Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Asthma is characterized by airway inflammation and hyperresponsiveness. The most common medications used to treat asthma are anti-inflammatory glucocorticoids and bronchodilators such as short- (albuterol) or long-acting (formoterol, salmeterol) ₂-adrenergic receptor agonists (-agonists).

Glucocorticoids act by binding and activating the cytosolic glucocorticoid receptor (GR). Activated GR then undergoes nuclear translocation, and is generally considered to interact in either a cis- or trans-repressive manner to inhibit cytokine gene expression (1). -Agonists bind to the G_s-coupled ₂-adrenergic receptors, which in turn activate adenylyl cyclase and increase cytosolic 3':5' cyclic adenosine monophosphate (cAMP). By virtue of its ability to activate the cAMP-dependent protein kinase (PKA), cAMP mediates relaxation of airway smooth muscle (ASM) through multiple mechanisms that either reduce the level of intracellular Ca²⁺ or mitigate the effects of Ca²⁺-dependent signaling (2). Recent evidence, however, suggests that cAMP-elevating agents (including -agonists) regulate other important processes in ASM cells.

Several recent studies (3), have implicated cAMP as a powerful modulator of ASM cell synthetic function. Agents that increase intracellular cAMP inhibit the secretion of normal T cells expressed and secreted (RANTES) (4, 6), eotaxin (5), and granulocyte-macrophage colony stimulating factor (GM-CSF) (6) induced by tumor necrosis factor (TNF)- in ASM cells, but augment TNF--induced interleukin (IL)-6 (4) and IL-8 (3, 6) secretion. Despite the abundance of data demonstrating the effects of TNF- and cAMP on the production of numerous cytokines, chemokines, and inflammatory molecules in ASM, transcriptional regulation of these products in ASM remains largely unexplored.

In this study, we examine the regulation of IL-6 and RANTES gene transcription in ASM cells stimulated with TNF-, glucocorticoids, and various cAMP elevating agents. IL-6 and RANTES have been found in increased amounts in the bronchoalveolar lavage of asthmatics (7, 8), and both factors have been implicated in allergic inflammatory processes (9, 10). IL-6, a pleiotropic cytokine, is considered to have both pro- (11) and anti-inflammatory (12) actions in asthma, whereas RANTES is involved in the recruitment of inflammatory cells such as eosinophils into the asthmatic airway after allergen challenge (13).

Regulation of gene expression is controlled by the binding of several transcription factors to known consensus sequences within the promoter region. In the IL-6 promoter, these include cis-acting binding elements for GR, activator protein (AP)-1, cAMP response element binding protein (CREB), CCAAT enhancer-binding protein- (C/EBP-), and nuclear factor (NF)-B (14). The RANTES promoter contains several important response elements, including a CD28-responsive element (CD28RE), two sets of AP-1 binding sites, individual binding elements for signal transducer and activator of transcription (STAT) protein, nuclear factor of activated T cells (NF-AT), and C/EBP, as well as two NF-B binding elements. Interestingly, no GR elements have been identified in the RANTES promoter (15).

Using a series of site-directed mutations within the human IL-6 (16) and RANTES (17) promoters, we demonstrate that TNF- promotes IL-6 gene expression in ASM cells via an NF-B-dependent pathway, whereas RANTES gene expression is mediated via AP-1 and NF-AT. In addition, we demonstrate that although TNF--induced IL-6 secretion is only partially inhibited by dexamethasone, RANTES secretion is totally inhibited by dexamethasone, despite the lack of a GR repressor element in the RANTES promoter. By activating the CRE region of the IL-6 promoter, -agonists induce IL-6 secretion from ASM and augment TNF--induced IL-6 secretion, reflecting an additive effect of NF-B and CRE response elements. In contrast, the inhibition of TNF--induced RANTES secretion by -agonists is via an AP-1 independent pathway. These data identify transcriptional mechanisms mediating TNF- and RANTES secretion in ASM, and demonstrate that glucocorticoids and -agonists regulate distinct cis- and trans- acting elements in each gene to modulate expression.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

ASM Cell Culture

Human tracheae were obtained from lung transplant donors in accordance with procedures approved by the University of Pennsylvania Committee on Studies Involving Human Beings at the University of Pennsylvania (Philadelphia, PA). ASM cells were dissected, purified, and cultured in Ham's F12 medium supplemented with 10% fetal bovine serum, 100 U/ml penicillin, and 0.1 mg/ml streptomycin (Gibco BRL Life Technologies, Grand Island, NY) as described previously (18). Confluent ASM cells were growth-arrested by incubating the monolayers in Ham's F12 with 0.1% bovine serum albumin for 48 h. A minimum of three different cell lines was used for each experiment.

Unless otherwise specified, all chemicals used in this study were purchased from the Sigma Chemical Co. (St. Louis, MO). Formoterol, salmeterol, and albuterol (racemic mixtures) were provided by Sepracor (Marlborough, MA).

IL-6 and RANTES Promoter Constructs

The IL-6 promoter constructs were kindly provided by Dr. Oliver Eickelberg (Yale University, New Haven, CT) with permission from Dr. Shigeru Katamine (Nagasaki University, Nagasaki, Japan) (16). The 651-bp fragment of the human IL-6 gene promoter located directly upstream of the transcriptional start site was subcloned (5'-Kpn I, 3'-Xho I) into pGL3 basic luciferase reporter gene vector plasmid (Promega, Madison, WI) to give the parental pIL-6-luc 651 construct (Figure 1A). Within pIL6-luc 651, the following transcription factor consensus binding sites were inactivated by site-directed mutagenesis (numbers indicate position relative to start site): AP-1 (276, pIL-6-luc 651 AP-1); C/EBP- (146, pIL-6-luc 651 C/EBP-); NF-B (63, pIL-6-luc 651 NF-B); and both NF-B and C/EBP- (63 and 146, pIL-6-luc 651 NF-B C/EBP-), as described previously (16). These mutations are known to inactivate the described consensus sequences and alter transcription factor binding (19). In addition, a 5'-deletion mutant (pIL-6-luc160) was created by Klenow treatment and blunt-end ligation following Kpn I/Aat II digestion of pIL-6-luc 651, effectively removing the AP-1 (276) and CREB (154) binding sites. These constructs are shown schematically in Figure 1A.

View larger version (30K): [in this window] [in a new window]   View larger version (13K): [in this window] [in a new window]   Figure 1.   NF-B mediates TNF--induced IL-6 gene expression. (A) Schematic of IL-6 promoter constructs used in this study (from Ref. 16). (B, C) ASM cells were transfected as described in MATERIALS AND METHODS. Cells were growth-arrested, then treated for 4 h with either (B) vehicle (control) or (C) 10 ng/ml TNF- at 37°C. Cells were then harvested and luciferase and -galactosidase activities assessed. Basal luciferase activities (after correction for -galactosidase) for each mutated or truncated promoter construct, compared with the full-length promoter, are shown in (B). (C) Fold difference in response to TNF- compared with control (indicated by dotted line at 1), where statistical analysis was performed using the Student's unpaired t test (* denotes a significant difference in luciferase activity compared with the full-length IL-6 promoter [P < 0.05]). Data represent mean ± SE values from 9-14 replicates.

The RANTES promoter constructs were a generous gift of Prof. Hiroyuki Moriuchi (Nagasaki University School of Medicine, Nagasaki, Japan) (17). The 1.4-kb 5'-noncoding sequence upstream of the RANTES gene was cloned into pGL2 basic luciferase reporter gene vector plasmid (Promega) using Sma I and Kpn I sites. Within pGL2-RANTES-1.4, the following transcription factor consensus binding sites were inactivated by site- directed mutagenesis (numbers indicate position relative to start site): B2 (30, pGL2-RANTES-1.4 B2); B1 (44, pGL2-RANTES-1.4 B1); C/EBP (92, pGL2-RANTES-1.4 C/ EBP); NF-AT (213, pGL2-RANTES-1.4 NF-AT); STAT (248, pGL2-RANTES-1.4 STAT); AP-1/AP-1 (327, pGL2-RANTES-1.4 TRE3/4); AP-1/AP-1 (345, pGL2-RANTES-1.4 TRE1/2); all AP-1 sites (327 and 345, pGL2-RANTES-1.4 TRE1-4); and a CD28-responsive element (579, pGL2-RANTES-1.4 CD28RE), as described previously (17). These constructs are shown schematically in Figure 2A.

View larger version (45K): [in this window] [in a new window]   View larger version (19K): [in this window] [in a new window]   View larger version (17K): [in this window] [in a new window]   Figure 2.   NF-AT and AP-1 mediate TNF--induced RANTES gene expression. (A) Schematic of RANTES promoter constructs used in this study (from Ref. 17). (B, C) ASM cells were transfected as described in MATERIALS AND METHODS. Cells were growth-arrested, then treated for 16 h with either (B) vehicle (control) or (C) 10 ng/ml TNF- at 37°C. Cells were then harvested and luciferase and -galactosidase activities assessed. Basal luciferase activities (after correction for -galactosidase) for each mutated promoter construct, compared with the full-length promoter, are shown in (B). (C ) Fold difference in response to TNF- compared with control (indicated by dotted line at 1), where statistical analysis was performed using the Student's unpaired t test (* denotes a significant difference in luciferase activity compared with the full-length RANTES promoter [P < 0.05]). Data represent mean ± SE values from 4-10 replicates.

JM109 competent cells (Promega) were transformed with ~ 10-20 ng of each plasmid DNA, according to the manufacturer's instruction (Promega). Resulting transformants were propagated for preparation of plasmid DNA and purified using the QIAGEN EndoFree Plasmid Maxi Kit (QIAGEN, Valencia, CA). The fidelity of all clones was confirmed by dideoxy sequencing.

ASM Cell Transfection, Luciferase and -Galactosidase Assays

Transfection of ASM cells was performed as described previously (20) using the calcium phosphate transfection system (Gibco BRL Life Technologies). Briefly, ASM cells were plated onto 100-mm dishes at a density of ~ 1 × 10⁶ cells/dish for 24 h, then transfected with 5 µg of either IL-6 or RANTES construct, as well as 2.5 µg of pSV--galactosidase control vector (Promega) to normalize transfection efficiencies. Following transfection, cells were cultured for 36 h, then growth-arrested for 24 h in Ham's F12 medium supplemented with 0.1% bovine serum albumin. To examine the underlying transcriptional mechanism for TNF--induced IL-6 or RANTES gene expression, cells were then treated with either vehicle or 10 ng/ml TNF- (Boehringer Mannheim, Indianapolis, IN) and incubated at 37°C for 4 h (IL-6) or 16 h (RANTES). To examine effect of cAMP-elevating agents on IL-6 gene expression, ASM cells were stimulated with TNF- in the presence or absence of 1 µM prostaglandin (PG) E₂ (Calbiochem, La Jolla, CA), 1 µM isoproterenol, or 1 µM formoterol for 4 h at 37°C. Cells were then harvested and luciferase and -galactosidase activities assessed according to manufacturer's instructions. Results were expressed as fold difference compared with control.

Measurement of IL-6 and RANTES Secretion from ASM Cells

To examine the effect of dexamethasone on TNF--induced IL-6 and RANTES, growth-arrested ASM cells were pretreated with or without dexamethasone (0.0001-1 µM) for 1 h. Cells were then stimulated with 10 ng/ml TNF-. After 24 h treatment at 37°C, cell supernatants were removed and frozen at -20°C for later analysis by enzyme-linked immunosorbent assay (ELISA). To examine the effects of cAMP-elevating agents on IL-6 or RANTES secretion, cells were pretreated for 30 min with isoproterenol (0.0001-10 µM), formoterol (0.0001-10 µM), salmeterol (0.01-10 µM), or albuterol (0.01-10 µM), and either left unstimulated or stimulated with TNF- (10 ng/ml) for 24 h at 37°C. To determine the combined effect of dexamethasone and salmeterol, cells were pretreated for 1 h with either vehicle or dexamethasone (0.0001-1 µM), in the absence or presence or 10 µM salmeterol for 30 min, before stimulation with TNF- (10 ng/ml) for 24 h. ELISA for IL-6 and RANTES were performed according to the manufacturer's instructions (R&D Systems, Minneapolis, MN).

Electrophoretic Mobility Shift Assay

To examine the effect of dexamethasone on TNF--induced AP-1, TRE3/4, or TRE1/2 binding, growth-arrested ASM cells were pretreated for 1 h with either vehicle or 1 µM dexamethasone, and either left unstimulated or stimulated with TNF- (10 ng/ml) at 37°C, for the indicated times. AP-1 binding was assessed using consensus and mutant gel shift oligonucleotides (Santa Cruz Biotechnology, Santa Cruz, CA). The oligonucleotides used for TRE3/4 and TRE1/2 were previously described (17). To assess the effect of cAMP-elevating agents on AP-1-binding, growth-arrested ASM cells were pretreated for 30 min with vehicle (control), 1 µM PGE₂, 1 µM isoproterenol, or 1 µM formoterol and either left without further stimulation, or stimulated with TNF- (10 ng/ml), for 1 h at 37°C. Nuclear extracts were prepared, and electrophoretic mobility shift assay (EMSA) performed to assess DNA binding, using methods described previously (21).

Statistical Analysis

Data was analyzed using the Student's unpaired t test. P values < 0.05 were sufficient to reject the null hypothesis for all analyses.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

NF-B Mediates TNF--Induced IL-6 Gene Expression

Regulation of IL-6 gene expression is controlled by the binding of several transcription factors to known consensus sequences within the promoter region of the IL-6 gene (14). ASM cells were transiently transfected with full-length, mutated, or truncated constructs of the human IL-6 promoter (Figure 1A). Figure 1B shows the basal luciferase activity for each mutated or truncated promoter construct, compared with the full-length promoter. TNF- (10 ng/ml) increased IL-6 promoter activity by 244.4 ± 52.3% (mean ± standard error (SE) after 4 h (Figure 1C). To identify cis-regulatory sequences responsible for the upregulation of IL-6 expression by TNF-, we tested promoter constructs with site-directed mutations in the AP-1 (pIL-6-luc 651 AP-1), NF-B (pIL-6-luc 651 NF-B), or C/EBP- (pIL-6-luc 651 C/EBP-) binding sequences, as well as a double mutant of the NF-B and C/EBP- sequence (pIL-6-luc 651 NF-B C/EBP-). We also transfected pIL-6-luc160, a 5'-truncated version of the parental construct deficient in both AP-1 and CRE binding sequences. Mutations within the AP-1 or C/EBP- binding sites in the IL-6 promoter had no effect on luciferase activity (Figure 1C). Similarly, cells transfected with pIL-6-luc160 had luciferase activity comparable to that of the full-length construct, suggesting that CRE binding is not responsible for TNF--induced IL-6 gene expression. However, after transfection with pIL-6-luc 651 NF-B, or the pIL-6-luc 651 NF-B C/EBP- double mutant, TNF--induced luciferase activity was significantly decreased (Figure 1C) (P < 0.05). These results suggest that the binding sites for NF-B, but not AP-1, C/EBP-, or CRE, mediate TNF-- induced IL-6 gene expression in ASM cells.

NF-AT and AP-1 Mediate TNF--Induced RANTES Gene Expression

ASM cells were transiently transfected with 1.4 kb of the RANTES promoter (pGL2-RANTES-1.4: Figure 2A). Figure 2B shows the basal luciferase activity for each mutated or truncated promoter construct, compared with the full-length promoter. When stimulated (for 16 h) with TNF-, promoter activity increased by 1,375.9 ± 258.4% (Figure 2C). To identify cis-acting elements required for TNF-- induced RANTES activity, we transiently transfected ASM cells with a series of promoter constructs (shown schematically in Figure 2A), in which binding elements had been mutated individually, or in combination. Mutations of the two NF-B binding elements, B1 (pGL2-RANTES-1.4 B1) and B2 (pGL2-RANTES-1.4 B2), had no effect on TNF--induced RANTES promoter activity (Figure 2C). Mutations in the STAT (pGL2-RANTES-1.4 STAT) or CD28RE (pGL2-RANTES-1.4 CD28RE) binding element also had no effect on cytokine-induced luciferase activity, although a mutation in the C/EBP site (pGL2-RANTES-1.4 C/EBP) had a modest, though nonsignificant, effect on promoter activity after TNF- stimulation. In contrast, mutation of the NF-AT binding site (pGL2-RANTES-1.4 NF-AT) significantly reduced (73.3 ± 17.2%, n = 6, P < 0.05) TNF--induced promoter activity. Mutation of the TRE1-4 domain (pGL2-RANTES-1.4 TRE1-4), which removes both sets of AP-1 response elements, decreased promoter activity by ~ 80% (Figure 2C). To determine which AP-1 response elements in the TRE1-4 domain contribute to TNF--induced activation of the RANTES promoter, ASM cells were transfected with constructs in which either TRE1/2 or TRE3/4 were modified by site-directed mutagenesis. Mutation of the TRE1/2 site had no effect on induction by TNF-, whereas mutation at the TRE3/4 site significantly decreased luciferase activity by ~ 70% (Figure 2C) (P < 0.05). Thus, NF-AT and AP-1 binding sites significantly contribute to TNF--induced RANTES promoter activity.

Dexamethasone Only Partially Inhibits TNF--Induced IL-6 Protein, whereas TNF--Induced RANTES Is Completely Abrogated

As previously shown, ASM cells stimulated with TNF- (10 ng/ml) for 24 h secrete 4,939.2 ± 177.0 pg/ml IL-6 (Figure 3A) and 13,838.6 ± 552.8 pg/ml RANTES (Figure 3B) (4). When cells were pretreated for 1 h with increasing concentrations of dexamethasone, RANTES secretion was completely abrogated, with an IC₅₀ of 1 nM (Figure 3B). In comparison, even at the highest concentration of dexamethasone, TNF--induced IL-6 secretion was only partially inhibited (Figure 3A).

View larger version (18K): [in this window] [in a new window]   Figure 3.   Dexamethasone completely inhibits TNF--induced RANTES, but not IL-6, secretion from ASM cells. Growth-arrested ASM cells stimulated with TNF- (10 ng/ml) were pretreated for 1 h with dexamethasone (0.0001-1 µM), and compared with cells stimulated with TNF- alone. After 24 h, secreted (A) IL-6 or (B) RANTES was measured by ELISA. Statistical analysis was performed using the Student's unpaired t test, where * denotes a significant effect of dexamethasone pretreatment on TNF-- induced IL-6 or RANTES secretion, compared with secretion from cells stimulated with TNF- alone (P < 0.05). Data represent mean ± SE values from more than six replicates.

Dexamethasone Inhibits TNF--Induced RANTES Gene Expression via an AP-1-Dependent Pathway

As demonstrated above, the AP-1 binding element TRE3/4, but not TRE1/2, significantly contributes to RANTES promoter activity after stimulation with TNF-. To examine whether the effect of dexamethasone on TNF--induced RANTES protein is via inhibition of an AP-1-mediated pathway in ASM cells, we initially performed EMSA with the consensus AP-1 binding sequence. An inducible AP-1 binding was seen as early as 15 min following stimulation with TNF-, which peaked at 1 h, and was maintained at 2 h (Figure 4A). Pretreatment of ASM cells with dexamethasone for 1 h significantly inhibited AP-1 binding following stimulation of ASM cells with TNF- (Figure 4A). We then performed further EMSA using oligonucleotides for the TRE3/4 and TRE1/2 sites in the RANTES promoter. As shown in Figure 4B, the binding of TRE1/2 is unaffected by TNF- stimulation; however, TRE3/4 is induced by TNF- and inhibited by dexamethasone. These assays confirm the results of the reporter assays and suggest that dexamethasone inhibits TNF--induced RANTES gene expression via an AP-1-dependent pathway.

View larger version (98K): [in this window] [in a new window]   View larger version (109K): [in this window] [in a new window]   Figure 4.   Dexamethasone inhibits TNF--induced AP-1 binding in ASM cells, specifically at the TRE3/4 site. (A) Growth-arrested ASM cells were stimulated with 10 ng/ml TNF- for the indicated times at 37°C. Nuclear extracts were prepared and assayed by EMSA for AP-1 binding using consensus gel shift oligonucleotides. The effect of dexamethasone pretreatment (1 µM for 1 h) on TNF-- induced AP-1 DNA binding (after 1 h) was also assessed. The results for mutant AP-1 oligonucleotides are included as control. (B) To examine the effect of dexamethasone on TNF--induced TRE1/2 or TRE3/4 binding, growth-arrested ASM cells were pretreated for 1 h with either vehicle or 1 µM dexamethasone, and either left unstimulated or stimulated for 1 h with TNF- (10 ng/ml) at 37°C. Nuclear extracts were prepared and assayed by EMSA for TRE1/2 or TRE3/4 binding using oligonucleotides described previously (17). Data shown are representative results from three experiments.

-Agonists are Modulators of IL-6 and RANTES Secretion from ASM Cells

Although cAMP is known to modulate cytokine production in a number of cell types (reviewed in Ref. 22), only a few recent studies have examined the effects of cAMP on the synthetic function of human ASM cells (3). Recently, we reported that cAMP-elevating agents, such as PGE₂, forskolin, or dibutryl cAMP, induce ASM cells to synthesize and secrete IL-6. Conversely, the secretion of RANTES in ASM treated with TNF- is potently inhibited (4). Therefore, we compared the effects of a variety of -agonists, including albuterol (short-acting, low intrinsic activity at the ₂-adrenergic receptors), isoproterenol (short-acting, high intrinsic activity), formoterol (long-acting, high intrinsic activity), and salmeterol (long-acting, low intrinsic activity), on IL-6 and RANTES protein expression.

Formoterol induced a 6.0-fold increase in IL-6 secretion at 0.1 µM, with an IC₅₀ of 0.01 µM (Figure 5A). The IC₅₀ for isoproterenol was an order of magnitude lower (~ 0.1 µM), although at 1 µM, isoproterenol also induced a 5.7-fold increase in IL-6 secretion. Salmeterol and albuterol, at higher concentrations, induced a ~ 2-fold increase in IL-6 secretion (P < 0.05). All of the -agonists were comparable in their ability to augment TNF--induced IL-6 secretion (Figure 5B).

View larger version (23K): [in this window] [in a new window]   View larger version (15K): [in this window] [in a new window]   Figure 5.   -Agonists induce IL-6 secretion, augment TNF-- induced IL-6 secretion, and inhibit TNF--induced RANTES secretion from ASM cells. (A) Growth-arrested cells were stimulated for 24 h with isoproterenol (square; 0.0001-10 µM), formoterol (diamond; 0.0001-10 µM), salmeterol (triangle; 0.01-10 µM), or albuterol (circle; 0.01-10 µM). IL-6 was measured by ELISA and statistical analysis was performed using the Student's unpaired t test, where * denotes a significant effect of -agonist pretreatment on IL-6 secretion, compared with secretion from unstimulated cells (167.3 ± 16.9 pg/ml) (P < 0.05). Data represent mean ± SE values from more than three replicates. (B) Cells stimulated with TNF- (10 ng/ml for 24 h) were pretreated for 30 min with isoproterenol (1 µM), formoterol (1 µM), salmeterol (10 µM), and albuterol (10 µM), and compared with cells stimulated with TNF- alone. IL-6 was measured by ELISA and statistical analysis was performed using the Student's unpaired t test, where * denotes a significant effect of -agonist pretreatment on TNF--induced IL-6 secretion, compared with secretion from cells stimulated with TNF- alone (P < 0.05). Data represent mean ± SE values from more than three replicates. (C) ASM cells were pretreated for 30 min with increasing doses of isoproterenol (square; 0.0001-10 µM), formoterol (diamond; 0.0001- 10 µM), salmeterol (triangle; 0.01-10 µM), or albuterol (circle; 0.01-10 µM), then stimulated with TNF- (10 ng/ml for 24 h). RANTES was measured by ELISA and compared with cells stimulated with TNF- alone. Statistical analysis was performed using the Student's unpaired t test, where * denotes a significant effect of each -agonist pretreatment on TNF--induced RANTES secretion, compared with secretion from cells stimulated with TNF- alone (13,835.7 ± 552.8 pg/ml) (P < 0.05). Data represent mean ± SE values from 3-40 replicates.

In contrast, -agonists inhibited TNF--induced RANTES secretion in a concentration-dependent manner (Figure 5C). Formoterol was the most effective inhibitor, with an IC₅₀ of 1 nM, inhibiting secretion of TNF--induced RANTES secretion by ~ 70% at 10 nM. Isoproterenol was 10-fold less potent than formoterol, but at 100 nM could abrogate TNF--induced RANTES by ~ 80%. At a concentration of 10 µM, salmeterol and albuterol inhibited RANTES secretion following TNF- stimulation by 46.2% and 31.0%, respectively (Figure 5C). cAMP-elevating agents alone had no effect on RANTES secretion (data not shown). These results show that the rank order of effectiveness of these compounds in their ability to inhibit TNF--induced RANTES secretion is similar to their ability to induce secretion of IL-6 and increase cytosolic levels of cAMP in ASM (Ref. 23; R. B. Penn and J. L. Benovic, unpublished data). As originally suggested in our previous study (4), however, RANTES appears to be more sensitive than IL-6 to small changes in cAMP concentration.

Transcriptional Regulation of IL-6 and RANTES by cAMP-Elevating Agents

In this study, and in our previous reports (4), treatment of ASM cells with cAMP-elevating agents potently augmented the secretion of IL-6. To define the underlying transcriptional mechanisms that modulate cAMP-induced IL-6 gene expression, ASM cells were transiently transfected with a series of IL-6 promoter constructs described in Figure 1A. Elevation of cAMP in ASM cells was achieved either by stimulation of G_s-coupled receptors using PGE₂ (Figure 6A), by nonselective activation of the -adrenergic receptor G_s-adenylyl cyclase pathway with isoproterenol (Figure 6B), or by treatment with the selective, long-acting -agonist, formoterol (Figure 6C). We found that 1 µM PGE_2, 1 µM isoproterenol, or 1 µM formoterol increased IL-6 promoter activity by 53.3 ± 24.4%, 156.2 ± 59.0%, or 105.3 ± 43.0%, respectively (Figures 6A-6C). Interestingly, single mutations in the AP-1, C/EBP-, or NF-B binding sequences, or a double mutant of the C/EBP- and NF-B sequence, had no significant effect on PGE₂-, isoproterenol-, or formoterol-induced IL-6 promoter activity (Figures 6A-6C). However, in cells transfected with pIL-6-luc160, a 5' truncated version of the parental construct deficient in both AP-1 and CRE binding sequences, cAMP-induced IL-6 promoter activity was significantly inhibited (Figures 6A-6C) (P < 0.05). Taken together, these results suggest that binding to CRE, but not to AP-1, NF-B, or C/EBP, mediates PGE₂-, isoproterenol-, or formoterol-induced IL-6 gene expression in ASM, via a cAMP-mediated pathway.

View larger version (22K): [in this window] [in a new window]   View larger version (13K): [in this window] [in a new window]   Figure 6.   CRE mediates PGE2-, isoproterenol- and formoterol-induced IL-6 gene expression. ASM cells were transfected with a luciferase-tagged IL-6 promoter construct as described in MATERIALS AND METHODS. Cells were growth-arrested, then treated for 4 h with either vehicle (control, indicated by dotted line at 1), or (A) 1 µM PGE2, (B) 1 µM isoproterenol, or (C) 1 µM formoterol. Results were expressed as fold difference in luciferase activity compared with control. Statistical analysis was performed using the Student's unpaired t test, where * denotes a significant effect of a particular mutation compared with results obtained following stimulation of the parental construct pIL-6-luc 651 (P < 0.05). Data represent mean ± SE values from 3-11 replicates.

Our data suggest that RANTES secretion after TNF- stimulation involves activation of AP-1 (Figure 2C). To investigate whether cAMP-elevating agents mediate their inhibitory effects on TNF--induced RANTES secretion by attenuating AP-1 DNA binding, EMSAs were performed. Surprisingly, we found that PGE₂, formoterol, and, to a lesser extent, isoproterenol, increase AP-1 DNA binding activity (Figure 7). Following stimulation with TNF-, there were no clear changes in AP-1 binding (Figure 7). These data suggest that cAMP-elevating agents mediate their effects on RANTES via an AP-1-independent pathway.

View larger version (90K): [in this window] [in a new window]   Figure 7.   Effect of PGE2, isoproterenol, and formoterol on TNF-- induced AP-1 binding in ASM cells. Growth-arrested ASM cells were pretreated for 30 min with vehicle (control), PGE2 (1 µM), isoproterenol (1 µM), or formoterol (1 µM), and either left without further stimulation, or stimulated with TNF- (10 ng/ml), for 1 h at 37°C. Nuclear extracts were prepared and assayed for AP-1 binding by EMSA. Data shown are representative results.

Effect of the Combination of Glucocorticoids and a -Agonist on TNF--Induced RANTES and IL-6 Secretion

We next examined the effectiveness of the -agonist salmeterol used in combination with the glucocorticoid dexamethasone on TNF--induced RANTES and IL-6 secretion from ASM cells. As previously shown, dexamethasone completely (95.4%) inhibited TNF--induced RANTES secretion (Figure 8A). In combination with 10 µM salmeterol, the same degree of inhibition could be achieved with a 100-fold lower concentration of dexamethasone (i.e., 0.01 µM). In contrast, TNF--induced IL-6 secretion (3,749.8 ± 520 pg/ml) was significantly augmented in the presence of salmeterol (5,919.3 ± 477.0 pg/ml) (Figure 8B). Further, IL-6 secretion following TNF- stimulation could not be completely inhibited by dexamethasone treatment (957.9 ± 232.7 pg/ml remaining after treatment with 1 µM dexamethasone), although the percent inhibition by dexamethasone compared with control was similar in salmeterol-treated and -untreated cells.

View larger version (26K): [in this window] [in a new window]   Figure 8.   Additive inhibition of TNF--induced RANTES, but not IL-6, by dexamethasone and salmeterol. Growth-arrested ASM cells were pretreated for 1 h with either vehicle or dexamethasone (0.0001-1 µM), in the absence (solid bars) or presence (shaded bars) of 10 µM salmeterol for 30 min, then stimulated with TNF- (10 ng/ml). After 24 h incubation, secreted RANTES (A) or IL-6 (B) was measured by ELISA. Statistical analysis was performed using the Student's unpaired t test, where * denotes a significant effect of salmeterol (P < 0.05). Data represent mean ± SE values from three replicates.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In this paper we explore the molecular basis for the induction of IL-6 and RANTES by TNF- in ASM cells, and determine the underlying transcriptional mechanisms for the effect of glucocorticoids and -agonists, used alone or in combination on these cytokines. We provide the first conclusive demonstration that TNF--induced IL-6 gene expression in ASM cells occurs via an NF-B-dependent pathway that is only partially inhibited by dexamethasone. In contrast, TNF--induced RANTES gene expression is mediated via AP-1 and NF-AT and is completely inhibited by dexamethasone. Further, we show that AP-1 DNA binding was effectively repressed by glucocorticoids, suggesting a molecular mechanism for the effect of dexamethasone on RANTES secretion induced by TNF-.

The IL-6 promoter region contains a variety of DNA binding sites, including those for GR, AP-1, CREB, C/EBP-, and NF-B. NF-B activation represents a major pathway mediating TNF--induced IL-6 secretion (reviewed in Ref. 24). In previous studies from our laboratory, we showed that not only does TNF- stimulate NF-B DNA binding activity in ASM cells, but also increases mRNA expression and protein levels for IL-6 (4, 21). Our collective data suggested, therefore, that TNF--induced IL-6 secretion was associated with NF-B activation. In this study, we confirm our initial hypothesis and, using a series of site-directed mutations within the human IL-6 promoter (16), we demonstrate that TNF- promotes IL-6 gene expression in ASM cells via an NF-B-dependent pathway. These results contrast with the study of McKay and colleagues (25), in which induced expression and translocation of AP-1 was suggested to play a role in IL-6 gene expression in human ASM cells. Although no direct assessment of gene expression was performed in this study (25), the authors showed an association between TNF- stimulation and fos and jun expression. However, whether these early immediate genes regulated cytokine secretion was not tested.

Numerous studies have demonstrated that TNF-- induced transcription of RANTES is NF-B-dependent (17). However, stimulus-specific differences exist, along with a synergistic cooperation between NF-B and other transcription factors, such as AP-1 (26) and STAT (27). In addition, RANTES expression following TNF- stimulation has been shown to occur via regulation of mRNA stability, as well an increase in transcription (28). In this study, we use a series of site-directed mutations within the human RANTES promoter (17) to show that AP-1 and NF-AT regulate the promoter activity of RANTES and that, in fact, NF-B is not necessary.

Glucocorticoids bind and activate cytosolic GR. Activated GR undergoes nuclear translocation where it can then either act directly (cis) by binding to the glucocorticoid response element (GRE) of gene promoters, or indirectly (trans) by binding to other transcription factors, to alter gene transcription. Whether cis-acting glucocorticoids are positive or negative regulators of transcription appears to be cell-type specific (reviewed in Ref. 29). Trans-acting GR exert negative regulation of gene transcription by physically binding to other transcription factors. We found that dexamethasone only partially abolished TNF--induced IL-6 secretion from ASM cells. These findings corroborate our previous study showing that cytokine-induced NF-B DNA-binding activity was unaffected by dexamethasone, whereas ICAM-1 expression was only partially inhibited (30). One explanation for our current results is that dexamethasone may exert its inhibitory effects on TNF--induced IL-6 gene expression via cis-repression at the GRE of the IL-6 promoter. Alternatively, activated GR may interfere with NF-B p65 subunit trans-activation, as has been postulated (31). In contrast, the RANTES promoter does not contain a GRE region. We found that AP-1 is necessary for TNF-- induced RANTES secretion and that TNF--induced AP-1 DNA binding is attenuated by dexamethasone. In other cells (32), as well as in human lung (33), AP-1 binding is significantly repressed by glucocorticoid treatment. Collectively, these results suggest that glucocorticoids exert their inhibitory effect on TNF--induced RANTES expression by trans-repression of AP-1.

cAMP exerts both direct and indirect effects on gene expression, in a cell- and stimulus-specific manner (reviewed in Ref. 22). cAMP-triggered expression of numerous genes is largely mediated by CREB, which interacts with CRE in the promoters of cAMP-responsive genes to activate transcription. In addition, TNF--induced AP-1 binding has been shown to be both upregulated (34) and downregulated (35) in response to cAMP, again suggesting that the effects of increased cAMP are cell type-specific. In our study, we demonstrate that -agonists, including short- (albuterol) or long-acting (formoterol, salmeterol) compounds, and the nonselective -adrenergic receptor agonist isoproterenol, have potent immunomodulatory effects on ASM cells. The stimulatory effect of these cAMP-elevating agents on IL-6 appears to occur via activation of the CRE region of the IL-6 promoter, but not AP-1, NF-B, or C/EBP-. Moreover, -agonists augmented TNF-- induced IL-6 secretion. This may reflect an additive effect of NF-B and CRE response elements on IL-6 gene expression. Previous work from our laboratory (4), and from Hallsworth and colleagues (6), demonstrated the inhibitory effect of elevated cAMP on cytokine-induced RANTES secretion from ASM cells. Surprisingly, our data suggest that the effects of cAMP elevation are not mediated via inhibition of AP-1 DNA binding. These differences may be due to the fact that agents such as PGE₂ have multiple receptors that are coupled to different G proteins. Further studies will be necessary to elucidate this pathway.

This is the first study to examine the combined use of glucocorticoids and -agonists on the secretion of IL-6 and RANTES from cytokine-stimulated ASM cells, although studies examining other smooth muscle-derived chemokines have shown similar results (3, 5). For example, Pang and Knox demonstrated that glucocorticoids and -agonists each partially inhibited TNF--induced eotaxin secretion from ASM cells and had an additive effect in combination (5). Although Pang and Knox did not examine the underlying transcriptional mechanisms, it is interesting to note that the eotaxin promoter does not contain a CRE (36). Although the IL-8 promoter does contain a CRE region (37), treatment of ASM cells with -agonists alone only minimally induced IL-8 secretion, and -agonists had no effect on induction by TNF- (3). Interestingly, TNF--induced IL-8 secretion was synergistically inhibited by combined treatment with glucocorticoids and -agonists (3). In contrast, we found that TNF--induced IL-6 secretion remained slightly upregulated following treatment with both glucocorticoids and -agonists (3). With respect to RANTES, -agonists did not alter the virtually complete inhibition of TNF--induced RANTES expression that occurred at high glucocorticoid concentrations. However, -agonist treatment did render low concentrations of glucocorticoids more efficacious in inhibiting TNF-- induced RANTES expression, which might be interpreted as a "glucocorticoid-sparing effect" consistent with that observed in patients with asthma undergoing combination therapy with inhaled corticosteroids and -agonists. Collectively, these studies show that the effect of glucocorticoids and -agonists on ASM gene expression represent a complex interplay between the transcription factors controlling cytokine gene expression.