Epstein-Barr Virus Interferes with the Amplification of IFNα Secretion by Activating Suppressor of Cytokine Signaling 3 in Primary Human Monocytes

Background Epstein-Barr virus is recognized to cause lymphoproliferative disorders and is also associated with cancer. Evidence suggests that monocytes are likely to be involved in EBV pathogenesis, especially due to a number of cellular functions altered in EBV-infected monocytes, a process that may affect efficient host defense. Because type I interferons (IFNs) are crucial mediators of host defense against viruses, we investigated the effect of EBV infection on the IFNα pathway in primary human monocytes. Methodology/Principal Findings Infection of monocytes with EBV induced IFNα secretion but inhibited the positive feedback loop for the amplification of IFNα. We showed that EBV infection induced the expression of suppressor of cytokine signaling 3 (SOCS3) and, to a lesser extent, SOCS1, two proteins known to interfere with the amplification of IFNα secretion mediated by the JAK/STAT signal transduction pathway. EBV infection correlated with a blockage in the activation of JAK/STAT pathway members and affected the level of phosphorylated IFN regulatory factor 7 (IRF7). Depletion of SOCS3, but not SOCS1, by small interfering RNA (siRNA) abrogated the inhibitory effect of EBV on JAK/STAT pathway activation and significantly restored IFNα secretion. Finally, transfection of monocytes with the viral protein Zta caused the upregulation of SOCS3, an event that could not be recapitulated with mutated Zta. Conclusions/Significance We propose that EBV protein Zta activates SOCS3 protein as an immune escape mechanism that both suppresses optimal IFNα secretion by human monocytes and favors a state of type I IFN irresponsiveness in these cells. This immunomodulatory effect is important to better understand the aspects of the immune response to EBV.


Introduction
Epstein-Barr virus (EBV), a human gamma-herpes virus, persists latently in over 90% of the adult population and is the cause of infectious mononucleosis in a small proportion of carriers. Viral reactivation is responsible for certain rare types of lymphoproliferative disorders and cancers [1]. Although its main target cells are B lymphocytes, EBV can spread to other cell types [2,3]. Particularly, efficient and sustained replication of EBV particles in primary human monocytes has been confirmed and shown to alter a number of cellular defense mechanisms. For example, EBV can negatively regulate monocyte secretion of TNF-a [4] and MIP-1a [5]. In addition, EBV infection reduces monocyte secretion of the antiviral lipid mediator prostaglandin E 2 (PGE 2 ) by targeting the enzyme cyclooxygenase-2 (COX-2) essential for prostaglandin synthesis [6]. We also reported that EBV infection impairs protein kinase C (PKC) function causing a reduction in monocyte phagocytic activity [6,7]. Because mono-cytes were shown to contribute to the spread of EBV infection [8], these data suggest that infection of these cells may have important implications in EBV pathogenesis.
Type I interferons (IFNs) critically contribute to host defense against viral invaders by inducing innate responses and subsequent adaptive immunity. Secreted IFNs function in an autocrine and paracrine fashion to potentiate cellular antiviral mechanisms and limit the replication and spread of the virus [9]. Upon viral sensing by host cells, two members of the interferon regulatory factor (IRF) family, IRF3 and IRF7, mainly activate IFN gene transcription and initiate the first wave of IFN secretion [10]. Subsequent binding of IFNs to their cognate receptor leads to the activation of the JAK/STAT pathway. JAK1 and Tyk2 kinases are constitutively associated with the IFN receptor subunits and upon activation, they phosphorylate each other at critical tyrosine residues within the intracellular domain of the receptor. STAT1 and STAT2 factors are then recruited via the phosphorylated tyrosines, bind the activated receptor and are in turn phosphor-ylated by JAK1 and Tyk2 [11]. Signaling downstream of the IFN receptors through the JAK/STAT pathway creates a positive feedback loop that prolongs activation of IFN-stimulated genes, mediates a second wave of IFN secretion and leads to the production of antiviral proteins such as 29-59-oligoadenylate synthetase and dsRNA-dependent protein kinase R (PKR) [12,13].
In order to avoid excessive host tissue injury whilst protecting effectively against infectious agents, the immune system features regulatory mechanisms to control the production and response to cytokines. The SOCS family of proteins comprises eight members (SOCS1-7 and CIS) critically involved in this process [14]. SOCS1 and SOCS3 are the best-characterized family members and have both been described to interfere with the response to IFNa [14,15]. The kinase inhibitory region (KIR) shared by SOCS1 and SOCS3 is sufficient to inhibit JAK tyrosine kinase activity [15]. In addition, SOCS1 has been proposed to target itself and JAK proteins to the microtubule organizing complex (MTOC)associated 20S proteasome for degradation [16]. Importantly, recent studies have shown that several viruses such as hepatitis C virus (HCV) [17], herpes simplex 1 virus (HSV-1) [18,19], enterovirus [20] and respiratory syncytial virus (RSV) [21] are capable of inducing expression of SOCS proteins and interfere with the IFN signaling pathway.
In the present study, we hypothesized that impairment in IFNa secretion by primary human monocytes infected with EBV involved the activation of SOCS proteins. We tested this hypothesis by examining SOCS1 and SOCS3 expression in parallel with several aspects of the IFNa pathway in infected cells. We showed that depletion of SOCS3 reduced the EBV-mediated suppression of the IFNa pathway and that the EBV protein Zta (also known as ZEBRA) was implicated in activating SOCS3 expression. Interference with the amplification of IFNa secretion caused by EBV infection may constitute an essential strategy that evolved to evade the antiviral response.

EBV interferes with IFNa secretion in human monocytes
Upon recognition of pathogen-associated molecular patterns (PAMPs), several pattern-recognition receptors (PRRs) activate the production and secretion of type I IFN. The synthetic doublestranded RNA analog poly(I:C) is an agonist of both TLR3 and MDA-5 and is a known activator of type I IFN [22]. To study the secretion of IFNa by human monocytes in the absence of potential pathogen-derived inhibitory factor, we stimulated these cells either once or twice with various concentrations of poly(I:C). As shown in Figure 1A, a single stimulation with increasing concentrations of the agonist led to the secretion of IFNa in a dose-dependent fashion. When cells were stimulated a second time with the same concentrations of poly(I:C), IFNa levels did not significantly differ from what was observed after a single stimulation ( Figure 1A). We repeated the experiment using live EBV and as observed with poly(I:C), a single monocyte treatment with increasing multiplicity of infection (m.o.i.) also led to increased IFNa secretion ( Figure 1B). However, cells stimulated a second time with EBV secreted significantly less cytokine at an m.o.i. of 0.1 ( Figure 1B). Given that monocytes did not become refractory to two stimulations with high concentrations of poly(I:C), these results are consistent with active interference on the IFNa secretion pathway caused by EBV infection.

EBV infection induces the expression of SOCS proteins
SOCS1 and SOCS3 are known to be involved in the negative feedback inhibition of IFNa signal transduction [14,15]. Since we measured a decrease in IFNa secretion following a second monocyte infection with EBV, we wanted to investigate whether SOCS protein induction upon primary EBV infection might contribute to this observation. Monocytes were infected with EBV for various times and expression of SOCS1 and SOCS3 was evaluated at both the mRNA and protein levels. Transcription of both SOCS1 and SOCS3 was increased following EBV infection, reaching maximum levels after 30 minutes (Figure 2A). Increased SOCS expression was also confirmed at the protein level since SOCS1 expression was increased at 60 minutes post-infection whilst SOCS3 expression progressively increased from 20 to 60 minutes post-infection ( Figure 2B). Thus, EBV infection causes the upregulation of two SOCS proteins involved in the modulation of the IFN pathway.

EBV-mediated SOCS3 activation causes the inhibition of the JAK/STAT pathway
The cellular response to IFNa occurs via the JAK/STAT pathway downstream of the IFNa receptor [11]. To further dissect the response of monocytes to IFNa, we first monitored the phosphorylation of STAT1 and STAT2 in response to single or dual IFNa stimulation in the absence of viral infection. As shown in Figure 3A, a 15-minute stimulation with IFNa caused an increase in phospho-STAT1 and phospho-STAT2 levels. The amounts of phospho-STAT1 and phospho-STAT2 were both reduced following a prolonged 20-hour exposure to IFNa, as compared to a 15-minute stimulation only. Importantly, a 15minute treatment with IFNa following a 20-hour exposure to IFNa caused an increase in both phospho-STAT1 and phospho-STAT2 compared to a 20-hour exposure only ( Figure 3A). These results demonstrate that uninfected monocytes are still responsive to IFNa stimulation following prolonged exposure to this cytokine and establish a model system that can then be used to study the effect of EBV infection on the IFNa pathway.
SOCS1 and SOCS3 suppress IFNa signaling downstream of the IFNa receptor by blocking signal transduction through the JAK/STAT pathway [14]. To investigate whether EBV infection correlates with an impairment in JAK/STAT signaling downstream of the IFNa receptor, we first monitored the phosphorylation of Tyk2, Jak1, STAT1 and STAT2 in monocytes infected with EBV alone or infected and restimulated with IFNa.
Triggering of the IFNa receptor through a short stimualtion with IFNa induced a strong activation of all members of the pathway ( Figure 3B). Whilst infection of monocytes with EBV for 20 hours led to the phosphorylation of Jak1, STAT1 but not Tyk2 nor STAT2, increased phosphorylation of these proteins could not be observed upon restimulation of infected cells with IFNa ( Figure 3B). Thus, EBV infection causes a blockage in the activation of the JAK/STAT pathway, a mechanism consistent with the action of SOCS1 and SOCS3.
To directly address the role of SOCS1 and SOCS3 in interfering with JAK/STAT signaling upon EBV infection, we used siRNA to silence the expression of these proteins. As shown in Figure 3C, both SOCS1-and SOCS3-targetting siRNA reduced expression levels of SOCS1 and SOCS3 respectively. Using both untransfected and transfected cells, we monitored phosphorylation of STAT1 and STAT2 under the same experimental conditions as in Figure 3B. In untransfected cells, akin to what was previously observed, a 15-minute IFNa stimulation of monocytes already infected with EBV for 20 hours did not increase levels of phospho-STAT1 and phospho-STAT2 as compared to EBV infected cells only ( Figure 3D). Transfection of cells with siRNA against SOCS1 had no major effect on the phosphorylation of both proteins, however, inhibition of SOCS3 caused a marked increase in phospho-STAT1 and phospho-STAT2 in dually stimulated cells. Overall, these results indicate that EBV infection of monocytes causes the inhibition of the JAK/STAT pathway via SOCS3.

Activation of IRF3 and IRF7 during EBV infection
Although SOCS1 and SOCS3 are known to inhibit the JAK/ STAT-mediated second wave of IFNa production, we were interested in whether IRF3 and IRF7, implicated in the first wave of type I IFN production, might also be affected by EBV infection. Indeed, it has been described that the JAK/STAT pathway modulates IRF7 expression via the formation of the interferonstimulated gene factor 3 (ISGF3) complex [23,24]. To verify the activation of both IRFs during primary EBV infection and upon restimulation, monocytes were stimulated once or twice with EBV and the presence of phosphorylated forms of IRF3 and IRF7 was evaluated by immunoblotting. Following a single stimulation with EBV, phospho-IRF3 and phospho-IRF7 were detected as early as 6 hours ( Figure 4). Whilst levels of phospho-IRF3 decreased thereafter, phospho-IRF7 levels progressively increase from 6 to 24 hours. Upon a second stimulation with EBV, phospho-IRF3 was further induced after 6 hours and then progressively decreased to a much greater extent to what was observed during the first stimulation. However, in the case of phospho-IRF7, whilst levels of phosphorylation were detected after a second EBV stimulation, no further increase of phosphorylation levels was observed. These results suggest that IRF3 and IRF7 can be activated by EBV but that IRF7 activity can be progressively affected after a prolonged stimulation with EBV.

SOCS3 plays a determinant role in the suppressive effect of EBV on IFNa secretion
Our results highlighted a putative role for SOCS3 in the EBVmediated suppression of IFNa secretion. To confirm its suppressive role, monocytes were transfected with siRNA directed against SOCS1 or SOCS3 and were infected once or twice with EBV. A first stimulation with EBV induced high levels of IFNa secretion, regardless of the siRNA transfected, as compared to poly(I:C) stimulation ( Figure 5). When cells were stimulated a second time with EBV, the suppressive effect of the virus was detectable in control siRNA-transfected monocytes. Although SOCS1-targetting siRNA did not impact IFNa secretion after the second   (5610 6 ) were treated with IFNa (1000 U/ml) for 15 minutes, with EBV for 20 hours or were pre-incubated for 20 hours in the presence of EBV followed by a stimulation with IFNa for 15 minutes. The expression of phospho(p)Tyk2, phospho(p)JAK1, phospho(p)STAT1, and phospho(p)STAT2 proteins was evaluated by Western blot analysis. Membranes were also probed with anti-Tyk2, JAK1, STAT1 and STAT2 as a loading stimulation with EBV, inhibition of SOCS3 significantly restored cytokine secretion. These results confirm the direct involvement of SOCS3 in the suppressive effect of EBV on IFNa secretion in human monocytes.

The viral protein Zta causes the transactivation of SOCS3
The EBV protein Zta is a basic leucine zipper (bZIP) transcription factor with many described functions including the interaction with host proteins and the modulation of cellular gene expression [25]. As shown in Figure 6A, Zta is strongly expressed in EBV-infected monocytes, further supporting the observation that EBV can efficiently infect this cell type [26]. In light of the many reports describing the modulation of immune-related host genes by Zta [25,27], we wanted to investigate whether this viral transactivator could induce the expression of SOCS3. To do so, HEK293 cells were co-transfected with a reporter vector driven by the SOCS3 promoter along with either a vector encoding wildtype Zta (Zta) or a vector encoding a mutated form of Zta (DZta) that has lost its normal transactivation activity [28]. In this system, SOCS3 promoter activity was enhanced proportionally to the amount of transfected Zta vector, however, such activation was not observed using the DZta vector ( Figure 6B). To confirm the ability of Zta to induce SOCS3 expression and modulate the JAK/ STAT pathway, we transfected human monocytes with the Zta or the DZta vector or with a mock control prior to stimulation with IFNa and monitored levels of SOCS3 and phospho-STAT2 by immunoblot. The amount of SOCS3 protein was enhanced in cells transfected with the Zta vector as compared to the cells transfected with the DZta vector or the mock control ( Figure 6C). In addition, increased SOCS3 expression in cells transfected with the Zta vector was accompanied by a marked decrease in phospho-STAT2 levels as compared to the cells transfected with the control vector ( Figure 6C). Finally, we observed a partial restoration of phospho-STAT2 levels in cells transfected with the DZta vector. Thus, EBV protein Zta can transactivate SOCS3 expression in order to interfere with the IFNa response pathway in human monocytes.

Discussion
In the present study, we demonstrated that infection of primary human monocytes with EBV leads to the inhibition of the IFNa signal transduction pathway and hence, to an impairment in the  control. (C) Monocytes (2610 6 cells) were transfected with 165 nM siRNA targeting SOCS1 or SOCS3 prior to EBV stimulation for 1 hour. Scramble siRNA was used as control. The expression of SOCS1 and SOCS3 was evaluated by Western blot analysis. Densitometry was performed and represents fold protein induction (relative to non-transfected cells) 6 std. dev. of experiments performed in duplicate. (D) Monocytes (2610 6 cells) were either left untransfected or were transfected with siRNA targeting SOCS1 or SOCS3 and stimulated as in (B). The expression of phospho(p)STAT1 and phospho(p)STAT2 proteins was evaluated by Western blot analysis. Membranes were also probed with anti-STAT1 and STAT2 as a loading control. Data are representative of three independent experiments. NS: non-stimulated; NT: non-transfected; SCR: scrambled siRNA. doi:10.1371/journal.pone.0011908.g003 amplification of IFNa secretion. Based on our results, we propose a hypothetical model of EBV-mediated negative regulation of IFN response and secretion in monocytes (Figure 7). According to this model, virion entry into the cell activates IRF3 and IRF7 leading to a first wave of type I IFN production. At the same time, EBV modulates SOCS3 expression in order to inhibit IFN receptormediated intracellular signaling through the JAK/STAT pathway. The latter results in a marked attenuation of the amplification loop initiated by the binding of type I IFNs to their cognate receptor. As a consequence, interferon-stimulated genes (ISGs) and IRF7 are negatively regulated and the second wave of IFNa secretion is impaired.
The importance of IFNs, originally discovered because of their ability to protect cells from viral infections, is highlighted by the observation that most viruses have evolved anti-IFN strategies [29]. Several studies have investigated mechanisms used by EBV to regulate the expression of IFNa and IFN-inducible genes. On one side, the early lytic EBV nuclear protein SM and the latent membrane protein 1 (LMP1) were reported to induce phosphorylation of STAT1 and the expression of ISGs [30,31]. To counteract this cellular recognition event, EBV was shown to downregulate IFN-induced transcription via the viral protein EBNA-2 [32,33] and to increase IFN receptor degradation via LMP2A and LMP2B [34] whilst EBV encoded EBER RNAs were found to be involved in IFN resistance by binding to PKR but failing to activate it [35,36]. Another interesting mechanism of interference with IFN secretion was demonstrated by Cohen and Lekstrom who showed that EBV BARF1 gene (known to encode a soluble colony-stimulating factor receptor) inhibits IFNa secretion by mononuclear cells [37]. To our knowledge, we are the first to report SOCS protein activation during EBV infection of monocytes.
Viral-mediated induction of SOCS proteins is currently emerging as a key mechanism of immune evasion. Indeed, HSV-1, another member of the herpes virus family, has been shown to activate SOCS3 in infected epithelial cells leading to the downregulation of the JAK/STAT cascade [19]. The authors of the study concluded that HSV-1-induced SOCS3 was mainly responsible for the suppression of IFN signaling. In the case of EBV, we also observed that SOCS1 was induced in infected monocytes. Whilst the use of siRNAs directed against SOCS3 confirmed its role in EBV-mediated suppression of the JAK/ STAT pathway and IFNa secretion, SOCS1 siRNA had no significant effect. The incomplete restoration of IFNa secretion with the use of SOCS3 siRNA shown in Figure 5 could either be explained by the difficulty to achieve high transfection efficiency in human monocytes coupled with the incomplete abolishment of SOCS expression by siRNA or by other viral-induced mechanisms targeting IFN signaling. Phosphatases such as protein tyrosine phosphatase 1B (PTP1B) [38] and SHP-2 [39] can interfere with the JAK/STAT pathway and represent candidate proteins potentially modulated by EBV. Other SOCS proteins such as CIS may also play a role. A recent study by Hashimoto et al., investigated the induction of all eight SOCS proteins during RSV infection and found that SOCS1, SOCS3 and CIS were activated [21]. Suppression of the three proteins by siRNA inhibited viral replication and activated type I IFN signaling. Although we do not conclude that SOCS3 activation is sufficient for EBV-mediated interference with IFNa secretion, it does represent an important mechanism as demonstrated for HCV, HSV-1, enterovirus and RSV [17,18,19,20,21].
The transactivation of SOCS3 by Zta puts forward a new role for this viral effector protein. Zta is composed of a C-terminal transactivation domain, a central basic region that mediates DNA contact and a characteristic bZIP domain extending towards the N-terminus. Expression of Zta on its own is sufficient to disrupt EBV latency and this protein has a major role in EBV-associated cell transformation by modulating cellular gene expression and interacting with host cell-cycle proteins [25]. In our study, Zta was sufficient to induce SOCS3 expression and inhibit STAT2 phosphorylation upon IFNa stimulation of monocytes. Whilst SOCS3 expression could not be recapitulated with mutated Zta, STAT2 phosphorylation was only partly restored following IFNa treatment in this context. The DZta vector encodes the full-lenght Zta protein with two amino acid substitutions in the transactivation domain, only affecting part of its transcriptional activity [28]. Thus, DZta-mediated activation of other IFN signaling modulatory factors may account for the incomplete restoration of STAT2 phosphorylation. A possible factor is IL-10, which is known to be activated by Zta [40] and to inhibit IFNa-induced phosphorylation of STAT proteins [41]. Certainly, the pleiotropic action of Zta during EBV infection is only beginning to be fully revealed and its dual effects (activation/suppression) may be clarified through future investigations.
One example of such suppressive effect is the modulation of IRF7 by Zta. In a study by Hahn and colleagues, IRF7 activation was negatively regulated by Zta [42]. Zta did not affect IRF7 levels but expression of both IRF7 and Zta were found to be directly associated. Since Zta is a nuclear protein and that phosphorylated IRF7 translocates to the nucleus, the authors postulated that interaction between Zta and activated IRF7 might be responsible for downmodulating the transcription of IRF7 target genes. In our study, we monitored the phosphorylation status of endogeneous IRF3 and IRF7 in human monocytes. As opposed to phospho-IRF3, which could still be induced upon secondary stimulation with EBV, phospho-IRF7 progressively decreased under this condition. Based on our results, we suggest that the effect of Zta on IRF7 is indirect and implicates the inhibition of the JAK/STAT pathway by SOCS, thereby causing a decrease in ISGF3-driven IRF7 expression. It is interesting to note that whilst our proposed mechanism differs from that stated by Hahn et al., both mechanisms are not mutually exclusive. As pointed out by the authors, IRF7, which was first cloned as a transcriptional regulator of the central EBV latency gene EBNA-1, is intricately associated with EBV infection [42]. Indeed, accumulating evidence highlights the use of different and/or redundant strategies by EBV to modulate IRFs expression and activity and interfere with the antiviral activity of type I IFNs [43,44,45]. Further research is needed to ask whether those strategies differ between cell types or upon primary EBV infection in comparison with reactivation from a latent infection.
Our study was performed using primary human monocytes in which productive EBV infection and viral-mediated alteration of several cellular functions have been demonstrated [3]. Here, we have shown that EBV infection induces SOCS3 activation via Zta and alters the IFNa signaling pathway. Using such a strategy, EBV might be able to survive longer within monocytes and optimize its dissemination. Furthermore, because monocytes are recognized as important antigen presenting cells linking the innate and adaptive immunity, suppression of their biological functions by EBV may thus affect the host immune response. Emerging therapeutic approaches aimed at downregulating SOCS gene expression [46] could possibly be beneficial against EBV infection by enhancing the innate antiviral activity of monocytes.

Ethics statement
Heparinized blood was obtained from healthy donors after written informed consent from all individuals in accordance with an Internal Review Board-approved protocol at CHUQ Research Center (Centre Hospitalier Université Laval).

Isolation, purification and culture of human monocytes
Peripheral blood mononuclear cells (PBMCs) were isolated by centrifugation of heparinized blood obtained from healthy donors over Lymphocyte Separation Medium (Wisent Inc., St-Bruno, QC, Canada). PBMCs were next allowed to adhere onto autologous serum-treated petri dishes in order to separate monocytes from the lymphocyte population. Monocytes were further enriched by cell sorting (FACSAria, BD Biosciences, MD, USA) which resulted in at least 99% pure monocyte suspension as determined by flow cytometry analysis using anti-CD14 monoclonal antibodies. Cell viability was more than 99% as tested by trypan blue dye exclusion procedure. Isolated monocytes were resuspended in RPMI-1640 supplemented with 10% fetal bovine serum (FBS).

Viral preparations
EBV strain B95-8 was produced as described previously [26]. Briefly, B95-8 cells were cultured in RPMI 1640 medium supplemented with 10% FBS in the presence of 20 ng/ml phorbol myristate acetate (PMA), a known inducer of viral reactivation. Cell-free supernatants were filtered through a 0.45 mm pore size filter and viral particles were concentrated by ultracentrifugation. Viral particles were resuspended in RPMI 1640 medium, titrated as described [47] and stored at 2150uC until use. Cell-free supernatants collected from B95-8-infected cells not exposed to PMA were processed as described above and used as mock controls.

Monocyte stimulation
Enriched monocytes were incubated with infectious EBV particles at the indicated multiplicity of infection (m.o.i.) or were transfected with poly(I:C) (Sigma-Aldrich, Oakville, ON, Canada) at indicated concentrations using lipofectamine reagent (Invitrogen, Burlington, ON, Canada) and cultured for 20 hours (first stimulation). Infected cells were then washed once in HBSS buffer and resuspended in fresh culture medium. Cells were then restimulated a second time with EBV, poly(I:C), or human IFNa (PBL Biomedical Laboratories, Piscataway, NJ) for indicated times (second stimulation). Following first and second stimulations, cellfree supernatants were harvested for IFNa quantitation by ELISA assay (PBL Biomedical Laboratories, Piscataway, NJ) or cells were lysed for Western blot or RT-PCR analyses as described below.

RNA isolation and RT-PCR amplification
Untreated and EBV-treated monocytes were cultured for various periods of time before RNA extraction. Total RNA from monocytes was isolated using Trizol reagent (Invitrogen, Burlington, ON, Canada) according to the manufacturer's instructions. One microgram of DNase-treated RNA was reverse transcribed to cDNA with oligo (pdT) primers in a 20 ml reaction containing 20 U of SuperScript II RNase H Reverse Transcriptase and 1 U of RNase inhibitor (Invitrogen, Burlington, ON, Canada). A volume of 5 ml cDNA samples was subjected to 35 cycles of PCR amplification in 50 ml of PCR mixture containing 0.5 U of Taq DNA Polymerase and 1.5 mg of the appropriate primers. Primers used in this study are depicted in Table 1. GAPDH was used as internal control.

Western blot analysis
Monocytes were incubated with appropriate agonists for indicated times, lysed (TAE buffer 16, 1 mM EDTA, 27 mM sucrose, 1% Triton X-100) and boiled for 5 minutes after addition of sample buffer (150 mM Tris pH 6.8, 1.2% SDS, 0.33% glycerol, 15% b-mercaptonethanol, 1% bromophenol blue). Proteins were separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE), under reducing conditions followed by transfer onto a nitrocellulose membrane. Membranes were pretreated in blocking solution containing 5% (w/v) dry milk in Tris-buffered saline-Tween 20 for 1 hour at room temperature and then incubated overnight at 4uC with anti-pIRF3, anti-phospho or total JAK1, anti-phospho or total Tyk2, anti-phospho or total STAT1, anti-phospho or total STAT2, (Cell Signaling, Danvers, MA), anti-IRF3, anti-IRF7, anti-SOCS-1, anti-SOCS-3, or anti-b-actin (Santa Cruz Biotechnology, Santa Cruz, CA). Membranes were washed four times with Tris-buffered saline-Tween 20 and incubated either with HRP-conjugated sheep anti-mouse Ig or donkey anti-rabbit IgG antibodies (Jackson ImmunoResearch Laboratories Inc., West Grove, PA) for 1 hour. Immunoreactive proteins were revealed by enhanced chemiluminescence (Perkin Elmer, Woodbridge, ON, Canada). Densitometry analysis was performed using the Image J software and relative protein levels were normalized to relative b-actin levels.

Small interfering RNA assay
Purified primary monocytes (2610 6 cells) were transfected with 165 nM of small interfering RNA against SOCS-1 (Sense: 59-GCAUUAACUGGGAUGCCGUtt-39 Antisense: 59-ACGGCA-UCCCAGUUAAU GCtg-39) or SOCS-3 (Sense: 59-GAAC-CUGCG CAUCCAGUGUtt-39 Antisense: 59-ACACUGGAU-GCGCAGGUUCtt-39) (Applied Biosystems/Ambion, Austin, TX) using lipofectamine according to the manufacturer's instruction. Scramble siRNA was used as control. Four hours posttransfection, cells were washed once in HBSS buffer and resuspended in culture medium in order to avoid cellular toxicity due to siRNA transfection. Twenty-four hours post-transfection, cells were stimulated as described and cell-free supernatants were harvested and tested for the presence of IFNa by ELISA or cells were lysed for Western blot analysis.

Statistical analysis
Data were analyzed by one-tailed analysis of variance (ANOVA) followed by Newman-Kheuls post-hoc test using PRISM3 software. Differences were considered significant at p#0.05.