Regulation of S100A8 by Glucocorticoids
http://www.100md.com
免疫学杂志 2005年第4期
Abstract
S100A8 (A8) has roles in inflammation, differentiation and development and is associated with oxidative defense. Murine A8 (mA8) is up-regulated in macrophages, fibroblasts, and microvascular endothelial cells by LPS. Glucocorticoids (GCs) amplified LPS-induced mA8 in these cells. Relative to stimulation by LPS, GCs increased mA8 gene transcription and mRNA half-life. Enhancement required new protein synthesis, IL-10 and products of the cyclooxygenase-2 pathway, and both ERK1/2 and p38 MAPK. Protein kinase A positively and protein kinase C negatively regulated this process. Promoter analysis indicated element(s) essential for LPS and dexamethasone enhancement colocated within the region –178 to 0 bp. In the absence of glucocorticoid response elements, NF1 motif at –58 is a candidate for mediation of enhancement. Gel shift analysis detected no differences between LPS- and LPS/dexamethasone-treated complexes within this region. GCs increased constitutive levels of A8 and S100A9 (A9) mRNA in human monocytes. The synovial membrane of rheumatoid patients treated with high dose i.v. methylprednisolone contained higher numbers of A8/A9-positive macrophages than pre- or posttreatment samples. Results support the proposal that A8 has anti-inflammatory properties that may be independent of hetero-complex formation with A9 and may also enable localized defense in the absence of overriding deleterious host responses.
Introduction
S100A8 (A8,5 MRP8, calgranulin A) is a member of the S100 family of calcium-binding proteins (22 members; Ref.1), generally coexpressed with S100A9 (A9, MRP14, calgranulin B). A8 and A9 accumulate in inflammatory fluids and are postulated to be involved in the pathogenesis of chronic human inflammatory diseases such as rheumatoid arthritis (RA), cystic fibrosis, Crohn’s disease, ulcerative colitis, allergic dermatitis, and infection (reviewed in Ref.2). A8 and A9 are expressed in macrophages in the lining and sublining layers of the rheumatoid synovium. Expression at the cartilage pannus junction, the site at which the inflamed synovial pannus invades cartilage and bone, had a different pattern from more distal sites, suggesting that macrophage activation or differentiation differ at various sites (3). A8 and A9 comprise 40% of the cytosolic protein of neutrophils and their intracellular expression has been associated with calcium sensing (4), myeloid cell differentiation, and a variety of inflammatory processes, particularly those involved in neutrophil activation and arachidonic acid transport (reviewed in Ref.5). A8 and A9 are expressed at low levels by monocytes but not normally by macrophages. Murine A8 (mA8) in elicited macrophages is moderately up-regulated by proinflammatory mediators such as LPS (6), TNF, and IFN- (7). LPS and IL-1 are key regulators of the gene in microvascular endothelial cells (MEC) (8) and fibroblasts. Induced A8 is mostly confined to the cytoplasm and/or nucleus in activated MEC (8) and keratinocytes (9) whereas it is cell-associated and secreted by activated macrophages (7, 10).
Several S100 proteins have important extracellular functions (11) and roles for A8 and A9 in regulating leukocyte transmigration are proposed (5, 12, 13). Neutrophils from A9-null mice are also deficient in A8 and have reduced/compromised chemotactic responses (4, 14). A9 and the A82-A9 complex (calprotectin) induce IL-8 production by airway epithelial cells, hence potentially amplifying neutrophilic inflammation in chronic pulmonary disease (15). Deletion of the A8 gene is embryonic lethal shortly after implantation, indicating distinct functions for A8 in development (16). Some functions of A8 are calcium-dependent and depend on hetero-dimerization with A9, e.g., arachidonic acid transport and antimicrobial defense (5).
We proposed a protective role for A8 in oxidative defense in acute inflammation (12). A8 is a potent scavenger of hypochlorite (17) and the A8/A9 complex may be an intracellular scaffold for the NADPH oxidase system and potentiate activity by acting as a reservoir for arachidonic acid (18). Thus these proteins may regulate oxidative events associated with inflammation. This is strongly supported by the fact that the A8 gene is up-regulated by anti-inflammatory mediators and oxidative stress such as UVA- and H2O2 (9, 10). IL-10 synergizes with LPS to markedly enhance A8 mRNA expression and protein secretion in stimulated murine macrophages, and endogenous IL-10 induced by LPS mediates induction by LPS. The synergy may depend on generation of mediators such as PGE2 and cAMP (10). We propose that the low levels of A8 induced by proinflammatory mediators may serve a homeostatic function, whereas higher amounts may reduce oxidative damage during resolution.
Glucocorticoids (GCs) are traditionally viewed as anti-inflammatory and immunosuppressive, and are commonly used to treat numerous inflammatory conditions, including allograft rejection, autoimmune diseases and allergic conditions (reviewed in Refs.19, 20). Monocyte/macrophage functions down-regulated by GCs include induction of cytokines such as IL-1, IL-6, IL-8, and TNF-, which generate positive inflammatory reactions, generally via transrepression of transcription factors such as NF-B and AP-1, by the GC-GC receptor (GR) complex (reviewed in Ref.21). Conversely, GCs may induce anti-inflammatory cytokines such as IL-4, IL-10, and TGF (21) and type II IL-1R (22), mostly mediated by GC response elements (GRE) in the promoter regions of these genes. GCs released via the neuroendocrine-immune network during stress may also act as selective immunomodulatory enhancers during infection and tissue invasion to induce genes such as macrophage migration inhibition factor (23), a critical modulator of innate immune responses.
Here we show that GCs have little direct effect, but markedly up-regulate LPS-induced A8 in murine exudate macrophages, MEC and fibroblasts. Moreover, dexamethasone (DEX) increased constitutive A8 and A9 mRNA in human monocytes. In support of data generated in vitro, increased numbers of macrophages expressing A8 and A9 were found in the synovial membrane of rheumatoid patients treated with high doses of i.v. ("pulse") methylprednisolone (MP), and these returned to pretreatment levels when treatment ceased. Our results indicate complex regulation of the A8 gene by GCs and increased mRNA stability and transcription both contribute. Promoter analysis indicated that NF1 may regulate GR transactivation of the mA8 promoter. Requirements for IL-10 and products of the cyclooxygenase-2 (COX-2) pathway are essential for A8 induction by LPS (10), and both were required for GC synergy. We suggest that interactions between the MAPK, protein kinase A (PKA), and protein kinase C (PKC) pathways may provide an additional level of control in this system. Our results support the proposal that A8 has anti-inflammatory properties that are independent of hetero-complex formation with A9 and may be important in Gram-negative sepsis, where high endogenous GCs levels and endotoxin coexist.
Materials and Methods
Cells and reagents
The murine monocyte-macrophage cell line RAW 264.7 (American Type Culture Collection; ATCC TIB 71) was a gift from Dr. D. Hume (University of Queensland, Brisbane, Australia). Cells were cultured at 37°C in 5% CO2 in air in RPMI 1640 (Invitrogen Life Technologies) supplemented with 127 U/ml penicillin, 127 μg/ml streptomycin (Sigma-Aldrich), and 10% heated (56°C, 30 min) bovine calf serum (HyClone) filtered through 0.2-μM Zetapore membranes (Cuno), hereafter referred to as cell culture medium (CM). Thioglycolate (TG)-elicited murine macrophages were obtained as described (7). RAW 264.7 cells (1.5 x 106) in 5 ml of CM were stimulated and harvested as indicated. Medium and mediators were only used if endotoxin levels were <20 pg/ml (chromogenic limulus amoebocyte assay; Associates of Cape Cod).
Plastic-adherent fibroblast short-term cell lines (splenic "primary" fibroblasts) were isolated and cultured as described (24) with some modification. Briefly, spleens from QS mice were minced, digested for 20 min at 37°C in collagenase (Sigma-Aldrich; 2 mg/ml), then passed through a tissue strainer (70 μM, BD Biosciences), and washed (3x CM). Nonadherent cells and debris were moved by replenishing CM after 24, 48, and 72 h. Cells from passages 2 through 8 were used in the experiments. Primary fibroblasts obtained by this method do not contain macrophages or other cells of hemopoietic origin (24) and had typical fibroblast morphology.
The murine microvascular brain endothelioma cell line (bEnd-3), a gift from Dr. W. Risau (Max-Planck Institute, Tübingen, Germany), was grown and used when postconfluent (6–7 days) as described (8).
Human blood monocytes were enriched using the monocyte isolation kit (Miltenyi Biotec). Briefly, PBMC from venous blood from healthy volunteers isolated on Lymphoprep (Nycomed) were suspended in CM. To avoid neutrophil contamination and stimulation of monocytes, the latter were isolated by negative selection. PBMC were incubated with a mixture of biotinylated CD3, CD7, CD16, CD19, CD56, CD123, and CD235a Abs and anti-biotin MicroBeads, and nonmonocytes were depleted using MACS (Miltenyi Biotec) with LS columns. Differential staining indicated that the population was 96% monocytes with viability >95% by trypan blue staining. Monocytes (5 x 105) were aliquoted into Nunc-minisorb tubes (Nunc) and incubated overnight at 37°C in 5% CO2 in air before stimulation.
Anti-mouse IL-10 mAb was from R&D Systems. DEX, hydrocortisone (HYD), corticosterone, cycloheximide (CHX), actinomycin D (ActD), calphostin C (CalC), PD98059, SB202190, and H89 were from Sigma-Aldrich. LPS (Escherichia coli, 055:B5) was from Difco.
Northern analysis
Northern blotting and preparation of mA8 riboprobe and 18S rRNA oligo probe were as described (7). Phosphorimage and densitometry analysis was performed using the Bio-Rad MultiAnalyst/Macintosh Version 1.0 software and the Bio-Rad Molecular Imager GS-525 system (Bio-Rad Laboratories).
Real time quantitative PCR analysis of A8 and A9
Total RNA (1 μg) from human monocytes was treated with DNase 1 (Ambion, Austin TX) and reverse transcribed using random hexamers and the SuperscriptIII First-Strand Synthesis System for RT-PCR (Invitrogen Life Technologies). Negative controls (no first strand synthesis) were prepared by performing reverse transcription reactions in the absence of reverse transcriptase. PCR amplification was performed with Platinum SYBR Green qPCR SuperMix UDG (Invitrogen Life Technologies). Reactions were performed in duplicate containing 2x SYBR Green qPCR SuperMix, 1 μl of template cDNA or control, 100 nM of human A8 primers (5' primer: GGGATGACCTGAAGAAATTGCTA, 3' primer: TGTTGATATCCAACTCTTTGAACCA) or human A9 primers (5' primer: GTGCGAAAAGATCTGCAAAATTT, 3' primer: GGTCCTCCATGATGTGTTCTATGA) in a final volume of 25 μl, and analyzed in 96-well optical reaction plates (Applied Biosystems). Reactions were amplified and quantified using an ABI 7700 sequence detector with standard cycle conditions and the Applied Biosystems software. Relative quantities of A8 and A9 mRNA were obtained using the comparative CT method and were normalized against human hypoxanthine-guanine phosphoribosyltransferase (HPRT: 5' primer: TCAGGCAGTATAATCCAAAGATGGT, 3' primer: AGTCTGGCTTATATCCAACACTTCG) as an endogenous control.
Detection of mA8 protein
A double-sandwich ELISA using rabbit polyclonal anti-mA8 IgG was performed using cell-free supernatants as described (7), with serial dilutions of recombinant mA8 as standard.
Nuclear run-on transcription assay
Nuclear run-on transcription was performed as described (25). Briefly, nuclei from 2 x 107 PBS-washed RAW cells were isolated in ice-cold lysis buffer (10 mM Tris-HCl (pH 7.4), 10 mM NaCl, 3 mM MgCl2, 0.5% Nonidet P-40). RNA transcription was initiated by the addition of 100 μl of reaction buffer (10 mM Tris-HCl (pH 8.3); 40% glycerol; 5 mM MgCl2; 100 μM EDTA; 2.5 mM DTT; 500 μM ATP, CTP, and GTP; and 3.3 μM [32P]UTP (50 μCi)), and incubation at 30°C for 30 min. The reaction was stopped with solution D, and total RNA was extracted. The aqueous phase was precipitated by the addition of tRNA and isopropanol. After washing in 70% ethanol, the pellet was resuspended in 10 μl of Tris-EDTA buffer (pH 7.3) then 200 μl of formamide. Equal amounts of radioactivity from each sample were hybridized with Hybond membranes to which 1 μg of mA8 cDNA, 1 μg of linearized plasmid containing the cDNA for HPRT, and 1 μg of linearized vector only, were immobilized by slot blot transfer. Hybridization was performed at 42°C for at least 36 h in 50% formamide, 5x standard saline citrate phosphate/EDTA, 5x Denhardt’s, 0.5% SDS, 50 μg/ml denatured herring sperm DNA, and 200 μg/ml tRNA. After washing filters to a stringency of 1x SSC, 0.1% SDS at 55°C, phosphorimage and densitometry analysis was performed using the Bio-Rad Molecular Imager GS-525 system.
Construction of luciferase reporter plasmids
Truncated promoter fragments of the mA8 gene were produced by nested deletion of a PCR-amplified product as described (10). A series of 5' and 3' deletion mutants was generated by nested-deletion using the double-stranded Nested Deletion Kit (Pharmacia Biotech). The 5' deletion mutants (with a common 3' end at +465) yielded the pCP-x/+465 constructs (x represents different 5' ends). The 3' deletion mutants (with a common 5' end at –316) yielded the pCP-316/±y constructs (y represents different 3' ends, see Fig. 5A). The extent of deletion in all clones was verified by sequencing.
FIGURE 5. A8 promoter analysis in response to LPS ± DEX. A, Luciferase activity of mA8 deletion constructs in the presence of LPS ± DEX. A8-luciferase chimeric gene constructs are shown below a partial restriction map of the 1053-bp A8 gene fragment. The name of each construct is given on the right. Solid lines represent A8 DNA fragments, open boxes (LUC) at the 3' end indicate the luciferase gene. The numbers below the map represent the positions of restriction sites relative to the transcription start site (+1). PCP-178/0-NF1 has the mutation at position –75 to –80. A8-luciferase fusion constructs were transfected into RAW cells in parallel with control plasmid pGL2-Control as described in Materials and Methods and stimulated with LPS (500 ng/ml) either alone or with DEX (10–6 M). Results are means of duplicate determinations from two experiments. Error bars represent mean ± SD. B, Gel shift analysis of the region –173 to –1 of the mA8 gene in the presence of LPS ± DEX. RAW cells were incubated with LPS (500 ng/ml) ± DEX (10–6 M) and nuclei prepared after 18 h, and extracts were used for gel shift analysis as described in Materials and Methods.
Site-directed mutagenesis
Site-directed mutagenesis was conducted with the QuickChange Kit (Stratagene). The sequence of the mutagenic primer (sense strand) for the NF1-like motif was: CGAGGCTGcataCAGCTGtaCgAGCTTTC (mutated bases are indicated in lowercase letters).
DNA transfection and luciferase assay
DNA transfection was according to Stacey et al. (25) with some modifications. RAW cells (6 x 106) cultured in CM and harvested in logarithmic growth phase were electroporated using a Bio-Rad Gene Pulser (260 V, 960 μF) in 0.4-cm electroporation cuvettes (Bio-Rad). Each recombinant plasmid was present at the same molar concentration as 10-μg pGL2-Control plasmid (Promega). To assess mA8 promoter inducibility, 2 h after electroporation cells were stimulated with 500 ng/ml LPS ± DEX and harvested after 18 h. Luciferase activity was measured using the Luciferase Assay System (Promega) in a TD-20/20 Luminometer (Turner Designs). Luciferase activity was normalized to total protein in cell lysates (bicinchoninic acid protein assay kit; Sigma-Aldrich); values generated were reproducible and constant, relative to pGL2-control.
Gel shift analysis
RAW cells (8 x 106) were stimulated with the appropriate mediators for 18 h and extracts for gel shift analysis isolated according to Osborn et al. (26). The 17-bp promoter region was produced using PCR and the pair primers (GCCGGTACCCATTCCTCAGACTCAGAAATGAAATGCTC and CGGGAGATCTATGTGAGGCTAAGTGTCAGCTGC) following purification on a QIAquick PCR purification column (Qiagen). The fragment was labeled using T4 polynucleotide kinase and purified on a NICK column (Amersham Pharmacia). Binding assays were performed in 20 mM HEPES (pH 7.9), 1 mM MgCl2, 4% Ficoll, 6% glycerol, 35 mM NaCl, 21 mM KCl, 0.5 mM DTT, 10 fmol of double stranded oligonucleotide, and appropriate amounts of nuclear proteins and herring sperm DNA. After 30 min on ice, the binding reaction was separated on discontinuous 6% PAGE with Tris-glycine buffer. Dried gels were phosphorimaged using the Bio-Rad system.
MP treatment and synovial biopsies
Twelve patients fulfilling the American College of Rheumatology criteria (1987) for RA were studied. All received 1000 mg of MP i.v. as the sodium hemisuccinate salt (Upjohn Kalamazoo) as described (27, 28). Clinical and laboratory assessments of inflammation were performed. Synovial membrane samples were obtained before and 24 h after MP in 12 patients and at disease relapse (between 6 and 12 wk later) in 6 patients, using needle arthroscopic techniques. This project was approved by the Repatriation General Hospital Ethics Committee with informed consent from each patient.
Immunostaining and analysis of synovial tissue
Serial sections were stained with monospecific affinity-purified rabbit anti-A8 or -A9 (29) using a standard three-stage immunoperoxidase method as described (30). Sections were coded and randomly analyzed by a blinded observer (P.Y.). Only synovium in which the lining layer was clearly evident was studied. Sections were examined at x400 magnification using a 1-mm graticule. All high power fields, except for those in which fibrous subsynovial conective tissue predominated, were included in the analysis. Positively stained cell counts were expressed as mean counts per square millimeter using a conversion factor of 0.0625–1. Results are given as mean (SEM - SEM) or mean (range). StatviewSE +Graphics statistical software (Abacus Concepts) was used. Means were compared using paired t tests. Differences were considered to be significant at p < 0.05.
Results
GCs enhance mA8 induction
mA8 transcript levels in macrophages, after exposure to corticosterone, HYD or DEX in the presence or absence of LPS were assayed by Northern blotting. DEX did not directly induce mA8 mRNA but markedly enhanced levels stimulated by LPS. Synergy was counteracted by RU486, a synthetic anti-GC that competes for binding of DEX to GR (Fig. 1A). Similar results were observed with corticosterone (data not shown) and HYD (Fig. 2, A and B). The steroid derivatives estradiol and 1,25-(OH)2 vitamin D3 neither directly induced expression nor enhanced expression induced by LPS (not shown). In contrast to mA8, LPS-induced TNF- mRNA was suppressed by DEX, and suppression was totally abolished by RU486 (Fig. 1A). Identical results were obtained with the RAW 267.4 macrophage cell line (not shown). A9 mRNA was present in bone marrow cells but never detected by Northern blotting (not shown). Quantitative RT-PCR detected very low levels of A9 mRNA in RAW cells and macrophages stimulated with LPS (>10–5-fold less than mRNA levels of A8); DEX neither induced nor enhanced the levels (three experiments, data not shown). To test whether synergy was cell-specific, primary fibroblasts and MEC were tested. mA8 mRNA levels in LPS-stimulated fibroblasts increased 7.3-fold with 10–7 M DEX, and 3.5-fold with 10–8 M (Fig. 1B). Similar results were obtained with the 3T3 cells, and synergy was suppressed with RU486 (not shown). When MEC were costimulated with LPS, up-regulation by DEX was optimal with 10–7–10–6 M DEX (Fig. 1C).
FIGURE 1. Induction of A8 mRNA by LPS and GCs. A, DEX (10–6 M) enhanced LPS-induced (100 ng/ml) A8 mRNA in TG-elicited macrophages incubated with stimulants for 20 h. RU486 (1 μM) counteracted the enhancing effect of DEX. Total RNA was analyzed by Northern blotting. Filters were rehybridized with the TNF probe and an 18S rRNA oligoprobe after stripping. Data quantitated by phosphorimaging are represented as bar charts. B, Fibroblasts were stimulated for 24 h with LPS (100 ng/ml) ± DEX at the doses indicated. C, MEC were incubated with LPS (500 ng/ml) ± DEX for 24 h. Experiments were performed at least three times with similar results. Ctrl, unstimulated control.
FIGURE 2. GCs enhance LPS-induced A8 up-regulation in macrophages. A, DEX (10–6 M) or HYD (10–6 M) enhanced A8 mRNA induced by different doses of LPS in TG-elicited macrophages stimulated for 20 h. B, TG-elicited macrophages were incubated with LPS (100 ng/ml) ± DEX or HYD at the concentrations indicated for 20 h. Total RNA was analyzed by Northern blotting. C, A8 protein (nanomolar) in supernatants of RAW cells incubated with LPS (100 ng/ml) ± DEX or HYD at the concentrations indicated, for 20 h quantitated by ELISA. At least two separate experiments yielded similar results. D, RAW cells were incubated with LPS (100 ng/ml) ± DEX (10–6 M) ± anti-IL-10 (IL10Ab, 2 ng/ml). Total RNA samples harvested at the times indicated were analyzed by Northern blotting. E, A8 in supernatants generated in D were quantitated by ELISA. Values represent means ± SD; n = 3.
Synergy was most evident when GCs were included with low concentrations of LPS. The low levels of mA8 mRNA induced with 20 ng/ml LPS markedly increased when DEX or HYD were included; enhancement was less obvious with high amounts of LPS (1000 ng/ml; Fig. 2A). Optimal up-regulation was between 10–7 and 10–6 M GCs (Fig. 2B). In accordance with these results, 3- to 4-fold more mA8 was secreted by macrophages when 10–6 or 10–7 M DEX or HYD were included with LPS (Fig. 2C). Similar sensitivity was observed with RAW cells and because these exhibited identical responses to exudate macrophages, these were used for more detailed analysis as they are a reliable source of large numbers of cells, and unlike primary macrophages, can be readily transfected.
Addition of GCs did not alter the kinetics of A8 mRNA induction by LPS; both were maximal at 20–24 h and decreased by 48 h (Fig. 2D). No detectable mA8 was secreted by unstimulated RAW cells but was evident after 12 h and reached peak levels of 35 ng/ml after 48 h stimulation with LPS. When DEX was included, mA8 levels peaked at 60 ng/ml (Fig. 2E). Taken together, these results show that GCs synergistically enhanced LPS-induced mA8 up-regulation at the mRNA and protein levels in macrophages, and the mRNA level in fibroblasts, and endothelial cells, in a dose-dependent manner via a GR-dependent mechanism.
GCs increase mA8 mRNA stability and transcription rate but do not alter duration of mRNA expression
To determine whether GCs affected mA8 mRNA stability, ActD, at a concentration that inhibits mA8 transcription (5 μg/ml) (6), was used in the presence/absence of GCs 20 h after addition of LPS, and cells harvested at 4 hourly intervals over 24 h. The mRNA half-life in the presence of LPS was 8 h, and in the presence of LPS and HYD (10–6 M) was maintained over 20 h, representing an approximate increase in stability of 2.5-fold (Fig. 3A). Similar results were obtained with DEX (not shown).
FIGURE 3. Mechanisms of enhancement by GCs on LPS-induced A8 in macrophages. A, RAW cells were incubated with LPS (500 ng/ml) ± HYD (10–6 M). ActD (5 μg/ml) was added 20 h after stimulation with LPS; samples harvested at 4 hourly intervals were analyzed by Northern blotting. The "LPS" samples were phosphorimaged for longer than the "LPS+HYD" samples to equalize the zero time-point band intensity. B, RAW cells were incubated for 18 h with LPS (500 ng/ml), DEX (10–6 M), or both, and subjected to nuclear run-on analysis. HPRT cDNA was used to normalize the overall rate of transcription; A8 cDNA vector alone was the negative control. C, Effect of time of addition of DEX on LPS-induced A8 mRNA. RAW cells were treated with LPS (100 ng/ml) for 24 h (). DEX (; 10–6M) was added at the indicated times relative to addition of LPS. The numbers represent 24, 3, or 1 h before, with (0 h), or 3, 6, 9, or 21 h after addition of LPS. Relative mRNA levels are presented as fold increases in A8 mRNA above those induced by LPS alone. Total RNA samples harvested at the times indicated were analyzed by Northern blotting. Data represents means ± SD of three experiments. *, p < 0.05; **, p < 0.01. D, Effects of CHX (2 μg/ml) on DEX (10–6 M)-mediated synergy of LPS (100 ng/ml)-induced A8 mRNA expression. Northern and nuclear run-on results are representative of three experiments.
The effects of LPS and DEX on the rate of mA8 gene transcription were assayed by nuclear run-on analysis (Fig. 3B). The very low basal transcription rate was not altered by DEX (10–6 M), but as reported (6), was increased by LPS (500 ng/ml). LPS and DEX together increased transcription which, from densitometry readings normalized to the rate of HPRT transcription, was 3-fold higher than that induced by LPS alone.
Enhancement depends on new protein synthesis and occurs via the COX-2-cAMP pathway
Many DEX-mediated effects occur via induction of new genes. We reasoned that if DEX increased the LPS response via new protein synthesis, synergy may be enhanced in macrophages pretreated with DEX, and blocked when costimulated with a protein synthesis inhibitor. Maximum mA8 mRNA induction in RAW cells occurred when cells were preincubated with DEX for 21 h before addition of LPS (3.8-fold greater increase than when DEX was added simultaneously with LPS), and this increment decreased when preincubation was reduced. The effect was biphasic, with a second small, but significant increase (1.7-fold compared with DEX and LPS added at the same time; p < 0.05) when DEX was added 6 h after LPS (Fig. 3C). RAW cells treated with LPS ± DEX and CHX, which efficiently inhibits protein synthesis and mA8 induction by LPS (10), did not express mA8 mRNA above control levels (Fig. 3D), confirming the requirement for new protein synthesis.
Because IL-10 strongly up-regulates LPS-stimulated mA8 expression in macrophages (10), and DEX induces IL-10 in some cell types (31, 32), we tested IL-10 involvement. When RAW cells were cultured with LPS+DEX in the presence of anti-IL-10, enhancement was suppressed to levels observed with LPS alone (Fig. 2D). Suppression was best reflected in secreted protein levels, which were only 30% of levels induced by LPS alone (Fig. 2E). Anti-IL-10 inhibited LPS-induced mA8 mRNA by 78% (Fig. 4A). More importantly, levels of IL-10 secreted by RAW cells treated with DEX+LPS or with LPS alone for 6 h were similar (54.3 vs 45.3 pg/ml, n = 2) and LPS-induced IL-10 at 24 h decreased by 40% when DEX was included (664.2 vs 429.2 pg/ml, respectively; n = 2). Thus, DEX synergy was unlikely to be due to increased levels of IL-10 in supernatants and the effect of anti-IL-10 may have been primarily on LPS-induced mA8 expression.
FIGURE 4. Effect of signal transduction inhibitors on DEX-enhanced A8 expression. A, RAW cells were incubated with LPS (100 ng/ml), DEX (10–6 M), or LPS + DEX for 20 h in the presence/absence of anti IL-10 (5 ng/ml) and/or COX inhibitors (indomethacin, Indo, 10 μM; NS389, 50 μM). B, Effect of DEX on LPS-induced A8 in RAW cells pretreated for 1 h with SB202190 (1 μM), PD98059 (10 μM), H89 (1 μM), and CalC (0.2 μM). Total RNA analyzed by Northern blotting. Values are means ± SD; n = 3. *, p < 0.05; **, p < 0.01 vs corresponding control (A) and LPS alone (B).
Because PGE2 and cAMP also synergize with LPS and are involved in mA8 induction (10), these pathways were tested. The COX inhibitor indomethacin, and the more specific COX-2 antagonist NS389, inhibited LPS-induced mA8 by 45%, whereas 60% of LPS+DEX-induced mRNA was abrogated, indicating that the COX-2 pathway may be involved in up-regulation by DEX. When anti-IL-10 and COX inhibitors were used together, inhibition was additive; 91% of LPS-induced mRNA was abolished and 82% of the LPS+DEX-induced mRNA levels were inhibited, suggesting the two signaling pathways may work independently.
Inhibition of enhancement by DEX was observed when H89, which decreases cAMP and PKA activity, was included. In contrast, the specific PKC antagonist CalC enhanced the DEX effect (Fig. 4B). These results implicate the cAMP pathway in the synergy and the PKC pathway may regulate this. The MAPK pathway is also involved in LPS induction of the mA8 gene in macrophages (7) and regulation by GCs can occur via p38 MAPK activation (33). The ERK1/2 (PD98059) and p38 (SB202190) inhibitors both abrogated LPS-induced A8 mRNA. The p38 inhibitor considerably reduced the LPS+DEX signal, although the degree of inhibition was less than the effect on LPS alone. The ERK1/2 inhibitor almost totally abrogated both responses (Fig. 4B). The results suggest overlapping pathways involved in induction of A8 by LPS and enhancement by DEX.
GC and LPS responsive elements colocalize
Promoter deletion analysis obtained after LPS stimulation indicated elements which enhanced or inhibited promoter function arranged in an alternating array within intron 1 and upstream of exon 1. Intron 1 contained an interesting tandem array of apparently interacting enhancer and repressor elements. Although consecutive removal of intronic regions +322 to +465 (pCP-316/+322), then +179 to +322 (pCP-316/+179) increased luciferase activity in response to LPS 3-fold then resulted in complete loss of activity, subsequent removal of the adjacent 51 bp (pCP-316/+129) restored activity, comparable to that obtained with pCP-316/+322 (Fig. 5A). This may be due to a repressor, enhancer, repressor array in which the enhancer in +179 to + 322 specifically counteracts the repressor in +129 to + 179. As agrees with our previous report (10), the region –173 to 0 (pCP-173/0) conferred LPS-responsive expression, and so contains the essential promoter of the mA8 gene. The relative level of expression induced by LPS+DEX, compared with LPS alone, was similar for all LPS-inducible constructs, indicating that none of the deleted regions contained elements mediating GC-enhanced transcription. These elements must be restricted to the region –178 to 0 (Fig. 5A), and although this contains no classical GRE, candidate elements include AP-1 motifs (–164, –102, –18) and an NF1 motif (–58). Because NF1 mediates induction of several genes by GCs (34, 35), the NF1 motif was mutated at –58 bp (pCP-178/0NF1). Fig. 5A shows that this construct completely abolished the signal induced by LPS+DEX, whereas the LPS-induced luciferase activity was reduced by only 38%, indicating involvement of this NF1 binding site in coinduction by DEX.
Gel shifts were performed on the region –173 to –1 in an attempt to locate elements responsible for GCs enhancement. Extracts from unstimulated cells produced two complexes and in the presence of LPS or LPS+DEX, the intensity of both complexes were reduced (Fig. 5B). No difference between LPS and LPS+DEX extracts was observed using oligonucleotide probes spanning the 173-bp region although at least five separate complexes were evident (data not shown). The lack of detectable change upon addition of DEX indicates that enhancement may occur either by a factor present in low abundance, or by modification of a factor already present in the LPS-induced complex (for example by phosphorylation).
GCs directly induce A8 and A9 mRNA in human monocytes and A8/A9-positive cells in the rheumatoid synovial membrane
Because there is strong homology between the human and murine proximal A8 promoters (1), we tested the effect of DEX on human monocytes which constitutively express A8 mRNA and low levels of A8 and A9 proteins (36). Fig. 6A confirms A8 and A9 mRNA in unstimulated cells. LPS increased mRNA expression of A8 (2.1-fold) and A9 (1.8-fold). DEX increased A8 and A9 mRNA 5-fold, but no synergy was observed when combined with LPS.
FIGURE 6. Effects of GCs on the expression of A8 and A9 in human monocytes and macrophages in RA synovium. A, Human monocytes incubated with LPS (100 ng/ml) ± DEX (10–6 M). After 24 h, total RNA was harvested and real-time RT-PCR performed in duplicate and HPRT was used as a control for amounts of RNA used in each reaction. Values are means ± SD; n = 4. Ctrl, unstimulated. *, p < 0.05; **, p < 0.01 vs absence of DEX. B, Synovial knee biopsies were taken from 12 patients with RA before and 24 h after treatment with 1000 mg of i.v. MP, and from 6 patients at disease relapse. Tissues were immunostained for A8 and A9, and numbers of positively stained cells were quantified. MP significantly increased expression of A8 within 24 h (p = 0.01) and increased expression of A9 (p = 0.06). A8- and A9-positive cells returned approximately to pretreatment levels at disease relapse. C, Synovial lining from a knee joint from a patient with RA before (left) and after treatment (right) with high dose MP. Infiltrating A8+ macrophages stain red. Sections were counterstained with hematoxylin (magnification, x40). An unrelated rabbit IgG, used as control, was unreactive (data not shown).
High doses of i.v. (pulse) steroids are used to treat inflammatory disorders, and rheumatoid patients treated with MP provide an excellent model to investigate their effects. A8 and A9 expression increased in macrophages within the rheumatoid synovial membrane 24 h after MP in 11 of 12 patients treated and returned to approximately pretreatment levels at the time of clinical disease relapse (Fig. 6B), which was used as an indicator of when the effects of MP had ceased. Fig. 6C shows increased numbers of A8-positive macrophages in synovial lining from a patient treated with MP compared with the biopsy taken before treatment. Extracellular A8 was obvious in both tissues, indicating secretion.
Discussion
GCs are anti-inflammatory drugs commonly used to treat allergic and autoimmune diseases by controlling expression of specific target genes. They limit expression of numerous inflammatory cytokines and other proteins associated with leukocyte migration and adhesion, primarily by suppressing gene transcription. GCs also enhance transcription of a limited number of genes, the products of which generally have anti-inflammatory effects (reviewed in Ref.37).
Here we show that GCs had little direct effect, but enhanced mA8 gene expression in murine macrophages stimulated with LPS, whereas TNF mRNA was reduced. MEC and fibroblasts also exhibited synergy, indicating that the effect was not cell-specific. GCs enhanced constitutive basal mRNA levels of A8 and A9 in human monocytes although there was no synergy with LPS. DEX and 1,25-(OH)2 vitamin D3 also induce A8 in the human macrophage cell line, U937 (38). Importantly, A8/A9-positive macrophages increased in biopsies of rheumatoid synovium from RA patients, whereas this treatment markedly reduced proinflammatory mediators such as TNF-, IL-8 (27), cell adhesion molecules (28), and chemokines (39). Increased expression occurred when patients were clinically improved, and thus when the effects of MP were clinically anti-inflammatory and decreased at clinical relapse, when the anti-inflammatory effect of MP had ceased and when we previously observed increases in proinflammatory mediators (27, 28, 39).
Enhancement in murine macrophages occurred with pharmacologically (40) and physiologically relevant concentrations of GCs (10–6–10–8 M). For example, in certain pathologic states such as Cushing’s disease, and in tissues at sites of synthesis, steroids can occur at micromolar levels (41). Costimulation was obligatory in the three cell types and marked synergy in macrophages was observed with DEX, HYD, or corticosterone, and correlated with elevated levels of secreted mA8 protein. Consistent with our finding that mA9 is barely induced in macrophages by LPS (6), the virtually undetectable A9 mRNA in LPS/DEX activated macrophages confirms that A8 does not require A9 for stability or secretion, and contradicts this proposal, generated from observations of neutrophil function in A9-null mice (4, 14). However, in contrast to murine macrophages, DEX increased A9 mRNA in human monocytes in vitro, to levels similar to A8, indicating likely differences in regulation of A9 in particular cell types/species.
In striking contrast to our observations, DEX repressed A8 and A9 expression associated with phorbol ester-induced inflammation in the skin of wild-type mice, but not c-fos–/– mice, indicating AP-1 target genes for which phorbol ester-induced transcription down-regulated by GCs depends on c-Fos (42). The positive and negative regulatory effects of GCs on constitutive S100B mRNA expression in hippocampal astrocytes are also distinct from those observed for other S100 proteins (43) and for LPS-inducible A8. Our studies contribute to the accumulating evidence that steroids are key modulators of the S100 gene family and that the nature of the response may vary according to the cell type and mediator responsible for induction.
Unlike the synergy of mA8 expression seen with IL-10, which reduces the time required for maximal gene induction (10), DEX did not alter the kinetics of the LPS response. However synergy was highest in macrophages preincubated with DEX for 21 h and the requirement for new protein synthesis suggests generation of an essential cofactor. The magnitude of A8 induction in macrophages activated with LPS is dependent on IL-10, and here we demonstrate marked reduction in synergy using anti-IL-10, although this treatment also reduces the LPS-induced gene. Moreover, supernatants generated from DEX+LPS treated RAW cells contained only marginally more IL-10 after 6 h compared with LPS alone, and levels decreased after 12 h. However under certain conditions, IL-10 may facilitate A8 induction by GCs. In humans, the effect of DEX on LPS-stimulated IL-10 secretion by monocytes is dependent on DEX and LPS concentrations (44). Furthermore, infusion of HYD in normal subjects directly before LPS administration increased LPS-induced plasma levels of IL-10, whereas induction of hypercortisolemia before LPS injection did not modify IL-10 levels (45). Gradual increases in serum IL-10 levels were observed in patients with multiple sclerosis (32) and patients undergoing cardiac surgery (46) and treated with GCs.
Here we show that increased mRNA stability and transcription are both likely to contribute to synergistic up-regulation of the mA8 gene. GCs can alter mRNA stability (21), but there are no reports demonstrating increased stability of inducible S100 transcripts, as described here for mA8. The mA8 transcript contains no regions with similarity to motifs reported to alter mRNA stability of other genes, suggesting a novel mechanism.
Elements responsible for mediating GCs enhancement were located in the region –173 to 0 of the essential promoter in which there are no obvious GREs. Promoters for many genes involved in inflammation have no demonstrated GRE sequences essential for binding activated GR and this is clearly indicated in dimerization-defective GR mice (47). Regulation of expression of GCs-responsive genes reflects a collaboration of multiple transcription factors (reviewed in Ref.21). NF1 is involved in GCs-dependent chromatin remodeling and also in transcriptional activation of the murine mammary tumor virus promoter (34), and is likely to regulate transcription of mA8 by DEX and LPS. The mA8 NF1 motif is a 13/14 palindrome contained within a 21-bp region and is identical in the murine and human A8 genes (–58 to –38). Its location immediately upstream of the TATA box suggests that the NF1 motif at position –58 may activate or repress formation of the basal transcription complex (48). Because the DEX response was found with a variety of cell types, it is possible that NF1 may regulate GR transactivation of the mA8 promoter in a manner similar to that described by Hebbar and Archer (34) as a classical transcription factor and in alteration of chromatin. AP-1 can either enhance or repress GCs-responsive genes depending on whether it is composed of c-jun homodimers or jun/fos heterodimers respectively (21). AP-1 binding motifs are located at –164, –102, and –18 before the transcriptional start site and are highly homologous in the murine and human genes. Because jun homodimers enhance LPS-induced expression of mA8 (R. J. Passey and C. L. Geczy, unpublished results), GCs may enhance A8 through a related mechanism.
Enhancement by DEX was also strongly reliant on the COX-2 pathway, which leads to increased PGE2 and cAMP that also enhance LPS induction of mA8 (7) and of human A8/9 mRNA in myelocytic differentiation (49). Mechanisms of GC enhancement via this pathway are unknown, as GR normally suppresses COX-2 induction (21). Its involvement is likely to be independent of IL-10, as inclusion of inhibitors of both pathways reduced synergy to a greater extent than when only a single inhibitor was added. MAPK pathways are involved in both LPS up-regulation and DEX enhancement, whereas PKC and PKA have opposite effect on modulating mA8 gene expression by DEX. GCs reduce PKC activity in rat liver parenchymal cells (50), and increased levels of cAMP and/or activation of PKA can potentiate GC-dependent gene expression by increasing GR affinity (51). Although the GR was traditionally viewed as ligand-activated, activation can also occur via alternate mechanisms including protein-protein interactions and phosphorylation pathways including PKA and PKC (52) and MAPK, with or without other transcription factors (21). Alterations to conformation or degree of phosphorylation of the complex may not be detected by gel shift analysis (Fig. 5B). Interactions between the MAPK, PKA, and PKC pathways involved in A8 induction may provide an additional level of control in this system.
The therapeutic effects of GCs are largely due to their ability to inhibit functions of macrophages and other APCs. Interestingly, IL-10 and GCs have several effects in common, particularly affecting macrophage function, possibly via their interference with transcriptional activation of AP-1 and NF-B (21, 53). Conversely, both agents synergized with LPS to up-regulate mA8. GCs and IL-10 both induce GCs-induced leucine zipper, a protein that prevents NF-B-mediated activation of transcription, in macrophages activated via Toll-like receptors and proposed to play a key anti-inflammatory role (54). Our results support the proposal that A8 has anti-inflammatory properties which could be manifested in functions such as its ability to down-regulate leukocyte adhesion (55), to regulate NADPH oxidase (18), to scavenge reactive oxygen intermediates (17) and, in combination with A9, to act as an anti-microbial agent. The role of GCs-induced leucine zipper in A8 gene regulation is worthy of investigation.
Apart from potential therapeutic effects, the synergistic induction in macrophage, endothelial cell, and fibroblast activation of A8 may be important in Gram-negative sepsis, where high endogenous GCs levels (micromolar (56)) and endotoxin coexist. mA8 is up-regulated by LPS and recruits neutrophils and monocytes but fails to activate these cells (57) and interestingly, neutrophils in LPS-DEX-treated rats have these properties and cause potentially less damage (58). Up-regulation of A8 in several cell types by GCs may enable positive host defense against invading organisms in the absence of overriding deleterious host responses. These may represent counterregulatory systems pivotal to maintaining host defense and may positively compensate for some of the anti-inflammatory consequences of GC action.
The significant increase in A8-positive macrophages in synovial lining from RA patients treated with MP suggests that the A8 may play an anti-inflammatory role. Human A8 and S100A12 (A12) appear to be more recently evolved from mA8 which exhibits some functions of both these proteins. A12 does not occur in the murine genome (1) and A12 and mA8 are chemotactic for monocytes in the picomolar range. Human (13) and murine (12) A8 recruit neutrophils and both can scavenge oxidants (17), a property not reliant on heterodimer formation (17). Conversely, A12 is not an efficient oxidant scavenger because it lacks cysteine and methionine residues and interestingly, preliminary studies show that A12 mRNA in monocytes is not regulated by GC. Thus high levels of A8 induced by GCs may downmodulate inflammation by processes involving oxidant defense, a property now suggested for several S100 proteins.
Acknowledgments
We are grateful to Dr. Zheng Yang and Dr. Shengping Hu for assistance with reporter analysis in initial experiments and to Dr. Johannes Roth (Institute of Experimental Dermatology and Department of Pediatrics, University of Munster, Munster, Germany) for performing immunostaining.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This study has been supported by grants from the National Health and Medical Research Council of Australia.
2 Current address: School of Photovoltaic Engineering, University of New South Wales, Sydney, New South Wales 2052, Australia.
3 Current address: Pacific Laboratory Medicine Services, Royal North Shore Hospital, Pacific Highway, St. Leonards, New South Wales 2065, Australia.
4 Address correspondence and reprint requests to Prof. Carolyn L. Geczy, Inflammatory Diseases Research Unit, School of Medical Sciences, University of New South Wales, Sydney, New South Wales 2052, Australia. E-mail address: c.geczy@unsw.edu.au
5 Abbreviations used in this paper: A8, S100A8; A9, S100A9; ActD, actinomycin D; CalC, calphostin C; CHX, cycloheximide; CM, culture medium; COX-2, cyclooxygenase-2; DEX, dexamethasone; GC, glucocorticoid; GR, GC-GC receptor; GRE, GC response elements; HPRT, hypoxanthine-guanine phosphoribosyltransferase; HYD, hydrocortisone; mA8, murine A8; MEC, microvascular endothelial cell; PKA, protein kinase A; PKC, protein kinase C; MP, methylprednisolone; TG, thioglycolate; RA, rheumatoid arthritis.
Received for publication July 16, 2004. Accepted for publication December 1, 2004.
References
Ravasi, T., K. Hsu, J. Goyette, K. Schroder, Z. Yang, F. Rahimi, L. P. Miranda, P. F. Alewood, D. A. Hume, C. Geczy. 2004. Probing the S100 protein family through genomic and functional analysis. Genomics 84:10.
Foell, D., M. Frosch, C. Sorg, J. Roth. 2004. Phagocyte-specific calcium-binding S100 proteins as clinical laboratory markers of inflammation. Clin. Chim. Acta 344:37.
Youssef, P., J. Roth, M. Frosch, P. Costello, O. Fitzgerald, C. Sorg, B. Bresnihan. 1999. Expression of myeloid related proteins (MRP) 8 and 14 and the MRP8/14 heterodimer in rheumatoid arthritis synovial membrane. J. Rheumatol. 26:2523
Hobbs, J. A., R. May, K. Tanousis, E. McNeill, M. Mathies, C. Gebhardt, R. Henderson, M. J. Robinson, N. Hogg. 2003. Myeloid cell function in MRP-14 (S100A9) null mice. Mol. Cell Biol. 23:2564.
Nacken, W., J. Roth, C. Sorg, C. Kerkhoff. 2003. S100A9/S100A8: myeloid representatives of the S100 protein family as prominent players in innate immunity. Microsc. Res. Tech. 60:569.
Hu, S. P., C. Harrison, K. Xu, C. J. Cornish, C. L. Geczy. 1996. Induction of the chemotactic S100 protein, CP-10, in monocyte/macrophages by lipopolysaccharide. Blood 87:3919.
Xu, K., C. L. Geczy. 2000. IFN- and TNF regulate macrophage expression of the chemotactic S100 protein S100A8. J. Immunol. 164:4916.
Yen, T., C. A. Harrison, J. M. Devery, S. Leong, S. E. Iismaa, T. Yoshimura, C. L. Geczy. 1997. Induction of the S100 chemotactic protein, CP-10, in murine microvascular endothelial cells by proinflammatory stimuli. Blood 90:4812
Grimbaldeston, M. A., C. L. Geczy, N. Tedla, J. J. Finlay-Jones, P. H. Hart. 2003. S100A8 induction in keratinocytes by ultraviolet A irradiation is dependent on reactive oxygen intermediates. J. Invest. Dermatol. 121:1168
Xu, K., T. Yen, C. L. Geczy. 2001. IL-10 up-regulates macrophage expression of the S100 protein S100A8. J. Immunol. 166:6358.
Donato, R.. 2001. S100: a multigenic family of calcium-modulated proteins of the EF-hand type with intracellular and extracellular functional roles. Int. J. Biochem. Cell Biol. 33:637
Passey, R. J., K. Xu, D. A. Hume, C. L. Geczy. 1999. S100A8: emerging functions and regulation. J. Leukocyte Biol. 66:549
Ryckman, C., K. Vandal, P. Rouleau, M. Talbot, P. A. Tessier. 2003. Proinflammatory activities of S100: proteins S100A8, S100A9, and S100A8/A9 induce neutrophil chemotaxis and adhesion. J. Immunol. 170:3233.
Manitz, M. P., B. Horst, S. Seeliger, A. Strey, B. V. Skryabin, M. Gunzer, W. Frings, F. Schonlau, J. Roth, C. Sorg, W. Nacken. 2003. Loss of S100A9 (MRP14) results in reduced interleukin-8-induced CD11b surface expression, a polarized microfilament system, and diminished responsiveness to chemoattractants in vitro. Mol. Cell Biol. 23:1034.
Ahmad, A., D. L. Bayley, S. He, R. A. Stockley. 2003. Myeloid related protein-8/14 stimulates interleukin-8 production in airway epithelial cells. Am. J. Respir. Cell Mol. Biol. 29:523
Passey, R. J., E. Williams, A. M. Lichanska, C. Wells, S. Hu, C. L. Geczy, M. H. Little, D. A. Hume. 1999. A null mutation in the inflammation-associated S100 protein S100A8 causes early resorption of the mouse embryo. J. Immunol. 163:2209.
Harrison, C. A., M. J. Raftery, J. Walsh, P. Alewood, S. E. Iismaa, S. Thliveris, C. L. Geczy. 1999. Oxidation regulates the inflammatory properties of the murine S100 protein S100A8. J. Biol. Chem. 274:8561.
Doussiere, J., F. Bouzidi, P. V. Vignais. 2002. The S100A8/A9 protein as a partner for the cytosolic factors of NADPH oxidase activation in neutrophils. Eur. J. Biochem. 269:3246.
McKay, L. I., J. A. Cidlowski. 1999. Molecular control of immune/inflammatory responses: interactions between nuclear factor-B and steroid receptor-signaling pathways. Endocr. Rev. 20:435.
Funder, J. W.. 1997. Glucocorticoid and mineralocorticoid receptors: biology and clinical relevance. Annu. Rev. Med. 48:231
De Bosscher, K., W. Van den Berghe, G. Haegeman. 2003. The interplay between the glucocorticoid receptor and nuclear factor-B or activator protein-1: molecular mechanisms for gene repression. Endocr. Rev. 24:488.
Re, F., M. Muzio, M. De Rossi, N. Polentarutti, J. G. Giri, A. Mantovani, F. Colotta. 1994. The type II "receptor" as a decoy target for interleukin 1 in polymorphonuclear leukocytes: characterization of induction by dexamethasone and ligand binding properties of the released decoy receptor. J. Exp. Med. 179:739.
Calandra, T., J. Bernhagen, C. N. Metz, L. A. Spiegel, M. Bacher, T. Donnelly, A. Cerami, R. Bucala. 1995. MIF as a glucocorticoid-induced modulator of cytokine production. Nature 377:68.(Kenneth Hsu, Robert J. Pa)
S100A8 (A8) has roles in inflammation, differentiation and development and is associated with oxidative defense. Murine A8 (mA8) is up-regulated in macrophages, fibroblasts, and microvascular endothelial cells by LPS. Glucocorticoids (GCs) amplified LPS-induced mA8 in these cells. Relative to stimulation by LPS, GCs increased mA8 gene transcription and mRNA half-life. Enhancement required new protein synthesis, IL-10 and products of the cyclooxygenase-2 pathway, and both ERK1/2 and p38 MAPK. Protein kinase A positively and protein kinase C negatively regulated this process. Promoter analysis indicated element(s) essential for LPS and dexamethasone enhancement colocated within the region –178 to 0 bp. In the absence of glucocorticoid response elements, NF1 motif at –58 is a candidate for mediation of enhancement. Gel shift analysis detected no differences between LPS- and LPS/dexamethasone-treated complexes within this region. GCs increased constitutive levels of A8 and S100A9 (A9) mRNA in human monocytes. The synovial membrane of rheumatoid patients treated with high dose i.v. methylprednisolone contained higher numbers of A8/A9-positive macrophages than pre- or posttreatment samples. Results support the proposal that A8 has anti-inflammatory properties that may be independent of hetero-complex formation with A9 and may also enable localized defense in the absence of overriding deleterious host responses.
Introduction
S100A8 (A8,5 MRP8, calgranulin A) is a member of the S100 family of calcium-binding proteins (22 members; Ref.1), generally coexpressed with S100A9 (A9, MRP14, calgranulin B). A8 and A9 accumulate in inflammatory fluids and are postulated to be involved in the pathogenesis of chronic human inflammatory diseases such as rheumatoid arthritis (RA), cystic fibrosis, Crohn’s disease, ulcerative colitis, allergic dermatitis, and infection (reviewed in Ref.2). A8 and A9 are expressed in macrophages in the lining and sublining layers of the rheumatoid synovium. Expression at the cartilage pannus junction, the site at which the inflamed synovial pannus invades cartilage and bone, had a different pattern from more distal sites, suggesting that macrophage activation or differentiation differ at various sites (3). A8 and A9 comprise 40% of the cytosolic protein of neutrophils and their intracellular expression has been associated with calcium sensing (4), myeloid cell differentiation, and a variety of inflammatory processes, particularly those involved in neutrophil activation and arachidonic acid transport (reviewed in Ref.5). A8 and A9 are expressed at low levels by monocytes but not normally by macrophages. Murine A8 (mA8) in elicited macrophages is moderately up-regulated by proinflammatory mediators such as LPS (6), TNF, and IFN- (7). LPS and IL-1 are key regulators of the gene in microvascular endothelial cells (MEC) (8) and fibroblasts. Induced A8 is mostly confined to the cytoplasm and/or nucleus in activated MEC (8) and keratinocytes (9) whereas it is cell-associated and secreted by activated macrophages (7, 10).
Several S100 proteins have important extracellular functions (11) and roles for A8 and A9 in regulating leukocyte transmigration are proposed (5, 12, 13). Neutrophils from A9-null mice are also deficient in A8 and have reduced/compromised chemotactic responses (4, 14). A9 and the A82-A9 complex (calprotectin) induce IL-8 production by airway epithelial cells, hence potentially amplifying neutrophilic inflammation in chronic pulmonary disease (15). Deletion of the A8 gene is embryonic lethal shortly after implantation, indicating distinct functions for A8 in development (16). Some functions of A8 are calcium-dependent and depend on hetero-dimerization with A9, e.g., arachidonic acid transport and antimicrobial defense (5).
We proposed a protective role for A8 in oxidative defense in acute inflammation (12). A8 is a potent scavenger of hypochlorite (17) and the A8/A9 complex may be an intracellular scaffold for the NADPH oxidase system and potentiate activity by acting as a reservoir for arachidonic acid (18). Thus these proteins may regulate oxidative events associated with inflammation. This is strongly supported by the fact that the A8 gene is up-regulated by anti-inflammatory mediators and oxidative stress such as UVA- and H2O2 (9, 10). IL-10 synergizes with LPS to markedly enhance A8 mRNA expression and protein secretion in stimulated murine macrophages, and endogenous IL-10 induced by LPS mediates induction by LPS. The synergy may depend on generation of mediators such as PGE2 and cAMP (10). We propose that the low levels of A8 induced by proinflammatory mediators may serve a homeostatic function, whereas higher amounts may reduce oxidative damage during resolution.
Glucocorticoids (GCs) are traditionally viewed as anti-inflammatory and immunosuppressive, and are commonly used to treat numerous inflammatory conditions, including allograft rejection, autoimmune diseases and allergic conditions (reviewed in Refs.19, 20). Monocyte/macrophage functions down-regulated by GCs include induction of cytokines such as IL-1, IL-6, IL-8, and TNF-, which generate positive inflammatory reactions, generally via transrepression of transcription factors such as NF-B and AP-1, by the GC-GC receptor (GR) complex (reviewed in Ref.21). Conversely, GCs may induce anti-inflammatory cytokines such as IL-4, IL-10, and TGF (21) and type II IL-1R (22), mostly mediated by GC response elements (GRE) in the promoter regions of these genes. GCs released via the neuroendocrine-immune network during stress may also act as selective immunomodulatory enhancers during infection and tissue invasion to induce genes such as macrophage migration inhibition factor (23), a critical modulator of innate immune responses.
Here we show that GCs have little direct effect, but markedly up-regulate LPS-induced A8 in murine exudate macrophages, MEC and fibroblasts. Moreover, dexamethasone (DEX) increased constitutive A8 and A9 mRNA in human monocytes. In support of data generated in vitro, increased numbers of macrophages expressing A8 and A9 were found in the synovial membrane of rheumatoid patients treated with high doses of i.v. ("pulse") methylprednisolone (MP), and these returned to pretreatment levels when treatment ceased. Our results indicate complex regulation of the A8 gene by GCs and increased mRNA stability and transcription both contribute. Promoter analysis indicated that NF1 may regulate GR transactivation of the mA8 promoter. Requirements for IL-10 and products of the cyclooxygenase-2 (COX-2) pathway are essential for A8 induction by LPS (10), and both were required for GC synergy. We suggest that interactions between the MAPK, protein kinase A (PKA), and protein kinase C (PKC) pathways may provide an additional level of control in this system. Our results support the proposal that A8 has anti-inflammatory properties that are independent of hetero-complex formation with A9 and may be important in Gram-negative sepsis, where high endogenous GCs levels and endotoxin coexist.
Materials and Methods
Cells and reagents
The murine monocyte-macrophage cell line RAW 264.7 (American Type Culture Collection; ATCC TIB 71) was a gift from Dr. D. Hume (University of Queensland, Brisbane, Australia). Cells were cultured at 37°C in 5% CO2 in air in RPMI 1640 (Invitrogen Life Technologies) supplemented with 127 U/ml penicillin, 127 μg/ml streptomycin (Sigma-Aldrich), and 10% heated (56°C, 30 min) bovine calf serum (HyClone) filtered through 0.2-μM Zetapore membranes (Cuno), hereafter referred to as cell culture medium (CM). Thioglycolate (TG)-elicited murine macrophages were obtained as described (7). RAW 264.7 cells (1.5 x 106) in 5 ml of CM were stimulated and harvested as indicated. Medium and mediators were only used if endotoxin levels were <20 pg/ml (chromogenic limulus amoebocyte assay; Associates of Cape Cod).
Plastic-adherent fibroblast short-term cell lines (splenic "primary" fibroblasts) were isolated and cultured as described (24) with some modification. Briefly, spleens from QS mice were minced, digested for 20 min at 37°C in collagenase (Sigma-Aldrich; 2 mg/ml), then passed through a tissue strainer (70 μM, BD Biosciences), and washed (3x CM). Nonadherent cells and debris were moved by replenishing CM after 24, 48, and 72 h. Cells from passages 2 through 8 were used in the experiments. Primary fibroblasts obtained by this method do not contain macrophages or other cells of hemopoietic origin (24) and had typical fibroblast morphology.
The murine microvascular brain endothelioma cell line (bEnd-3), a gift from Dr. W. Risau (Max-Planck Institute, Tübingen, Germany), was grown and used when postconfluent (6–7 days) as described (8).
Human blood monocytes were enriched using the monocyte isolation kit (Miltenyi Biotec). Briefly, PBMC from venous blood from healthy volunteers isolated on Lymphoprep (Nycomed) were suspended in CM. To avoid neutrophil contamination and stimulation of monocytes, the latter were isolated by negative selection. PBMC were incubated with a mixture of biotinylated CD3, CD7, CD16, CD19, CD56, CD123, and CD235a Abs and anti-biotin MicroBeads, and nonmonocytes were depleted using MACS (Miltenyi Biotec) with LS columns. Differential staining indicated that the population was 96% monocytes with viability >95% by trypan blue staining. Monocytes (5 x 105) were aliquoted into Nunc-minisorb tubes (Nunc) and incubated overnight at 37°C in 5% CO2 in air before stimulation.
Anti-mouse IL-10 mAb was from R&D Systems. DEX, hydrocortisone (HYD), corticosterone, cycloheximide (CHX), actinomycin D (ActD), calphostin C (CalC), PD98059, SB202190, and H89 were from Sigma-Aldrich. LPS (Escherichia coli, 055:B5) was from Difco.
Northern analysis
Northern blotting and preparation of mA8 riboprobe and 18S rRNA oligo probe were as described (7). Phosphorimage and densitometry analysis was performed using the Bio-Rad MultiAnalyst/Macintosh Version 1.0 software and the Bio-Rad Molecular Imager GS-525 system (Bio-Rad Laboratories).
Real time quantitative PCR analysis of A8 and A9
Total RNA (1 μg) from human monocytes was treated with DNase 1 (Ambion, Austin TX) and reverse transcribed using random hexamers and the SuperscriptIII First-Strand Synthesis System for RT-PCR (Invitrogen Life Technologies). Negative controls (no first strand synthesis) were prepared by performing reverse transcription reactions in the absence of reverse transcriptase. PCR amplification was performed with Platinum SYBR Green qPCR SuperMix UDG (Invitrogen Life Technologies). Reactions were performed in duplicate containing 2x SYBR Green qPCR SuperMix, 1 μl of template cDNA or control, 100 nM of human A8 primers (5' primer: GGGATGACCTGAAGAAATTGCTA, 3' primer: TGTTGATATCCAACTCTTTGAACCA) or human A9 primers (5' primer: GTGCGAAAAGATCTGCAAAATTT, 3' primer: GGTCCTCCATGATGTGTTCTATGA) in a final volume of 25 μl, and analyzed in 96-well optical reaction plates (Applied Biosystems). Reactions were amplified and quantified using an ABI 7700 sequence detector with standard cycle conditions and the Applied Biosystems software. Relative quantities of A8 and A9 mRNA were obtained using the comparative CT method and were normalized against human hypoxanthine-guanine phosphoribosyltransferase (HPRT: 5' primer: TCAGGCAGTATAATCCAAAGATGGT, 3' primer: AGTCTGGCTTATATCCAACACTTCG) as an endogenous control.
Detection of mA8 protein
A double-sandwich ELISA using rabbit polyclonal anti-mA8 IgG was performed using cell-free supernatants as described (7), with serial dilutions of recombinant mA8 as standard.
Nuclear run-on transcription assay
Nuclear run-on transcription was performed as described (25). Briefly, nuclei from 2 x 107 PBS-washed RAW cells were isolated in ice-cold lysis buffer (10 mM Tris-HCl (pH 7.4), 10 mM NaCl, 3 mM MgCl2, 0.5% Nonidet P-40). RNA transcription was initiated by the addition of 100 μl of reaction buffer (10 mM Tris-HCl (pH 8.3); 40% glycerol; 5 mM MgCl2; 100 μM EDTA; 2.5 mM DTT; 500 μM ATP, CTP, and GTP; and 3.3 μM [32P]UTP (50 μCi)), and incubation at 30°C for 30 min. The reaction was stopped with solution D, and total RNA was extracted. The aqueous phase was precipitated by the addition of tRNA and isopropanol. After washing in 70% ethanol, the pellet was resuspended in 10 μl of Tris-EDTA buffer (pH 7.3) then 200 μl of formamide. Equal amounts of radioactivity from each sample were hybridized with Hybond membranes to which 1 μg of mA8 cDNA, 1 μg of linearized plasmid containing the cDNA for HPRT, and 1 μg of linearized vector only, were immobilized by slot blot transfer. Hybridization was performed at 42°C for at least 36 h in 50% formamide, 5x standard saline citrate phosphate/EDTA, 5x Denhardt’s, 0.5% SDS, 50 μg/ml denatured herring sperm DNA, and 200 μg/ml tRNA. After washing filters to a stringency of 1x SSC, 0.1% SDS at 55°C, phosphorimage and densitometry analysis was performed using the Bio-Rad Molecular Imager GS-525 system.
Construction of luciferase reporter plasmids
Truncated promoter fragments of the mA8 gene were produced by nested deletion of a PCR-amplified product as described (10). A series of 5' and 3' deletion mutants was generated by nested-deletion using the double-stranded Nested Deletion Kit (Pharmacia Biotech). The 5' deletion mutants (with a common 3' end at +465) yielded the pCP-x/+465 constructs (x represents different 5' ends). The 3' deletion mutants (with a common 5' end at –316) yielded the pCP-316/±y constructs (y represents different 3' ends, see Fig. 5A). The extent of deletion in all clones was verified by sequencing.
FIGURE 5. A8 promoter analysis in response to LPS ± DEX. A, Luciferase activity of mA8 deletion constructs in the presence of LPS ± DEX. A8-luciferase chimeric gene constructs are shown below a partial restriction map of the 1053-bp A8 gene fragment. The name of each construct is given on the right. Solid lines represent A8 DNA fragments, open boxes (LUC) at the 3' end indicate the luciferase gene. The numbers below the map represent the positions of restriction sites relative to the transcription start site (+1). PCP-178/0-NF1 has the mutation at position –75 to –80. A8-luciferase fusion constructs were transfected into RAW cells in parallel with control plasmid pGL2-Control as described in Materials and Methods and stimulated with LPS (500 ng/ml) either alone or with DEX (10–6 M). Results are means of duplicate determinations from two experiments. Error bars represent mean ± SD. B, Gel shift analysis of the region –173 to –1 of the mA8 gene in the presence of LPS ± DEX. RAW cells were incubated with LPS (500 ng/ml) ± DEX (10–6 M) and nuclei prepared after 18 h, and extracts were used for gel shift analysis as described in Materials and Methods.
Site-directed mutagenesis
Site-directed mutagenesis was conducted with the QuickChange Kit (Stratagene). The sequence of the mutagenic primer (sense strand) for the NF1-like motif was: CGAGGCTGcataCAGCTGtaCgAGCTTTC (mutated bases are indicated in lowercase letters).
DNA transfection and luciferase assay
DNA transfection was according to Stacey et al. (25) with some modifications. RAW cells (6 x 106) cultured in CM and harvested in logarithmic growth phase were electroporated using a Bio-Rad Gene Pulser (260 V, 960 μF) in 0.4-cm electroporation cuvettes (Bio-Rad). Each recombinant plasmid was present at the same molar concentration as 10-μg pGL2-Control plasmid (Promega). To assess mA8 promoter inducibility, 2 h after electroporation cells were stimulated with 500 ng/ml LPS ± DEX and harvested after 18 h. Luciferase activity was measured using the Luciferase Assay System (Promega) in a TD-20/20 Luminometer (Turner Designs). Luciferase activity was normalized to total protein in cell lysates (bicinchoninic acid protein assay kit; Sigma-Aldrich); values generated were reproducible and constant, relative to pGL2-control.
Gel shift analysis
RAW cells (8 x 106) were stimulated with the appropriate mediators for 18 h and extracts for gel shift analysis isolated according to Osborn et al. (26). The 17-bp promoter region was produced using PCR and the pair primers (GCCGGTACCCATTCCTCAGACTCAGAAATGAAATGCTC and CGGGAGATCTATGTGAGGCTAAGTGTCAGCTGC) following purification on a QIAquick PCR purification column (Qiagen). The fragment was labeled using T4 polynucleotide kinase and purified on a NICK column (Amersham Pharmacia). Binding assays were performed in 20 mM HEPES (pH 7.9), 1 mM MgCl2, 4% Ficoll, 6% glycerol, 35 mM NaCl, 21 mM KCl, 0.5 mM DTT, 10 fmol of double stranded oligonucleotide, and appropriate amounts of nuclear proteins and herring sperm DNA. After 30 min on ice, the binding reaction was separated on discontinuous 6% PAGE with Tris-glycine buffer. Dried gels were phosphorimaged using the Bio-Rad system.
MP treatment and synovial biopsies
Twelve patients fulfilling the American College of Rheumatology criteria (1987) for RA were studied. All received 1000 mg of MP i.v. as the sodium hemisuccinate salt (Upjohn Kalamazoo) as described (27, 28). Clinical and laboratory assessments of inflammation were performed. Synovial membrane samples were obtained before and 24 h after MP in 12 patients and at disease relapse (between 6 and 12 wk later) in 6 patients, using needle arthroscopic techniques. This project was approved by the Repatriation General Hospital Ethics Committee with informed consent from each patient.
Immunostaining and analysis of synovial tissue
Serial sections were stained with monospecific affinity-purified rabbit anti-A8 or -A9 (29) using a standard three-stage immunoperoxidase method as described (30). Sections were coded and randomly analyzed by a blinded observer (P.Y.). Only synovium in which the lining layer was clearly evident was studied. Sections were examined at x400 magnification using a 1-mm graticule. All high power fields, except for those in which fibrous subsynovial conective tissue predominated, were included in the analysis. Positively stained cell counts were expressed as mean counts per square millimeter using a conversion factor of 0.0625–1. Results are given as mean (SEM - SEM) or mean (range). StatviewSE +Graphics statistical software (Abacus Concepts) was used. Means were compared using paired t tests. Differences were considered to be significant at p < 0.05.
Results
GCs enhance mA8 induction
mA8 transcript levels in macrophages, after exposure to corticosterone, HYD or DEX in the presence or absence of LPS were assayed by Northern blotting. DEX did not directly induce mA8 mRNA but markedly enhanced levels stimulated by LPS. Synergy was counteracted by RU486, a synthetic anti-GC that competes for binding of DEX to GR (Fig. 1A). Similar results were observed with corticosterone (data not shown) and HYD (Fig. 2, A and B). The steroid derivatives estradiol and 1,25-(OH)2 vitamin D3 neither directly induced expression nor enhanced expression induced by LPS (not shown). In contrast to mA8, LPS-induced TNF- mRNA was suppressed by DEX, and suppression was totally abolished by RU486 (Fig. 1A). Identical results were obtained with the RAW 267.4 macrophage cell line (not shown). A9 mRNA was present in bone marrow cells but never detected by Northern blotting (not shown). Quantitative RT-PCR detected very low levels of A9 mRNA in RAW cells and macrophages stimulated with LPS (>10–5-fold less than mRNA levels of A8); DEX neither induced nor enhanced the levels (three experiments, data not shown). To test whether synergy was cell-specific, primary fibroblasts and MEC were tested. mA8 mRNA levels in LPS-stimulated fibroblasts increased 7.3-fold with 10–7 M DEX, and 3.5-fold with 10–8 M (Fig. 1B). Similar results were obtained with the 3T3 cells, and synergy was suppressed with RU486 (not shown). When MEC were costimulated with LPS, up-regulation by DEX was optimal with 10–7–10–6 M DEX (Fig. 1C).
FIGURE 1. Induction of A8 mRNA by LPS and GCs. A, DEX (10–6 M) enhanced LPS-induced (100 ng/ml) A8 mRNA in TG-elicited macrophages incubated with stimulants for 20 h. RU486 (1 μM) counteracted the enhancing effect of DEX. Total RNA was analyzed by Northern blotting. Filters were rehybridized with the TNF probe and an 18S rRNA oligoprobe after stripping. Data quantitated by phosphorimaging are represented as bar charts. B, Fibroblasts were stimulated for 24 h with LPS (100 ng/ml) ± DEX at the doses indicated. C, MEC were incubated with LPS (500 ng/ml) ± DEX for 24 h. Experiments were performed at least three times with similar results. Ctrl, unstimulated control.
FIGURE 2. GCs enhance LPS-induced A8 up-regulation in macrophages. A, DEX (10–6 M) or HYD (10–6 M) enhanced A8 mRNA induced by different doses of LPS in TG-elicited macrophages stimulated for 20 h. B, TG-elicited macrophages were incubated with LPS (100 ng/ml) ± DEX or HYD at the concentrations indicated for 20 h. Total RNA was analyzed by Northern blotting. C, A8 protein (nanomolar) in supernatants of RAW cells incubated with LPS (100 ng/ml) ± DEX or HYD at the concentrations indicated, for 20 h quantitated by ELISA. At least two separate experiments yielded similar results. D, RAW cells were incubated with LPS (100 ng/ml) ± DEX (10–6 M) ± anti-IL-10 (IL10Ab, 2 ng/ml). Total RNA samples harvested at the times indicated were analyzed by Northern blotting. E, A8 in supernatants generated in D were quantitated by ELISA. Values represent means ± SD; n = 3.
Synergy was most evident when GCs were included with low concentrations of LPS. The low levels of mA8 mRNA induced with 20 ng/ml LPS markedly increased when DEX or HYD were included; enhancement was less obvious with high amounts of LPS (1000 ng/ml; Fig. 2A). Optimal up-regulation was between 10–7 and 10–6 M GCs (Fig. 2B). In accordance with these results, 3- to 4-fold more mA8 was secreted by macrophages when 10–6 or 10–7 M DEX or HYD were included with LPS (Fig. 2C). Similar sensitivity was observed with RAW cells and because these exhibited identical responses to exudate macrophages, these were used for more detailed analysis as they are a reliable source of large numbers of cells, and unlike primary macrophages, can be readily transfected.
Addition of GCs did not alter the kinetics of A8 mRNA induction by LPS; both were maximal at 20–24 h and decreased by 48 h (Fig. 2D). No detectable mA8 was secreted by unstimulated RAW cells but was evident after 12 h and reached peak levels of 35 ng/ml after 48 h stimulation with LPS. When DEX was included, mA8 levels peaked at 60 ng/ml (Fig. 2E). Taken together, these results show that GCs synergistically enhanced LPS-induced mA8 up-regulation at the mRNA and protein levels in macrophages, and the mRNA level in fibroblasts, and endothelial cells, in a dose-dependent manner via a GR-dependent mechanism.
GCs increase mA8 mRNA stability and transcription rate but do not alter duration of mRNA expression
To determine whether GCs affected mA8 mRNA stability, ActD, at a concentration that inhibits mA8 transcription (5 μg/ml) (6), was used in the presence/absence of GCs 20 h after addition of LPS, and cells harvested at 4 hourly intervals over 24 h. The mRNA half-life in the presence of LPS was 8 h, and in the presence of LPS and HYD (10–6 M) was maintained over 20 h, representing an approximate increase in stability of 2.5-fold (Fig. 3A). Similar results were obtained with DEX (not shown).
FIGURE 3. Mechanisms of enhancement by GCs on LPS-induced A8 in macrophages. A, RAW cells were incubated with LPS (500 ng/ml) ± HYD (10–6 M). ActD (5 μg/ml) was added 20 h after stimulation with LPS; samples harvested at 4 hourly intervals were analyzed by Northern blotting. The "LPS" samples were phosphorimaged for longer than the "LPS+HYD" samples to equalize the zero time-point band intensity. B, RAW cells were incubated for 18 h with LPS (500 ng/ml), DEX (10–6 M), or both, and subjected to nuclear run-on analysis. HPRT cDNA was used to normalize the overall rate of transcription; A8 cDNA vector alone was the negative control. C, Effect of time of addition of DEX on LPS-induced A8 mRNA. RAW cells were treated with LPS (100 ng/ml) for 24 h (). DEX (; 10–6M) was added at the indicated times relative to addition of LPS. The numbers represent 24, 3, or 1 h before, with (0 h), or 3, 6, 9, or 21 h after addition of LPS. Relative mRNA levels are presented as fold increases in A8 mRNA above those induced by LPS alone. Total RNA samples harvested at the times indicated were analyzed by Northern blotting. Data represents means ± SD of three experiments. *, p < 0.05; **, p < 0.01. D, Effects of CHX (2 μg/ml) on DEX (10–6 M)-mediated synergy of LPS (100 ng/ml)-induced A8 mRNA expression. Northern and nuclear run-on results are representative of three experiments.
The effects of LPS and DEX on the rate of mA8 gene transcription were assayed by nuclear run-on analysis (Fig. 3B). The very low basal transcription rate was not altered by DEX (10–6 M), but as reported (6), was increased by LPS (500 ng/ml). LPS and DEX together increased transcription which, from densitometry readings normalized to the rate of HPRT transcription, was 3-fold higher than that induced by LPS alone.
Enhancement depends on new protein synthesis and occurs via the COX-2-cAMP pathway
Many DEX-mediated effects occur via induction of new genes. We reasoned that if DEX increased the LPS response via new protein synthesis, synergy may be enhanced in macrophages pretreated with DEX, and blocked when costimulated with a protein synthesis inhibitor. Maximum mA8 mRNA induction in RAW cells occurred when cells were preincubated with DEX for 21 h before addition of LPS (3.8-fold greater increase than when DEX was added simultaneously with LPS), and this increment decreased when preincubation was reduced. The effect was biphasic, with a second small, but significant increase (1.7-fold compared with DEX and LPS added at the same time; p < 0.05) when DEX was added 6 h after LPS (Fig. 3C). RAW cells treated with LPS ± DEX and CHX, which efficiently inhibits protein synthesis and mA8 induction by LPS (10), did not express mA8 mRNA above control levels (Fig. 3D), confirming the requirement for new protein synthesis.
Because IL-10 strongly up-regulates LPS-stimulated mA8 expression in macrophages (10), and DEX induces IL-10 in some cell types (31, 32), we tested IL-10 involvement. When RAW cells were cultured with LPS+DEX in the presence of anti-IL-10, enhancement was suppressed to levels observed with LPS alone (Fig. 2D). Suppression was best reflected in secreted protein levels, which were only 30% of levels induced by LPS alone (Fig. 2E). Anti-IL-10 inhibited LPS-induced mA8 mRNA by 78% (Fig. 4A). More importantly, levels of IL-10 secreted by RAW cells treated with DEX+LPS or with LPS alone for 6 h were similar (54.3 vs 45.3 pg/ml, n = 2) and LPS-induced IL-10 at 24 h decreased by 40% when DEX was included (664.2 vs 429.2 pg/ml, respectively; n = 2). Thus, DEX synergy was unlikely to be due to increased levels of IL-10 in supernatants and the effect of anti-IL-10 may have been primarily on LPS-induced mA8 expression.
FIGURE 4. Effect of signal transduction inhibitors on DEX-enhanced A8 expression. A, RAW cells were incubated with LPS (100 ng/ml), DEX (10–6 M), or LPS + DEX for 20 h in the presence/absence of anti IL-10 (5 ng/ml) and/or COX inhibitors (indomethacin, Indo, 10 μM; NS389, 50 μM). B, Effect of DEX on LPS-induced A8 in RAW cells pretreated for 1 h with SB202190 (1 μM), PD98059 (10 μM), H89 (1 μM), and CalC (0.2 μM). Total RNA analyzed by Northern blotting. Values are means ± SD; n = 3. *, p < 0.05; **, p < 0.01 vs corresponding control (A) and LPS alone (B).
Because PGE2 and cAMP also synergize with LPS and are involved in mA8 induction (10), these pathways were tested. The COX inhibitor indomethacin, and the more specific COX-2 antagonist NS389, inhibited LPS-induced mA8 by 45%, whereas 60% of LPS+DEX-induced mRNA was abrogated, indicating that the COX-2 pathway may be involved in up-regulation by DEX. When anti-IL-10 and COX inhibitors were used together, inhibition was additive; 91% of LPS-induced mRNA was abolished and 82% of the LPS+DEX-induced mRNA levels were inhibited, suggesting the two signaling pathways may work independently.
Inhibition of enhancement by DEX was observed when H89, which decreases cAMP and PKA activity, was included. In contrast, the specific PKC antagonist CalC enhanced the DEX effect (Fig. 4B). These results implicate the cAMP pathway in the synergy and the PKC pathway may regulate this. The MAPK pathway is also involved in LPS induction of the mA8 gene in macrophages (7) and regulation by GCs can occur via p38 MAPK activation (33). The ERK1/2 (PD98059) and p38 (SB202190) inhibitors both abrogated LPS-induced A8 mRNA. The p38 inhibitor considerably reduced the LPS+DEX signal, although the degree of inhibition was less than the effect on LPS alone. The ERK1/2 inhibitor almost totally abrogated both responses (Fig. 4B). The results suggest overlapping pathways involved in induction of A8 by LPS and enhancement by DEX.
GC and LPS responsive elements colocalize
Promoter deletion analysis obtained after LPS stimulation indicated elements which enhanced or inhibited promoter function arranged in an alternating array within intron 1 and upstream of exon 1. Intron 1 contained an interesting tandem array of apparently interacting enhancer and repressor elements. Although consecutive removal of intronic regions +322 to +465 (pCP-316/+322), then +179 to +322 (pCP-316/+179) increased luciferase activity in response to LPS 3-fold then resulted in complete loss of activity, subsequent removal of the adjacent 51 bp (pCP-316/+129) restored activity, comparable to that obtained with pCP-316/+322 (Fig. 5A). This may be due to a repressor, enhancer, repressor array in which the enhancer in +179 to + 322 specifically counteracts the repressor in +129 to + 179. As agrees with our previous report (10), the region –173 to 0 (pCP-173/0) conferred LPS-responsive expression, and so contains the essential promoter of the mA8 gene. The relative level of expression induced by LPS+DEX, compared with LPS alone, was similar for all LPS-inducible constructs, indicating that none of the deleted regions contained elements mediating GC-enhanced transcription. These elements must be restricted to the region –178 to 0 (Fig. 5A), and although this contains no classical GRE, candidate elements include AP-1 motifs (–164, –102, –18) and an NF1 motif (–58). Because NF1 mediates induction of several genes by GCs (34, 35), the NF1 motif was mutated at –58 bp (pCP-178/0NF1). Fig. 5A shows that this construct completely abolished the signal induced by LPS+DEX, whereas the LPS-induced luciferase activity was reduced by only 38%, indicating involvement of this NF1 binding site in coinduction by DEX.
Gel shifts were performed on the region –173 to –1 in an attempt to locate elements responsible for GCs enhancement. Extracts from unstimulated cells produced two complexes and in the presence of LPS or LPS+DEX, the intensity of both complexes were reduced (Fig. 5B). No difference between LPS and LPS+DEX extracts was observed using oligonucleotide probes spanning the 173-bp region although at least five separate complexes were evident (data not shown). The lack of detectable change upon addition of DEX indicates that enhancement may occur either by a factor present in low abundance, or by modification of a factor already present in the LPS-induced complex (for example by phosphorylation).
GCs directly induce A8 and A9 mRNA in human monocytes and A8/A9-positive cells in the rheumatoid synovial membrane
Because there is strong homology between the human and murine proximal A8 promoters (1), we tested the effect of DEX on human monocytes which constitutively express A8 mRNA and low levels of A8 and A9 proteins (36). Fig. 6A confirms A8 and A9 mRNA in unstimulated cells. LPS increased mRNA expression of A8 (2.1-fold) and A9 (1.8-fold). DEX increased A8 and A9 mRNA 5-fold, but no synergy was observed when combined with LPS.
FIGURE 6. Effects of GCs on the expression of A8 and A9 in human monocytes and macrophages in RA synovium. A, Human monocytes incubated with LPS (100 ng/ml) ± DEX (10–6 M). After 24 h, total RNA was harvested and real-time RT-PCR performed in duplicate and HPRT was used as a control for amounts of RNA used in each reaction. Values are means ± SD; n = 4. Ctrl, unstimulated. *, p < 0.05; **, p < 0.01 vs absence of DEX. B, Synovial knee biopsies were taken from 12 patients with RA before and 24 h after treatment with 1000 mg of i.v. MP, and from 6 patients at disease relapse. Tissues were immunostained for A8 and A9, and numbers of positively stained cells were quantified. MP significantly increased expression of A8 within 24 h (p = 0.01) and increased expression of A9 (p = 0.06). A8- and A9-positive cells returned approximately to pretreatment levels at disease relapse. C, Synovial lining from a knee joint from a patient with RA before (left) and after treatment (right) with high dose MP. Infiltrating A8+ macrophages stain red. Sections were counterstained with hematoxylin (magnification, x40). An unrelated rabbit IgG, used as control, was unreactive (data not shown).
High doses of i.v. (pulse) steroids are used to treat inflammatory disorders, and rheumatoid patients treated with MP provide an excellent model to investigate their effects. A8 and A9 expression increased in macrophages within the rheumatoid synovial membrane 24 h after MP in 11 of 12 patients treated and returned to approximately pretreatment levels at the time of clinical disease relapse (Fig. 6B), which was used as an indicator of when the effects of MP had ceased. Fig. 6C shows increased numbers of A8-positive macrophages in synovial lining from a patient treated with MP compared with the biopsy taken before treatment. Extracellular A8 was obvious in both tissues, indicating secretion.
Discussion
GCs are anti-inflammatory drugs commonly used to treat allergic and autoimmune diseases by controlling expression of specific target genes. They limit expression of numerous inflammatory cytokines and other proteins associated with leukocyte migration and adhesion, primarily by suppressing gene transcription. GCs also enhance transcription of a limited number of genes, the products of which generally have anti-inflammatory effects (reviewed in Ref.37).
Here we show that GCs had little direct effect, but enhanced mA8 gene expression in murine macrophages stimulated with LPS, whereas TNF mRNA was reduced. MEC and fibroblasts also exhibited synergy, indicating that the effect was not cell-specific. GCs enhanced constitutive basal mRNA levels of A8 and A9 in human monocytes although there was no synergy with LPS. DEX and 1,25-(OH)2 vitamin D3 also induce A8 in the human macrophage cell line, U937 (38). Importantly, A8/A9-positive macrophages increased in biopsies of rheumatoid synovium from RA patients, whereas this treatment markedly reduced proinflammatory mediators such as TNF-, IL-8 (27), cell adhesion molecules (28), and chemokines (39). Increased expression occurred when patients were clinically improved, and thus when the effects of MP were clinically anti-inflammatory and decreased at clinical relapse, when the anti-inflammatory effect of MP had ceased and when we previously observed increases in proinflammatory mediators (27, 28, 39).
Enhancement in murine macrophages occurred with pharmacologically (40) and physiologically relevant concentrations of GCs (10–6–10–8 M). For example, in certain pathologic states such as Cushing’s disease, and in tissues at sites of synthesis, steroids can occur at micromolar levels (41). Costimulation was obligatory in the three cell types and marked synergy in macrophages was observed with DEX, HYD, or corticosterone, and correlated with elevated levels of secreted mA8 protein. Consistent with our finding that mA9 is barely induced in macrophages by LPS (6), the virtually undetectable A9 mRNA in LPS/DEX activated macrophages confirms that A8 does not require A9 for stability or secretion, and contradicts this proposal, generated from observations of neutrophil function in A9-null mice (4, 14). However, in contrast to murine macrophages, DEX increased A9 mRNA in human monocytes in vitro, to levels similar to A8, indicating likely differences in regulation of A9 in particular cell types/species.
In striking contrast to our observations, DEX repressed A8 and A9 expression associated with phorbol ester-induced inflammation in the skin of wild-type mice, but not c-fos–/– mice, indicating AP-1 target genes for which phorbol ester-induced transcription down-regulated by GCs depends on c-Fos (42). The positive and negative regulatory effects of GCs on constitutive S100B mRNA expression in hippocampal astrocytes are also distinct from those observed for other S100 proteins (43) and for LPS-inducible A8. Our studies contribute to the accumulating evidence that steroids are key modulators of the S100 gene family and that the nature of the response may vary according to the cell type and mediator responsible for induction.
Unlike the synergy of mA8 expression seen with IL-10, which reduces the time required for maximal gene induction (10), DEX did not alter the kinetics of the LPS response. However synergy was highest in macrophages preincubated with DEX for 21 h and the requirement for new protein synthesis suggests generation of an essential cofactor. The magnitude of A8 induction in macrophages activated with LPS is dependent on IL-10, and here we demonstrate marked reduction in synergy using anti-IL-10, although this treatment also reduces the LPS-induced gene. Moreover, supernatants generated from DEX+LPS treated RAW cells contained only marginally more IL-10 after 6 h compared with LPS alone, and levels decreased after 12 h. However under certain conditions, IL-10 may facilitate A8 induction by GCs. In humans, the effect of DEX on LPS-stimulated IL-10 secretion by monocytes is dependent on DEX and LPS concentrations (44). Furthermore, infusion of HYD in normal subjects directly before LPS administration increased LPS-induced plasma levels of IL-10, whereas induction of hypercortisolemia before LPS injection did not modify IL-10 levels (45). Gradual increases in serum IL-10 levels were observed in patients with multiple sclerosis (32) and patients undergoing cardiac surgery (46) and treated with GCs.
Here we show that increased mRNA stability and transcription are both likely to contribute to synergistic up-regulation of the mA8 gene. GCs can alter mRNA stability (21), but there are no reports demonstrating increased stability of inducible S100 transcripts, as described here for mA8. The mA8 transcript contains no regions with similarity to motifs reported to alter mRNA stability of other genes, suggesting a novel mechanism.
Elements responsible for mediating GCs enhancement were located in the region –173 to 0 of the essential promoter in which there are no obvious GREs. Promoters for many genes involved in inflammation have no demonstrated GRE sequences essential for binding activated GR and this is clearly indicated in dimerization-defective GR mice (47). Regulation of expression of GCs-responsive genes reflects a collaboration of multiple transcription factors (reviewed in Ref.21). NF1 is involved in GCs-dependent chromatin remodeling and also in transcriptional activation of the murine mammary tumor virus promoter (34), and is likely to regulate transcription of mA8 by DEX and LPS. The mA8 NF1 motif is a 13/14 palindrome contained within a 21-bp region and is identical in the murine and human A8 genes (–58 to –38). Its location immediately upstream of the TATA box suggests that the NF1 motif at position –58 may activate or repress formation of the basal transcription complex (48). Because the DEX response was found with a variety of cell types, it is possible that NF1 may regulate GR transactivation of the mA8 promoter in a manner similar to that described by Hebbar and Archer (34) as a classical transcription factor and in alteration of chromatin. AP-1 can either enhance or repress GCs-responsive genes depending on whether it is composed of c-jun homodimers or jun/fos heterodimers respectively (21). AP-1 binding motifs are located at –164, –102, and –18 before the transcriptional start site and are highly homologous in the murine and human genes. Because jun homodimers enhance LPS-induced expression of mA8 (R. J. Passey and C. L. Geczy, unpublished results), GCs may enhance A8 through a related mechanism.
Enhancement by DEX was also strongly reliant on the COX-2 pathway, which leads to increased PGE2 and cAMP that also enhance LPS induction of mA8 (7) and of human A8/9 mRNA in myelocytic differentiation (49). Mechanisms of GC enhancement via this pathway are unknown, as GR normally suppresses COX-2 induction (21). Its involvement is likely to be independent of IL-10, as inclusion of inhibitors of both pathways reduced synergy to a greater extent than when only a single inhibitor was added. MAPK pathways are involved in both LPS up-regulation and DEX enhancement, whereas PKC and PKA have opposite effect on modulating mA8 gene expression by DEX. GCs reduce PKC activity in rat liver parenchymal cells (50), and increased levels of cAMP and/or activation of PKA can potentiate GC-dependent gene expression by increasing GR affinity (51). Although the GR was traditionally viewed as ligand-activated, activation can also occur via alternate mechanisms including protein-protein interactions and phosphorylation pathways including PKA and PKC (52) and MAPK, with or without other transcription factors (21). Alterations to conformation or degree of phosphorylation of the complex may not be detected by gel shift analysis (Fig. 5B). Interactions between the MAPK, PKA, and PKC pathways involved in A8 induction may provide an additional level of control in this system.
The therapeutic effects of GCs are largely due to their ability to inhibit functions of macrophages and other APCs. Interestingly, IL-10 and GCs have several effects in common, particularly affecting macrophage function, possibly via their interference with transcriptional activation of AP-1 and NF-B (21, 53). Conversely, both agents synergized with LPS to up-regulate mA8. GCs and IL-10 both induce GCs-induced leucine zipper, a protein that prevents NF-B-mediated activation of transcription, in macrophages activated via Toll-like receptors and proposed to play a key anti-inflammatory role (54). Our results support the proposal that A8 has anti-inflammatory properties which could be manifested in functions such as its ability to down-regulate leukocyte adhesion (55), to regulate NADPH oxidase (18), to scavenge reactive oxygen intermediates (17) and, in combination with A9, to act as an anti-microbial agent. The role of GCs-induced leucine zipper in A8 gene regulation is worthy of investigation.
Apart from potential therapeutic effects, the synergistic induction in macrophage, endothelial cell, and fibroblast activation of A8 may be important in Gram-negative sepsis, where high endogenous GCs levels (micromolar (56)) and endotoxin coexist. mA8 is up-regulated by LPS and recruits neutrophils and monocytes but fails to activate these cells (57) and interestingly, neutrophils in LPS-DEX-treated rats have these properties and cause potentially less damage (58). Up-regulation of A8 in several cell types by GCs may enable positive host defense against invading organisms in the absence of overriding deleterious host responses. These may represent counterregulatory systems pivotal to maintaining host defense and may positively compensate for some of the anti-inflammatory consequences of GC action.
The significant increase in A8-positive macrophages in synovial lining from RA patients treated with MP suggests that the A8 may play an anti-inflammatory role. Human A8 and S100A12 (A12) appear to be more recently evolved from mA8 which exhibits some functions of both these proteins. A12 does not occur in the murine genome (1) and A12 and mA8 are chemotactic for monocytes in the picomolar range. Human (13) and murine (12) A8 recruit neutrophils and both can scavenge oxidants (17), a property not reliant on heterodimer formation (17). Conversely, A12 is not an efficient oxidant scavenger because it lacks cysteine and methionine residues and interestingly, preliminary studies show that A12 mRNA in monocytes is not regulated by GC. Thus high levels of A8 induced by GCs may downmodulate inflammation by processes involving oxidant defense, a property now suggested for several S100 proteins.
Acknowledgments
We are grateful to Dr. Zheng Yang and Dr. Shengping Hu for assistance with reporter analysis in initial experiments and to Dr. Johannes Roth (Institute of Experimental Dermatology and Department of Pediatrics, University of Munster, Munster, Germany) for performing immunostaining.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This study has been supported by grants from the National Health and Medical Research Council of Australia.
2 Current address: School of Photovoltaic Engineering, University of New South Wales, Sydney, New South Wales 2052, Australia.
3 Current address: Pacific Laboratory Medicine Services, Royal North Shore Hospital, Pacific Highway, St. Leonards, New South Wales 2065, Australia.
4 Address correspondence and reprint requests to Prof. Carolyn L. Geczy, Inflammatory Diseases Research Unit, School of Medical Sciences, University of New South Wales, Sydney, New South Wales 2052, Australia. E-mail address: c.geczy@unsw.edu.au
5 Abbreviations used in this paper: A8, S100A8; A9, S100A9; ActD, actinomycin D; CalC, calphostin C; CHX, cycloheximide; CM, culture medium; COX-2, cyclooxygenase-2; DEX, dexamethasone; GC, glucocorticoid; GR, GC-GC receptor; GRE, GC response elements; HPRT, hypoxanthine-guanine phosphoribosyltransferase; HYD, hydrocortisone; mA8, murine A8; MEC, microvascular endothelial cell; PKA, protein kinase A; PKC, protein kinase C; MP, methylprednisolone; TG, thioglycolate; RA, rheumatoid arthritis.
Received for publication July 16, 2004. Accepted for publication December 1, 2004.
References
Ravasi, T., K. Hsu, J. Goyette, K. Schroder, Z. Yang, F. Rahimi, L. P. Miranda, P. F. Alewood, D. A. Hume, C. Geczy. 2004. Probing the S100 protein family through genomic and functional analysis. Genomics 84:10.
Foell, D., M. Frosch, C. Sorg, J. Roth. 2004. Phagocyte-specific calcium-binding S100 proteins as clinical laboratory markers of inflammation. Clin. Chim. Acta 344:37.
Youssef, P., J. Roth, M. Frosch, P. Costello, O. Fitzgerald, C. Sorg, B. Bresnihan. 1999. Expression of myeloid related proteins (MRP) 8 and 14 and the MRP8/14 heterodimer in rheumatoid arthritis synovial membrane. J. Rheumatol. 26:2523
Hobbs, J. A., R. May, K. Tanousis, E. McNeill, M. Mathies, C. Gebhardt, R. Henderson, M. J. Robinson, N. Hogg. 2003. Myeloid cell function in MRP-14 (S100A9) null mice. Mol. Cell Biol. 23:2564.
Nacken, W., J. Roth, C. Sorg, C. Kerkhoff. 2003. S100A9/S100A8: myeloid representatives of the S100 protein family as prominent players in innate immunity. Microsc. Res. Tech. 60:569.
Hu, S. P., C. Harrison, K. Xu, C. J. Cornish, C. L. Geczy. 1996. Induction of the chemotactic S100 protein, CP-10, in monocyte/macrophages by lipopolysaccharide. Blood 87:3919.
Xu, K., C. L. Geczy. 2000. IFN- and TNF regulate macrophage expression of the chemotactic S100 protein S100A8. J. Immunol. 164:4916.
Yen, T., C. A. Harrison, J. M. Devery, S. Leong, S. E. Iismaa, T. Yoshimura, C. L. Geczy. 1997. Induction of the S100 chemotactic protein, CP-10, in murine microvascular endothelial cells by proinflammatory stimuli. Blood 90:4812
Grimbaldeston, M. A., C. L. Geczy, N. Tedla, J. J. Finlay-Jones, P. H. Hart. 2003. S100A8 induction in keratinocytes by ultraviolet A irradiation is dependent on reactive oxygen intermediates. J. Invest. Dermatol. 121:1168
Xu, K., T. Yen, C. L. Geczy. 2001. IL-10 up-regulates macrophage expression of the S100 protein S100A8. J. Immunol. 166:6358.
Donato, R.. 2001. S100: a multigenic family of calcium-modulated proteins of the EF-hand type with intracellular and extracellular functional roles. Int. J. Biochem. Cell Biol. 33:637
Passey, R. J., K. Xu, D. A. Hume, C. L. Geczy. 1999. S100A8: emerging functions and regulation. J. Leukocyte Biol. 66:549
Ryckman, C., K. Vandal, P. Rouleau, M. Talbot, P. A. Tessier. 2003. Proinflammatory activities of S100: proteins S100A8, S100A9, and S100A8/A9 induce neutrophil chemotaxis and adhesion. J. Immunol. 170:3233.
Manitz, M. P., B. Horst, S. Seeliger, A. Strey, B. V. Skryabin, M. Gunzer, W. Frings, F. Schonlau, J. Roth, C. Sorg, W. Nacken. 2003. Loss of S100A9 (MRP14) results in reduced interleukin-8-induced CD11b surface expression, a polarized microfilament system, and diminished responsiveness to chemoattractants in vitro. Mol. Cell Biol. 23:1034.
Ahmad, A., D. L. Bayley, S. He, R. A. Stockley. 2003. Myeloid related protein-8/14 stimulates interleukin-8 production in airway epithelial cells. Am. J. Respir. Cell Mol. Biol. 29:523
Passey, R. J., E. Williams, A. M. Lichanska, C. Wells, S. Hu, C. L. Geczy, M. H. Little, D. A. Hume. 1999. A null mutation in the inflammation-associated S100 protein S100A8 causes early resorption of the mouse embryo. J. Immunol. 163:2209.
Harrison, C. A., M. J. Raftery, J. Walsh, P. Alewood, S. E. Iismaa, S. Thliveris, C. L. Geczy. 1999. Oxidation regulates the inflammatory properties of the murine S100 protein S100A8. J. Biol. Chem. 274:8561.
Doussiere, J., F. Bouzidi, P. V. Vignais. 2002. The S100A8/A9 protein as a partner for the cytosolic factors of NADPH oxidase activation in neutrophils. Eur. J. Biochem. 269:3246.
McKay, L. I., J. A. Cidlowski. 1999. Molecular control of immune/inflammatory responses: interactions between nuclear factor-B and steroid receptor-signaling pathways. Endocr. Rev. 20:435.
Funder, J. W.. 1997. Glucocorticoid and mineralocorticoid receptors: biology and clinical relevance. Annu. Rev. Med. 48:231
De Bosscher, K., W. Van den Berghe, G. Haegeman. 2003. The interplay between the glucocorticoid receptor and nuclear factor-B or activator protein-1: molecular mechanisms for gene repression. Endocr. Rev. 24:488.
Re, F., M. Muzio, M. De Rossi, N. Polentarutti, J. G. Giri, A. Mantovani, F. Colotta. 1994. The type II "receptor" as a decoy target for interleukin 1 in polymorphonuclear leukocytes: characterization of induction by dexamethasone and ligand binding properties of the released decoy receptor. J. Exp. Med. 179:739.
Calandra, T., J. Bernhagen, C. N. Metz, L. A. Spiegel, M. Bacher, T. Donnelly, A. Cerami, R. Bucala. 1995. MIF as a glucocorticoid-induced modulator of cytokine production. Nature 377:68.(Kenneth Hsu, Robert J. Pa)