Recruitment of inflammatory cells, including neutrophils, monocyte/macrophage, eosinophils, and T cells, is central to inflammation and is driven by increases in the expression of cytokines, chemokines, adhesion molecules, and other inflammatory proteins. For example, activation of structural cells in the airways by inflammatory insult or cytokines, such as interleukin (IL)-1β or TNF-α, leads to the expression and production of IL-8 (CXCL8) and granulocyte macrophage–colony-stimulating factor (GM-CSF) (Laberge and El Bassam, 2004). Because IL-8 is primarily a neutrophil chemoattractant (Mukaida, 2003), whereas GM-CSF stimulates myeloid-derived cells, including basophils, eosinophils, neutrophils, monocyte/macrophages, as well as subsets of dendritic cells, these two proteins differentially contribute to inflammatory cell influx (Martinez-Moczygemba and Huston, 2003).

In terms of the regulation of gene expression, transcription factors, including nuclear factor-κB (NF-κB) and activator protein-1, are central to the up-regulation of inflammatory gene expression. Indeed, both IL-8 and GM-CSF are regulated by NF-κB and activator protein-1 in human airway epithelial cells (Zhou et al., 2003; Newton et al., 2007). However, the modulation of mRNA stability, especially that of adenine- and uridine-rich element (ARE)-containing messages, is also a major controller of inflammatory gene expression (Stellato, 2004; Barreau et al., 2005). Thus, the proinflammatory cytokines IL-1β and TNF-α stabilize ARE-containing mRNAs, and this may increase gene expression by virtue of both elevated mRNA levels and enhanced translation. In contrast, anti-inflammatory glucocorticoids promote mRNA decay of many inflammatory genes, and this contributes substantially to the profound anti-inflammatory effects of such compounds (Stellato, 2004; Newton and Holden, 2007).

A key feature of many unstable inflammatory mRNAs is the presence of multiple copies of the pentameric sequence AUUUA (Barreau et al., 2005). This sequence is bound by a variety of ARE-binding proteins, and its function is subject to control via various inflammatory signaling pathways, of which the p38 mitogen-activated protein kinase (MAPK) pathway is possibly the most important (Dean et al., 2004). Indeed, ARE-dependent mRNA stabilization of both GM-CSF and IL-8 can occur via the p38 MAPK-activated protein kinase 2 pathway (Winzen et al., 1999). Conversely, the zinc finger protein ZFP36, also known as tristetraprolin (TTP), is an acute phase protein that binds to AREs and negatively regulates ARE-containing mRNAs, including TNF-α (Carballo et al., 1998; Mahtani et al., 2001), GM-CSF (Carballo et al., 2000), IL-2 (Ogilvie et al., 2005), IL-3 (Ming et al., 2001), and IL-8 (Winzen et al., 2007). The importance of TTP as a negative regulator of inflammatory gene expression is illustrated by the fact that mice in which the TTP gene is lacking demonstrate severe and chronic inflammation via an effect that involves increased TNF-α expression (Taylor et al., 1996). In addition to this role in feedback control, TTP expression is also reported to be up-regulated by glucocorticoids, and this may contribute to the post-transcriptional repression of inflammatory gene expression (Smoak and Cidlowski, 2006; Ishmael et al., 2008; Kaur et al., 2008). However, this point remains unclear because TTP expression induced by an inflammatory stimulus is reduced by dexamethasone (Jalonen et al., 2005).

After the induction of TTP protein, it seems that nucleocytoplasmic shuttling of TTP occurs and initially ARE-containing mRNAs are targeted for rapid degradation (Brook et al., 2006), probably in the processing-bodies that contain many of the enzymes involved in mRNA decay (Sandler and Stoecklin, 2008). Certainly, TTP can associate with various components of the mRNA deadenylation and decay pathways that are localized together in the processing-bodies (see Sandler and Stoecklin, 2008, and references therein). In addition, TTP has been found to associate, via components of the microRNA processing machinery, with the microRNA miR16, and this is proposed to promote ARE-mediated mRNA degradation (Jing et al., 2005).

In diseases such as asthma, inflammation of the lung can be targeted by anti-inflammatory glucocorticoids (Barnes, 2004). However, in some individuals, and in chronic obstructive pulmonary disease, glucocorticoids are poorly effective. Consequently, there is considerable interest in developing alternative anti-inflammatory strategies, such as small-molecule inhibitors of inflammatory cascades, including the MAPK and NF-κB signal transduction pathways (Barnes, 2004). Therefore, we have selected A549 pulmonary cells as a well characterized model of the pulmonary epithelium to examine TTP expression in response to the proinflammatory cytokine IL-1β and the glucocorticoid dexamethasone. Using this model, we examine roles for the p38 and ERK MAPK pathways as well as the dependence of TTP expression on NF-κB.

Materials and Methods

Cell Culture. A549 cells were grown to confluence in Dulbecco's modified Eagle's medium (DMEM) (Invitrogen, Burlington, ON, Canada) supplemented with 10% fetal calf serum (Canadian sourced; Invitrogen). Before experiments, cells were incubated overnight in serum-free media. Cells were then changed to fresh serum-free media containing cytokine and drugs. IL-1β (R&D Systems, Hornby, ON, Canada) was dissolved in phosphate-buffered saline plus 0.1% bovine serum albumin; dexamethasone (Sigma-Aldrich Canada, Oakville, ON, Canada) was dissolved in Hanks' balanced salt solution; and PD098059, U0126, SB203580 (EMD Biosciences, San Diego, CA), PS-1145, and ML120B (Millennium Pharmaceuticals, Cambridge, MA) were all dissolved in dimethyl sulfoxide (DMSO). Final concentrations of DMSO added to cells were <0.1% unless otherwise stated.

Adenoviral Infection. As described previously (Catley et al., 2005), A549 cells were grown to preconfluence (∼70%) and then incubated for 24 h with the indicated multiplicity of infection (MOI) of adenoviral (Ad) serotype 5 expression vector or an empty Ad5 vector (null). As described previously, Ad5-IκBαΔN encodes a dominant version of IκBα, whereas Ad5-IKK1(KM) and Ad5-IKK2(KA) produce dominant-negative versions of IKK1 and IKK2, respectively (Catley et al., 2005). In each case, cells were then incubated in serum-free media overnight before treatment with cytokine or drugs.

Reporter Cells and Luciferase Assay. A549 cells harboring the NF-κB-dependent luciferase reporter 6κBtkluc.neo were grown to confluence in medium containing 0.5 mg/ml G-418 (Geneticin; Promega, Madison, WI) as described previously (Catley et al., 2005). Cells were incubated overnight in serum-free media without G-418 before experiments in fresh serum-free media containing drugs and cytokine. After 6 h, cells were harvested in 1× reporter lysis buffer (Promega), and luminescence was measured using a BD Monolight luminometer (BD Biosciences, San Jose, CA).

Western Blotting. Cells were lysed in 1× reporter lysis buffer (Promega) containing 1× Complete protease inhibitor cocktail (Roche Diagnostics, Laval, QC, Canada) and phosphatase inhibitors (Sigma-Aldrich Canada). Total cellular lysates were size-fractionated on 4 to 12% gradient NuPAGE bis-Tris acrylamide gels (Invitrogen) and electroblotted to Hybond-ECL membranes (GE Healthcare, Baie d'Urfé, QC, Canada). Membranes were incubated with primary antibodies according to the manufacturers' instructions. The following anti-sera were used: anti-phospho-p44/42 (9101), anti-p44/42 (9102), anti-phospho-p38 (9211), and anti-p38 (9212) (Cell Signaling Technology Inc., Danvers, MA); anti-FLAG (F3165) and anti-HA (H3663) (Sigma-Aldrich Canada); TTP (sc-14030) (Santa Cruz Biotechnology, Inc., Santa Cruz, CA); TTP (SAK21A) (gift from A. R. Clark, Kennedy Institute of Rheumatology Division, Imperial College London, UK); TTP (ab36558) (Abcam Inc., Cambridge, MA); and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) [4699–9555(ST)] (AbD Serotec, Raleigh, NC). After washing, membranes were incubated with horseradish peroxidase-linked anti-rabbit or anti-mouse immunoglobulins (Dako Diagnostics Canada Inc., Mississauga, ON, Canada). Immune complexes were detected using enhanced chemiluminescence (GE Healthcare) and visualized by autoradiography.

RNA Isolation, cDNA Synthesis, and SYBR Green Real-Time PCR. Total RNA was isolated using the RNeasy mini kit (QIAGEN, Mississauga, ON, Canada), and 0.5 μg was used in reverse transcription reactions using avian myeloblastosis virus reverse transcriptase (Promega). Resultant cDNA was diluted 1:5 in RNase-free water and stored at 4°C. Real-time PCR analysis was carried out using an ABI 7900HT instrument (Applied Biosystems, Foster City, CA) on 2.5 μl of cDNA using SYBR GreenER Master Mix (Invitrogen) in a 20-μl reaction volume. Relative cDNA concentrations were obtained from a cDNA standard curve of serial dilutions of an IL-1β-stimulated cDNA sample. Amplification conditions of 50°C for 2 min, 95°C for 10 min, 40 cycles of 95°C for 15 s, and 60°C for 1 min were used. Primers specific to GAPDH (NM_002046) (forward, 5′-TTCACCACCATGGAGAAGGC-3′ and reverse, 3′-AGGAGGCATTGCTGATGATCT-5′) or TTP (NM_003407) (TTPS forward, 5′-CATGGCCAACCGTTACACC-3′ and TTPS reverse, 3′-AGCGACAGGAGGCTCTCGTAC-5′ or TTPAb forward, 5′-GCGGGAGTTTTTGCACCA-3′ and TTPAb reverse, 3′-GACCGGGCAGTCACTTTGTC-5′) were used (see Supplemental Fig. S1 for relative positions of primer binding sites in TTP mRNA). Primer specificity was determined by analysis of dissociation (melt) curves (95°C for 15 s, 60°C for 20 s, and 95°C for 15 s with ramping to 95°C over 20 min). A single peak in the rate of change of fluorescence with temperature indicated primer specificity.

Calf Intestinal Alkaline Phosphatase Treatment of Cell Lysates. Protein (25 μg) from cell lysates was incubated with 6.25 U of calf intestinal alkaline phosphatase (CIAP) (Invitrogen) at 37°C for 15 min in a total volume of 30 μl. Reactions were stopped by addition of SDS-containing sample loading buffer, and samples were subject to Western blot analysis as described above.

siRNA-Mediated Gene Silencing. Cells were grown in 12-well plates to ∼60 to 70% confluence, washed with serum-free DMEM, and then incubated with siRNA-containing serum-free media at 37°C for 24 h before the addition of drugs and cytokine. siRNA was prepared by mixing Lipofectamine 2000 (1 μl of 1 μg/μl; Invitrogen) and siRNA (25 nM; QIAGEN or Dharmacon RNA Technologies, Chicago, IL) together in 100 μl of serum-free DMEM, and this mixture was then incubated at room temperature for 30 min. The sequences for siRNA targeting were as follows: TTP siRNA 1 (5′-ACCGACGATATAATTATTATA-3′), TTP siRNA 2 (5′-ACGACTTTATTTATTCTAATA-3′), TTP siRNA 5 (5′-TAGCATATTTAAGGGAGGCAA-3′), TTP siRNA 4 (5′-TAGAATCTTATGTGCTGTGAA-3′), scrambled TTP siRNA 1a (a fully scrambled sequence of TTP siRNA 1) (custom designed, 5′-AATTCACATGATTAATAGTAA-3′), scrambled TTP siRNA 1b [a partially scrambled sequence of TTP siRNA 1 in which three pairs of bases are switched (base changes underlined)] (custom designed, 5′-AACGACAATTTATATAATAGA-3′) (all from QIAGEN) (see Supplemental Fig. S1 for relative locations), GFP siRNA (control siRNA) (P-002048-03-20, 5′-GGCAAGCTGACCCTGAAGTTC-3′) (Dharmacon RNA Technologies), GAPDH siRNA (5′-CCGAGCCACATCGCTCAGACA), and lamin A/C siRNA (5′-AACTGGACTTCCAGAAGAACA-3′) (QIAGEN).

Data Presentation and Statistical Analysis. Densitometric analysis of immunoblot data was performed on unsaturated images using TotalLab software (Nonlinear Dynamics, Newcastle, UK). All graphical data are presented as mean ± S.E. Statistical analysis between groups was performed using one-way analysis of variance (ANOVA) with a Bonferroni post-test or paired t test as indicated. Significance between groups was assumed where P < 0.05 (*), P < 0.01 (**), or P < 0.001 (***).

Results

Expression of TTP in A549 Cells Treated with IL-1β and Dexamethasone. Treatment of A549 cells with IL-1β for 1 h initially led to the appearance of a single immunoreactive band of approximately 40 kDa (Fig. 1A), which correlated with the 10-fold induction of TTP mRNA observed at this time (Fig. 1B). By 2 h post-IL-1β stimulation, there was a modest increase in the intensity of the lower immunoreactive band as well as the appearance of a less mobile band of approximately 45 kDa (upper band) (Fig. 1A). By 6 h, IL-1β-induced expression of both upper and lower immunoreactive bands was diminished until at 18 h this had essentially returned to basal levels. After the initial peak of IL-1β-induced TTP mRNA at 1 h, expression was reduced to approximately 3-fold at 18 h (Fig. 1B). Compared with IL-1β alone, dexamethasone at 1 μM (the maximally effective concentration for both the repression of inflammatory gene expression and the induction of glucocorticoid response element-dependent transcription in these cells; Chivers et al., 2006) produced a 3- to 4-fold induction of TTP mRNA at all time points analyzed (Fig. 1B; see Supplemental Fig. S2 for data analyzed using a second set of primers). This corresponded to a barely detectable level of TTP protein at approximately 40 kDa at 1 h, whereas a modestly induced single approximate 40-kDa band was visualized by 2 h after treatment (Fig. 1A). The intensity of this immunoreactive band represented only 11.0 ± 1.5% of the total immunoreactivity induced by IL-1β at 2 h and was virtually undetectable by 6 h after dexamethasone treatment. In combination, dexamethasone had no significant effect on IL-1β-induced TTP mRNA, although a slight reduction was apparent (Fig. 1B; also see Supplemental Fig. S2). Likewise, the induction of the lower TTP band by IL-1β at 1 h was repressed by 54.1 ± 3.6% (P < 0.05) in the presence of dexamethasone. By 2 h, a 27.9 ± 10.0% reduction in the intensity of both the lower and upper TTP bands was apparent in the presence of IL-1β plus dexamethasone compared with IL-1β alone. However, this did not reach significance. By 6 h, the presence of dexamethasone markedly reduced, by 61.6 ± 8.6% (P < 0.05), expression of TTP induced by IL-1β (Fig. 1A).

To confirm that the upper and lower immunoreactive bands observed by Western blot analysis were both TTP, we tested additional antibodies raised against TTP (see Supplemental Fig. S1 for the relative position of antibody epitopes). Initially, SAK21A and sc-14030 were examined using protein lysates run in parallel on one gel to allow a direct comparison of immunoreactive bands. This analysis revealed that although both antibodies detect a number of apparently nonspecific background bands, they both detected two bands of the same molecular mass (∼40–45 kDa) that were induced by IL-1β (see Supplemental Fig. S3 for representative blots). These data suggest that both bands were indeed TTP. This conclusion was confirmed by the analysis of a third antibody (ab36558), identical to the antibody reported by Brooks et al. (2002), which detected the same two immunoreactive bands at 40 and 45 kDa, but without many of the nonspecific bands (data not shown; see Supplemental Fig. S3B).

To examine the relationship between the two TTP bands, we tested whether the upper band may, as described previously (Mahtani et al., 2001), represent a phosphorylated version of TTP. A549 cells were stimulated for 2 h with IL-1β to induce both major forms of TTP. The cellular extracts were then not treated, incubated at 37°C for 15 min (control) or incubated at 37°C for 15 min in the presence of the phosphatase CIAP. As described in Fig. 1A, IL-1β treatment for 2 h induced both the upper and lower TTP bands. However, in samples treated with CIAP, the presence of the upper band was reduced (Fig. 1C), suggesting that this band represents phosphorylated TTP.

Overexpression of IκBαΔN Prevents NF-κB-Dependent Transcription and Inhibits IL-1β-Induced TTP Expression. To explore the role of NF-κB in the expression of TTP, A549 cells harboring the stably integrated NF-κB-dependent luciferase reporter 6κBtk.luc.neo, were infected with increasing MOIs of an adenoviral expression vector coding for the dominant variant of IκBα IκBαΔN. As described previously (Catley et al., 2005), IL-1β robustly induced luciferase activity by 13-fold. Increasing the MOI of Ad5-IκBαΔN resulted in a dose-dependent inhibition of the IL-1β-induced luciferase activity to near basal levels at a MOI of 100 (Fig. 2A). By contrast, the empty Ad5 vector at a MOI of 100 showed no effect on the induction of luciferase activity by IL-1β (Fig. 2B). These data confirm the utility of using Ad5-IκBαΔN as a probe to explore the role of NF-κBin the induction of TTP by IL-1β. Therefore, A549 cells were infected with various MOIs of both Ad5-IκBαΔN and empty vector before stimulation of cells with IL-1β for 1 h. As described in Fig. 1B, TTP mRNA expression was highly induced after 1 h of IL-1β treatment (Fig. 2C). Induction of TTP mRNA was profoundly inhibited by increasing the MOI of Ad5-IκBαΔN but not by empty Ad5 vector (Fig. 2C). This pattern of inhibition correlated closely with the inhibition of NF-κB-dependent transcription in which total inhibition was achieved at a MOI of 100 (Fig. 2A).

Cell lysates from the above-mentioned and parallel experiments were prepared at 6 h post-IL-1β stimulation (Fig. 2D). Concentration-dependent overexpression of IκBαΔN was verified by Western blot analysis, and this was found to correlate with reduced expression of TTP protein (Fig. 2D). By contrast, the empty Ad5 vector was without effect (Fig. 2D). In these experiments, it was the upper, modified, TTP band, that was predominantly inhibited by overexpression of IκBαΔN (Fig. 2D). However, analysis of IL-1β-induced TTP protein at 1 h in the absence and presence of 100 MOI IκBαΔN virus revealed profound inhibition (90.7 ± 2.0%) of the lower TTP band (Fig. 2E). Likewise, IL-1β-induced expression of both upper and lower TTP bands was markedly reduced by Ad5-IκBαΔN at both the 2- and 6-h post-IL-1β treatment (89.9^* ± 1.9 and 81.9 ± 3.9% respectively, where ^* represents P < 0.05), and these data are consistent with NF-κB dependence.

Effect of Dominant-Negative IKK1 and IKK2 Overexpression on NF-κB-Dependent Transcription and IL-1β-Induced TTP Expression. IKK1 and IKK2 dominant-negative viruses [Ad5-IKK1(KM) and Ad5-IKK2(KA), respectively] were used to further probe the role of the NF-κB signaling pathway on TTP expression. A549 6κBtk.luc.neo reporter cells were initially infected with increasing MOIs of IKK1(KM). IL-1β induced reporter activity by 6-fold, and this was partially prevented by increasing the MOI of Ad5-IKK1(KM) (Fig. 3A). Parallel Western blot analysis for the HA tag verified the overexpression of IKK1(KM) (Fig. 3A). Western blot analysis for TTP demonstrated inhibition of expression at the higher MOIs, and this was consistent with the effects on the NF-κB reporter (Fig. 3A).

To explore the role of IKK2, A549 6κBtk.luc.neo reporter cells were infected with increasing MOIs of IKK2(KA). Again, IL-1β induced luciferase activity by 6-fold (Fig. 3B). However, in contrast to IKK1(KM), IKK2(KA) revealed a considerably enhanced efficacy and potency of inhibition, with dose-dependent repression of reporter activity occurring at MOIs of 0.1 to 30 (Fig. 3B). Indeed, at an MOI of 30, NF-κB reporter activity was below the basal level. Western blot analysis of the FLAG epitope verified overexpression of IKK2(KA), and this correlated with both the inhibition of NF-κB-dependent transcription as well expression of TTP (Fig. 3B). These data support the previous finding that IKK2, and not IKK1, is the principal kinase involved in the activation of NF-κB (Catley et al., 2005) and further indicate that this kinase is predominantly responsible for the induction of TTP expression by IL-1β.

Effect of Small-Molecule IKK2 Inhibitors on IL-1β-Induced TTP Expression. To further test the role of IKK2, the expression of IL-1β-induced TTP was examined in the presence of the IKK2-selective small-molecule inhibitors PS-1145 and ML120B (Castro et al., 2003; Wen et al., 2006). Analysis of IL-1β-induced TTP expression at 6 h revealed a dose-dependent, but partial, inhibition of TTP expression by PS-1145 (Fig. 3C). This effect was maximal at ∼10 μM and is consistent with the previously reported inhibition of NF-κB-dependent transcription in A549 cells (Newton et al., 2007). The vehicle DMSO showed no effect on IL-1β-induced expression of TTP (Fig. 3C). Subsequent analysis of 10 μM PS-1145 revealed robust inhibition (65.1 ± 2.6%, P < 0.01; 32.1 ± 12.8%, P < 0.05; 58.2 ± 8.3%) of IL-1β-induced TTP expression at 1, 2, and 6 h, respectively (Fig. 3D). Likewise, the structurally distinct compound ML120B, at 10 μM, a concentration that inhibits NF-κB-dependent transcription in A549 cells (Newton et al., 2007), also caused inhibition (54.2 ± 8.9%, P < 0.01; 32.2 ± 12.8%, P < 0.05; 54.8 ± 10.7%) of IL-1β-induced TTP at 1, 2, and 6 h, respectively (Fig. 3D). Taken together, these data provide convincing support for the dominant-negative IKK2 adenovirus data and indicate a role for IKK2 in the up-regulation of TTP expression by IL-1β.

ERK and p38 MAPK Are Transiently Phosphorylated in Response to IL-1β. To confirm that ERK and p38 MAPK are phosphorylated in response to IL-1β, A549 cells were stimulated with IL-1β for 0.25, 0.5, 1, 2, or 6 h. Western blot analysis for phospho-ERK indicated that p42/44 ERK phosphorylation reached a peak in the region of 15 to 30 min (Fig. 4A). By 1 h, phospho-ERK levels were considerably reduced, and these subsequently reappeared at a low level at 120 min (Fig. 4A). Western blot analysis for phospho-p38 revealed a similar pattern of activation, with a peak of p38 phosphorylation at approximately 15 to 30 min. This then declined to basal levels at 60 min and then reappeared at approximately 120 min (Fig. 4B). In each case, variability in the time of maximal phosphorylation after IL-1β treatments suggests that the true peak of phosphorylation, and therefore activity, lies between these times.

PD098059 and U0126 Inhibit ERK Phosphorylation, and U0126 Partially Inhibits TTP Expression. To examine a role of the MEK-ERK pathway in the regulation of TTP expression, the effect of the MEK1/2 inhibitors PD098059 and U0126 were tested on the expression of TTP. As shown in Fig. 4, ERK phosphorylation is maximally induced by IL-1β by 15 to 30 min. Treatment of A549 cells with increasing concentrations of either PD098059 or U0126 before IL-1β stimulation resulted in a concentration-dependent inhibition of ERK phosphorylation to, or below, basal levels (Fig. 5A). These data validate the effectiveness of both PD098059 and U0126 at inhibiting the MEK1/2-ERK pathway.

Subsequently, A549 cells were treated with maximally effective concentrations (10 and 30 μM) of PD098059 or U0126 before stimulation with IL-1β for 1 h. In these experiments, IL-1β induced TTP mRNA by approximately 14-fold, and this was unaffected by treatment with PD098059 (Fig. 5B). In contrast, these concentrations of U0126 revealed a partial inhibitory effect on TTP mRNA expression, and this reached significance at 30 μM (Fig. 5B). Analysis of TTP protein expression in cells treated with increasing concentrations of PD098059 or U0126 before IL-1β stimulation indicated that there was again no effect of PD098059, even at 30 μM, on TTP expression. However, U0126 seemed to partially inhibit TTP expression at concentrations higher than 10 μM (Fig. 5C). Likewise, PD098059 (30 μM) had no effect on IL-1β-induced TTP protein expression at 1, 2, or 6 h (Fig. 5D), whereas U0126 (10 μM) partially inhibited TTP protein expression by 75.6 ± 2.5% (P < 0.05), 55.1 ± 8.5% (P < 0.05), and 61.0 ± 11.8% (P < 0.01) at 1, 2, and 6 h, respectively (Fig. 5E). Because PD098059 inhibits MEK1 in the 1 to 10 μM range (Alessi et al., 1995), and this is consistent with the prevention of ERK phosphorylation, our data do not support a role of the MEK1-ERK pathway in the induction of TTP by IL-1β. By contrast, partial inhibition of TTP expression by U0126 could be the result of either off-target effects, or because U0126 is equipotent for inhibition of MEK1 and MEK2 (Favata et al., 1998). However, given that PD09859 also inhibits MEK2 by 50% at 50 μM (Alessi et al., 1995), we suggest that off-target effects may explain the U0126 data.

SB203580 Inhibits p38 Activity and TTP Expression. To investigate the possible regulation of TTP expression by p38, the effect of the p38 inhibitor SB203580 was examined. Thus, A549 cells were treated with increasing concentrations of SB203580 before stimulation with IL-1β for 15 min. Western blot analysis showed that phosphorylation of heat shock protein-27, a known downstream target of p38 (Kyriakis and Avruch, 2001), is induced by IL-1β (Fig. 6A). This response was inhibited to basal levels by increasing concentrations of SB203580 (Fig. 6A), validating the effectiveness of SB203580 at inhibiting p38 activity. A549 cells were therefore treated with maximally effective concentrations (10 and 30 μM) of SB203580 before stimulation with IL-1β for 1 h. TTP mRNA was induced approximately 15-fold in response to IL-1β, and this was partially inhibited by SB203580 at both 10 and 30 μM (Fig. 6B). Western blot analysis for TTP protein revealed a profound concentration-dependent inhibition of the upper TTP band but little effect on the lower TTP band (Fig. 6C). However, time course analysis of IL-1β-induced TTP protein in the absence or presence of 10 μM SB203580, revealed that both the upper and lower TTP bands were considerably inhibited at 1 and 2 h, whereas by 6 h only the upper band was inhibited (Fig. 6D). Percentage of repression of both TTP bands combined at these time points were 58.3 ± 20.2, 80.8 ± 7.0 (P < 0.05), and 63.2 ± 6.4% (P < 0.05), respectively. These data suggest that p38 MAPK plays a role in the regulation of both TTP expression and its subsequent modification.

In addition, analysis of A549 cells treated with IL-1β, dexamethasone, or a combination of IL-1β plus dexamethasone showed that p38 is only phosphorylated in response to IL-1β and not dexamethasone (Fig. 6E). Indeed, dexamethasone inhibits IL-1β-induced p38 phosphorylation by 36.1 ± 21.7% at 15 min, by 59.6 ± 7.8% at 30 min, by 63.9 ± 10.5% at 1 h, and by 63.2 ± 10.2% at 2 h (P < 0.05) (Fig. 6E). These data suggest that although p38 is involved in the induction of IL-1β-induced TTP expression, there can be no role for p38 in dexamethasone-induced TTP expression because there is no p38 phosphorylation, and therefore activation, under these conditions. Furthermore, repression of the p38 MAPK pathway by dexamethasone (Fig. 6E) provides one compelling explanation for the reduction in IL-1β-induced TTP observed in the presence of dexamethasone (Fig. 1A).

Effect of siRNA-Mediated Knockdown of TTP Expression on GM-CSF and IL-8 Release. To examine a functional role for TTP in the regulation of inflammatory protein expression, A549 cells were transfected with TTP-specific siRNA or a control siRNA for 24 h before stimulation with IL-1β. Initially, cells were stimulated for 6 h and Western blot analysis was performed for TTP and GAPDH. TTP protein expression was induced by IL-1β and was dramatically reduced by the TTP-specific siRNA, but not the unrelated control siRNA (Fig. 7A).

Having established conditions for siRNA-mediated knockdown of TTP protein, the supernatants from IL-1β-treated cells were harvested at 6, 12, and 18 h and analyzed for GM-CSF and IL-8 release. In the case of GM-CSF, protein expression was induced to 62.0 ± 9.3 pg/ml at 6 h, 67.4 ± 9.0 pg/ml at 12 h, and 113.5 ± 23.6 pg/ml at 18 h by IL-1β. This response was significantly increased by 50 to 150% in the presence of siRNA 1 and siRNA 2 at 6, 12, and 18 h (Fig. 7B). Although siRNA 5 showed a trend toward increased GM-CSF release at 12 and 18 h, this did not reach statistical significance. Therefore, to further examine the role of TTP in regulating GM-CSF expression, the effect of an additional siRNA, siRNA 4, that targets TTP was analyzed (Supplemental Fig. S4A). This siRNA resulted in a robust reduction in IL-1β-induced TTP expression and also significantly increased GM-CSF release by approximately 100% at 6 h (Supplemental Fig. S4A). These data are consistent with the effect of siRNA 1 and 2, and taken together they provide strong support for a role for TTP in the inhibition of GM-CSF release.

IL-8 protein expression was also induced by IL-1β (6073 ± 784 pg/ml at 6 h, 5778 ± 1276 pg/ml at 12 h, and 7383 ± 995 pg/ml at 18 h). However, a significant (20–30%) enhancement by siRNA 1 and siRNA 2 only occurred at 6 h (Fig. 7B). At 12 h, these two siRNAs also seemed to increase IL-8 release, but this failed to reach statistical significance. As with GM-CSF, siRNA 5 had little effect on IL-8 release. Crucially, transfection of cells with the control siRNA had no apparent effect on IL-1β-induced GM-CSF or IL-8 expression. Likewise, transfection of control and TTP-specific siRNA in the absence of IL-1β did not produce any significant increases in cytokine release (data not shown).

In separate experiments, the effect of TTP knockdown was analyzed in the context of the repression by dexamethasone of IL-1β-induced GM-CSF expression. In this analysis, IL-1β-induced GM-CSF expression was strongly repressed (86.8 ± 2.8%; P < 0.05) by the addition of dexamethasone (Fig. 7C). Although this repressive effect was unaffected (85.6 ± 3.0; P < 0.01) by the control siRNA, the presence of TTP-specific siRNA significantly enhanced GM-CSF expression, suggesting that TTP was exerting a repressive effect in this system. However, because siRNA to TTP leads to enhanced GM-CSF release (Fig. 7B), the overall repression by dexamethasone is in fact similar to that in the controls. Thus, with siRNA 1-treated cells, IL-1β-induced GM-CSF release was increased by 137 ± 28% (and this is comparable with data in Fig. 7B). Likewise, in the presence of siRNA 1, the repression exerted by dexamethasone on IL-1β-induced GM-CSF was 85.3 ± 1.7% (P < 0.001). Because this is not significantly different from the repression observed in either naive (untransfected) or control siRNA-transfected cells, these data reveal clear roles for feedback control by TTP in both cells treated with either IL-1β alone or IL-1β and dexamethasone. As additional controls for the selectivity of targeting TTP, the effect of fully and partially scrambled versions of siRNA 1, scrambled siRNA 1a and 1b, were also examined (Fig. 7C, right). In each case, these scrambled siRNAs showed no effect on the expression of TTP (Fig. 7D) and resulted in no significant change to either the IL-1β-induced expression of GM-CSF or to the repression exerted by dexamethasone. Finally, to explore the possibility of nonspecific effects due to activating the RNA-induced silencing complex pathway, the effect of using siRNA to knock down the expression of GAPDH and lamin A/C was assessed on the expression of GM-CSF (Supplemental Fig. S4B). In these experiments, there was no significant effect of siRNA targeted to either GAPDH or lamin A/C on the release of GM-CSF induced by IL-1β. Taken together, these data support the concept that the effects of the TTP-targeting siRNA are due to the repression of TTP expression and not due to off-target effects. It is therefore possible that the lack of a significant response to siRNA 5 may be due to an off-target effect.

Discussion

The regulation of TTP expression and function is complex, and we do not have a complete picture of the signaling events that are involved. Once the protein is made, it is subject to regulatory events, including phosphorylation, intracellular compartmentalization, and possible degradation to control both the amount of TTP and its activity in regulating ARE-dependent mRNA turnover and translation (Sandler and Stoecklin, 2008). In the current study, we have used human A549 pulmonary cells to explore the regulation and role of TTP expression after IL-1β and dexamethasone treatment. After stimulation with IL-1β, TTP mRNA was strongly induced, peaking at 1 h, and declined thereafter. This kinetic was also observed at the level of TTP protein expression and suggests tight coupling between mRNA and protein expression. Initially, TTP protein is represented by a single immunoreactive band of ∼40 kDa at 1 h post-IL-1β treatment. By 2 h after stimulation, TTP expression is observed as a doublet of the ∼40-kDa band plus an ∼45-kDa band. This lower mobility band has previously been attributed to the presence of phosphorylation, and its loss, or reduction, after phosphatase (CIAP) treatment is consistent with these reports (Carballo et al., 2001; Mahtani et al., 2001).

To understand the up-regulation of TTP expression by IL-1β, we evaluated the possible role of the acute phase transcription factor NF-κB. These data revealed profound repression of TTP expression by IκBαΔN, a dominant inhibitor of NF-κB, suggesting that TTP is a highly NF-κB-dependent gene. The subsequent finding that TTP expression is more sensitive to inhibition by dominant-negative IKK2 than dominant-negative IKK1 and is inhibited by the IKK2-selective inhibitors PS-1145 and ML120B is consistent with NF-κB dependence and indicates that IKK2 is the main kinase mediating this effect. This is consistent with data showing IKK2 to be primarily responsible for the induction of NF-κB-dependent transcription and the expression of inflammatory genes such as cyclooxygenase-2, GM-CSF, and IL-8 in pulmonary epithelial cells (Catley et al., 2005; Newton et al., 2007). In terms of the induction of TTP protein, it is noteworthy that despite inhibition of NF-κB, the expression of the lower TTP band was still induced at 6 h post-IL-1β stimulation. Thus, other mechanisms of induction may be responsible for the production of TTP protein at longer time points. Interestingly, in the presence of NF-κB blockade, and despite some appearance of the ∼40-kDa protein, no phosphorylated TTP was observed. Thus, it is possible that NF-κB may play a role in the events leading to the phosphorylation of TTP. Although further experimentation would be required to examine this possibility, the finding that TTP is a NF-κB-dependent gene and is inhibited by small-molecule IKK inhibitors raises a concern in respect of the potential use of IKK, or other, inhibitors of NF-κB as anti-inflammatory agents. For example, in patients with chronic obstructive pulmonary disease, responsiveness to steroid treatment tends to be poor; thus, the use of IKK inhibitors as a potentially more effective therapy has been proposed (Caramori et al., 2004). However, the current data raise a potential concern with this strategy. Thus, inhibition of NF-κB in the context of inflammatory diseases may reduce transcription of inflammatory genes. However, the parallel block of TTP expression may promote mRNA stabilization of certain ARE-containing transcripts and this could, in some instances, counteract the transcriptional inhibition of such genes.

In terms of the regulation of TTP by the p38 MAPK pathway, there is considerable data indicating phosphorylation of TTP by the downstream kinase p38/MAPK-activated protein kinase (Dean et al., 2004). In A549 cells, the appearance of the upper, phosphorylated form of TTP was inhibited by SB203580 over a concentration range that is consistent with the inhibition of p38 MAPK. In addition, the induction of TTP protein by IL-1β was dramatically, but not completely, prevented by p38 inhibition. Thus, the p38 MAPK plays a role in both the expression and post-translational modification of TTP. Functionally, TTP phosphorylation is suggested to promote translocation to the cytoplasm and reduce the destabilizing activity of TTP for ARE-containing mRNAs (Johnson et al., 2002; Brook et al., 2006; Hitti et al., 2006). These phosphorylation events occur on serine 52 and 178 of TTP, and this seems to promote the binding of 14-3-3 proteins to TTP (Chrestensen et al., 2004; Brook et al., 2006). Furthermore, the binding of 14-3-3 proteins may also prevent the dephosphorylation of TTP by protein phosphatase 2A, which may in turn reduce TTP-dependent ARE destabilizing activity (Stoecklin et al., 2004; Sun et al., 2007). Thus, reduced TTP phosphorylation after blockade of p38 MAPK may, in A549 cells, reduce the expression of ARE-containing transcripts. Conversely, a generalized reduction in TTP expression, after siRNA-mediated knockdown is predicted to up-regulate ARE-containing transcripts. Thus, our finding that both GM-CSF and IL-8 release were increased after knockdown of TTP, confirms TTP as an overall negative feedback regulator of these genes in A549 cells (Sandler and Stoecklin, 2008). These data are supported by previous reports that TTP can increase the degradation of IL-8 mRNA (Winzen et al., 2007; Suswam et al., 2008). Collectively, the above-mentioned studies may be taken to indicate more of a Ying-Yang or Janus-like role for TTP. In the initial unphosphorylated state, TTP is a negative regulator of ARE-containing genes, whereas the ability to reduce, or switch off, this destabilizing activity by phosphorylation may facilitate mRNA stabilization. Indeed, TTP has been shown to stabilize inducible nitric-oxide synthase mRNA (Fechir et al., 2005). Altering the phosphorylation state of TTP may therefore provide a key mechanism for regulating the expression of inflammatory genes. Consequently, the exact timing and extent of siRNA-mediated repression of TTP may differentially impact on target transcripts, and this may provide some explanation for the different levels of effect imparted by the three TTP siRNA molecules tested on GM-CSF and IL-8 in the current study.

Our prior RNA analysis identified TTP as a glucocorticoid-inducible gene, and this, at least initially, is consistent with both the established destabilizing activity of TTP and the anti-inflammatory effects of glucocorticoids (Newton and Holden, 2007; Kaur et al., 2008). Thus, Smoak and Cidlowski (2006) document a potential role for TTP in the dexamethasone-induced destabilization of TNF-α. Indeed, our data showing that siRNA-mediated knockdown of TTP leads to a reduced repressive effect of dexamethasone on IL-1β-induced GM-CSF (Fig. 7C) may at first seem to support a role for TTP in the repression exerted by dexamethasone. However, it is clear from our present study that the level of dexamethasone-induced TTP protein is very modest compared with the level induced by IL-1β alone. Furthermore, our data indicate that in the presence of dexamethasone, IL-1β-induced levels of TTP are considerably reduced, and this observation is consistent with the repression of TTP by dexamethasone in lipopolysaccharide-treated macrophages (Jalonen et al., 2005). Because, in the presence of IL-1β alone, TTP is a negative regulator of GM-CSF, and in the presence of IL-1β plus dexamethasone TTP expression is reduced, our data indicate that TTP is predominantly a feedback control gene rather than primarily an effector of glucocorticoid-dependent inhibition. Thus, in the presence of siRNA 1, TTP expression is knocked down considerably, but the overall repression by dexamethasone is not affected. This issue is not addressed by the Smoak and Cidlowski (2006) study, so the relative contribution of TTP to feedback control versus glucocorticoid-dependent inhibition of TNF-α remains open. Likewise, the study of Ishmael et al. (2008) used TTP knockout fibroblasts to show a major role for TTP in the glucocorticoid-dependent down-regulation of inflammatory genes. However, complete knockout of TTP induces an inflammatory phenotype by increasing TNF-α expression and mRNA stabilization and presumably production of other inflammatory genes (Carballo et al., 1998, 2000; Lai et al., 2006). This effect is especially apparent in the presence of proinflammatory cytokines or serum, which normally induces TTP expression, and indeed, serum was present throughout the Ishmael study. Thus, it is possible that indirect consequences of TTP loss (e.g., increased TNF-α expression leading to secondary mRNA stabilization or reduced glucocorticoid receptor activity) may result in an over-representation of its role in the presence of glucocorticoid. This effect may account for the modest induction of TTP by dexamethasone in our current study (where all experiments were conducted in a serum-free environment).

Notwithstanding the above-mentioned points, TTP is modestly induced by dexamethasone in A549 cells, and it is certainly possible that glucocorticoid-induced TTP may help maintain TTP-dependent feedback control in a situation in which it would otherwise become shut off. Thus, glucocorticoids repress activation of the p38 MAPK pathway (Newton and Holden, 2007), and we now report reduced levels of phosphorylated p38 MAPK after dexamethasone treatment of A549 cells. Given the role of p38 MAPK in TTP expression, reduced activation of the p38 pathway in the presence of dexamethasone may explain the reductions in IL-1β-induced TTP observed at 1 and 6 h. However, at 2 h post-IL-1β plus dexamethasone, the repression of TTP expression was marginal and not significant. It is therefore tempting to speculate that a contribution from the dexamethasone-induced TTP, which was maximal at 2 h, may prevent a more profound loss of TTP protein and thereby help maintain feedback inhibition of ARE-containing genes.

In conclusion, we present evidence that TTP is induced by IL-1β in a NF-κB-dependent and p38 MAPK-dependent manner. We show that TTP is involved in the negative feedback control of both GM-CSF and IL-8 and suggest that this activity may be adversely affected by anti-inflammatory strategies that inhibit NF-κB. We further show that TTP is modestly induced by the glucocorticoid dexamethasone and that in the context of IL-1β, dexamethasone reduces TTP expression.