By repressing inflammatory gene expression, glucocorticoids are the most effective treatment for chronic inflammatory diseases such as asthma. However, in some patients with severe disease, or who smoke or suffer from chronic obstructive pulmonary disease, glucocorticoids are poorly effective. Although many investigators focus on defects in the repression of inflammatory gene expression, glucocorticoids also induce (transactivate) the expression of numerous genes to elicit anti-inflammatory effects. Using human bronchial epithelial (BEAS-2B) and pulmonary (A549) cells, we show that cytokines [tumor necrosis factor α (TNFα) and interleukin 1β], mitogens [fetal calf serum (FCS) and phorbol ester], cigarette smoke, and a Gq-linked G protein-coupled receptor agonist attenuate simple glucocorticoid response element (GRE)-dependent transcription. With TNFα and FCS, this effect was not overcome by increasing concentrations of dexamethasone, budesonide, or fluticasone propionate. Thus, the maximal ability of the glucocorticoid to promote GRE-dependent transcription was reduced, and this was shown additionally for the glucocorticoid-induced gene p57KIP2. The long-acting β2-adrenoceptor agonists (LABAs) formoterol fumarate and salmeterol xinafoate enhanced simple GRE-dependent transcription to a level that could not be achieved by glucocorticoid alone. In the presence of TNFα or FCS, which repressed glucocorticoid responsiveness, these LABAs restored glucocorticoid-dependent transcription to levels that were achieved by glucocorticoid alone. Given the existence of genes, such as p57KIP2, which may mediate anti-inflammatory actions of glucocorticoids, we propose that repression of transactivation represents a mechanism for glucocorticoid resistance and for understanding the clinical benefit of LABAs as an add-on therapy in asthma and chronic obstructive pulmonary disease.
Asthma is a chronic disease characterized by nonspecific airway hyperresponsiveness, reversible bronchoconstriction, airway remodeling, and inflammation, which may include both eosinophilic and, particularly in severe asthma, neutrophilic components (Jatakanon et al., 1999; Louis et al., 2000). Currently, the most effective treatments for all but the mildest asthmatic patients are glucocorticoids, which are taken in an inhaled form and are known clinically as inhaled corticosteroids (ICSs) (Barnes, 2006). By down-regulating inflammatory gene expression, ICSs reduce lung inflammation to improve symptoms and reduce exacerbation frequency (Barnes, 2006). Although ICSs control symptoms in the majority of asthmatic patients, 25 to 35% may show reduced or suboptimal responsiveness to therapy (Mjaanes et al., 2006; Zeiger et al., 2006; Martin et al., 2007). In this respect, it appears that the nature of the inflammation may play a determining role. Thus, glucocorticoid sensitivity is reduced in patients with severe disease who often show predominantly neutrophilic airway inflammation, whereas individuals with eosinophilic asthma are generally glucocorticoid responsive (Barnes and Adcock, 2009; Newton et al., 2010b). In addition, asthmatic patients who smoke or individuals who suffer from chronic obstructive pulmonary disease, a smoking-induced condition that is characterized by an irreversible decline in lung function, typically show poor responsiveness to glucocorticoids (Braganza et al., 2008). Although there are multiple reasons for this, which may include genetic factors, it is equally clear that proinflammatory stimuli reduce responsiveness to glucocorticoids (Barnes and Adcock, 2009; Newton et al., 2010b). Thus, there is an unmet need to understand and improve glucocorticoid efficacy in these patient groups.
Binding of glucocorticoid to the glucocorticoid receptor (GR) (NR3C1) induces translocation of the receptor from the cytoplasm to the nucleus, where GR modulates gene transcription in both positive and negative fashions (Newton and Holden, 2007). Thus, GR interacts with proinflammatory transcription factor complexes, such as nuclear factor κB or activator protein 1 (AP-1), to recruit histone deacetylases (HDACs) and decrease the transcription of inflammatory genes (an effect referred to as transrepression) (Barnes, 2006). In contrast, GR induces transcriptional responses from a variety of simple and complex glucocorticoid response elements (GREs) (Newton and Holden, 2007). Despite this, the key anti-inflammatory activities of glucocorticoids are attributed generally to direct transrepression by GR (Barnes, 2006). However, in the context of inflammation, proinflammatory cytokines, for example, tumor necrosis factor α (TNFα), may attenuate the ability of glucocorticoids to inhibit proinflammatory transcription factor activity by both increasing activation of c-Jun N-terminal kinase (JNK) and other mitogen-activated protein kinases (MAPKs) and reducing the expression of HDAC2 (Barnes and Adcock, 2009). In contrast, the repressive effect of glucocorticoids acting on core inflammatory genes, including cyclooxygenase 2, interleukin (IL)-6, IL-8, and granulocyte/macrophage colony-stimulating factor, can be blocked by inhibitors of transcription or translation (Newton and Holden, 2007; Newton et al., 2010a). Such data imply that glucocorticoid-dependent repression of inflammatory genes involves the expression of glucocorticoid-induced genes (Newton and Holden, 2007). Certainly, glucocorticoids induce the expression of many hundreds of target genes, and many of these display properties that are consistent with anti-inflammatory effects (Newton and Holden, 2007). Thus MAPK phosphatase 1, also called dual-specificity phosphatase 1, is induced profoundly and rapidly by glucocorticoids to switch off MAPK pathways, reduce inflammatory gene transcription, and destabilize inflammatory gene mRNAs (Issa et al., 2007; Clark et al., 2008; Diefenbacher et al., 2008; King et al., 2009; Newton et al., 2010a). Likewise, glucocorticoid-induced leucine zipper (GILZ) (TSC22D3) mRNA and protein are induced highly by glucocorticoids and act to reduce transcriptional responses via factors including AP-1 and nuclear factor κB (Mittelstadt and Ashwell, 2001; Eddleston et al., 2007). Another protein with putative anti-inflammatory properties is cyclin-dependent kinase inhibitor 1c, also known as p57KIP2, which regulates cell cycle progression and can switch off the JNK/MAPK pathway (Samuelsson et al., 1999; Chang et al., 2003).
Taken together, it is increasingly apparent that transactivation by GR contributes toward the anti-inflammatory activity of glucocorticoids. Indeed, although some patterns of steroid resistance may reflect altered GR translocation, a number of patients show changes in histone modification that suggest a failure of GR transactivation rather than transrepression (Matthews et al., 2004). Given this, it is plausible that transactivation of anti-inflammatory genes also may be the subject of glucocorticoid resistance. In the current study, our objective was to determine whether glucocorticoid-induced gene expression was reduced by inflammatory stimuli. Consequently, we have investigated the effects of various stimuli on the ability of glucocorticoids to induce transcription from a simple GRE-dependent reporter and on the expression of two glucocorticoid-inducible genes, p57KIP2 and GILZ, that have anti-inflammatory potential.
Materials and Methods
Cell Culture and Drugs.
Type II epithelial (A549) and bronchial epithelial (BEAS-2B) cells were cultured to confluence in Dulbecco's modified Eagle's medium or Dulbecco's modified Eagle's medium/F12 (Invitrogen, Carlsbad, CA) with 14 mM NaHCO3 added, respectively, supplemented with 10% fetal calf serum (FCS) (Canadian sourced; Invitrogen) and 2 mM l-glutamine. Before all of the experiments, cells were incubated overnight in serum-free medium before replacing with fresh serum-free medium containing drugs and cytokines. Dexamethasone was dissolved in Hanks' balanced salt solution (Sigma-Aldrich, St. Louis, MO). Budesonide, formoterol furoate (formoterol), fluticasone propionate, salmeterol xinafoate (salmeterol) (Sigma-Aldrich), 9,11-dideoxy-11a,9a-epoxymethano-prostaglandin F2α (U46619) (Cayman Chemical, Ann Arbor, MI), and phorbol 12-myristate 13-acetate (PMA) (Enzo Life Sciences, Farmingdale, NY) were dissolved in dimethyl sulfoxide (DMSO) (Sigma-Aldrich) (final DMSO concentration <0.1%). Recombinant human IL-1β and TNFα (R&D Systems, Minneapolis, MN) were dissolved in phosphate-buffered saline containing 0.1% bovine serum albumin. H2O2 (Sigma-Aldrich) and FCS (Invitrogen) were diluted in serum-free medium before experiments. Cigarette smoke extract-conditioned medium (CSE) was prepared by bubbling the smoke from one research-grade cigarette (3R4F 12/2006; University of Kentucky, Lexington, KY) through 4 ml of serum-free medium (Hudy et al., 2010). Optical density at 320 nm (OD320) was measured, and fresh medium added to give an initial OD320 of 1.
Reporter Constructs and Luciferase Assay.
A549 and BEAS-2B cells stably transfected with the 2×GRE-containing luciferase reporter, pGL3.neo.TATA.2GRE (Chivers et al., 2006), were cultured to confluence in 24-well plates. Luciferase assays were performed according to the manufacturer's guidelines (Firefly Luciferase Assay Kit; Biotium, Hayward, CA).
Cell Viability Assay.
Cells in 24-well plates were analyzed for cell viability using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. After aspiration, MTT solution (1 mg/ml in phosphate-buffered saline) (200 μl/well) was added for 30 min at 37°C. Wells were aspirated, and DMSO (200 μl) was added before the measurement of absorbance at 584 nm.
Confluent BEAS-2B cells were loaded for 2 h with Fluo-4 (Fluo-4 NW Calcium Assay Kit; Invitrogen) at 37°C, according to the manufacturer's guidelines. With excitation at 494 nm, basal fluorescence (F0) was measured at 516 nm for 1.6 s in a FLUOstar OPTIMA plate reader (BMG Labtech GmbH, Offenburg, Germany). The thromboxane receptor agonist U46619 (Cayman Chemical) was injected, and fluorescence was measured for a further 48.4 s. Values were expressed as F/F0, which is a proxy for intracellular calcium concentration ([Ca2+]i).
RNA Isolation and Reverse Transcriptase Polymerase Chain Reaction.
Total RNA was extracted from cells in 12-well plates using the RNeasy Mini Kit (QIAGEN, Valencia, CA). cDNA was prepared from 0.5 μg of RNA using a qScript cDNA synthesis kit (Quanta, Gaithersburg, MD) before being diluted 1:5 in DNase-free water. Real-time polymerase chain reaction (PCR) was performed using a 7900HT instrument (Applied Biosystems Inc., Foster City, CA) on 2.5 μl of cDNA in a reaction volume of 10 μl using SYBR Green chemistry (Invitrogen). Amplification conditions of 50°C, 2 min; 95°C, 10 min; then 40 cycles of 95°C, 15 s; 60°C, 1 min were used before dissociation (melt curve) analysis to confirm primer specificity (King et al., 2009). PCR primers were: GILZ (TSC22D3), forward 5′-GGC CAT AGA CAA CAA GAT CG-3′, reverse 5′-ACT TAC ACC GCA GAA CCA CCA-3′; cyclin-dependent kinase inhibitor 1c/p57KIP2, forward 5′-CGG CGA TCA AGA AGC TGT C-3′, reverse 5′-GGC TCT AAA TTG GCT CAC CG-3′; and glyceraldehyde 3-phosphate dehydrogenase, forward 5′-TTC ACC ACC ATG GAG AAG GC-3′, reverse 5′-AGG AGG CAT TGC TGA TGA TCT-3′.
Western Blot Analysis.
Cells were lysed in 1× Laemmli buffer supplemented with 1× complete protease inhibitor cocktail (Roche Diagnostics, Indianapolis, IN) and phosphatase inhibitors (Sigma-Aldrich). Total cell lysates were size-fractionated by Tris-glycine SDS polyacrylamide gel electrophoresis on 10 or 15% gels before electrotransfer to Hybond ECL membranes (GE Healthcare, Chalfont St. Giles, Buckinghamshire, UK). Membranes were blocked in 5% milk/phosphate-buffered saline before being probed with GILZ antiserum (provided by Bruce Zuraw) (Eddleston et al., 2007), cyclin-dependent kinase inhibitor 1c/p57KIP2 (2557; Cell Signaling Technology, Danvers, MA), or glyceraldehyde 3-phosphate dehydrogenase (4699–9555(ST); AbD Serotec, Raleigh, NC). After incubation with horseradish peroxidase-conjugated anti-rabbit or anti-mouse Ig, immune complexes were detected using Pierce ECL Western Blotting Substrate (Thermo Fisher Scientific, Waltham, MA). Densitometry was performed on Western blot images acquired below the saturation level of the film, using TotalLab software (Nonlinear Dynamics, Newcastle upon Tyne, UK).
Data are presented as means ± S.E. Statistical analysis was performed in GraphPad Prism, version 5.03 (GraphPad Software, San Diego, CA), using repeated measures one-way analysis of variance (ANOVA) with Bonferroni's or Dunnett's post-tests or paired t tests as appropriate.
Effects of TNFα and FCS on Simple GRE-Dependent Transcription.
BEAS-2B cells that had been transfected stably with the 2×GRE luciferase reporter, pGL3.neo.TATA.2GRE were treated with 10 ng/ml TNFα for various periods, both before and after stimulation with a maximally effective concentration (1 μM) of the synthetic glucocorticoid dexamethasone (Kaur et al., 2008; King et al., 2009) (Fig. 1A). Dexamethasone induced simple GRE-dependent transcription, and this was significantly and maximally reduced to 54% of the dexamethasone-induced response with a 1-h preincubation with TNFα (Fig. 1A). Near maximal inhibition was observed with 2- and 0.5-h preincubations with TNFα (Fig. 1A). This effect was concentration-dependent, with maximal repression being achieved at a TNFα concentration of 10 ng/ml (Fig. 1B). To examine whether the effect was generic or simply specific to BEAS-2B cells, the effects of both TNFα and IL-1β were tested on the pGL3.neo.TATA.2GRE reporter in A549 type II pulmonary cells (Fig. 1, C and D). In each case, dexamethasone-induced luciferase activity was repressed significantly and progressively by increasing concentrations of TNFα or IL-1β.
Serum is an established mitogen and activates many kinase cascades including MAPKs. BEAS-2B cells harboring the 2×GRE reporter therefore were treated with 10% FCS for various times before or after stimulation with 1 μM dexamethasone (Fig. 2A). As with TNFα, preincubation with 10% FCS for 1 and 2 h significantly reduced 2×GRE reporter activity. This was not apparent with a 6-h preincubation but was again apparent after an 18-h preincubation with FCS. With a 1-h preincubation, the effect of FCS was found to be concentration dependent and occurred in both BEAS-2B and A549 cells (Fig. 2, B and C). Because FCS is a complex mixture of growth factors, hormones, and other biological entities, BEAS-2B cells were treated with epidermal growth factor to determine whether this alone was able to induce glucocorticoid resistance. However, despite activation of the JNK pathway being induced, there was no obvious effect of epidermal growth factor on simple 2×GRE-dependent transcription induced by dexamethasone (Supplemental Fig. 1).
Because proinflammatory cytokines, such as TNFα, may induce different signaling pathways to FCS, we tested whether these two agents could combine to yield a greater overall repression (Fig. 2D). As above, 1-h preincubation with TNFα resulted in a concentration-dependent reduction in dexamethasone-induced GRE-dependent transcription. Likewise, 10% FCS produced a 40% drop in dexamethasone-induced luciferase activity, which in the presence of TNFα further declined to the level that was achieved by the highest concentrations of TNFα. These data do not support the concept of separate pathways of repression but suggest that both TNFα and FCS may act on a common pathway, which once maximally activated cannot produce greater repression of GRE-dependent transcription.
It is possible that increasing the concentration of glucocorticoid may overcome the repressive effect that is exerted by TNFα or FCS. Therefore, BEAS-2B 2×GRE reporter cells were pretreated with either TNFα (10 ng/ml) or FCS (10%) before stimulation with various concentrations of dexamethasone. Dexamethasone produced maximal activation of the reporter at a concentration of approximately 1 μM with EC50 values in the 26 to 50 nM range (Fig. 3A; Supplemental Table 1). Although this is consistent with previous data (Kaur et al., 2008), the presence of TNFα or FCS significantly reduced the maximal response to dexamethasone, without significantly altering the EC50 value (Fig. 3A; Supplemental Table 1). Qualitatively similar data were obtained with respect to budesonide and fluticasone propionate (Fig. 3A; Supplemental Table 1). Likewise, the repression of GRE-dependent transcription that was exerted by TNFα, IL-1β, and FCS in A549 cells also was not overcome by elevated concentrations of dexamethasone (Fig. 3). However, both TNFα and FCS modestly increased (by 35 and 28%, respectively) the EC50 values for dexamethasone (Fig. 3; Supplemental Table 1). Thus, the repressive effect that occurs with all of the glucocorticoids is common to at least two different cell lines and is not surmountable with increased concentrations of glucocorticoid.
Effect of Protein Kinase C Activation and Signaling from a Gq-Linked G Protein-Coupled Receptor.
Mediators, including histamine, leukotrienes, and prostanoids such as prostaglandin D2 and thromboxane, are released from multiple cell types and can elicit bronchospasm (Barnes et al., 1998). Such agonists act on G protein-coupled receptors that couple via the G protein Gqα to phospholipase C. This results in hydrolysis of phosphatidylinositol (4,5)-bisphosphate to yield inositol (1,4,5)-trisphosphate and diacylglycerol, which promotes intracellular calcium release ([Ca2+]i) and activates protein kinase C, respectively. BEAS-2B 2×GRE reporter cells therefore were pretreated with a maximally effective concentration (100 nM) of the protein kinase C activator PMA (Fig. 4A). Both 1- and 2-h pretreatments with PMA resulted in significant reductions in dexamethasone-induced 2×GRE reporter activity. This effect was concentration dependent, with the maximum observed repression being achieved at 1 μM PMA (Fig. 4B).
To test whether a bona fide G protein-coupled receptor agonist could elicit a similar effect, BEAS-2B cells were treated with various concentrations of the thromboxane analog U46619. Effects on [Ca2+]i were monitored using the calcium-sensing dye Fluo-4, which revealed a maximal response in the presence of 1 μM U46619 that occurred within seconds of adding the drug (Fig. 4C). However, 1-h pretreatment of BEAS-2B cells harboring the 2×GRE reporter revealed no effect of U46619 on dexamethasone-induced luciferase activity (Fig. 4D). In contrast, preincubations of 2, 6, and 18 h with U46619 produced significant repression of GRE-dependent transcription induced by dexamethasone (Fig. 4E). Likewise, the acetylcholine receptor agonist carbachol robustly induced [Ca2+]i in BEAS-2B cells with a maximally effective concentration of 100 μM (Supplemental Fig. 2A). However, no significant effects on dexamethasone-induced GRE-dependent transcription were elicited at any time tested (Supplemental Fig. 2B).
Effects of CSE and H2O2 on Simple GRE-Dependent Transcription.
CSE, diluted to different OD320 values in serum-free medium, was added to both BEAS-2B and A549 cells harboring the 2×GRE reporter (Fig. 5, A and B). In each case, dexamethasone-induced reporter activity was reduced significantly in a concentration-dependent fashion. As determined by the MTT assay, these treatments resulted in no significant effect on cell viability at CSE concentrations of OD 0.3 or below in BEAS-2B cells but produced minimal effects at OD 0.3 in A549 cells and some loss of cell viability at the highest concentration tested (OD 1.0) in both cell lines. Thus, although high concentrations of CSE are clearly toxic to cells, the ability to inhibit GRE-dependent transcription occurs at lower, nontoxic concentrations of CSE. In contrast, whereas the addition of H2O2 to BEAS-2B cells also resulted in significant repression of dexamethasone-induced reporter activity, this closely correlated with a substantial loss in cell viability (Fig. 5C). Thus, any repressive effects simply may be secondary to cell death.
Rescue of Repression by Long-Acting β2-Adrenoceptor Agonists.
In asthmatic patients whose symptoms are controlled poorly by inhaled glucocorticoids, the addition of a LABA is recommended as a step-up treatment option (Barnes, 2006). LABAs enhance the anti-inflammatory effect of the inhaled glucocorticoid to a level that cannot be achieved by the glucocorticoid alone (Giembycz et al., 2008; Newton et al., 2010b). Accordingly, we assessed the effects of LABAs on TNFα- and FCS-induced resistance to glucocorticoids. BEAS-2B cells harboring the simple 2×GRE reporter were pretreated with TNFα or FCS before stimulation with dexamethasone in the absence or presence of the LABAs salmeterol and formoterol (Fig. 6). As described under Effects of CSE and H2O2 on Simple GRE-Dependent Transcription, dexamethasone induced 2×GRE activity, and this was reduced significantly by TNFα or FCS preincubation (Fig. 6; Tables 1 and 2). In the presence of dexamethasone, increasing concentrations of salmeterol or formoterol progressively increased reporter activity to approximately three times the level achieved by dexamethasone alone. In the presence of either FCS or TNFα, which significantly repressed dexamethasone-induced 2×GRE activity, both salmeterol and formoterol increased dexamethasone-induced 2×GRE activity back to or slightly above the level that was achieved by dexamethasone alone (Fig. 6; Tables 1 and 2). Thus, LABAs restored the level of GRE-dependent transcription to that seen in the absence of proinflammatory or mitogenic stimuli. In each case, the sensitivity to the LABAs was unaltered, with EC50 values for the enhancement of 2×GRE-dependent transcription in the absence or presence of 10% FCS of: salmeterol, 5.99 ± 1.21 × 10−10 M (no FCS) and 7.69 ± 1.73 × 10−10 M (+FCS); formoterol 7.30 ± 1.21 × 10−11 M (no FCS) and 8.32 ± 1.86 × 10−11 M (+FCS). Likewise, EC50 values for the enhancement of 2×GRE-dependent transcription in the absence and presence of TNFα (10 ng/ml) were: salmeterol, 9.58 ± 1.20 × 10−10 (no TNFα), 8.54 ± 1.22 × 10−10 (+TNFα); formoterol, 6.06 ± 1.16 × 10−11 (no TNFα), 5.93 ± 1.56 × 10−11 (+TNFα). Despite the unaltered sensitivity, it is apparent that the overall enhancement of the response to dexamethasone that was achieved by the LABAs in the presence of FCS or TNFα was less than that achieved in the absence of FCS or TNFα (Tables 1 and 2).
Effects of Proinflammatory Stimuli on the Expression of p57KIP2 and GILZ.
A549 cells were treated with various concentrations of dexamethasone in the presence or absence of IL-1β (1 ng/ml). Total RNA was harvested after 6 h, and the expression of p57KIP2 and GILZ was examined by real-time PCR. As described previously (Kaur et al., 2008), both genes were induced highly by dexamethasone (Fig. 7A). Alone IL-1β had no significant effect on the expression of either gene. However, the presence of IL-1β dramatically reduced the p57KIP2 expression that was induced by dexamethasone. Like the simple GRE reporter, IL-1β did not affect sensitivity to dexamethasone (EC50 value without IL-1β = 1.45 ± 2.64 × 10−8 M; EC50 value with IL-1β = 2.17 ± 1.93 × 10−8 M), but the maximal response (efficacy) was attenuated significantly (Fig. 7A). In contrast, there was no obvious effect of IL-1β on the induction of GILZ mRNA by dexamethasone. To explore the generality of this result, the effect of TNFα was tested on the dexamethasone-induced mRNA expression of p57KIP2 and GILZ in BEAS-2B cells (Fig. 7B). As in the A549 cells, mRNAs for both genes were induced highly by dexamethasone. However, whereas dexamethasone-induced p57KIP2 again was repressed markedly by TNFα, there was no significant effect on GILZ expression. In the presence of formoterol, which had no effect alone, the ability of dexamethasone to induce p57KIP2 mRNA was potentiated significantly (Fig. 7B). Likewise, in the presence of TNFα, formoterol restored the dexamethasone-induced level of p57KIP2 mRNA to a level that was similar to that observed in the absence of TNFα (Fig. 7B). Thus, formoterol functionally restored this aspect of the response to glucocorticoids that was attenuated previously by TNFα. Despite this, it is salient to note that the induction of p57KIP2 mRNA by dexamethasone plus formoterol nevertheless was reduced significantly by the presence of TNFα. GILZ expression was not altered significantly in the presence of formoterol. These findings were corroborated by Western blot analysis for both p57KIP2 and GILZ proteins (Fig. 7C).
Numerous cytokines including IL-1, IL-2, IL-4, TNFα, and interferon γ are described as reducing various aspects of GR activity (Kam et al., 1993; Sher et al., 1994; Pariante et al., 1999; Irusen et al., 2002; Tliba et al., 2008). For example, in peripheral blood mononuclear cells, IL-2 plus IL-4 reduces the affinity of glucocorticoid for GR (Kam et al., 1993; Irusen et al., 2002). Likewise, a combination of TNFα plus interferon γ reduces the response to glucocorticoids in human airway smooth muscle cells (Tliba et al., 2008). Similarly, IL-1α may reduce GR translocation and transcriptional activation in the mouse fibroblast line L929 (Pariante et al., 1999). With respect to IL-2, this effect may involve the mitogen-activated protein kinase kinase/extracellular signal-regulated kinase pathway (Tsitoura and Rothman, 2004). In peripheral blood mononuclear cells from steroid-resistant patients, c-fos (FBJ murine osteosarcoma viral oncogene homolog) expression is reported to be up-regulated, leading to increased AP-1 activity and reduced GR activity (Lane et al., 1998). More recently, many authors have focused on how these changes may relate to the ability of GR to repress the expression of inflammatory genes. Thus, the recruitment of HDAC2 to the promoters of inflammatory genes forms a key component of the transrepressive ability of GR (Barnes, 2006; Barnes and Adcock, 2009). Furthermore, reduced HDAC2 expression, possibly due to enhanced degradation (Adenuga et al., 2009), may explain the glucocorticoid resistance that is observed in chronic obstructive pulmonary disease (Ito et al., 2005). Equally, smoke inhalation in mice or cells treated with CSE show reduced HDAC2 activity, and this is consistent with reports of reduced repressive abilities of glucocorticoids in alveolar macrophages from chronic obstructive pulmonary disease patients or after treatment with CSE (Culpitt et al., 2003; Adenuga et al., 2009; Marwick et al., 2009).
However, the repression of inflammatory gene expression by glucocorticoids also involves the up-regulation of anti-inflammatory genes (Newton and Holden, 2007; Clark et al., 2008). Therefore, we examined a number of common stimuli, including proinflammatory cytokines, mitogens, such as FCS and PMA, as well as CSE and oxidative stress, on the ability of glucocorticoids to induce transcriptional responses. We found significant reductions in the levels of simple GRE-dependent transcription after pretreatment with IL-1β, TNFα, PMA, FCS, and CSE. These effects were observed in two models of lung epithelial cells, suggesting that the current data are generally applicable. Indeed, similar responses using GRE reporters also are reported in human airway smooth muscle cells and mouse fibroblasts (Pariante et al., 1999; Tliba et al., 2008). Furthermore, although a modest increase in the EC50 value to glucocorticoid was found after TNFα and FCS treatment in A549 cells, the biological significance of this is questionable. More importantly, we show that the reduced responses described with the simple GRE reporter also apply to the bona fide glucocorticoid-inducible gene p57KIP2. Given biological properties, including the repression of cell proliferation and the attenuation of JNK signaling (Samuelsson et al., 1999; Chang et al., 2003), which are consistent with the effects of glucocorticoids, it is not unreasonable to predict that loss of glucocorticoid-inducible gene expression may affect functional responses. In certain patients, this may contribute to reduced responses to ICS therapy.
More importantly, our data predict that simply increasing the dose of glucocorticoid will not overcome this resistance, because the maximal inducibilities of the GRE reporter and of p57KIP2 expression were repressed. This correlates with the clinical situation, where increasing the concentration of the ICS is only of modest benefit. Thus, clinical guidelines (http://www.ginasthma.com) recommend the addition of a LABA to an ICS rather than simply increasing the dose of an ICS (Giembycz et al., 2008). Because LABAs can enhance GR-dependent gene transcription to levels that cannot be achieved by glucocorticoid alone (Kaur et al., 2008), we were prompted to explore the effects of LABAs in the context of induced reductions in steroid responsiveness. We now show that, whereas TNFα and FCS repress dexamethasone-dependent transcription, the clinically relevant LABAs formoterol or salmeterol restored GRE-dependent transcription back to the level that was achieved by dexamethasone alone (in the absence of an inflammatory stimulus). Although there was some evidence that the presence of TNFα or FCS also reduced the level of simple GRE enhancement produced by the LABAs, this still represents a significant improvement in glucocorticoid action. More importantly, the effect was confirmed at the level of p57KIP2 expression and therefore offers considerable mechanistic insight as to how LABAs could, by increasing the expression of otherwise repressed glucocorticoid-induced genes, act to overcome glucocorticoid resistance in patients. Thus, concentration-dependent increases in glucocorticoid-dependent transcription that are produced by LABAs may contribute to both enhanced protection from severe exacerbations and steroid-sparing in patients whose ICS + LABA combination therapy is intensified, as needed, on top of regular maintenance therapy. Such advantages were shown using a budesonide/formoterol combination inhaler (Symbicort) for Maintenance and Reliever Therapy (Symbicort SMART), which was found to produce the greatest therapeutic benefit in patients with severe asthma and with a high risk of exacerbation (Peters, 2009). Notwithstanding such effects, the fact that expression of p57KIP2, induced by dexamethasone plus formoterol, was significantly reduced by inflammatory stimuli also indicates that there may be considerable further ability to improve the therapeutic potential of ICS + LABA combination therapies.
Although the mechanisms underlying the repression of glucocorticoid-induced transcription were not addressed specifically in the current study, other investigators have variously indicated reduced binding affinity of ligand for GR, reduced expression and/or translocation of GR, as well as reduced ability to interact with AP-1, phosphorylation of GR, or the induction of GRβ, as potential mechanisms to account for resistance (Adcock et al., 1995ab; Webster et al., 2001; Irusen et al., 2002; Matthews et al., 2004). These possibilities require investigation in the context of the current findings. Despite this, the current study is important, because it provides a valid alternative to effects on transrepression in explaining steroid resistance. Thus, reduced expression of key glucocorticoid-induced effector genes may lead to clinically relevant attenuation of glucocorticoid efficacy. Likewise, the ability to enhance the expression of these same genes with a LABA, in the context of a glucocorticoid, may account for the enhanced clinical benefit of combining a LABA with a regular ICS. An important point to come from these data is that it should be possible to indentify inhibitors of signaling pathways that prevent the repression of GR-dependent gene expression by inflammatory stimuli such as TNFα or possibly cigarette smoke. The data reported in the current study should encourage the continued analysis of glucocorticoid-inducible genes as well as the pathways that may lead to reductions in their expression in the context of inflammation.
Participated in research design: Rider, Giembycz, and Newton.
Conducted experiments: Rider, King, Holden, and Newton.
Performed data analysis: Rider, King, Holden, and Newton.
Wrote or contributed to the writing of the manuscript: Rider, King, Holden, Giembycz, and Newton.
We are grateful to David Proud and Suzanne Traves for their assistance in setting up the CSE protocols and Bruce Zuraw and Jane Eddleston for the kind gift of the GILZ antibody.
This research was supported by a grant from AstraZeneca (to R.N. and M.A.G.); operating grants from the Canadian Institutes of Health Research [Grant 68828] (to R.N.); a studentship from the Lung Association of Alberta and Northwest Territories (to C.F.R.); studentships from Alberta Innovates-Health Solutions (to C.F.R. and E.M.K.); and an Izaak Walton Killam postdoctoral Fellowship (to N.S.H.). A grant from the Canadian Fund for Innovation and the Alberta Science and Research Authority provided equipment and infrastructure for conducting real-time PCR. R.N. and M.A.G. also are supported by grants from Gilead Sciences and GlaxoSmithKline.
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
- inhaled corticosteroid
- analysis of variance
- activator protein 1
- cigarette smoke extract
- dimethyl sulfoxide
- fetal calf serum
- fluticasone propionate
- glucocorticoid-induced leucine zipper
- glucocorticoid receptor
- glucocorticoid response element
- histone deacetylase
- c-Jun N-terminal kinase
- long-acting β2-adrenoceptor agonist
- mitogen-activated protein kinase
- 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide
- optical density
- polymerase chain reaction
- 9,11-dideoxy-11a,9a-epoxymethano-prostaglandin F2α
- phorbol 12-myristate 13-acetate
- tumor necrosis factor α.
- Received February 24, 2011.
- Accepted May 23, 2011.
- Copyright © 2011 by The American Society for Pharmacology and Experimental Therapeutics