-8 gene expression has been reported to be regulated at both the transcriptional and posttranscriptional levels (25, 46), CAM did not change the half-life of IL-8 mRNA transcripts. This suggests that repression of TNF--induced IL-8 gene expression by CAM is not modulated by changes in transcript stability.

The study data showed that stimulation by TNF- clearly activated the IL-8 promoter in BET-1A cells. This suggests that the TNF--induced IL-8 gene expression observed in the study was regulated mainly at a transcriptional level in BET-1A cells, in accord with our previous findings with cells of the bronchogenic carcinoma cell line HS24 (25). We also demonstrated remarkable induction of luciferase activity after TNF- stimulation with plasmids including the AP-1 or NF-B-like binding sites, but not with a plasmid lacking the NF-B-like binding site. These results accord well with the finding that both AP-1-DNA and NF-B-DNA complexes were markedly induced in response to stimulation with TNF- as demonstrated by EMSA. It has been reported that the IL-8 promoter can be activated by several stimuli in various kinds of cells, and that the responsive elements may differ according to the type of cells (46). For example, cooperative interaction of C/EBP (NF-IL-6) and NF-B is required for the response to IL-1 or hepatitis B virus X protein in a human fibrosarcoma cell line (50, 51). The same sequence is necessary for activation of the IL-8 promoter in Jurkat T cells (52). In epithelial cells, the AP-1 binding site has an important role in activation of the IL-8 promoter (25). In cells of the gastric cancer cell line MKN 45, TNF- and interferon- synergistically induce IL-8 promoter activity through the AP-1 and NF-B binding sequences (53). The same sequences are required for activation of the IL-8 promoter in response to Helicobacter pylori stimulation in MKN 45 cells (42), and paclitaxel stimulation in ovarian cancer cell lines (54). Thus, our finding in BET-1A cells that both AP-1 and NF-B are important factors for TNF--induced IL-8 gene transcription is quite consistent with previous findings.

In the present study we showed that CAM inhibited TNF--induced IL-8 gene transcription in BET-1A cells. CAM caused a significant decrease in TNF--induced IL-8 promoter activity in the construct pNAF (1,481 to +44 bp). The construct pN130 (130 to +44 bp), including the AP-1 binding site, produced a significant decrease in TNF--induced luciferase activity. In contrast, the further-deleted construct pN112 (112 to +44 bp), including an NF-B-like but not the AP-1 binding site, did not show significant inhibition by CAM. In addition, mutation of the AP-1 binding site in the promoter of the IL-8 gene abrogated the repression by CAM of TNF--induced luciferase activity. EMSA also revealed that CAM inhibited the TNF--induced binding of AP-1 to its proper sequence in the IL-8 gene, but that CAM had no apparent effect on TNF--induced NF-B binding. These findings suggested that the AP-1 binding site is crucial for the repression of TNF--induced activation of the IL-8 promoter. However, the NF-B binding site did not seem to mediate this repression in BET-1A cells. In the luciferase reporter gene assay, the constructs pN391 and pN335, containing sequences upstream of the AP-1 binding site (391 to 131 bp), did not show a significant decrease in IL-8 promoter activity. This suggests that there are other regulatory elements in the 5'-flanking region upstream of the the AP-1 binding site.

Several studies of IL-8 gene repression have been reported (55). FK506, an immunosuppressive agent, inhibited PMA- and ionomycin-induced IL-8 promoter activity through either AP-1 or NF-B binding sequences in Jurkat cells. However, EMSA showed that FK506 did not alter formation of the AP-1 binding complex (55). It has also been reported that glucocorticoids (GCs) have an inhibitory effect on IL-8 promoter activity, through either the glucocorticoid response element (GRE) in a fibrosarcoma cell line (56) or the NF-B binding site in a glioblastoma cell line (57). Interferon has been reported to inhibit IL-8 promoter activity through C/EBP (NF-IL-6) and NF-B-like sequences (58). Thus, the mechanism for IL-8 gene repression by CAM found in the present study has never previously been reported, and seems to be unique to macrolides.

The inhibition of IL-8 gene transcription by CAM required a long period of incubation, and the degree of inhibition of IL-8 gene expression seems to have been mild as compared with that produced by FK506 or GCs (55). This suggests that CAM itself cannot interact directly with the IL-8 promoter, but that the effects of CAM are mediated by unknown factors. We have therefore hypothesized that CAM induces either unknown regulatory factors or negative transcription factors to repress AP-1 binding and/ or its effect on the IL-8 promoter of BET-1A cells (59).

We have reported a novel mechanism of macrolide regulation of IL-8 gene transcription in human bronchial epithelial cells, in which repression of IL-8 gene expression by CAM, mediated by the AP-1 binding site, is a result of macrolide activity. Our findings provide a clue for designing new drugs for treating chronic airway inflammation such as that in chronic bronchitis, DPB, and bronchial asthma.