Synthesis of glucocorticoid-induced leucine zipper (GILZ) by macrophages: an anti-inflammatory and immunosuppressive mechanism shared by glucocorticoids and IL-10

Glucocorticoids (GCs) are potent anti-inflammatory and immunosuppressive drugs. Their therapeutic effects are largely due to their ability to inhibit many functions of macrophages and of other antigen-presenting cells. Interleukin 10 (IL-10) is an anti-inflammatory cytokine that has a number of effects in common with GCs, particularly those affecting macrophage functions. Both GCs1-13 and IL-10 (reviewed in Stordeur and Goldman14) inhibit antigen processing, the expression of HLA, CD80, and CD86, and the synthesis of nitric oxide, cyclo-oxygenase 2, adhesion molecules, cytokines, and chemokines. The intracellular events induced by the binding of GCs and IL-10 to their respective receptors are not fully understood, but they also share certain characteristics. In particular, both GCs and IL-10 interfere with the function of the transcriptional activators AP-1 and NF-κB (reviewed in Stordeur and Goldman,14 Karin,15 and Barnes and Karin16). Pathogen-associated molecular patterns (PAMPs) of bacterial components activate macrophages by binding to the toll-like receptors (TLRs), which trigger the nuclear factor κB (NF-κB) pathway and stimulate the production of inflammatory proteins, including cytokines, chemokines, CD80 and CD86 (reviewed in Aderem and Ulevitch,17 Akira et al,18 Golenbock and Fenton,19 Zhang and Ghosh,20 and Wagner21). The involvement of NF-κB in this process may explain why both GCs and IL-10 inhibit the activation of macrophages by PAMPs.

While comparing cDNAs from untreated and dexamethasone (DXM)–treated murine T cells, we identified a new member of the leucine zipper family, glucocorticoid-induced leucine zipper (GILZ).22 The GILZ gene is expressed in the lymphocytes of healthy mice and in human lymphohemopoietic cells23(murine GILZ GenBank accession number, AF024519; human GILZ GenBank accession number, AF228339). GILZ interferes with the function of AP-1 and NF-κB in T lymphocytes,22,24,25 and it prevents the FasL expression and apoptosis induced by cross-linking of the antigen receptor.22,25 GILZ inhibits NF-κB by directly binding to the p65 subunit of this factor, thereby preventing its interaction with specific sites on gene regulatory elements.25 

As NF-κB and AP-1 play a critical role in macrophage activation, we investigated whether GILZ was produced by macrophages and whether it mediated the immunosuppressive and anti-inflammatory effects of GCs in these cells. We also investigated whether GILZ production was stimulated by IL-10, because this could account for the similarities in the effects of GCs and IL-10. We found that macrophages produced GILZ in noninflamed tissues in humans and mice. GILZ production by macrophages was stimulated by GCs and by IL-10, both in vitro and in vivo. In addition, GILZ inhibited the expression of TLRs and of molecules involved in antigen presentation and in inflammation. GILZ gene expression was down-regulated in macrophages from inflammatory lesions of delayed-type hypersensitivity (DTH) reactions, whereas it persisted in tumor-infiltrating macrophages from Burkitt lymphomas.

Human monocytes were isolated from healthy human volunteers by Percoll centrifugation. Monocyte pellets (> 90% pure, as assessed by cytology and CD68 staining) were resuspended in macrophage serum-free medium (SFM) (Gibco/BRL, Cergy-Pontoise, France) and cultured at a density of 3 × 10⁶ cells/mL in polypropylene tubes (Labortechnik, Poitiers, France). Cells were recovered 24 hours later and cultured for 4 hours in macrophage SFM with various stimulants: rhIL-4, rhIL-10 (a gift from K. Moore, DNAX, Palo Alto, CA), rhIL-13 (a gift from A. Minty, Sanofi Elf Bio Recherches, Labege, France), each used at a concentration of 20 ng/mL, or rhIFNγ (PrepoTech, Rocky Hill, NJ), used at a concentration of 100 U/mL. DXM (Sigma, L'isle d'Abeau, France) was used at a concentration of 10⁻⁷ M. The THP-1 cell line was cultured in RPMI medium supplemented with 10% fetal calf serum (Life Technologies, Cergy-Pontoise, France) and was stimulated with rhIFNγ (100 U/mL), lipopolysaccharide (LPS; 1 μg/mL, Sigma), or an anti-CD40 mAb (1 μg/mL, clone G28.5, American Type Culture Collection, Manassas, VA). In some experiments, cycloheximide (CHX; 20 mM, Sigma) was added to the cells 30 minutes before stimulation with DXM.

Mouse macrophages were isolated from the peritoneal cavity of 5- to 7-week-old C3H/HeN mice (Iffa Credo, L'Abresle, France), 4-5 days after the injection of 0.5 mL endotoxin-free 10% thioglycollate medium (Difco Laboratories, Detroit, MI). Endotoxin was eliminated from all solutions with Detoxi-gel (Pierce, Rockford, IL). The cells were allowed to adhere to a plastic surface for 2 hours, and nonadherent cells were then removed. On flow cytometric analysis, ∼99% of adherent cells expressed Mac-1 but not B220, CD3, or TCRα/β. For in vitro experiments, cells were incubated for 4 hours at 37°C alone or with 10⁻⁷ M DXM.

The expression of CD80, CD86, and TLR2 on THP-1 cells was evaluated by flow cytometry. THP-1 cells were incubated for 1 hour in the presence of human serum. We assessed the expression of CD80 and CD86 using phycoerythrin (PE)–conjugated anti-CD80 and anti-CD86 mAbs (Becton Dickinson, San Jose, CA). TLR2 expression was evaluated by incubation with a goat polyclonal anti-TLR2 Ab (Santa Cruz, Santa Cruz, CA), followed by a PE-conjugated anti–goat F(ab')₂ (Jackson Immunoresearch, West Grove, PA). As controls, we used a PE-conjugated IgG1 mAb (Immunotech, Marseilles, France) and a normal goat polyclonal IgG (Santa Cruz). Flow cytometry was performed using a FACScan machine (Becton Dickinson).

The human GILZ cDNA sequence was inserted into theEcoRI site of pcDNA3 (Invitrogen, San Diego, CA) (GILZ-pcDNA3). THP-1 cells were cotransfected by electroporation (250 volts, 960 μF) with either 20 μg pcDNA3 (control) or 20 μg of GILZ-pcDNA3 and 2 μg of pGFP (Clontech, Palo Alto, CA). IFNγ, LPS, or anti-CD40 mAb was added to the culture 24 hours later. Viability, cell recovery, and the fraction of green fluorescence protein–producing cells were similar in the control and GILZ-pcDNA3–transfected cells. The expression of CD80, CD86, and TLR2 was evaluated 48 hours after transfection. Regulated on activation normal T-cell expressed and secreted (RANTES) and macrophage inflammatory protein 1α (MIP-1α) concentrations were determined in the supernatants 72 hours after transfection by enzyme-linked immunosorbent assay (R&D, Minneapolis, MN). The effect of GILZ on NF-κB function was determined by cotransfection of 12 μg of either GILZ-pcDNA3 or the pcDNA3 empty vector and 8 μg of either the NF-κB–luciferase reporter plasmid26 or pcDNA3. Luciferase activity was tested 4 hours after addition of LPS or medium alone to the cells.

Extracted proteins were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotted, as previously described.27 The primary antibody was a rabbit polyclonal antiserum recognizing GILZ.22 The secondary antibody was a horseradish peroxidase–labeled goat anti–rabbit IgG (Pierce). An anti–β tubulin mAb (Calbiochem, San Diego, CA) was used as the control. The antigen-antibody complexes were detected by enhanced chemiluminescence (SuperSignal, Pierce).

Whole THP-1 cell extracts were prepared, and immunoprecipitation was performed in radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris [tris(hydroxymethyl)aminomethane] pH 7.5, 150 mM NaCl, 1% NP-40, 0.5% deoxycholate, 0.1% SDS, and 5 mM EDTA [ethylenediaminetetraacetic acid]) supplemented with 1 mM phenylmethylsulfonyl fluoride. Antigen-antibody complexes were precipitated with protein A bound to Sepharose beads (Pharmacia, Uppsala, Sweden) prior to SDS-PAGE. A rabbit polyclonal antiserum recognizing GILZ was prepared using a fusion protein containing the full GILZ amino-acid sequence. Protein lysates (500 μg in RIPA buffer) from untreated, DXM-treated, or cytokine-treated cells were immunoprecipitated with 3 μL of anti–p65 subunit rabbit polyclonal Ab (Biomol Research Laboratories, Plymouth Meeting, PA). For coimmunoprecipitation experiments, Western blots were carried out with anti-p65 and anti-GILZ Abs, as previously described.24 

Expression of the human GILZ gene was analyzed by RT-PCR. β-actin mRNA was first quantified by competitive reverse transcription–polymerase chain reaction (RT-PCR).28 hGILZ mRNA was amplified in samples containing equivalent numbers of β actin mRNA molecules. Forty cycles of PCR (each consisting of 60 seconds at 94°C, 90 seconds at 62°C, and 60 seconds at 72°C) were performed with the 5′-TCTGCTTGGAGGGGATGTGG-3′ sense primer and the 5′-ACTTGTGGGGATTCGGGAGC-3′ antisense primer. For the quantification of murine GILZ gene expression, the cDNA produced from 100 ng of total RNA was amplified by 40 cycles of PCR (each cycle consisting of 60 seconds at 94°C, 90 seconds at 62°C, and 60 seconds at 72°C) using the 5′-CAGCAGCCACTCAAACCAGC-3′ sense primer and the 5′-ACCACATCCCCTCCAAGCAG-3′ antisense primer. The amplified products were subjected to gel electrophoresis and analyzed by computed-assisted densitometry with National Institutes of Health Image 1.62 software (Bethesda, MD).

Total RNA was extracted using Trizol (Gibco-BRL, Cergy-Pontoise, France). A murine GILZ probe for RNase protection assays was constructed by PCR. The PCR product (210 bp) was inserted into the vector pCRII-TOPO (Invitrogen). The cloning product was sequenced to confirm that no point mutation had been introduced. The GILZ probe was linearized with XhoI and transcribed with T7 RNA polymerase (Gibco/BRL) in the presence of α³²[P]UTP; the transcription products were purified on a Sephadex G50 column (Roche, Meylan, France). The assay was performed using the RPA III kit (Ambion, Austin, TX). Total RNA (5 μg) was hybridized overnight at 60°C with 2 × 10⁵ cpm of GILZ antisense probe and 4 × 10⁴ cpm of β-actin antisense probe. A mouse β-actin antisense probe (Ambion) was used as an internal control to standardize expression levels between samples. Samples were run on a 6% polyacrylamide/8 M urea denaturing gel. The gel was placed against x-ray film for 1 day at −70°C. The intensity of each band was measured with a PhosphorImager (Molecular Dynamics, Sunnyvale, CA).

In situ hybridization experiments were performed on frozen tissue sections as previously described.29 For the detection of hGILZ, a digoxigenin (DIG)–labeled antisense probe was obtained by linearizing GILZ-pcDNA3 with HindIII and synthesizing a cRNA under control of the SP6 promoter. The sense probe was obtained by linearizing GILZ-pcDNA3 with ApaI and synthesizing a cRNA under control of the T7 promoter. Normal human tissues were studied using 3 samples, each from a different individual, for every organ. Small intestine samples containing granulomas from 3 patients with Crohn disease and lymph nodes from 2 patients with active tuberculosis were studied for GILZ gene expression. Also analyzed were 2 lymph nodes from patients with classical Burkitt lymphomas.

For double-labeling experiments, in situ hybridization was performed as described above. Immunohistochemistry was then performed, using the 3-stage peroxidase technique with an anti-CD68 mAb (Dako, Glastrup, Denmark).

For the detection of mGILZ, sense and antisense probes were synthesized from the GILZ gene–containing pCRII-TOPO vector. The level of expression of the mGILZ gene was determined in 7-week-old female BALB/C and IL-10-transgenic mice (IL-10-Tg),30 kindly provided by H. Groux (Nice, France). These mice are transgenic for the IL-10 gene under the control of a murine major histocompatibility complex class II promoter.

Seven-week-old C57/BL6 mice were injected intraperitoneally with 10 μg/g DXM or phosphate-buffered saline (PBS). Tissue samples were collected 4 hours later. Also injected intraperitoneally were 5- to 7-week-old C3H/HeN mice, with 10 μg/g DXM 4-5 days after thioglycollate treatment, and peritoneal macrophages were isolated as described above.

GILZ gene expression was studied by in situ hybridization in normal human tissues. All tissues examined contained GILZ gene–expressing cells, the distribution and morphology of which were similar to those of macrophages and related cells. In the lung, GILZ gene–expressing cells were located mostly within the alveolar spaces and in the subpleural lymphatic sinus (Figure1A). No labeling was detected with the sense probe, used as a negative control (Figure 1B). In the colon, labeled cells were dispersed within the lamina propria. In the liver, labeled cells were detected within sinusoids; this and their morphology identified them as Küpffer cells (Figure 1C). In the kidney, all GILZ gene–expressing cells were located in the glomeruli; their distribution and morphology were consistent with those of mesangial cells (Figure 1D). Expression of the GILZ gene by macrophages was demonstrated directly in the lungs and liver by double-labeling experiments combining in situ hybridization with the GILZ antisense probe and immunohistochemistry with an anti-CD68 antibody. Most GILZ gene–expressing cells were CD68 positive (Figure 1F), and most CD68-positive cells expressed the GILZ gene. Therefore, in the absence of inflammation, the GILZ gene is constitutively expressed in nonlymphoid tissues, in which the major source of GILZ is the macrophage lineage.

We carried out in vitro experiments to confirm that macrophages produced GILZ and to study the regulation of GILZ gene expression. We analyzed GILZ gene expression in monocytes by semiquantitative RT-PCR. Weak expression of the GILZ gene was observed in unstimulated monocytes. The level of expression of the GILZ gene increased strongly and rapidly in cells cultured in the presence of DXM (Figure2A-B). Western blotting with an anti-GILZ antibody gave similar results: untreated monocytes (Figure 2C, lane 1) produced small amounts of GILZ, and the incubation of cell cultures with DXM for 4 hours resulted in a large increase in the amount of GILZ detected (lane 2).

The kinetics of GILZ gene induction was determined using the THP-1 monocytic cell line. Expression of the GILZ gene was induced as early as 30 minutes following DXM addition. It peaked at 90 minutes and decreased thereafter (Figure 3A). The requirement for protein synthesis to stimulate GILZ gene expression was analyzed. Expression of the GILZ gene was stronger in CHX-treated THP-1 cells than in controls (Figure 3B) 4 hours after DXM addition. Similar findings were observed with freshly isolated monocytes (data not shown). Thus, the in vitro production of GILZ by monocytes is strongly and rapidly increased by the addition of GCs, and it does not require protein synthesis.

We investigated whether IL-10, like GCs, stimulated GILZ production. Monocytes from healthy individuals were cultured either alone or in the presence of IL-10, and GILZ gene expression was monitored by RT-PCR. IL-10 up-regulated GILZ gene expression. We also analyzed the effect of IL-10 on GILZ production by Western blot analysis. The results obtained were consistent with RT-PCR studies, showing that IL-10 stimulated GILZ production by monocytes (Figure 4).

We investigated whether GILZ gene expression in monocytes could be stimulated by cytokines other than IL-10. IFNγ tested alone or in the presence of tumor necrosis factor α (TNFα) did not influence the level of GILZ gene expression (data not shown). In contrast, the Th2 cytokines IL-4 and IL-13 stimulated GILZ gene expression and GILZ production to an extent similar to that achieved with IL-10 (Figure 4).

We then evaluated the ability of GCs to stimulate the synthesis of GILZ in vivo in mice. Semiquantitative RT-PCR studies showed that, as in humans, GILZ was constitutively produced in the lungs, intestine, kidneys, and liver of mice. Four hours after the intraperitoneal administration of DXM, GILZ production strongly increased in the lungs and kidneys. It also increased in the intestine and liver, but to a lesser extent. These results were confirmed at the protein level: Western blotting showed a higher concentration of GILZ in the lungs of DXM-treated mice than in those of controls (Figure5A).

We carried out ribonuclease protection assay to investigate GILZ gene expression by purified macrophages from the peritoneal cavity. The GILZ gene was constitutively expressed by peritoneal macrophages, and the addition of GCs in vitro increased GILZ production (Figure 5B). We found that the level of GILZ expression by macrophages was higher in peritoneal cells collected 4 hours after DXM treatment than in peritoneal cells collected 4 hours after PBS treatment (Figure 5C).

We investigated whether IL-10 also stimulated GILZ gene expression in vivo by comparing GILZ gene expression in wild-type mice and mice specifically overproducing IL-10 in antigen-presenting cells. Expression of the GILZ gene was monitored by in situ hybridization in the lungs, liver, kidneys, and lymph nodes. In nonlymphoid organs, the density and distribution of GILZ gene–expressing cells were similar in IL-10 transgenic mice and in controls (data not shown). In contrast, the pattern of GILZ gene expression in lymph nodes differed greatly between the transgenic and control mice. In controls, only a few cells expressed the GILZ gene. Most of these cells were present in subcapsular lymphatic sinuses, a location consistent with their macrophage origin. The number of GILZ gene–expressing cells was much larger in IL-10 transgenic mice than in controls. Most positive cells were located in lymphatic sinuses. Their distribution extended beyond the subcapsular sinuses because numerous GILZ gene–expressing cells also were present in the trabecular sinuses. Positive cells also were found in parafollicular T-cell zones. Comparison, on serial sections, of Mac-1⁺ and GILZ gene–expressing cells showed that the vast majority of intrasinusal cells were macrophages in IL-10 transgenic mice and that macrophages and GILZ gene–expressing cells were identical in terms of density and location (Figure6).

Overall, these results show that, in mice, GC treatment and IL-10 overproduction both stimulate GILZ production in vivo, and that in both cases, macrophages are involved in this increase in GILZ production.

We then investigated whether the induction of GILZ production in macrophages was involved in the anti-inflammatory and immunosuppressive effects of GCs and IL-10.

In the absence of stimulation, THP-1 cells weakly expressed CD80 and CD86. IFNγ increased the expression of both molecules. This effect of IFNγ was prevented by the addition of either GCs or IL-10 (Figure7A). We then compared the expression of CD80 and CD86 in THP-1 cells transfected with a GILZ gene–expressing vector (GILZ-THP-1 cells) or with the pcDNA3 empty vector (control transfected cells). The increase in CD80 and CD86 expression induced by IFNγ persisted in control transfected cells but was abolished in GILZ-THP-1 cells (Figure 7B).

We also analyzed the effect of GILZ on the production of inflammatory chemokines by THP-1 cells. Control transfected cells constitutively produced the chemokines RANTES (CCL5) and MIP-1α (CCL3). IFNγ up-regulated the production of both molecules. In GILZ-THP-1 cells, the level of production of RANTES and of MIP-1α was lower than in the controls at baseline and after stimulation with IFNγ (Figure8A-B). These findings were confirmed at the mRNA level. RT-PCR experiments showed that the pattern of expression of the RANTES gene mirrored that of the GILZ gene in THP-1 cells: the RANTES gene was expressed at a low level in GILZ-THP-1 cells, and IFNγ did not greatly stimulate RANTES gene expression in these cells (Figure 8C). When THP-1 cells were stimulated by LPS instead of IFNγ, transfection of the GILZ gene also inhibited expression of RANTES, MIP-1α, CD80, and CD86 (data not shown).

Both GCs and IL-10 inhibit the stimulation of macrophages by PAMPs. We investigated whether this inhibition involved an effect of GCs and IL-10 on TLR expression and, if so, whether GILZ could reproduce such an effect. We used flow cytometry to assess the effects of GCs, IL-10, and GILZ on the expression of TLR2. Unstimulated THP-1 cells weakly expressed TLR-2 at their surface, and this expression was up-regulated if the cells were stimulated with LPS, IFNγ, or an anti-CD40 mAb. GCs and IL-10 prevented the increase in TLR2 expression induced by these 3 stimulants (Figure9A). We then compared TLR2 expression in GILZ-THP-1 cells and control transfected cells. An increase in TLR-2 expression was induced by LPS, IFNγ, and the anti-CD40 mAb in control transfected cells, but no such increase was observed in GILZ-THP-1 cells (Figure 9B).

Thus, GILZ gene expression in monocytes reproduces several of the effects of GCs and IL-10: it inhibits the production of inflammatory chemokines of the costimulatory molecules CD80 and CD86 and of TLR2.

We tried to unravel the intracellular mechanism of action of GILZ in THP-1 cells. We investigated whether GILZ interacted with the p65 subunit of NF-κB. Cells were cultured for 4 hours, either alone or in the presence of DXM or IL-10. Cellular proteins were immunoprecipitated with an anti-p65 antiserum and separated by SDS-PAGE. Immunoblotting was then performed with either an anti-p65 or an anti-GILZ antiserum. In the absence of DXM or cytokines, no GILZ was coimmunoprecipitated with p65. In contrast, GILZ was coimmunoprecipitated with p65 in THP-1 cells treated with either DXM or IL-10. Treatment of the cells with IL-4 or IL-13 also induced the binding of GILZ to the p65 subunit of NF-κB (Figure 10A-B).

To directly test the inhibitory effect of GILZ on NF-κB function, THP-1 cells were cotransfected with the GILZ-expressing vector and/or with a NF-κB–luciferase reporter plasmid. There was a spontaneous activity of NF-κB in THP-1 cells. Activation of the cells for 4 hours increased this NF-κB activity approximately 2.8-fold. Expression of GILZ inhibited both the spontaneous and LPS-induced NF-κB activity (Figure 10C).

We analyzed the in vivo expression of the GILZ gene by macrophages in humans in 2 contrasting clinical situations: DTH reactions, in which macrophages play an essential role in the inflammatory process, and tumors, in which macrophages take up apoptotic malignant cells by phagocytosis without promoting significant local inflammation or an antitumor immune reaction.

We studied 2 types of DTH reaction by in situ hybridization: Crohn disease and tuberculosis. In all cases, CD68⁺ cells were abundant within the granulomas. The macrophages in the granulomas were activated, as reflected by the strong expression of the RANTES gene at this site (Figure 11A). In contrast, little if any RANTES gene expression was detected outside granulomas. The pattern of GILZ gene expression mirrored that of the RANTES gene. Numerous GILZ gene–expressing cells were found in the tissue surrounding the granulomas. In contrast, granulomas contained no GILZ gene–expressing cells (Figure 11B). A similar pattern was observed in the 2 types of DTH reactions studied.

Two Burkitt lymphomas were analyzed. In tumors of this type, macrophages are easy to identify because they lead to the starry sky pattern of the tumor and because they contain apoptotic bodies (Figure 11C-D). In situ hybridization experiments showed that the GILZ gene was expressed by most macrophages infiltrating the tumor and particularly by those containing apoptotic bodies. In contrast, malignant cells did not express the GILZ gene at a detectable level (Figure 11E).

Thus, GILZ gene expression was down-regulated in activated macrophages from DTH reactions, whereas, in Burkitt lymphomas, it persisted in macrophages involved in the phagocytosis of apoptotic malignant cells.

GILZ is produced by B and T lymphocytes, and its production is increased by GCs.22 31 In this work, we showed that macrophages constitutively produced GILZ in normal tissues of humans and mice. We also showed that GCs and IL-10 stimulated the production of GILZ by macrophages, both in vitro and in vivo, in mice and humans. The ability of both GCs and IL-10 to stimulate GILZ production by macrophages is particularly interesting given the large number of anti-inflammatory and immunosuppressive properties shared by these 2 agents. The induction of GILZ production may be a common pathway used by GCs and IL-10 to inhibit inflammation and immune activation. Consistent with this hypothesis, we found that GILZ reproduce several of the effects of GCs and IL-10 on cells of the monocyte/macrophage lineage. The transfection of THP-1 cells with a plasmid encoding GILZ affected the production by these cells of 2 inflammatory chemokines, RANTES and MIP-1α, of the costimulatory molecules CD80 and CD86, and of the PAMP receptor TLR2.

The mechanism of action of GILZ in macrophages is probably the same as that in T lymphocytes, in which GILZ inhibits AP-1 and NF-κB functions.22,24,25 We show that in THP-1 cells, GILZ binds to the p65 subunit of NF-κB and prevents transcription from NF-κB–dependent regulatory elements. LPS or CD40 triggering activates the NF-κB pathway in macrophages. The interaction between GILZ and NF-κB may thus account for the inhibition by GILZ of macrophage activation by these 2 stimulants. IFNγ activates neither the NF-κB nor the AP-1 pathways, but many of the effects of IFNγ are observed only in the presence of TNFα, which itself activates the NF-κB pathway.9,32 By interfering with NF-κB, and thus with effects of TNFα produced by THP-1 cells,9 GILZ may indirectly inhibit the response of these cells to IFNγ.

The Th2 cytokines IL-4 and IL-13 also stimulate the production of GILZ and its binding to p65. This result is consistent with the ability of IL-4 and IL-13 to inhibit the response of macrophages to LPS, an effect involving inhibition of the NF-κB pathway.33-38 We indeed observed that IL-4 and IL-13 inhibited LPS-induced CD80 and CD86 expression in THP-1 cells (data not shown). However, several other effects of IL-4 and IL-13 significantly differ from those of IL-10 and GCs. For example, IL-4 and IL-13 stimulate the production by macrophages of several inflammatory chemokines, including macrophage-derived chemokine and RANTES,9 39reflecting important differences in the mechanism of action of these 2 cytokines with respect to those of GCs and IL-10.

By deactivating macrophages and antigen-presenting cells, GCs and IL-10 abolish inflammation and favor immune tolerance. Our results indicate that these effects may be mediated by GILZ. Expression of the costimulatory molecules CD80 and CD86 during antigen presentation to T lymphocytes is required for T-lymphocyte activation: in the absence of CD80 and CD86, antigen presentation results in T-lymphocyte anergy or apoptosis.40-43 Activation of antigen-presenting cells by CD40 is another important step in the balance between immune activation and tolerance induction.44 45 Thus, by inhibiting the expression of CD80 and CD86 by macrophages and the activation of such cells through CD40, GILZ should promote tolerance rather than T-lymphocyte activation.

The production of inflammatory chemokines leads to the local recruitment of leukocytes to inflamed tissues. The transcriptional activation of many chemokine genes requires the binding of NF-κB to their regulatory sequences. The inhibition of NF-κB activity by GILZ may thus account for the inhibition by GCs and IL-10 of cell recruitment in inflammatory lesions. The binding of PAMPs to TLRs activates the NF-κB pathway, which in turn stimulates the production of inflammatory mediators. IL-10 inhibits the expression of TLR4 in dendritic cells.46 We found that GCs and IL-10 also inhibited the expression of TLR2 in macrophages and that this effect was mediated by GILZ. The ability of GILZ to inhibit both TLR expression and NF-κB signaling may explain why GCs and IL-10 limit the extent of inflammation induced by infectious agents.

The activation of T lymphocytes through the T-cell receptor/CD3 complex inhibits GILZ production in T lymphocytes.25 Similarly, the activation of B lymphocytes via the antigen receptor down-regulates GILZ expression in these cells.31 It is interesting to note that in granulomas of DTH reactions, macrophages are strongly activated but produce no GILZ. This suggests that the inhibition of GILZ production may be a general phenomenon during the activation of immune cells, including macrophages. The mechanism of inhibition of GILZ production in activated cells is unknown, but it may involve the presence of inhibitory factors. We indeed observed that CHX enhances the induction of GILZ gene expression in THP-1 cells, suggesting that this expression is down-regulated by inhibitory factors. In DTH reactions, down-regulation of GILZ production in macrophages would in turn increase the expression of TLR and inflammatory chemokines. Among these latter, RANTES is critical in the development of DTH granulomas.47 48 

The activation pathway involving PAMPs, TLR, and NF-κB has been implicated in the induction of Th1 responses,49 thus linking innate immune responses to DTH reactions. IL-10 prevents the Th1-inducing adjuvant property of LPS.50 The ability of both IL-10 and GCs to up-regulate GILZ gene expression presumably accounts for the efficiency of both agents in preventing and treating exacerbated Th1 responses such as Crohn disease, in which stimulation of intestinal immune cells by the gut flora is critical.51 

Unlike macrophages in DTH reactions, tumor-infiltrating macrophages in Burkitt lymphomas do not display down-regulation of GILZ production. This may reflect the strategy by which the tumor escapes immune surveillance. The phagocytosis of apoptotic cells stimulates IL-10 production by macrophages,52,53 and tumor-associated macrophages produce IL-10.54 This may explain why tumor-infiltrating macrophages, which are clearly involved in the phagocytosis of malignant cells in Burkitt lymphomas, continue to produce GILZ. IL-10 produced in tumors inhibits NF-κB activation in macrophages, prevents inflammation, and favors immune tolerance.54-56 Our findings suggest that GILZ production by tumor-infiltrating macrophages helps the tumor to avoid detection by the immune system.

Overall, our findings show that GCs and IL-10 stimulate the production of GILZ by macrophages and that GILZ mediates macrophage deactivation by these 2 agents. Therefore, GILZ appears to be a critical actor in the anti-inflammatory and immunosuppressive effects of GCs and IL-10. The development of DTH reactions is associated with the down-regulation of GILZ gene expression within lesions, which may be essential for Th1-related inflammation. In contrast, GILZ gene expression persists in the tumor-infiltrating macrophages of Burkitt lymphomas, possibly contributing to the failure of the immune system to reject the tumor. Therefore, expression of the GILZ gene by macrophages affects the balance between exacerbated immune reactions and immune tolerance. This property may be a key element in the anti-inflammatory and immunosuppressive effects of GCs and IL-10.

We thank Anne Marfaing-Koka for technical advice, Yolande Richard for the gift of the anti-CD40 mAb, and Régine Paris and Agnès Florentin for technical assistance. Hervé Groux is acknowledged for the gift of IL-10 transgenic mice, and Mercedes Rincon for the gift of the NF-κB–luciferase plasmid.