Induction of Glucocorticoid-induced Leucine Zipper (GILZ) Contributes to Anti-inflammatory Effects of the Natural Product Curcumin in Macrophages*

GILZ (glucocorticoid-induced leucine zipper) is inducible by glucocorticoids and plays a key role in their mode of action. GILZ attenuates inflammation mainly by inhibition of NF-κB and mitogen-activated protein kinase activation but does not seem to be involved in the severe side effects observed after glucocorticoid treatment. Therefore, GILZ might be a promising target for new therapeutic approaches. The present work focuses on the natural product curcumin, which has previously been reported to inhibit NF-κB. GILZ was inducible by curcumin in macrophage cell lines, primary human monocyte-derived macrophages, and murine bone marrow-derived macrophages. The up-regulation of GILZ was neither associated with glucocorticoid receptor activation nor with transcriptional induction or mRNA or protein stabilization but was a result of enhanced translation. Because the GILZ 3′-UTR contains AU-rich elements (AREs), we analyzed the role of the mRNA-binding protein HuR, which has been shown to promote the translation of ARE-containing mRNAs. Our results suggest that curcumin treatment induces HuR expression. An RNA immunoprecipitation assay confirmed that HuR can bind GILZ mRNA. In accordance, HuR overexpression led to increased GILZ protein levels but had no effect on GILZ mRNA expression. Our data employing siRNA in LPS-activated RAW264.7 macrophages show that curcumin facilitates its anti-inflammatory action by induction of GILZ in macrophages. Experiments with LPS-activated bone marrow-derived macrophages from wild-type and GILZ knock-out mice demonstrated that curcumin inhibits the activity of inflammatory regulators, such as NF-κB or ERK, and subsequent TNF-α production via GILZ. In summary, our data indicate that HuR-dependent GILZ induction contributes to the anti-inflammatory properties of curcumin.

The polyphenol curcumin (diferuloylmethane) is the principal bioactive compound of turmeric (Curcuma longa) prepara-tions, which are commonly used as traditional remedies or spices. Recently, extensive research has shown that curcumin has a broad range of therapeutic effects, including anti-inflammatory, anti-oxidant, anti-proliferative, hypoglycemic, lipidlowering, anti-thrombotic, and anti-coagulant activity. At the same time, no dose-limiting toxicity was observed in clinical trials, indicating pharmacological safety. Thus, curcumin administration has been suggested as a potential therapeutic approach for the treatment of several pathologic conditions, especially inflammatory diseases such as cardiovascular pathologies, arthritis, asthma, ulcerative colitis, inflammatory bowel disease, and type II diabetes (1)(2)(3)(4)(5). Several in vivo studies suggested a profound effect of curcumin on cells of the mononuclear phagocyte system. In a mouse model of abdominal aortic aneurysms, oral administration of curcumin efficiently suppressed mononuclear inflammation in the aortic walls, i.e. proinflammatory cytokine expression and adverse connective tissue remodeling (6). In addition, curcumin decreased the number of proinflammatory M1 macrophages but increased the number of anti-inflammatory M2 macrophages in the myocardium in a rat model of experimental autoimmune myocarditis (7). In diabetic nephropathy, curcumin administration reduced the infiltration of macrophages into the glomeruli and impaired the expression of proinflammatory cytokines, such as tumor necrosis factor-␣ (TNF-␣) (8). Likewise, oral administration of nano-emulsion curcumin diminished macrophage recruitment in a mouse model of peritonitis (9). Curcumin has been shown to interact with various molecular targets, including transcription factors, enzymes, and receptors (1,5,10). The anti-inflammatory effects of curcumin in macrophages and other mononuclear cells are associated with its ability to suppress nuclear factor B (NF-B) signaling (8 -13).
Over the last years the endogenous anti-inflammatory mediator glucocorticoid-induced leucine zipper (GILZ) 2 has attracted increasing attention. GILZ mediates its anti-inflammatory activity mainly by direct binding to the proinflamma-tory transcription factors NF-B and activator protein AP-1, thereby inhibiting their translocation (14 -17). Additionally, GILZ has also been reported to be a negative regulator of mitogen-activated protein kinase (MAPK) signaling (18 -20).
Due to its strong induction by glucocorticoids (GCs) in the thymus, early studies on GILZ concentrated on its effects on thymocytes and T-lymphocytes (21). Several reports indicated that the effects of GILZ overexpression are similar to GC treatment regarding the induction of thymocyte and T-lymphocyte apoptosis (14,17,(21)(22)(23). Moreover, lack of GILZ impairs GCinduced apoptosis in B-lymphocytes (24).
In monocytes and macrophages, GILZ can be induced by anti-inflammatory agents, such as GCs and interleukin 4 (IL-4) or IL-10 (15,(25)(26)(27). GILZ overexpression in macrophage-like THP-1 cells resulted in GC-like effects, i.e. reduced expression of macrophage activation markers, chemokine expression, and NF-B activity upon treatment with the Toll-like receptor (TLR) 4 ligand lipopolysaccharide (LPS) (15). In the present study we aimed to examine the influence of curcumin on GILZ expression in macrophages and to elucidate whether GILZ plays a role in the anti-inflammatory properties of curcumin in this cell type.

Results
In the first series of experiments we examined whether curcumin influences the expression of GILZ in the macrophagelike cell lines RAW264.7 and U937. After incubation with curcumin at non-toxic concentrations, increased GILZ protein levels were observed in murine RAW264.7 and human U937 cells as well as primary human monocyte-derived macrophages (HMDMs), as assessed by Western blot (Fig. 1, A-D) and/or flow cytometric analysis (Fig. 1, H and I). The specificity of the GILZ-PE antibody used for flow cytometry was evaluated using untreated and dexamethasone-treated wild-type (WT) and GILZ knock-out (KO) MPI macrophages ( Fig. 1, E-G).
To analyze the molecular mechanism underlying curcuminmediated GILZ induction, we first examined whether curcumin was able to activate the glucocorticoid receptor (GR). HEK 293T cells were transfected with a vector encoding a GR-GFP fusion protein followed by treatment with either dexamethasone or curcumin for up to 40 min. As expected, live microscopy revealed translocation of the GR into the nucleus upon dexamethasone treatment. In contrast, GR translocation was absent in curcumin-treated cells ( Fig. 2A and supplemental Fig. 1), indicating that the up-regulation of GILZ by curcumin was not associated with GR activation.
Next, we examined GILZ mRNA levels in RAW264.7 and U937 cells after curcumin exposure for up to 3 h. No difference in GILZ mRNA expression was detected in either cell line (Fig.  2, B and C) nor primary HMDMs (data not shown), suggesting that neither the level nor the stability of the GILZ transcript was significantly altered in response to curcumin. Analysis of GILZ mRNA levels after the addition of the transcription inhibitor actinomycin D in the absence or presence of curcumin confirmed that curcumin did not affect GILZ mRNA stability (Fig. 2D).
The influence of curcumin on GILZ protein stability was assessed by incubating cells with cycloheximide, a protein syn-thesis inhibitor, either alone or in combination with curcumin, for different time points. Cycloheximide treatment resulted in a significant decrease of GILZ protein levels after 4 h. However, this effect was not influenced by curcumin co-treatment (Fig.  2E), indicating that curcumin did not stabilize GILZ protein. Thus, we determined the impact of curcumin on the translation of GILZ. To this end we treated U937 cells with curcumin and subjected the resulting lysates to polysomal fractionation. The absorbance at 254 nm, reflecting the RNA content, decreased in the later fractions of the polysome profile, whereas the 60S and 80S peaks increased in response to curcumin, thus indicating a suppression of global translation upon curcumin treatment (Fig. 3A). This is in line with a previous report showing that curcumin inhibited cap-dependent translation by suppressing the activation of translation initiation factors (28). Interestingly, although global translation appeared to be inhibited, the distribution of GILZ mRNA shifted from the subpolysomal fractions (fractions 1-3) into the late polysomal fractions (fractions 6 -9) (Fig. 3B). This indicates that curcumin enhances GILZ translation.
Adenylate-and uridylate-rich elements, commonly found in the 3Ј-untranslated region of mRNAs, are central elements of gene regulation and have been reported to affect the translation process. The mRNA-binding protein HuR binds to mature mRNAs through adenylate-and uridylate-rich element recognition motifs in their 3Ј-UTRs, typically via U-rich sites, thereby stabilizing the respective mRNAs and modulating their translation (29 -32). As the 3Ј-UTR of GILZ mRNA contains a U-rich region (33) and might, therefore, be regulated by HuR, we examined whether curcumin affects HuR expression. Western blot analysis showed that HuR levels were indeed increased in curcumin-treated cells (Fig. 4A). Similar results were obtained with primary HMDMs (Fig. 4B).
We then investigated whether HuR can bind to GILZ mRNA. Cell lysates were subjected to HuR immunoprecipitation followed by quantification of HuR-bound mRNAs by real-time RT-PCR. In contrast to the negative control GAPDH, GILZ mRNA was enriched to a similar degree as the positive control CCNB1 (Fig. 4C) (34), demonstrating that HuR can bind to GILZ mRNA. To characterize the influence of enhanced HuR expression on GILZ, we overexpressed HuR in HEK 293T cells. GILZ protein levels were indeed increased in HuR-overexpressing cells, whereas GILZ mRNA was not (Fig. 4, D-F). These observations strengthen our notion that curcumin increases GILZ mRNA translation via HuR induction.
To examine the functional consequences of increased GILZ levels after curcumin treatment, cells were treated with curcumin followed by LPS exposure. Subsequently, NF-B activity was measured by luciferase reporter gene assay. As shown in Fig. 5A, curcumin strongly inhibited LPS-induced NF-B activation. NF-B activity was restored upon siRNA-mediated GILZ knockdown, which supports the assumption that GILZ induction contributes to the anti-inflammatory actions of curcumin. In accordance, the inhibitory effect of curcumin on LPS-induced NF-B-dependent inducible nitric-oxide synthase activity was abrogated upon GILZ knockdown (Fig. 5, C and D). However, GILZ knockdown did not fully rescue either luciferase activity or nitrite accumulation. To evaluate whether the reasons underlying this observation involve residual GILZ expression or GILZ-independent mechanisms, primary murine bone marrow-derived macrophages (BMMs) obtained from wild-type (WT) and GILZ KO mice were employed. Curcumin was able to induce GILZ in BMMs, as indicated by flow cytometric analysis (Fig. 6A).
The effect of curcumin on NF-B activity in GILZ-depleted BMMs was comparable with that seen in siGILZ-treated RAW264.7 cells (Fig. 6B), suggesting that GILZ, although considerably contributing to the anti-inflammatory actions of curcumin, is not the only factor involved.
Loss of GILZ has also been shown to enhance extracellularsignal regulated kinase (ERK) signaling in murine bone marrow-derived macrophages (20). Therefore, we examined LPSinduced ERK activation in curcumin-treated WT and GILZ KO BMMs. LPS-mediated ERK activation was strongly inhibited by curcumin in WT macrophages at concentrations that up-regulated GILZ expression, whereas ERK signaling was preserved in equally treated GILZ KO BMMs (Fig. 6, C-F). Both ERK and NF-B are involved in LPS-induced TNF-␣ production in murine BMMs (20). Thus, we also examined the inhibitory effects of curcumin on TNF-␣ secretion in WT and GILZ KO cells. Preincubation of WT BMMs with curcumin decreased the induction of TNF-␣ by LPS in a dose-dependent manner. In contrast, curcumin failed to show any effect on GILZ KO cells under most of the tested conditions ( Fig. 6G), thereby supporting the notion that curcumin exerts its anti-inflammatory actions at least in part via GILZ.

Discussion
For centuries the natural compound curcumin has been used for the treatment of inflammatory conditions in Eastern medicine, and various reports have confirmed its ethnopharmacological relevance in recent years (1,3). Several in vitro and in vivo studies suggested that curcumin may be used to treat distinct pathological conditions, including chronic inflammatory processes and neoplastic diseases (10,35).
The mechanisms underlying the anti-inflammatory properties of curcumin involve the regulation of different molecular targets, including transcription factors, growth factors, inflammatory cytokines, and protein kinases (10). The ability of curcumin to interfere with NF-B signaling is central to its antiinflammatory action. In general, the effect has been attributed to an attenuated nuclear translocation of NF-B, which has previously been explained by an induction of its inhibitor IB or down-regulation of molecules required for TLR signaling (8 -13, 36). In the present study, we demonstrated for the first time that curcumin up-regulates the anti-inflammatory mediator GILZ. GILZ can directly interact with NF-B, thereby inhibiting its translocation. Our data suggest that GILZ induction considerably contributes to the inhibition of NF-B by curcumin in macrophages.
In addition, GILZ has been reported to attenuate MAPK signaling, e.g. by binding to Ras/Raf, which then leads to the inhi-bition of downstream mitogen-activated protein kinases, such as ERK (18 -20). It was recently shown that curcumin not only diminishes NF-B activity but also interferes with other proinflammatory signaling pathways in macrophages, including LPS-induced ERK signaling (36). Our experiments with LPSactivated BMMs from wild-type and GILZ knock-out mice demonstrated that curcumin inhibits the activity of ERK, at least in part, via GILZ. In line with these findings, the production of proinflammatory effectors, such as NO and TNF, was decreased by curcumin in a GILZ-dependent manner.
We previously observed high constitutive GILZ expression levels in primary human in vitro differentiated and pulmonary macrophages (33). Interestingly, GILZ levels were significantly reduced in human alveolar macrophages as well as in vivo in mouse lungs upon TLR activation (20,33). Both siRNAmediated GILZ knockdown in human macrophages and GILZ knock-out in murine BMMs increased the responsiveness toward LPS, suggesting that repression of endogenous GILZ expression is a regulatory mechanism that enhances the proinflammatory response.
In addition to its up-regulation by curcumin demonstrated in the present study, GILZ can be induced in monocytes and macrophages by GCs and anti-inflammatory cytokines, such as IL-10 (15,25). The potent inhibitory action of GCs on cytokine and chemokine production in human monocytes has been shown to be abrogated by siRNA-mediated GILZ silencing (25), indicating that GILZ is essential for anti-inflammatory GC actions. In accordance, GILZ overexpression in the macrophage cell line THP-1 resulted in GC-like effects upon LPS challenge (15). Macrophages from the inbred LPS-resistant mouse strain SPRET/Ei have been reported to highly express GILZ. SPRET/Ei macrophages, therefore, showed reduced IL-6 and IL-12 levels after LPS treatment, whereas GILZ knockdown by siRNA compensated the effect. In summary, these reports indicate that GILZ up-regulation efficiently prevents inflammatory responses in myeloid cells. Of note, specific up-regulation of GILZ might avoid the metabolic side effects of GC therapy, as GILZ and GCs have opposing consequences for the differentiation of mesenchymal stem cells: whereas GCs favor adipocyte differentiation and suppress osteoblast formation, GILZ overexpression induces osteogenic differentiation (37,38).
Our findings implicate the curcumin-inducible RNA-binding protein (RBP) HuR as a novel positive regulator of GILZ expression. RBPs can bind to coding as well as noncoding regions of mRNAs. Upon binding, HuR is able to modulate the translation of its target transcripts. HuR has been shown to promote the translation of many target mRNAs, such as HIF1A, TP53, DUSP1, CYCS, and HMOX1 (32, 39 -43). The molecular mechanisms by which HuR modulates translation are poorly understood but may involve competition or cooperation with microRNAs. Alternatively, HuR might interfere with internal ribosome entry sites (IRESs) of target transcripts (32).
Unlike most other RBPs that induce mRNA destabilization, ELAV/Hu proteins preferentially stabilize U-rich transcripts. HuR has been reported to increase the stability of several mRNAs, including VEGFA, DUSP1, and cyclins CCNA2, CCNB1, CCNE, and CCND1 (32, 34, 44 -47). The underlying mechanisms are not entirely characterized, but it has been suggested that HuR competes with other RBPs that facilitate mRNA recruitment to sites of mRNA degradation, such as the exosome and P-bodies (32). However, the interaction of HuR with GILZ mRNA did not affect its stability in our hands. The influence of other HuR targets on the anti-inflammatory effects of curcumin presently remains elusive and might represent an interesting subject for future studies.
In summary, our data demonstrate that curcumin induces GILZ in macrophages, which contributes to its anti-inflammatory actions. These new insights into the effects of curcumin in the suppression of inflammatory responses suggest that curcumin treatment may be beneficial in various inflammatory diseases associated with excessive macrophage activation. The clinical use of curcumin is restricted by its low solubility and poor systemic bioavailability. Therefore, orally administered curcumin has mainly been used for the treatment of gastrointestinal disorders, including inflammatory bowel disease and colon cancer (48 -50). New therapeutic approaches aiming to overcome these limitations, e.g. nano-emulsion or silica nanoparticle formulations, have recently been described (9, 51) and might expand the number of pathologic conditions that may benefit from curcumin treatment in the near future. Alternatively, curcumin might serve as a lead for GILZ-inducing derivates with a better bioavailability.
HMDMs-Buffy coats were obtained from healthy adult blood donors (Blood Donation Center, Saarbrücken, Germany). The use of human material for the isolation of primary cells was approved by the local ethics committee (permission no. 130/08). Peripheral blood mononuclear cells were isolated by density gradient centrifugation using Lymphocyte Separation Medium (PAA) and LeucoSep tubes (Greiner). After washing with PBS, monocytes were purified from peripheral blood mononuclear cells by plastic adherence for 1 h. Monocytes were differentiated into macrophages in RPMI 1640 supplemented with 10% FCS, 100 units/ml penicillin G, 100 mg/ml streptomycin, 2 mM glutamine, and 20 ng/ml GM-CSF at 37°C and 5% CO 2 for 8 days. Medium was changed every other day.
BMMs-BMMs were obtained from 8 -12-week-old male WT or GILZ KO mice as described previously (20). Briefly, femurs and tibias were flushed with standard medium. Erythrocytes contained in the cell pellet after centrifugation were lysed with hypotonic buffer. Cells were resuspended in standard medium supplemented with M-CSF (50 ng/ml, Biomol), transferred into a culture flask, and allowed to adhere overnight. Non-adherent cells were collected and cultivated for another 5 days in M-CSF-containing medium. Subsequently, cells were detached with accutase (Sigma), suspended in standard medium with 10 ng/ml M-CSF, and seeded in 96-well plates (7.5 ϫ 10 4 cells/well in 150 l of medium) for MTT assays and TNF-␣ quantification or 12-well plates (0.5 ϫ 10 6 cells/well in 1 ml of medium) for Western blot analyses. BMMs were found to be Ͼ95% pure as indicated by flow cytometric analysis using an antibody against F4/80 (data not shown).
MPI Cells-MPI cells, i.e. non-transformed self-renewing primary murine macrophages, were obtained from C57BL/6 mice based on a method described previously (20,53). Briefly, MPI cells were prepared from fetal livers of male 15-day-old mouse embryos and grown in standard medium supplemented with 30 ng/ml murine GM-CSF (Biomol). Proliferating cells were subcultured by splitting them 1:5 after 6 -8 days. The purity of the cell preparations was Ն95%, as assessed by flow cytometric analysis of F4/80 expression (data not shown).
Determination of Cell Viability-To ensure that non-toxic concentrations of curcumin were used, the MTT colorimetric assay was performed as described previously (20,54). Absorbance measurements were carried out at 550 nm with 630 nm as the reference wavelength using a microplate reader (Tecan Sunrise). The cell viability obtained from at least three independent experiments was calculated relative to untreated and solvent controls (data not shown).  OCTOBER 28, 2016 • VOLUME 291 • NUMBER 44

JOURNAL OF BIOLOGICAL CHEMISTRY 22955
HuR Overexpression-HEK 293T cells were seeded at a density of 2 ϫ 10 5 cells per well into a 12-well plate and were grown for 24 h. Cells were either transfected with the pZeoSV2(Ϫ)HuR or the empty pZeoSV2(Ϫ) control vector using PolyFect (Qiagen) according to the manufacturer's recommendations. Cells were harvested for RNA or protein isolation 48 h after transfection.  n ϭ 4, duplicates). ***, p Ͻ 0.001 compared with solvent control. B, NF-B activity in LPS-stimulated BMMs was measured by luciferase reporter gene assay. Cells were left untreated (Co) or pretreated either with solvent control (0.05% DMSO) or with curcumin at the indicated concentrations for 60 min followed by LPS stimulation (1 g/ml) for 4 h. Values for LPS-induced activation in DMSO-pretreated cells were set as 100%, and data are presented as the means ϩ S.E. (n ϭ 3, triplicates). C-F, WT and GILZ KO BMMs were pretreated with solvent control (0.05% DMSO, Co) or curcumin for 3 h followed by LPS stimulation for 15 min (100 ng/ml). ERK phosphorylation was examined by Western blot analysis. C and D, one representative blot is shown. E and F, signal intensities were quantified, normalized to the indicated loading controls, and expressed as % curcumin-induced inhibition. Data represent the means ϩ S.E. (n ϭ 5). *, p Ͻ 0.05 compared with equally treated WT BMMs. G, BMMs were pretreated with curcumin (12.5 or 25 M) or vehicle (0.05% DMSO) for 1 h followed by the addition of LPS at the indicated concentrations for 4 h. TNF was measured by bioassay. Data represent the means ϩ S.E. (n ϭ 4, triplicates). B and E-G: o , p Ͻ 0.05; oo , p Ͻ 0.01; ooo , p Ͻ 0.001 versus equally treated cells of the same genotype without curcumin pretreatment; ϩ , p Ͻ 0.05; ϩϩ , p Ͻ 0.01; ϩϩϩ , p Ͻ 0.001 versus equally treated cells of the same genotype with 12.5 M curcumin pretreatment; *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001 versus equally treated WT cells.
Flow Cytometry-Staining was essentially performed as described previously (20,56,57). Briefly, cells were detached from the plates using accutase (Sigma). For extracellular staining of F4/80, cells were washed with PBS and resuspended in flow cytometry buffer (FCB; PBS containing 2.5% (v/v) bovine calf serum and 0.05% (w/v) NaN 3 ). 5 ϫ 10 5 cells were incubated with mouse BD Fc Block (BD Biosciences) as recommended by the supplier followed by the addition of F4/80 or isotype control antibody (0.3 g in 100 l of FCB) for 15 min on ice. The cells were washed in FCB and resuspended in 1% (w/v) cold paraformaldehyde in PBS, pH 7.6. To detect intracellular GILZ, cells were incubated with the Fc Blocking reagent and fixed for 10 min in 1% (w/v) paraformaldehyde in PBS, pH 7.6, followed by permeabilization in SAP (FCB with 0.2% (w/v) saponin and blocking for 30 min in 20% FCS (v/v, diluted in SAP). Cells were stained with the GILZ-specific antibody or the matching isotype control (0.5 g in 100 l of SAP) for 30 min on ice. The stained cells were examined on a BD LSRFortessa™ cell analyzer, and results were analyzed using the FACS Diva software (BD Biosciences). Curcumin itself did not interfere with flow cytometric analysis (data not shown).
Determination of NF-B Activity-For reporter gene assays, a vector containing NF-B responsive elements driving firefly luciferase expression (pGL4.32[luc2P/NF-B-RE/Hygro], Promega) was used. The pRL-TK vector provided constitutive expression of renilla luciferase and served as an internal control value, to which expression of the firefly luciferase reporter gene was normalized. To assess the influence of GILZ on curcumininduced NF-B-inhibition, RAW 264.7 macrophages were cotransfected with the firefly/renilla luciferase vectors (3.33 and 0.67 g/ml, respectively) and either Gilz-targeting (siGILZ) or control siRNA (siCo, 10 nM) using the JetPrime reagent (PolyPlus transfection) as recommended by the supplier. Knockdown efficiency was assessed by flow cytometry after 24 h (siGILZ: 50.5 Ϯ 10.7% GILZ expression when compared with siCo (ϭ100%), p Ͻ 0.05).
For luciferase assays, RAW 264.7 cells were seeded at 10 4 cells per well into a 96-well plate, treated as indicated, and harvested by the addition of 1ϫ passive lysis buffer (Promega). BMMs were transfected and processed as reported previously (20). RNA Immunoprecipitation-Immunoprecipitation of HuRassociated RNA was performed using the magnetic SureBeads system (Bio-Rad) according to the manufacturer's specifications. For four samples, 4.5 ϫ 10 7 U937 cells were centrifuged for 5 min at 500 ϫ g and washed with cold PBS (137 mM NaCl, 2.7 mM KCl, 10.1 mM Na 2 HPO 4 , 1.8 mM KH 2 PO 4 , pH 7.4). The cell pellet was lysed in 2 ml of radioimmune precipitation assay (RIPA) buffer (50 mM Tris-HCl, pH 7.5, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.05% sodium dodecyl sulfate, 1 mM EDTA, 150 mM NaCl) and centrifuged for 10 min at 16,000 ϫ g and 4°C. All buffers were supplemented with RNaseOut (Invitrogen) and a protease inhibitor mixture (Complete; Roche Diagnostics) as suggested by the suppliers. The beads were pretreated according to the manufacturer's recommendations and incubated in antibody dilutions. The concentration of the HuR antibody or the IgG control was 5 g in 200 l of 150 mM NaCl solution per sample. 500 l of the samples were added to the pretreated beads, incubated on a rotator for 1 h at room temperature, and washed according to the manufacturer's protocol. RNA extraction was performed using the Qiazol reagent (Qiagen). 500 l of Qiazol were added to the beads. After adding 100 l of chloroform, samples were thoroughly vortexed and centrifuged at 16,000 ϫ g and 4°C for 10 min. The aqueous phase was transferred to a tube containing 6 l of linear acrylamide (5 mg/ml, Ambion), 60 l of 5 M ammonium acetate, and 600 l of isopropyl alcohol followed by vortexing and precipitation at Ϫ80°C overnight. Tubes were thawed on ice and centrifuged at 16,000 ϫ g and 4°C for 10 min. The pellet was washed once with 0.5 ml 70% ethanol, dried, and dissolved in 20 l RNase-free water.
RNA Isolation, Reverse Transcription, and Real-time RT-PCR-Total RNA from cultured cells was extracted using either Qiazol (Qiagen) or the High Pure RNA Isolation kit (Roche Diagnostics) as recommended by the manufacturers. After removing residual DNA (DNA-free kit, Applied Biosystems), 1 g of total RNA was reverse-transcribed using the High Capacity cDNA Reverse Transcription kit (Applied Biosystems). DNase treatment was skipped for IP samples. Quantitative realtime RT-PCR was performed using the CFX96 Touch™ Real-Time PCR Detection System (Bio-Rad), gene-specific primers (Eurofins MWG Operon, Table 1), and 5 ϫ HOT FIREPol EvaGreen qPCR Mix Plus. Standards from 60 to 0.0006 amol of the PCR product cloned into pGEMTeasy (Promega) were run alongside the samples to generate a standard curve. Alternatively, a cDNA dilution series were used. All samples and standards were analyzed in triplicate, except for IP samples, which were measured in duplicate. The PCR conditions were 95°C for 15 min followed by 45 cycles of 20 s at 95°C, 20 s at the primer-specific annealing temperature, and 20 s at 72°C.
Griess Assay-RAW 264.7 cells were cultured in 96-well plates and treated with LPS (0.1 g/ml) in the presence or absence of curcumin (3.1 and 6.3 M) or DMSO as a solvent control. After 20 h, the concentration of nitrite, a stable metabolite of NO, was measured in the supernatants by Griess assay as described previously (58). Briefly, 90 l of 1% sulfanilamide in 5% H 3 PO 4 and 90 l of 0. 1% N-(1-naphthyl)ethylenediamine dihydrochloride in H 2 O were added to 100 l of the cell culture supernatant followed by absorbance measurement at 550 nm and the reference wavelength 630 nm using a Tecan Sunrise microplate reader. A standard curve of sodium nitrite dissolved in medium was used to determine nitrite concentrations. GILZ knockdown was performed by transfecting RAW264.7 cells with control siRNA or siGILZ (40 nM) using the jetPrime reagent (PolyPlus transfection) as recommended by the supplier. Knockdown efficiency was assessed by flow cytometry 24 h after transfection (siGILZ: 29.7 Ϯ 7.8% GILZ expression when compared with siCo (ϭ100%); p Ͻ 0.001).
TNF Bioassay-The TNF bioassay was performed as previously described (20,59). In brief, L929 cells were seeded at a density of 3 ϫ 10 4 cells per well into a 96-well microplate. After 24 h, medium was replaced by actinomycin D solution (1 g/ml in standard medium). After 1 h of preincubation with actinomycin D, BMM supernatants were added. Dilution series of recombinant murine TNF-␣ (100 -2500 pg/ml, Biomol) in standard medium were run alongside the samples to generate a standard curve. The plates were incubated for an additional 24 h at 37°C. The MTT assay was used to assess cell viability.
GR Translocation Assay-HEK 293T cells were transfected with pEGFP GR vector (glucocorticoid receptor tagged with GFP) by electroporation using the Amaxa nucleofection device (Lonza) and were grown in -Slide VI 0.4 (ibidi). After 24 h, cells were analyzed by live microscopy (60). The cells were analyzed with an Axio Observer Z1 epifluorescence microscope equipped with an AxioCam Mrm, Incubator Xlmulti S1, Temp-Modul S1, and CO 2 Modul S1 (Zeiss). All cell images were obtained using a Fluar 40ϫ/1.30 Oil M27 objective. All videos were recorded in a humidified atmosphere at 37°C and 5% CO 2 . Data were obtained and analyzed by the AxioVision software (Zeiss). For nuclei staining, Hoechst 33342 (2 M) was added 5 min before the acquisition. Cells were either treated with 50 M curcumin or 100 nM dexamethasone. Video recording was started directly after drug application.
Polysomal Fractionation-8 ϫ 10 6 U937 cells were seeded into 75-cm 2 flasks. After 1 day, the cells were treated with 6.25 M curcumin or vehicle control for 3 h followed by polysomal fractionation as described previously (61). Briefly, after incubation with 100 g/ml cycloheximide (CHX) for 10 min at 37°C, cells were harvested in PBS/CHX (100 g/ml) and lysed in 750 l of polysome buffer (140 mM KCl, 20 mM Tris-HCl, pH 8.0, 5 mM MgCl 2 , 0.5% Nonidet P-40, 0.5 mg/ml heparin, 1 mM DTT, 100 units/ml RNasin (Promega), 100 g/ml CHX). After pelleting, the cytoplasmic lysates were layered onto 11 ml of 10 -50% continuous sucrose gradients. The gradients were centrifuged at 35,000 rpm for 2 h at 4°C without brake using an SW40 rotor in a Beckman ultracentrifuge. Thereafter, the gradients were collected in 1-ml fractions using a Gradient Station (Biocomp). Absorbance was measured at 254 nm. RNA was precipitated by the addition of sodium acetate (3 M) and isopropyl alcohol and further purified using the Nucleospin RNA kit (Macherey-Nagel) according to the manufacturer's manual. mRNA derived from polysomal fractions or total RNA extracted from the gradients was reverse-transcribed using the Maxima First Strand cDNA synthesis kit (Thermo Fisher Scientific). GILZ mRNA was subsequently quantified by real-time RT-PCR using the iQ SYBR Green Supermix (Bio-Rad).
Statistical Analysis-Data are expressed as the means Ϯ S.E. Statistical significances between multiple groups were determined either by one-way analysis of variance with Bonferroni's post hoc test for normally distributed values or the Mann-Whitney test for data that were not normally distributed. Comparison of normally distributed data in data sets that only contained two groups was performed using Student's t test. Values with p Ͻ 0.05 were considered significant. ., and C. R. contributed to drafting the manuscript. S. B. and C. R. provided materials and discussed the data. A. K. K. initiated the study and participated in data interpretation and manuscript preparation. All authors read and approved the final manuscript.