SOCS1 can regulate TLR-mediated signal transduction, yet mechanistic studies in murine macrophages have been confusing and contradictory. This study has used an adenoviral transfection system to determine the role of SOCS1 in the regulation of TNF-α production by activated human monocytes. Monocytes were infected with AdV-SOCS1 or with an empty vector control, AdV-GFP, for 24 h before activation with the TLR4 ligand, LPS. SOCS1 did not regulate TNF-α mRNA or protein production within the first two hours of TLR4 activation. However, SOCS1 suppressed the sustained production of TNF-α by primary human monocytes and synovial fluid macrophages ex vivo. In addition, SOCS1 regulated the production of IL-6, but not IL-10, by monocytes. Analysis of the early signaling pathway downstream of TLR4 demonstrated that SOCS1 had no regulatory effect on the activation or on the DNA binding capacity of NFκB. The late effects of LPS are mediated in part through the MyD88-independent pathway activating IRF3 and initiating the production of IFN-β. In response to adenoviral infection and before LPS exposure, monocytes expressed enhanced levels of IFN-β and Myxovirus A mRNA, an anti-viral molecule characterizing IFN-β activity. These two genes were reduced in AdV-SOCS1-infected cells. Further, SOCS1 regulated IFN-dependent pathways in LPS-activated cells as evidenced by reduced IFN-β production and STAT1 phosphorylation. Using AdV-infection to dissect SOCS1 control of IFN-dependent pathways, this study suggests that SOCS1-regulation of the IFN-dependent component of the LPS-induced TLR4 signaling pathway may contribute to the down-regulation of inflammatory cytokine production by AdV-SOCS1-infected human monocytes.
Chronic inflammatory diseases such as rheumatoid arthritis, Crohn’s disease, and inflammatory bowel disease are characterized by prolonged and deregulated proinflammatory cytokine production. TNF-α is arguably the most important proinflammatory cytokine driving the inflammatory response, initiating a cascade of events that controls the sequential production of other inflammatory cytokines such as IL-1 and IL-6. Although anti-TNF-α therapies have been successful at reducing inflammation, the side effects associated with such treatments, such as the risk of sepsis, have prompted the development of therapeutic strategies to specifically target the intracellular signaling intermediates responsible for regulating TNF-α production. Development of this approach requires a better understanding of the signaling pathways initiating and regulating TNF-α production in human cells.
TLRs are a family of membrane bound pattern recognition receptors that respond to invading pathogens to initiate an immune response (reviewed in Ref. 1). TLR4 is expressed on macrophages, and expression levels and activation through this receptor is increased in synovial fluid macrophages from patients with rheumatoid arthritis (2). TLRs recognize many endogenous ligands (such as fibronectin/hyaluronan fragments, self-mRNA, and high mobility group box 1) (3). It has been hypothesized that once joint damage occurs, a self-perpetuating process mediated by endogenous ligands through TLR-activation is established, resulting in chronic progressive disease.
The signaling pathways initiated downstream of the TLR, as a consequence of ligand binding, are responsible for driving the production of TNF-α. Specificity of signaling is determined by ligand-specific receptor usage and the differential recruitment of the adaptor proteins MyD88, MyD88-adaptor-like (Mal),3 toll/interleukin-1 receptor domain-containing adaptor protein inducing IFN-β (TRIF), and TRIF-related adaptor molecule (TRAM) which are required for signal amplification (4, 5). For example, the TLR4 ligand LPS initiates two signal transduction pathways, the MyD88-dependent and a MyD88-independent, TRIF-dependent pathway. The early production of TNF-α, i.e., within the first few hours of receptor activation, is due to the rapid activation of the NFκB pathway. NFκB is sequestered in an inactive form in the cytoplasm via its association with the inhibitory κB (IκB) complex. Following activation, NFκB is released into the cytoplasm and translocates to the nucleus where it initiates gene transcription. TLR4 signaling via the MyD88-independent pathway requires the recruitment of TRIF and TRAM and contributes to the sustained production of TNF-α. Activation of this pathway culminates in the activation of IRF3, a member of the IRF family of transcription factors which regulate the expression of type I IFNs such as IFN-α/β (6). The induction of IFN-β production by IRF3 is central to the initiation of the anti-viral immune response (7, 8, 9, 10).
A number of recent reports have suggested that TLR-mediated signaling can be regulated by members of the suppressor of cytokine signaling (SOCS) family of proteins. The SOCS proteins are a family of eight proteins that include cytokine-inducible SH2-domain containing protein and SOCS1–7. SOCS1 was first characterized as a regulator of IFN-γ-induced inflammatory responses (11, 12), but more recent studies have suggested that the regulatory capacities of SOCS1 are not restricted to the regulation of IFN. Recently there has been much interest in the ability of SOCS1 to regulate TLR-mediated signal transduction and the subsequent production of proinflammatory cytokines (13, 14, 15, 16, 17). SOCS1 expression is increased in macrophages following LPS exposure (13). Furthermore, studies in bone marrow-derived macrophages suggest negative regulation of TLR-mediated signaling by SOCS1. The mechanisms underlying SOCS1 regulation of proinflammatory cytokine production remain unclear. Studies in murine macrophages have suggested control of NFκB signaling by SOCS1 (14, 15, 16, 17), whereas others suggest that SOCS1 regulates IRF3-driven processes (13, 18, 19, 20).
The capacity of SOCS1 to regulate TLR-induced proinflammatory cytokine production by activated human monocytes and macrophages has not been investigated. In this study, infection of human monocytes with a SOCS1-expressing adenovirus did not regulate TNF-α mRNA or protein production within the first two hours of TLR activation. However, by 24 h, TLR-induced TNF-α levels were significantly reduced. Our results suggest that this effect may be due to regulation of the MyD88-independent TRIF-dependent pathway. Furthermore, AdV-SOCS1 control of TNF-α production was confirmed ex vivo in primary human macrophages obtained from the synovial fluid of patients with inflammatory arthritis.
Materials and Methods
Generation of SOCS1 expressing adenovirus
A human SOCS1-containing construct was donated by Prof. N. Nicola and Dr. T. Willson (Walter and Eliza Hall Institute of Medical Research, Parkville, Victoria, Australia). SOCS1 cDNA was amplified using gene specific primers designed to generate an N-terminal hemagglutinin tag and restriction sites XhoI and HindIII at the 5′ and 3′ cDNA ends, respectively. Following amplification and purification, hemagglutinin-SOCS1 cDNA was subcloned into a GFP-expressing adenoviral transfer vector, pAdTrack-CMV (Qbiogene). Positive clones were selected using kanamycin selective medium and confirmed by restriction digest and sequence analysis. pAdTrack-CMV constructs containing no insert (AdV-GFP) or SOCS1 (AdV-SOCS1) were linearized by digesting with PacI and transformed into BJ5183 AD-1 cells (Stratagene) containing pAdEasy-1 (Qbiogene). Recombinants containing SOCS1 and GFP were selected on the basis of colony size and confirmed by restriction digest and sequence analysis. pAdEasy-1 constructs containing SOCS1 or control constructs were transfected into complementing human embryonic kidney (HEK) 293 cells using Lipofectamine 2000 (Invitrogen). HEK 293 cells were purchased from BD Clontech and were maintained in DMEM (Sigma) supplemented with 5% FCS (ThermoTrace), 5 μg/ml gentamicin (Invitrogen), 2 mM l-glutamine (Invitrogen) and 1 mM sodium pyruvate and incubated at 37°C in 5% CO2. Recombinant viruses were purified from both the HEK 293 cell lysates and culture supernatants using a modified chromatography-based system, based on BD Clontech’s Adeno-X virus purification kit. The generation of the SOCS3 expressing adenovirus (AdV-SOCS3) by this laboratory has been previously published (21).
Isolation of primary human monocytes and adenoviral infection
Human monocytes were purified to ∼85% by centrifugal elutriation (Beckman JE-6B; Beckman Coulter) of mononuclear cells isolated from buffy coats on density gradients (Lymphoprep-Nyegaard). Buffy coats from human blood were provided by the Australian Red Cross Blood Service (Perth, Australia). All studies were performed in accordance with National Health and Medical Research Council (Australia) guidelines and approved by the institutional ethics committee. Alternatively, monocytes were isolated from PBMC on the basis of CD14 expression using CD14 microbeads (Miltenyi Biotec) and the autoMACS separation system (Miltenyi Biotec). Isolated monocytes were cultured in RPMI 1640 medium containing 2 mM glutamine, 50 μM 2-ME, 5 μg/ml gentamicin (Invitrogen) and 2 mM MOPS. After overnight incubation in Teflon pots (Savillex) containing RPMI 1640 supplemented with 10% FCS and 25 ng/ml recombinant human M-CSF (PeproTech) to induce αvβ5 expression, the cells were harvested and plated at a density of 0.5 × 106 cells/100 μl in polypropylene culture tubes (Minisorb; Nunc). AdV-GFP (control virus), AdV-SOCS1, or AdV-SOCS3 were added at a multiplicity of infection (MOI) of 50, 30 and 50, respectively, unless otherwise indicated, and centrifuged at 1000 × g for 60 min at 37°C as previously described (22). An additional 400 μl of culture medium was added to each tube before incubation for 24 h at 37°C before stimulation with the TLR4 ligand LPS (50 or 500 ng/ml), or the TLR2 ligand Pam3CSK4 (300 ng/ml; InvivoGen).
Determination of infection efficiency by flow cytometry
Uninfected cells and cells incubated with AdV-GFP, AdV-SOCS1, and AdV-SOCS3 were harvested by centrifugation 24 h after AdV-infection and washed once in PBS containing 0.2% BSA and 0.02% sodium azide. Cells were resuspended in FACS fixative (1% formaldehyde in PBS) and stored at 4°C until analysis. CD14 staining was performed using anti-human CD14-PE Ab (BD Biosciences). Analysis of CD14 expression by Flow cytometry (FACSCalibur; BD Biosciences) confirmed that all larger cells defined by forward and side scatter were monocytes. Infection efficiency was determined as the percentage of GFP-positive monocytes. The amount of virus per cell, estimated by GFP expression, was assessed as mean fluorescence intensity (MFI) by flow cytometry and expression levels analyzed using FlowJo software (version 4.6.1).
Western Blot Analysis
After isolation and overnight culture with M-CSF, monocytes were isolated and infected with AdV-GFP, AdV-SOCS1, or left uninfected (no virus) for 24 h. They were then stimulated with 500 ng/ml LPS for 0–120 min. Cells were harvested by centrifugation and monocytes lysed in protein lysis buffer (10 mM Tris, 50 mM NaCl, 5 mM EDTA, 1% Triton X-100, pH 7.6) supplemented with 5 mM sodium fluoride, 10 mM sodium molybdate, 1 mM sodium pyrophosphate, 2 mM sodium orthovanadate and 1× protease inhibitors (complete mini, protease inhibitor cocktail tablets; Roche). Approximately 7.5 μg of protein lysate was resolved per lane of a SDS-PAGE gel and transferred to nitrocellulose membrane (Pall Gelman Laboratory). Membranes were blocked for at least 1 h in 5% skim milk in TBS/0.05% Tween 20 (block buffer) followed by overnight incubation at 4°C with primary Abs; Anti-phospho-IκBα (Ser32) and anti-phospho-STAT1 (Tyr 701) were from Cell Signaling Technology. Anti-IκBα (C-21) and anti-STAT1 Abs were from Santa Cruz Biotechnology. Anti-SOCS1 Abs were obtained from Imgenex or Santa Cruz Biotechnology (SOCS1, H-93). An anti-β-tubulin Ab was purchased from Abcam. Abs were either diluted in block buffer or 5% BSA in TBS/0.05% Tween 20 according to the manufacturer’s guidelines. Following four sequential 5 min washes in TBS/0.05% Tween 20, membranes were incubated with HRP-conjugated rabbit-anti-mouse or donkey-anti-goat secondary Abs (Rockland Immunochemicals) diluted in block buffer. Bound secondary Ab was detected using chemiluminescence (Roche Diagnostics) and visualized using CL-XPosure film (Pierce).
Real-Time PCR Analysis of TNF-α mRNA
Total RNA was isolated using TRIzol reagent (Invitrogen) from ∼106 monocytes. Total RNA was reverse transcribed to cDNA using omniscript II reverse transcriptase (Qiagen) and oligo-dT primers (Promega) in the presence of RNase Inhibitor (PerkinElmer). Single stranded cDNA was diluted 1:5 and real time PCR performed using the QuantiTect SYBR green PCR kit (Qiagen). Intron-spanning gene-specific primers were designed to recognize TNF-α or Myxovirus resistance-A (MxA) and Ubiquitin conjugating enzyme ED2 (UBE2D2) transcripts. PCR was performed using ABI-PRISM 7900HT (Applied Biosystems). PCR conditions were initial denaturation at 95°C for 15 min followed by 40 cycles of 95°C for 15 s, then 60°C for 1 min. Melting curve analysis was used to assess the specificity of the PCR. Copy numbers were determined by 10-fold serial dilutions of plasmid standards and normalized to the reference gene UBE2D2 (23). Alternatively, expression levels were determined by a standard curve created from serial dilutions of the PCR product and normalized to UBE2D2.
TNF-α, IL-6, and IL-10 levels were assayed in culture supernatants using human TNF-α and IL-6 Elisa sets (BD OptEIA; BD Biosciences) or IL-10 Ab pairs (BD Pharmingen) and the DELFIA assay system (PerkinElmer). IFN-β levels were determined in culture supernatants using a human IFN-β ELISA kit (PBL Biomedical Laboratories). For each experiment samples were assayed in triplicate. The assays were sensitive to levels >10 pg/ml TNF-α, IL-6, and IL-10.
Nuclear extraction, NFκB EMSA, and supershift assays
Nuclear extracts were prepared from uninfected, AdV-GFP- and AdV-SOCS1-infected monocytes. Briefly, following stimulation with the appropriate TLR ligand for 30 min, 1 ml of ice-cold RPMI 1640 was added to each tube and cells harvested by centrifugation at 1200 rpm for 7 min at 4°C. Cells were washed in ice-cold PBS and subsequently lysed in protein lysis buffer (10 mM HEPES.KOH, pH7.9, 10 mM KCl, 1.5 mM MgCl2) supplemented with 1× protease inhibitor (complete mini, protease inhibitor cocktail tablets; Roche), 0.5 mM DTT and 2 mM sodium orthovanadate before use. Lysates were incubated on ice for 20 min, before 0.6% Nonidet P-40 was added with vortexing to mix. Samples were centrifuged at 12,500 rpm for 30 s at 4°C. The cytoplasmic fraction (supernatant) was transferred to a new microfuge tube while nuclear extraction buffer (420 mM NaCl, 20 mM HEPES.KOH, pH 7.9, 1.5 mM MgCl2, 0.2 mM EDTA, 25% glycerol supplemented with 1× protease inhibitor (complete mini, protease inhibitor cocktail tablets; Roche), 0.5 mM DTT and 2 mM sodium orthovanadate) was added to the pellet. Following 20 min incubation on ice with intermittent agitation, the nuclear fraction was collected by centrifugation (12,500 rpm, 5 min, 4°C) and quantified using Bradford protein assay. The NFκB oligonucleotides were obtained from Santa Cruz Biotechnology. Approximately 1 μg of nuclear extract was incubated with 20 fmols 32P-labeled NFκB probe and resolved on a non-denaturing 5% polyacrylamide/0.5× Tris borate-EDTA gel at 180 V for up to 2 h. Specificity of NFκB DNA binding was determined by preincubating the nuclear lysates with a 50-fold excess of either unlabelled NFκB probe (NFκB CP) or an unlabelled mutant NFκB probe (mNFκB CP). Gels were dried onto 3MM Whatmann filter paper and complexes visualized using Super RX x-ray film (Fuji Film). For supershift assays, 1 μg of nuclear extract was preincubated with 1 μl of either anti-p65 or anti-p50 supershift Ab (Santa Cruz Biotechnology) for 30 min on ice before incubation with the 32P-labeled NFκB probe. Samples were resolved on a non-denaturing 5% polyacrylamide/0.5× Tris borate-EDTA and complexes visualized as previously described.
PCR Array analysis of genes associated with TLR-signaling pathways
Total RNA was isolated using TRIzol reagent (Invitrogen) from ∼3 × 106 uninfected, AdV-GFP- or AdV-SOCS1-infected monocytes. Following DNase treatment to eliminate contaminating genomic DNA (Turbo DNase, Ambion) total RNA was further purified using the RT2 qPCR-Grade RNA Isolation kit (SABiosciences) and quantified using a Nanodrop ND-1000 spectrophotometer (BioLab). For PCR array analysis ∼250–500 ng of total RNA was reverse transcribed to cDNA using the RT2 First Strand kit as per manufacturer’s instructions (SABiosciences). The PCR was performed on an ABI 7000 (Applied Biosystems). For each 96-well plate, 25 μl of mastermix containing cDNA and a Human TLR-signaling pathway RT2 Profiler PCR plate (PAHS-018, SABiosciences) was prepared. Amplification was performed in accordance with the manufacturer’s guidelines. For data analysis, four housekeeping genes, hypoxanthine phosphoribosyltransferase 1, ribosomal protein L13a, glyceraldehyde-3-phosphate dehydrogenase, and β actin, as present on the PCR array, were used for normalization. The cycle threshold (CT) was determined for each sample and normalized to the average CT of the four housekeeping genes. Comparative CT method was used to calculate relative gene expression. Data are represented as fold change relative to control (either AdV-GFP vs uninfected or AdV-SOCS1 vs AdV-GFP). Statistical analysis was performed on Log2-fold change data using a one sample t test.
All values have been expressed as mean ± SEM. Significance of the results has been evaluated using one-way ANOVA or a paired Student’s t test. A p value of <0.05 was considered significant.
SOCS1 expression by primary human monocytes following LPS exposure or adenoviral transfection
SOCS1 mRNA levels were significantly increased in human monocytes 2 h following LPS exposure (Fig. 1,A). The ability of SOCS1 to regulate LPS-induced TNF-α production was investigated in human monocytes infected with AdV-SOCS1. To determine the optimal infection efficiency, human monocytes were infected at different MOI for 24 h with AdV-SOCS1 and levels of GFP measured by flow cytometry. An infection efficiency at MOI 30 was optimal for AdV-SOCS1 (Fig. 1,B, three different batches of AdV-SOCS1 are shown) and MOI 50 for the empty vector control, AdV-GFP (data not shown). MOI 50 and MOI 30 were used in all subsequent experiments of infection with AdV-GFP and AdV-SOCS1, respectively. After 24 h, mean infection efficiencies were 55 ± 3% and 66 ± 3% (± SEM, n = 27 experiments) for monocytes infected with AdV-GFP and AdV-SOCS1, respectively. Twenty-four hours following infection, SOCS1 was detected in AdV-SOCS1-infected monocytes. Endogenous SOCS1 in AdV-GFP-infected cells remained below the level of detection (Fig. 1 C).
SOCS1 regulates the late, but not early production of TNF-α by LPS-activated human monocytes
TNF-α production was assayed from uninfected, AdV-GFP-, AdV-SOCS1- and AdV-SOCS3-infected monocyte cultures isolated from 18 independent donors (n = 12 for AdV-SOCS3). There was no significant regulation of TNF-α production by AdV-SOCS1-infected monocytes when stimulated with LPS for 2 h (Fig. 2,A). As previously published (21, 24), there was minimal TNF-α produced in the absence of LPS (Figs. 3,C and 6).
In contrast, there was a significant reduction in TNF-α levels 24 h after LPS stimulation. Furthermore, the regulatory effects observed at 24 h were specific for SOCS1 as no significant difference was observed in TNF-α production between AdV-GFP- and AdV-SOCS3-infected cells. These data suggest that LPS-induced TNF-α production may occur via two mechanisms, an early mechanism not susceptible to SOCS1-regulation and a later pathway that can be regulated by SOCS1.
As an initiator of the immune response, TNF-α induces the sequential production of other cytokines including IL-6. The effect of AdV-SOCS1 on LPS-induced IL-6 and IL-10 production was examined (Fig. 2 B). AdV-SOCS1 significantly reduced LPS-induced IL-6 production at 24 h without significantly altering IL-10 production.
Control of TLR4-induced TNF-α production by SOCS1 in human monocytes
To ensure that the regulatory effects of SOCS1 on TNF-α production were not an artifact of the cell system used, the effects of AdV-SOCS1 on LPS-induced TNF-α production were titrated (Fig. 3,A). The amount of SOCS1 virus/cell was altered by varying the MOI with cells infected at MOI of 1, 5, 10, 20, 30, 50 and 100 (Fig. 3,A). Increasing the level of AdV-SOCS1 infection augmented the percentage of GFP positive cells and the corresponding MFI (Fig. 3,B) in a dose-dependent manner. Furthermore, the level of AdV-SOCS1-mediated suppression of LPS-induced TNF-α production was dose-dependent at LPS concentrations of 50 and 500 ng/ml (Fig. 3 C).
SOCS1 does not modulate the TLR-induced NFκB pathway
In unstimulated cells, NFκB is sequestered in the cytoplasm by IκB. Upon cell activation, phosphorylation of the IκBα subunit targets it for proteasomal degradation. The effects of AdV-SOCS1-infection on IκBα-phosphorylationSer32 and IκBα-degradation (n = 4 donors) were determined for 2 h following LPS exposure (Fig. 4,A). Monocytes were uninfected or infected with AdV-GFP or AdV-SOCS1 for 24 h before LPS stimulation. IκBα-phosphorylationSer32 was detected 15 min after stimulation with LPS, peaked at 60 min and decreased by 120 min after LPS exposure. However, the extent and kinetics of IκBα-phosphorylationSer32 did not significantly differ between AdV-GFP- or AdV-SOCS1-infected cells (Fig. 4 A). In addition, no significant difference was observed in the rate of degradation of IκBα. These data suggest that SOCS1 does not modulate this pathway at the level of, or upstream of, IκBα activation.
The capacity of AdV-SOCS1 to regulate the transcriptional potential of NFκB was determined by EMSA (Fig. 4,B). Nuclear lysates were prepared from uninfected (no virus), AdV-GFP- and AdV-SOCS1-infected monocytes following incubation with or without LPS for 30 min. Stimulation of monocytes with LPS (Fig. 4, B and D) or Pam3CSK4 (Fig. 4,B) increased NFκB DNA binding activity. Quantitative analysis demonstrated no significant difference in NFκB-DNA binding capacity by lysates from AdV-GFP- and AdV-SOCS1-infected cells from three donors exposed to LPS for 30 min (Fig. 4,C). Further analysis of SOCS1 effects on NFκB transcriptional activity was performed using lysates from monocytes purified according to CD14 expression (Fig. 4,D). Elutriated monocytes were incubated overnight in M-CSF (25 ng/ml) before enrichment of CD14+ cells (>99% purity) and infection for 24 h with AdV-GFP or AdV-SOCS1. Nuclear extracts were prepared from cells incubated with or without LPS for 30 min. Some NFκB DNA binding activity was detected in nuclear extracts from unstimulated CD14-positive monocytes, possibly due to CD14 ligation by the Ab (Fig. 4,D). Specificity of NFκB DNA binding was determined by incubating nuclear lysates with a 50-fold excess of either an NFκB CP or an mNFκB CP, containing a G to C substitution in the NFκB DNA binding region, in addition to the radiolabeled NFκB probe (Fig. 4,D). NFκB binding was blocked in cells incubated with NFκB CP, while the mNFκB CP had no effect on NFκB binding capacity (Fig. 4,D). A partial shift in the position of the NFκB complex was observed in the presence of an anti-p65 Ab, whereas the complex completely shifted with an anti-p50 Ab, demonstrating that both p65-p50 heterodimer and p50 homodimer NFκB complexes are present in unstimulated cells (Fig. 4 D). For uninfected (n = 1 cell donor), AdV-GFP-infected (n = 3 cell donors), and AdV-SOCS1-infected cells (n = 3 cell donors), 21%, 22 ± 3%, and 20 ± 6% (mean ± SEM) of the NFκB complex supershifted with the anti-p65 Ab. Following LPS exposure for 30 min, this increased to 28%, 32 ± 7% and 28 ± 5% (mean ± SEM) of the NFκB complex supershifted with the anti-p65 Ab for uninfected, AdV-GFP-, and AdV-SOCS1-infected cells, respectively. These observations suggest that SOCS1 does not regulate the NFκB pathway in human monocytes.
Regulation of the AdV-induced IFN response by SOCS1
The possibility that the significantly reduced TNF-α levels produced by AdV-SOCS1-infected cells could be, at least in part, due to SOCS1 regulation of the IFN-β pathway was investigated. AdV-infection per se was used to study the regulation of type I IFN-mediated responses. IFN-β, IRF1, IRF3, MxA mRNA, and mRNA for a panel of inflammatory cytokines were determined in total RNA from uninfected, AdV-GFP-, and AdV-SOCS1-infected monocytes from three donors 24 h following initial AdV exposure and before LPS exposure (Fig. 5,A–C). IFN-β mRNA levels were significantly increased in response to AdV-infection compared with uninfected cells (Fig. 5,A). In contrast, there was a trend toward reduced IFN-β mRNA and a significant reduction in IRF1 mRNA levels in AdV-SOCS1-infected monocytes. IRF3 mRNA levels remained unaffected by AdV exposure (Fig. 5,A). The mRNA levels of MxA, a commonly used biomarker for type I IFN activity, were elevated in response to AdV-infection (AdV-GFP-infected monocytes) and significantly reduced in AdV-SOCS1-infected cells, suggesting that SOCS1 blocked the actions of AdV-induced IFN. In addition, TNF-α mRNA and IFN-β production were determined in uninfected, AdV-GFP-, AdV-SOCS1-, and AdV-SOCS3-infected monocytes activated with LPS for up to 6 h (Fig. 5, D and E). TNF-α mRNA levels were significantly increased in AdV-GFP- and AdV-SOCS3-infected monocytes at 2 h. In contrast, there was no significant difference in TNF-α mRNA levels detected in uninfected and AdV-SOCS1-infected monocytes (Fig. 5 D). LPS-induced IFN-β levels were similar in uninfected and AdV-GFP infected cells, whereas an AdV-SOCS1-specific reduction in IFN-β was observed 6 h post-LPS exposure. IFN-β levels remained below the level of detection in cell cultures not exposed to LPS.
Many of the effects of type I IFN are mediated via the transcription factor STAT1. STAT1 activation in response to LPS exposure for 2 h was investigated in uninfected cells and cells infected with AdV-GFP and AdV-SOCS1 (Fig. 5,F). STAT1 phosphorylation was increased in AdV-GFP-infected cells before LPS exposure. In contrast, the levels of STAT1-phosphorylation in AdV-SOCS1-infected cells at time 0 were not different to those detected in uninfected cells. LPS-induced STAT1-phosphorylation, detected at 120 min, was similar in both uninfected and AdV-GFP-infected cells. Levels of phosphorylated STAT1 were lower at 120 min in AdV-SOCS1-infected cells (Fig. 5 F).
SOCS1 regulates TNF-α production by human synovial fluid macrophages ex vivo
Mononuclear cells were isolated from synovial fluids drained from the knees of two patients with inflammatory arthritis. Macrophage purity was 74% for donor 1 and 49% for donor 2. After overnight incubation with M-CSF, cells were infected (or left uninfected) with AdV-GFP or AdV-SOCS1 for 24 h. Cells were then exposed to LPS for a further 24 h and production of TNF-α examined (Fig. 6). For cells from donor 1, infection efficiencies of 63% and 59% were obtained for AdV-GFP- and AdV-SOCS1-infected cells, respectively. For cells from donor 2, infection efficiencies were 69% and 79% for AdV-GFP- and AdV-SOCS1-infected cells, respectively. In comparison to uninfected cells, AdV-GFP infection significantly enhanced LPS-induced TNF-α production in donor 2, although this did not reach significance for cells from donor 1. These results suggest that in inflammatory macrophages, AdV-infection may prime the TLR-mediated immune response through elevation of IFN-associated genes (Fig. 5). Overexpression of SOCS1 controlled AdV-driven and LPS-stimulated TNF-α production (Fig. 6).
In this study, TNF-α production by human monocytes was controlled by several interacting pathways. In the absence of adenoviral infection, two mechanisms for LPS-induction of TNF-α synthesis exist: an early MyD88-dependent and a later MyD88-independent, TRIF-dependent, pathway (25). These processes interact to enhance levels of TNF-α produced after several hours of LPS exposure. In cells infected with the adenoviral vectors for 24 h before TLR activation, a significant increase in IFN-β and MxA mRNA was detected in the absence of LPS. IFN-β protein expression was detected 6 h after LPS exposure. As previously reported (26), IRF3 mRNA levels were not induced in human monocytes in response to adenoviral infection, presumably due to constitutive expression. The contribution of IFN to TLR-induced proinflammatory cytokine production is unclear. It is notable that minimal TNF-α was produced by AdV-infected cells in the absence of TLR signaling. The AdV-induced pathway, like the MyD88-independent pathway, was dependent on the activity of IFN-β. It has been recently suggested that IRF3, like STAT1, can also bind to and transactivate the human TNF-α promoter (27). In one study, the MyD88-independent pathway required IRF3-dependent expression of TNF-α to activate NFκB (28).
Enhanced SOCS1 expression following exposure of murine macrophages to LPS suggests that SOCS1 may control LPS-mediated signaling in a negative feedback loop (13, 14, 18, 29, 30). Up-regulated IκB-, p38- and STAT1-phosphorylation in SOCS1-deficient macrophages, together with reduced NO synthesis and elevated TNF-α production, had implicated SOCS1 in direct regulation of TLR4 signaling (14). There were reports that SOCS1 could block MyD88-mediated signaling by binding to the adaptor protein Mal following its tyrosine phosphorylation, thus targeting it for proteosomal degradation (15). An additional study reported that SOCS1 interacted directly with the p65 subunit of the NFκB complex (17).
Using an adenoviral vector for SOCS1 overexpression in human monocytes, SOCS1 suppressed LPS-induced TNF-α production at 24 h without affecting TNF-α levels after 2 h. These data suggest that SOCS1 targets in human monocytes the later MyD88-independent (TRIF-dependent) pathway (summarized in Fig. 7), although some later regulation of an NFκB pathway cannot be ruled out. The SOCS1-mediated reduction in AdV-induced IFN-β, IRF1, MxA mRNA, and IFN-β production suggested that the SOCS1-control of TNF-α production by primary human monocytes occurred, at least in part by an indirect mechanism notably, via regulation of type I IFN production. Furthermore, in the PCR arrays, there was no change in the mRNA expression of components of the NFκB pathway (p50, p65) in AdV-infected cells before LPS exposure and suggested that the NFκB pathway was not a target for AdV-actions (data not shown). No supportive data of SOCS1 control of NFκB signaling were found in a study in human keratinocytes (31). Similar to our investigation, reduced inflammatory cytokine production by SOCS1-infected cells was observed and SOCS1 did not alter the NFκB pathway over the 24 h time course studied (31). These observations support a series of studies performed by Dalpke and colleagues (reviewed in Ref. 32) which propose both direct and indirect mechanisms of TLR-associated IFN-β-dependent signaling, each displaying different susceptibility to SOCS1 regulation. Effects of IFN-β on the duration, and not the magnitude of the NFκB activation were also questioned (32). In another study in CpG DNA-stimulated macrophages that failed to produce IFN-β, STAT1 did not promote the production of inflammatory cytokines but instead limited their production (33). This mechanism of altered transcription factor activity may help to explain the dramatic decrease in TNF-α production by IFN-β-compromised, AdV-SOCS1-infected monocytes. However, we cannot rule out the possibility that regulation of IFN-signaling alone is not sufficient for this effect. Perhaps the combined actions of SOCS1 on IFN-mediated pathways and on unknown targets may be required for the potent regulatory effect on TNF-α production observed (Fig. 7).
Our data using primary human monocytes suggest that SOCS1 expression has no direct effect on the NFκB pathway early after LPS stimulation. There was no effect on IκBα-phosphorylation and in keeping with a recent study of GM-CSF and M-CSF treated murine bone marrow-derived macrophages (34), both p50 homodimer and p65-p50 heterodimer complexes were identified in nuclear extracts from M-CSF cultured human monocytes. Furthermore, the relative amounts of these complexes did not alter between uninfected, AdV-GFP- or AdV-SOCS1-infected cells before or after LPS exposure. It had been reported that the nuclear translocation of the p65-p50 heterodimer and p50 homodimer complexes were important for LPS-induced TNF-α and IL-10 production, respectively (35). However, with regulation of TNF-α but not IL-10 in this study, and the regulation of TNF-α occurring independently of the NFκB pathway, this conclusion could not be confirmed in human monocytes.
These data suggest that control of TLR signaling differs between primary human and murine macrophages. This is not the first study with a similar conclusion; the adaptor protein Mal has been reported to be involved in p65-NFκB activation in manipulated cell lines and mouse but not in human macrophages (36). Different cell types may also have different adaptor protein use by TLR receptors (36).
The regulatory effects of SOCS1 on TNF-α and IL-6 production were specific and were not observed in AdV-GFP- or AdV-SOCS3-infected cells. Furthermore, AdV-SOCS1-mediated effects on TNF-α production could be titrated. The regulation of IL-6 production in AdV-SOCS1-infected cells may be due to impaired TNF-α signaling as a consequence of reduced TNF-α production.
IL-10 production in response to TLR-stimulation was not modulated in monocytes by overexpression of SOCS1. Furthermore, in the PCR arrays, the IFN-dependent pathway operating in response to AdV-infection did not affect IL-10 levels. This result suggests that IRF3 and type I IFNs do not regulate IL-10 expression and would explain outcomes of acute inflammatory arthritis in Socs1−/−Ifnγ−/− mice (37). The severity of synovial inflammation and joint destruction at the peak of disease was greater in the absence of SOCS1, although disease resolution, a process dependent on IL-10 activity, occurred normally.
In this study we propose that the effects of SOCS1 on proinflammatory cytokine production by human monocytes reflect the control by SOCS1 of type I IFN production. A role for SOCS1 in regulating the production and actions of type I IFN is not new (38, 39). In murine macrophages, SOCS1 can control the production and actions of IFN-β by interacting with the IFN (α, β, and ω) receptor-1 chain blocking downstream signaling (13, 18, 19, 20). SOCS1 control of the MyD88-independent pathway, rather than the NFκB-dependent pathway, is further supported by the observation that SOCS proteins target tyrosine phosphorylated proteins. The TLR signaling cascades are predominately regulated by serine/threonine phosphorylation (4), which potentially limits possible SOCS1-interacting partners within the TLR pathway signaling intermediates (19). Furthermore, TLR4 signaling was not modulated in Socs1−/− mice (19).
Recently, 4–1BBL/TNF-α superfamily member-9 was implicated in the regulation of LPS-induced TNF-α production, by interacting with the TLR complex responsible for sustained TNF-α production, i.e., via regulation of type I IFN (40). SOCS1 may interact with 4–1BBL and target 4–1BBL for degradation. Alternatively, SOCS1 may be associated with the receptor itself and prevent association of the 4–1BBL-TLR complex required for the late production of TNF-α.
It has been suggested that SOCS1-regulation of IFN-β is important for keeping the innate immune response in check and provides a mechanism for preventing an exaggerated response to activation (32). In a study comparing the effects of an empty vector control and an adenoviral vector encoding SOCS1, Sakurai and colleagues demonstrated the capacity of SOCS1 to limit adenoviral-mediated activation of the innate immune response (41). These effects of SOCS1 are not limited to immune cells. SOCS1 regulates the type I IFN-mediated anti-viral effects triggered by influenza A virus exposures in human respiratory epithelial cells (42).
Low levels of expression combined with rapid turnover have made studies investigating the role of SOCS1 in human monocytes and macrophages difficult. Further human monocytes as non-proliferative, phagocytic cells are difficult to transfect to high efficiency. In this study primary human monocytes were infected to greater than 50% efficiency with an AdV encoding human SOCS1, introduced into the cells by spinofection (21, 22, 24). The adenoviral infection approach enhanced the type I IFN levels but fortuitously provided a mechanism to examine the IFN-dependent pathway of TLR. Use of adenoviral vectors to express a small interfering RNA to SOCS1 would have had the same vector-induced, IFN-dependent effects. Further, while successful silencing of the Socs1 gene has been previously demonstrated in murine dendritic cells (43, 44), small interfering RNA are now known to nonspecifically activate TLR7 and TLR8, initiating an immune response which makes analysis of TLR-ligand-specific responses impossible (45).
This study demonstrated that SOCS1 regulates the MyD88-independent, TRIF-dependent pathway in human monocytes and macrophages. The capacity of SOCS1 to regulate sustained TNF-α production is of potential therapeutic benefit for the treatment of chronic inflammatory conditions, such as rheumatoid arthritis and Crohn’s disease. This study makes an important contribution to our understanding of the regulatory capacity of SOCS1 in human cells at a time when SOCS1 mimetics are being considered for therapeutic use (46, 47).
We acknowledge the technical assistance of Dr. Tulene Kendrick, and Drs. Janet Roddy and Nicola Cook and their patients for the provision of synovial fluid.
The authors have no financial conflict of interest.
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.
This work was supported by National Health and Medical Research Council (NHMRC) Grant 275546 (to P.H.H.), a State and Territory Affiliate grant, Arthritis Australia and the Adam Gilchrist Trading Challenge Project grant from Arthritis Australia (to C.M.P.) and NIH CA22556 (to S.E.N.). S.E.N. was supported by an NHMRC Biomedical Career Development Award.
Abbreviations used in this paper: Mal, MyD88-adaptor-like; CT, cycle threshold; HEK, human embryonic kidney; IRF, IFN regulatory factor; IκB, inhibitory κB; MFI, mean fluorescence intensity; MOI, multiplicity of infection; MxA, myxovirus resistance-A; NFκB CP, unlabelled NFκB probe; mNFκB CP, unlabelled mutant NFκB probe; SOCS, suppressor of cytokine signaling; TRAM, TRIF-related adaptor molecule; TRIF, toll/interleukin-1 receptor domain-containing adaptor inducing IFN-β; UBE2D2; ubiquitin conjugating enzyme E2D2.