Triggering of Toll-like receptors modulates IFN-γ signaling: involvement of serine 727 STAT1 phosphorylation and suppressors of cytokine signaling

Authors


Abstract

Microbial stimuli activate cells of the innate immune system by triggering Toll-like receptors (TLR). Activation of macrophages and dendritic cells is further enhanced by secondary signals like IFN-γ. Here we analyzed the interplay of IFN-γ and TLR signaling in cells of the innate immune system. Using a STAT1-dependent reporter construct we show that IFN-γ signaling can be enhanced as well as inhibited by simultaneous stimulation with either defined TLR agonists or whole-bacterial lysates. Short costimulation resulted in the amplification of IFN-γ signaling and was attributable to the p38 mitogen-activated protein kinase (MAPK)-dependent phosphorylation of signal transducer and activator of transcription (STAT)1 on serine 727. In contrast, prolonged co-incubation as well as pre-incubation with TLR agonists led to an inhibition of IFN-γ signaling. TLR triggering induced expression of suppressor of cytokine signaling (SOCS)-1, SOCS-3 and cytokine-inducible SH2 domain-containing protein (CIS). Overexpression of SOCS-1 and, to a lesser extend, of SOCS-3 and CIS inhibited IFN-γ signaling as measured by activation of STAT1. Moreover, pre-incubation with TLR-dependent stimuli impaired IFN-γ-induced MHC class II regulation but enhanced CD40 and CD86 expression. Taken together, the results indicate a tight interplay between TLR and IFN-γ signaling pathways which involve induction of SOCS proteins and serine phosphorylation of STAT1.

Abbreviations:
CIS:

Cytokine-inducible SH2 domain-containing protein

Jak:

Janus kinase

LTA:

Lipoteichoic acid

MAPK:

Mitogen-activated protein kinase

ODN:

Oligodeoxynucleotide

TLR:

Toll-like receptor

RT:

Reverse transcription

SOCS:

Suppressors of cytokine signaling

STAT:

Signal transducer and activator of transcription

1 Introduction

Innate immunity provides a sophisticated system to protect the host from infectious danger. Recognition of a wide variety of microbes is achieved by a limited set of receptors which interact with conserved microbial structures 1. Among these receptors, Toll-like receptors (TLR) play an important role. TLR selectively interact with conserved microbial structures. Prototypical stimuli are lipoteichoic acid (LTA), lipopolysaccharide (LPS) and bacterial CpG-DNA, which are recognized by TLR-2, TLR-4 and TLR-9, respectively. Stimulation of TLR on innate immune cells results in secretion of pro-inflammatory cytokines, chemokines, increased effector functions such as phagocytosis and enhanced capacity to present antigen to T cells.

Although innate immunity by itself represents a powerful system to combat exogenous invaders, many infections can only be cleared in combination with the adaptive immune system 2. This is reflected by the fact that activation of innate immunity is a prerequisite for the induction of adaptive immune responses. Indeed, MyD88-deficient mice show impaired antigen-specific activation of Th1 responses 3. In turn, adaptive immunity induces positive feedback loops which increase the effectiveness of innate immunity. In this respect IFN-γ plays a major role and is needed to prime macrophages to defend certain microbes, e.g. mycobacteria 4. In addition, signal transducer and activator of transcription (STAT)1-deficient mice, which are defective in IFN-γ signaling, show an impaired innate immune response 5. Hence, full activation of innate immune cells is only achieved by the combinatorialaction of IFN-γ and TLR-dependent signals. Besides cooperative effects of IFN-γ and TLR signaling on macrophage function, opposite outcomes have also been reported recently 6.

During initiation of a response, costimulation with LPS enhanced IFN-γ signaling. The molecular mechanisms underlying these cooperative effects have been analyzed recently. It is well-known that IFN-γ interacts with a heterodimeric type II cytokine receptor resulting in activation of Janus kinase (Jak)1 and Jak2 as well as Y701 STAT1 phosphorylation. Tyrosine-phosphorylated STAT1 dimerizes and translocates into the nucleus were it initiates gene transcription. For complete transcriptional activity, however, STAT1 has also to be phosphorylated on S727, which is situated within a mitogen-activated protein kinase (MAPK) phosphorylation motif. Indeed, after costimulation with LPS an increased phosphorylation of STAT1 S727 was found, probably mediated by the MAPK p38 7. Similar enhancing effects on IFN-γ signaling were also seen after infection with Listeria monocytogenes8.

In contrast, prolonged incubation with the TLR-dependent stimuli LPS 9, 10 and CpG-DNA 11 resulted in a decrease in IFN-γ responsiveness which correlated with the induction of proteins of the family of suppressors of cytokine signaling (SOCS). SOCS, also termed STAT-induced STAT inhibitor (SSI) or Jak-binding protein (JAB), and the related cytokine-inducible SH2 domain-containing protein (CIS) comprise a family of eight proteins which contain differing N termini, a central SH2 domain and a C-terminal SOCS box 1215. SOCS proteins are induced by a variety of cytokines, growth factors and hormones which signal via the Jak/STAT pathway. SOCS proteins subsequently inhibit further receptor signaling 16. The N-terminal extended SH2 domain is crucial during this process. Therefore SOCS proteins are classical feedback inhibitors. However, recent data show that induction of SOCS can also be brought around by Jak/STAT-independent stimuli. LPS, CpG-DNA, TNF-α and insulin have been shown to induce various SOCS proteins 911, 1719. Moreover, forced MAPK activation also resulted in SOCS-3 induction with subsequent inhibition of IL-6 signaling 20. Jak/ STAT-independent SOCS induction thus may generate an intracellular cross-talk of signaling pathways that regulates IFN-γ sensitivity.

These observations indicate that IFN-γ signaling and TLR signaling are interconnected and severely affect the activation status of innate immune cells. To study whether the particular observations represent a general mechanism of IFN-γ and TLR interplay, we examined the effects of TLR ligands as well as whole bacteria on IFN-γ signaling. Co-incubation of TLR ligands and IFN-γ resulted in an early enhancement and later inhibition of IFN-γ signaling. At early phases TLR stimuli in general induced S727 phosphorylation of STAT1 thereby increasing STAT signaling. TLR stimuli also induced SOCS-1, SOCS-3 and CIS proteins. Especially SOCS-1 was able to inhibit IFN-γ signaling at later time points. Furthermore, SOCS proteins inhibited IFN-γ-induced up-regulation of MHC class II, CD40 and CD86, demonstrating the complex interconnectivity of TLR and IFN-γ signaling pathways.

2 Results

2.1 TLR-dependent microbial stimuli modulate IFN-γ signaling

We generated RAW 264.7 macrophages stably overexpressing a luciferase reporter under the control of the IFN-γ-inducible, STAT1-dependent IFP53 promoter. Six different clones were initially examined with similar results concerning IFN-γ responsiveness. IFN-γ dose-dependently induced the activation of the reporter gene reaching a maximum after 16 h and subsequently declining again (Fig. 1A). Co-incubation with either CpG-oligodeoxynucleotide (ODN), LPS or LTA resulted in a marked enhancement of IFN-γ responsiveness at early time points. Thus the maximum of reporter activity was already achieved after 4 h of co-incubation; however, the maximal promoter activity did not change. Surprisingly, co-incubation for longer than 8 h then led to an inhibition of IFN-γ responsiveness. The same results were obtained when using heat-inactivated whole-cell lysates of different bacteria (Fig. 1B).

TLR-dependent stimuli alone failed to induce the IFN-γ reporter (Fig. 1C), underlining the IFN-γ selectivity of the reporter gene construct. Pre-incubation with TLR-dependent stimuli for 2 h and subsequent stimulation with IFN-γ only slightly enhanced the reporter gene induction (Fig. 1D). Hence, for optimal induction TLR-dependent stimuli and IFN-γ have to be present in parallel. However, prestimulation for 16 h resulted in inhibition of IFN-γ responsiveness (Fig. 1D). To verify the results obtained with the reporter construct, we also determined the induction of the IFN-γ-inducible chemokine MIG/ CXCL9. This chemokine was exclusively inducible by IFN-γ but not by TLR stimuli. Corroborating the reporter experiments, pre-incubation for 2 h with CpG-ODN resulted in an enhanced induction by IFN-γ whereas this response was severely diminished after prestimulation for 14 h (Fig. 1E).

Figure 1.

Influence of TLR-dependent stimuli on IFN-γ signaling. (A) RAW 264.7 cells stably overexpressing a STAT1-dependent luciferase reporter were stimulated with 30 U/ml IFN-γ either alone or in the presence of 1 μM CpG-ODN, 0.5 μg/ml LPS or 10 μg/ml LTA. After the indicated incubation periods luciferase activity was determined. (B) The same experiment as in (A) was performed with heat-inactivated lysates (108 CFU/ml) of S. aureus, E. coli or P. aeruginosa. (C) Similarly, CpG-ODN, LPS and LTA were tested for STAT1 reporter activation alone. (D) Cells were pretreated with the indicated stimuli for different time periods. Subsequently cells were washed once and stimulated with 30 U/ml IFN-γ for additional 4 h. (E) Cells were pretreated with 1 μM CpG-ODN as indicated. Subsequently, cells were stimulated with 10 U/ml IFN-γ for 4 h. mRNA for MIG/CXCL9 was measured by quantitative RT-PCR with SybrGreen.

2.2 TLR-dependent stimuli induce serine phosphorylation of STAT1

Tyrosine phosphorylation of STAT1 is necessary for dimerization and subsequent transcriptional activation. In addition, serine phosphorylation on residue 727 was reported to unfold the full transcriptional activity of STAT1. We therefore tested tyrosine and serine phosphorylation of STAT1 after triggering cells with IFN-γ or TLR-dependent stimuli. Paralleling the results obtained with the STAT1 reporter, only IFN-γ but not TLR-dependent stimuli were able to induce rapid tyrosine phosphorylation of STAT1 within 15 min (Fig. 2A). However, IFN-γ and all of the tested TLR stimuli induced phosphorylation of S727 of STAT1 (Fig. 2A). When stimulation was performed for 3 h, TLR-dependent stimuli alone were able to induce a weak tyrosine phosphorylation of STAT1 (Fig. 2A). Delayed STAT1 tyrosine phosphorylation by TLR agonists alone correlated with IFN-β synthesis as measured by quantitative reverse transcription (RT)-PCR (data not shown). In addition we also evaluated the phosphorylation status of STAT1 after combined activation with LPS and IFN-γ (Fig. 2B). Costimulation for 2 h resulted in tyrosine and serine phosphorylation of STAT1. In contrast, after 16 h only IFN-γ induced tyrosine phosphorylation while serine phosphorylation was still observed in all of the tested conditions.

Figure 2.

TLR-dependent stimuli induce serine phosphorylation of STAT1. (A) RAW 264.7 cells were stimulated with 10 U/ml IFN-γ, 1 μM CpG-ODN, 0.5 μg/ml LPS or 10 μg/ml LTA for 15 min (first panel), 3 h (second panel) or 45 min (third and fourth panel). Cell lysates were blotted and probed with the indicated antibodies. (B) RAW 264.7 cells were stimulated with 0.5 μg/ml LPS and 10 U/ml IFN-γ for 2 h or 16 h. Western Blots were done as above. (C, D) RAW 264.7 cells stably overexpressing a STAT1-dependent luciferase gene were prestimulated either with PD98059 (C) or SB203580 (D) in the indicated concentration for 45 min. Subsequently, cells were stimulated with 10 U/ml IFN-γ and 1 μM CpG-ODN for further 4 h and then luciferase activity was determined. (E) RAW 264.7 cells were prestimulated with SB203580 for 30 min and then additionally activated with 0.1 μg/ml LPS for 45 min. Equal amounts of cell lysates were probed with phosphoserine-specific STAT1 antibody.

2.3 Enhancement of IFN-γ responsiveness by TLR stimuli is p38-dependent

To study the mechanisms of TLR-dependent enhancement of IFN-γ responsiveness, we used inhibitors of MAPK signaling. PD98059, which blocks activation of MAPK/ERK kinase (MEK)1/2, slightly reduced STAT1 reporter activation by IFN-γ (Fig. 2C). In parallel, reporter activity after co-incubation with CpG-ODN and IFN-γ was also reduced; however, the enhancing effect of CpG-ODN on IFN-γ signaling was preserved (Fig. 2C). In contrast, inhibition of p38 by SB203580 did not affect IFN-γ signaling yet abolished the costimulatory effect of CpG-ODN in a dose-dependent manner (Fig. 2D). Since p38 was involved in TLR-dependent enhancement of IFN-γ signaling, we next examined whether p38 influences serine phosphorylation of STAT1. We observed that the p38 inhibitor SB203580 indeed diminished serine phosphorylation in response to LPS; however, no complete inhibition was seen (Fig. 2E).

2.4 TLR mRNA regulation by IFN-γ 

In addition to collaborative effects of TLR ligands on IFN-γ signaling, IFN-γ in turn also regulates sensitivity to TLR stimuli like LPS. We asked whether IFN-γ could affect expression of TLR and thus influence TLR-signaling. Due to the lack of well-characterized mouse TLR antibodies, peritioneal macrophages were assayed for their quantitative TLR mRNA expression (Fig. 3). Macrophages showed marked differences in TLR mRNA expression profile. TLR-2 was transcribed at highest rates, followed by TLR-9 and TLR-6 (Fig. 3A). Upon activation with IFN-γ TLR were up-regulated within 1 to 16 h post-stimulation. However, transcription rate returned to baseline levels after 24 h. In addition, regulative effects were only moderate (Fig. 3B).

Figure 3.

Regulation of TLR mRNA expression in macrophages. Peritoneal macrophages were stimulated with 10 U/ml IFN-γ for the indicated time periods. Expression of TLR mRNA was determined by quantitative RT-PCR. (A) TLR mRNA expression in non-stimulated cells as expressed relative to β-actin. (B) TLR mRNA regulation after stimulation with IFN-γ as expressed in comparison to non-stimulated cells.

2.5 TLR-dependent stimuli induce SOCS

Next we analyzed the mechanisms responsible for negative regulatory effects of TLR stimuli on IFN-γ signaling after prolonged time periods. To this we examined the induction of mRNA for SOCS, which are known inhibitors of Jak/STAT signaling. SOCS-1, SOCS-3 and CIS were induced upon stimulation with either pure TLR ligands or heat-inactivated bacterial lysates (Fig. 4A, B) in RAW 264.7 macrophages. When using bone marrow-derived dendritic cells (BMDC), only SOCS-1 and SOCS-3 were induced (Fig. 4C). However, BMDC showed higher basal expression rates of SOCS-1, SOCS-2, SOCS-3 and especially CIS. In both cell types a slight down-regulation of SOCS-5 could be observed upon stimulation. Regarding the kinetics of SOCS induction, SOCS-3 was induced most rapidly, becoming measurable and peaking already 1 h after LPS stimulation (Fig. 4D). In contrast, SOCS-1 and CIS induction were more delayed with expression starting after 4 h. SOCS-2 and SOCS-5 were initially down-regulated. After 8 h SOCS induction decreased again; however, it stayed elevated in cell culture for up to 24 h (data not shown). SOCS-2 was slightly induced after 8 h.

Figure 4.

SOCS induction through microbial stimuli. (A, B) RAW 264.7 cells were stimulated with 0.5 μg/ml LPS, 1 μM CpG-ODN, 10 μg/ml LTA or heat-inactivated lysates of S. aureus, E. coli, P. aeruginosa for 4 h. mRNA expression was determined by quantitative RT-PCR and is expressed as induction relative to non-stimulated cells. (C) BMDC were assayed for SOCS induction as in (A). (D) Kinetics of SOCS mRNA induction were examined in RAW 264.7 cells after stimulation with 0.5 μg/ml LPS.

2.6 SOCS-1 is the most potent inhibitor of IFN-γ signaling

To test whether the induced SOCS proteins indeed modulated IFN-γ signaling, we transiently cotransfected RAW 264.7 macrophages with the IFN-γ reporter construct and differing amounts of SOCS (Fig. 5A). When these cells were stimulated with IFN-γ we observed that SOCS-1 was the most potent inhibitor of IFN-γ signaling. SOCS-1 did not significantly affect viability (data not shown). In addition, SOCS-3 and CIS were effective at inhibiting IFN-γ when using high amounts of plasmids. In contrast, SOCS-2 did not inhibit but even slightly enhanced IFN-γ signaling.

Moreover we generated clones that stably overexpressed either of the SOCS proteins. Clones were screened for SOCS mRNA expression of the respective SOCS by quantitative RT-PCR. Clones which showed expression patterns similar to the induction observed in RAW 264.7 cells after stimulation were selected. These cells were stimulated with IFN-γ and then stained with phosphotyrosine-specific STAT1 antibodies (Fig. 5B). While control cells showed prominent nuclear staining, SOCS-1-overexpressing cells were almost completely negative. SOCS-3 and CIS overexpression resulted in a slightly reduced number and intensity of nuclear staining.

In addition, SOCS-1 overexpression resulted in a nearly complete loss of nitrite induction after IFN-γ stimulation (Fig. 6A). In a costimulation assay with CpG-ODN and IFN-γ administration, again SOCS-1 overexpression abolished the IFN-γ-induced nitrite induction while CpG-ODN-induced secretion was unaffected (Fig. 6B).

Figure 5.

Inhibition of IFN-γ signaling by SOCS. (A) RAW 264.7 cells were transiently transfected by electroporation with decreasing amounts of SOCS or CIS expression plasmids (20, 10, 5 μg) together with 15 μg of IFN-γ-inducible IFP53 luciferase reporter. DNA quantity was hold constant by addition of control plasmid DNA. After 16 h cells were stimulated with 10 U/ml IFN-γ for 6 h and thereafter luciferase activity was determined. Displayed are typical results of six independent experiments. (B) RAW macrophages stably overexpressing either SOCS-1, SOCS-2, SOCS-3 or CIS were stimulated with 100 U/ml IFN-γ for 30 min. Cells were fixed in methanol, stained with phosphotyrosine-specific STAT1 antibody and photographed.

Figure 6.

SOCS expression reduces IFN-γ-induced nitric oxide generation. RAW 264.7 cells stably overexpressing either SOCS-1, SOCS-2, SOCS-3 or CIS were assayed for nitrite generation in the supernatant after stimulation for 26 h. (A) Cells were activated with 30 U/ml IFN-γ. Mean values and SD of three experiments are given. (B) Cells were incubated with either 1 μM CpG-ODN alone or in combination with 30 U/ml IFN-γ. Shown are the results of one of three independent experiments.

2.7 Pretreatment with TLR-dependent stimuli reduces IFN-γ-induced STAT1 phosphorylation

To correlate SOCS induction with loss of IFN-γ responsiveness after TLR triggering, we examined the effects of pretreatment of cells with TLR stimuli on IFN-γ signaling. Pretreatment for 14 h with subsequent IFN-γ stimulation resulted in a complete loss or marked reduction in STAT1 tyrosine phosphorylation as determined by immunocytochemistry (Fig. 7A) and Western blot (Fig. 7B). However, pre-incubation for 1 h was not sufficient for inhibition of IFN-γ signaling.

Figure 7.

Pretreatment with TLR-dependent stimuli reduces IFN-γ responsiveness. (A) RAW macrophages were pretreated with 1 μM CpG-ODN, 0.5 μg/ml LPS or 10 μg/ml LTA for 14 h. Subsequently, cells were stimulated with 10 U/ml IFN-γ for 30 min. Cells were then fixed, stained with phosphotyrosine-specific STAT1 antibody and photographed. (B) RAW macrophages were prestimulated as in (A) for either 1 h or 12 h. Cells were then stimulated with 3 U/ml IFN-γ for 20 min. Cell lysates were probed with either phosphotyrosine-specific STAT1 antibody or total STAT1 antibody.

2.8 Pretreatment with TLR stimuli differentially affects IFN-γ-induced expression of MHC class II and costimulatory molecules

In order to examine the interplay of TLR-dependent stimuli and IFN-γ on further functions of macrophages, we pretreated cells with either CpG-ODN, LPS or LTA and subsequently stimulated with IFN-γ and detected surface expression of antigen-presenting molecules (Fig. 8). Corroborating the above results, we also observed a marked inhibition of MHC class II up-regulation by IFN-γ after TLR pretreatment. However, when we analyzed the expression of CD40 and CD86, pretreatment with all of the tested stimuli increased IFN-γ-induced CD40 expression. CD86 was also increased with LPS and CpG-ODN but not with LTA pretreatment. LPS and CpG-ODN alone were effective in slightly up-regulating CD86 after 48 h of stimulation, and all of the stimuli up-regulated CD40 after 72 h (data not shown). Nevertheless all of the TLR stimuli were only weak inductors of MHC class II expression when stimulated for prolonged time periods.

Figure 8.

Prestimulation with TLR-dependent stimuli differentially affects IFN-γ-induced surface molecules. RAW 264.7 cells were prestimulated with 1 μM CpG-ODN, 0.5 μg/ml LPS or 10 mg/ml LTA for 12 h. Cells were further incubated with either medium (solid gray) or 5 U/ml IFN-γ (black line) for 48 h. Cells were examined for surface expression of MHC class II, CD40 and CD86 by flow cytometry. Displayed is the percentage of positive cells. Controls were performed with isotype staining.

2.9 SOCS-1 overexpression diminishes IFN-γ-induced up-regulation of MHC class II, CD40 and CD86

Finally we studied the effects of SOCS molecules on IFN-γ-induced up-regulation of antigen presentation (Fig. 9). SOCS-1 overexpression resulted in a complete loss of IFN-γ responsiveness in terms of MHC class II, CD40 and CD86 regulation. In contrast, SOCS-3 overexpression led to a slight reduction of MHC class II and CD40 regulation while CIS affected regulation of CD40 and CD86.

Figure 9.

SOCS expression inhibits IFN-γ-induced up-regulation of antigen-presenting molecules. RAW cells stably overexpressing SOCS or CIS were incubated for 48 h with either medium (solid gray) or 10 U/ml IFN-γ (black line). Subsequently, expression of MHC class II, CD40 and CD86 was determined by flow cytometry. Displayed is the percentage of positive cells of one representative out of five experiments.

3 Discussion

In functional terms, complete activation of macrophages requires two independent signals: IFN-γ and microbial patterns which are sensed via TLR. Here we provide evidence that the intracellular signaling pathways which are distinct in terms of the involved molecules are functionally interconnected.

A short contact with both stimuli results in a marked enhancement of IFN-γ responsiveness. S727 phosphorylation of STAT1, which is known to enhance transcriptional activity, is induced by each of the tested TLR-dependent stimuli. This is in accordance with previous results showing S727 phosphorylation of STAT1 after stimulation with LPS or L. monocytogenes8, 10. However, our results extend these observations towards a general mechanism of TLR. TLR-induced serine phosphorylation was dependent on the p38 MAPK (Fig. 2D). Similar results were also reported by Decker et al. 7 yet have been controversially discussed in literature 21, 22.

Extracellular signal-regulated kinase (ERK)2 has also been proposed to mediate this activation 21; however, our results do not support these findings. Interestingly, we observed that ERK activation contributed to IFN-γ-mediated activation of the IFP promoter. Indeed it has been reported that ERK activation in response to IFN-γ  contributes to the control of gene transcription 23. We did neither observe rapid tyrosine phosphorylation of STAT1 nor IFP promoter activation by TLR stimulation alone. However, prolonged stimulation for 3 h resulted in tyrosine phosphorylation of STAT1 after TLR triggering (Fig. 2A) as has been reported by other groups for CpG-DNA 24 or LPS 25. It is therefore improbable that early enhancement of IFN-γ by TLR-stimuli is due to a direct activation of the Jak/STAT pathway by TLR. It can be speculated that autocrine stimulation via IFN-β synthesis results in STAT1 tyrosine phosphorylation as a secondary signaling event 25. In addition to the amplification of IFN-γ signaling through S727 phosphorylation, it has recently been reported that p38 MAPK can enhance IFN-γ signaling also independently of serine phosphorylation 26.

Besides the positive effects of TLR stimuli on IFN-γ signaling we also observed an early yet transient up-regulation of TLR mRNA after IFN-γ triggering. This in turn would induce a kind of intracellular amplification loop and is consistent with findings which show early up-regulation of TLR also to TLR stimuli themselves in dendritic cells 27.

However, prolonged activation of macrophages could also result in detrimental effects, thus innate cells have to be turned off again. Here we provide evidence that one general mechanism of down-regulation in response to microbial stimuli is the induction of SOCS proteins. SOCS-1 and SOCS-3 were consistently induced by TLR-dependent stimuli. This confirms and extends earlier reports of SOCS induction by LPS and CpG-DNA 10, 11. Concerning the mode of induction we previously reported that CpG-DNA was able to directly induce SOCS-1 and SOCS-3. Using cycloheximide we were able to show that intermediate protein synthesis was not necessary. Also the addition of neutralizing type I interferon antiserum did not alter CpG-DNA-mediated SOCS induction 11. This was in accordance with similar observations made for SOCS-3 induction by LPS 10. Contrasting results were communicated by Crespo et al. who proposed that LPS-mediated SOCS-1 induction was partly mediated by IFN-β 9. All of the tested TLR-dependent stimuli induced similar profiles of SOCS protein while IFN-β induction by TLR is reported to differ considerably 25.

SOCS family members show no absolute specificity for individual signaling pathways. SOCS-1 plays an important role in the inhibition of IFN-γ signaling 28, 29. Corroborating the importance of SOCS-1 for IFN-γ signaling we also observed that SOCS-1 was the most efficient inhibitor in comparison to CIS or SOCS-3. In addition, SOCS-3, which has also been implicated in LPS-mediated reduction of IFN-γ responsiveness, seems to play a more important role in IL-6 inhibition 30. Recently, it has been reported that SOCS-3 induction contributes to IL-10-mediated inhibition of LPS and that TNF-α-induced apoptosis in fibroblasts could be inhibited by SOCS-1 31, 32. This would expand the spectrum of SOCS-mediated inhibition to signaling pathways beyond of Jak/STAT signaling.

Moreover, it has been reported that SOCS-1 also exerts inhibitory effects on LPS signaling 33, 34. SOCS-1+/– mice and SOCS-1–/– mice on an IFN-γ-deficient background were more sensitive to LPS actions including lethality and cytokine induction. Furthermore, forced expression of SOCS-1 resulted in diminished NFκB activation after LPS triggering. In contrast, using an alternative approach with SOCS overexpression we did not observe major differences in nitric oxide induction after triggering with the TLR-9 stimulus CpG-DNA. Further studies using this approach are currently performed. However, it could be that deficiency of SOCS-1 during development markedly alters the general activation status of innate immunity, and thus the discrepancies between knockout studies and overexpression experiments could be explained. It also could be that intermediate cytokines which are secreted in an auto- or paracrine fashion are susceptible to SOCS inhibition, and in this respect IFN-β is an interesting candidate. Moreover, only one report was able to show a direct interaction of SOCS-1 with IL-1R-associated kinase (IRAK) but not with MyD88 34.

Consequently we were able to show that pretreatment with TLR ligands decreases IFN-γ sensitivity due to SOCS induction. However, when examining expression of MHC class II, CD40 and CD86, we only observed a reduction of IFN-γ-induced MHC class II regulation. In contrast, CD40 and CD86 expression were even up-regulated. Multiple facts could explain these differences. First, MHC class II expression is mainly regulated by IFN-γ. In contrast, CD40 and CD86 are also regulated by TLR-dependent stimuli, especially when performing longer stimulation. Thus it could be that for CD40 and CD86 synergistic effects overcome reduced but not abolished IFN-γ signaling. Second, additional IFN-γ-mediated signals could exist which are not susceptible to SOCS inhibition. In this respect Johnston et al. have reported that SOCS-3 can be tyrosine-phosphorylated and that this modification leaves intact activation of ERK but inhibits STAT5 signaling in response to IL-2 35. Strengthening the assumption of differential effects of SOCS on IFN-γ signaling it has been reported that some genes can be induced by IFN-γ independently of STAT1 36.

In conclusion, TLR-dependent stimuli severely modify IFN-γ signaling and the net result can be either enhancement or inhibition. This in turn impacts the activation status of macrophages and emphasizes the temporal component in cellular activation. In addition, the microenvironment will modulate the cell's responsiveness towards cytokines. Inhibition of IFN-γ signaling could contribute to a general down-modulation as a protection of overwhelming inflammation. On the other hand, SOCS expression could help in sharpening the cell's response to microbes by inhibiting various cytokines which within the microenvironment act on the cell. Thus SOCS expression in innate immune cells could inhibit further developmental signals delivered by GM-CSF or G-CSF 37 but leaves intact the activation through microbial stimuli via TLR. This would contribute to the undisturbed generation of the indispensable pro-inflammatory program after microbial encounter.

4 Materials and methods

4.1 Reagents and plasmids

Phosphorothioate-modified ODN 1668 (TCCATGACGTTCCTGATGCT) was custom synthesized by TIB Molbiol (Berlin, Germany). S-form LPS from the wild-type strain Salmonella enterica sv. Minnesota was obtained from U. Seydel (Research Center Borstel, Germany). Highly purified (>99%) biologically active LTA from Staphylococcus aureus was prepared as described 38. Heat-inactivated lysates were prepared from S. aureus (ATCC 29213), Escherichia coli (ATCC 25922) and Pseudomonas aeruginosa (ATCC 27853) by incubation at 98°C for 30 min.

MAPK inhibitors SB203580 and PD98059 were purchased from Calbiochem (Schwalbach, Germany). IFN-γ was obtained from Dr. Adolf (Vienna, Austria). Antibodies to STAT1 were purchased from New England Biolabs (Frankfurt, Germany).

The IFN-γ-inducible, STAT1-dependent IFP53 luciferase reporter plasmid was donated by T. Decker (Vienna, Austria). Mammalian expression vectors for SOCS-1, SOCS-2 and SOCS-3 were described previously 12 and were obtained from D. Hilton (Victoria, Australia). A neomycin resistance cassette was a kind gift of H. Haecker (Munich, Germany). CIS expression plasmid was from A. Yoshimura (Kurume, Japan).

4.2 Cells

RAW 264.7 cells were cultured in Clicks/RPMI 1640. BMDC from BALB/c mice were prepared as described 39. Peritoneal macrophages were obtained 4 days after injection of 1.5 ml4% thioglycolate intraperitoneally. Cells were plated in culture medium and washed 4 h after plating to remove remaining nonadherent cells.

4.3 Determination of nitric oxide

Cells (1.5×105) were stimulated in 96-well plates in 300 μl of medium for 26 h. Nitric oxide accumulation was measured photometrically (550 nm) by mixing equal parts of supernatant and Griess reagent (1:1 mixture of 1 g% sulfanilamide/5% H3PO4 and 0.1% naphthyl-ethylenediamine dihydrochloride).

4.4 Transfection experiments

RAW 264.7 cells (4×106) were transfected by electroporation at 290 V, 1050 μF in an EasyjecT plus gene pulser (PeqLab, Erlangen, Germany). Cells (1×106) were plated and used for stimulation.

Stable transfections were established by cotransfection of SOCS, CIS or IFP53 plasmids and the neomycin resistance plasmid in a ratio of 10:1. Cells were plated and 24 h after transfection 0.8 μg/ml G418 (Gibco-BRL, Karlsruhe, Germany) was added. G418-resistant clones were picked, expanded and tested for expression of SOCS or CIS mRNA by quantitative RT-PCR. IFP53-transfected cells were tested for IFN-γ inducibility. Luciferase induction was measured with LucLite (Packard, The Netherlands) in a TopCount scintillation device (Packard) and is either expressed as cpm or as inducibility (ratio of luciferase activity in stimulated cells divided by activity in control cells).

4.5 Quantitative RT-PCR

Total RNA from the cells was isolated by using HighPureTMRNA-kit (Roche, Mannheim, Germany) which included DNase I digestion. One microgram of total RNA preparation was reverse-transcribed with cDNA synthesis kit (MBI Fermentas, St. Leon-Rot, Germany). cDNA diluted 1:4 was used as template in the TaqMan-PCR-mix according to the manufacturer's standard protocol (Eurogentec, Seraign, Belgium; ABI Prism 7700, Applied Biosystems, Germany).

Primer for β-actin, SOCS-1, SOCS-3 and TNF-α have been described 11. Further primers: TLR-2 (sense: GCCAAGAGGAAGCCCAAGA, antisense: AAGGGCGGGTCAGAGTTCTC), TLR-4 (sense: GCAGGTGGAATTGTATCGCC, antisense: TTCGAGGCTTTTCCATCCAA), TLR-6 (sense: CGAGCCTGAGGCATCTAGACC, antisense: GAGCAACTGGGAGCAGATCC), TLR-9 (sense: GGGCCCATTGTGATGAACC, antisense: GCTGCCACACTTCACACCAT), SOCS-2 (sense: AGGCCCAGAAGCCCCAC, antisense: GTTGGTAAAGGCAGTCCCCA), SOCS-5 (sense: TTGGTAGAACTCGAAGCGGC, antisense: AACACAAACCCACCGTGTCC), CIS (sense: GAACCGAAGGTGCTAGACCCT, antisense: TGTACCCTCCGGCATCTTCT) and MIG/CXCL9 (sense: TCCTTTTGGGCATCATCTTC, antisense: CAGTGTAGCAATGATTTCAGTTTTG) were purchased from MWG (Munich, Germany). Fluorogenic probes (FAM/TAMRA) were obtained from Eurogentec, Belgium: TLR-2: CCTGCAGGGACGTTTGCTATGATGC, TLR-4: TCTTAGCAGAAACACCTACCTGGAATGGGAG, TLR-6: TCATTCAATGACTTTGATGTACTGCCTGTGTGT, TLR-9: CAGTTCTAGACGTGAGAAGCAACCCTCTGC, SOCS-2: AATGGGACTGTTCACCTGTACCTGACCAAA, SOCS-5: TTCAGAGGCGAGAGCGGCGC, CIS: ATCTGCTGTGCATAGCCAAGACGTTCTCC.

Specificity of RT-PCR was controlled by no template and no RT controls. PCR efficiencies for all reactions were determined and were similar (0.94–1.0). Threshold values were normalized to expression of β-actin. Quantitative PCR results are expressed either as relative induction towards the housekeeping gene β-actin (1/2ΔCt) or were set in reference to non-stimulated control cells. The calculated values then represent the 2n-fold induction of mRNA.

4.6 Western blot

RAW 264.7 cells (10×106) were stimulated as indicated in the experiment. Cell lysates were prepared directly in 1× SDS buffer according to a standard protocol of New EnglandBiolabs (Frankfurt, Germany). Protein content was determined and equal amounts of lysates were fractionated by 12% SDS-PAGE and electrotransferred to polyvinylidene difluoride membranes. Detection was accomplished by enhanced chemiluminescence system (Amersham, Freiburg, Germany).

4.7 Immunocytochemistry

Cells were grown in medium containing 0.5% FCS in eight-well chamber slides (Nunc, Wiesbaden, Germany) for 2 days. Cells were then treated as indicated. Subsequently cells were fixed in methanol and incubated with the indicated antibody at 4°C overnight. Staining was done with ABC staining system (Santa Cruz Biotechnologies, Heidelberg, Germany).

4.8 Flow cytometry analysis

Cells were washed in PBS/2% FCS. Fc-block was done with anti-FcγRII/III monoclonal antibodies (clone 2.4G2, PharMingen) and 10% normal mouse serum. Cells were stained with fluorescein isothiocyanate (FITC)-conjugated anti I-Ad/Ed monoclonal antibodies (clone 2G9, PharMingen), FITC-CD40 (clone HM40–3, PharMingen) or phycoerythrin-CD86 (clone GL1, PharMingen). Cells were analyzed on a Partec PAS flow cytometer (Dako, Germany).

Acknowledgements

We thank Nadine Schelberg and Helena Bykow for excellent technical assistance. We appreciate the helpful support by T. Decker and H. Häcker. This work was supported by grants of the Deutsche Forschungsgemeinschaft DFG He1452/2, He1452/4 and by the European Community (QLK2-CT-2000–00336).

Footnotes

  1. 1

    WILEY-VCH

  2. 2

    WILEY-VCH

  3. 3

    WILEY-VCH

  4. 4

    WILEY-VCH

  5. 5

    WILEY-VCH

  6. 6

    WILEY-VCH

  7. 7

    WILEY-VCH

  8. 8

    WILEY-VCH

  9. 9

    WILEY-VCH

Ancillary