Abstract
IFN-β induces the production of secreted IL-1R antagonist (sIL-1Ra) without triggering synthesis of the agonist IL-1β in human monocytes. This might account for its anti-inflammatory properties. Canonically, IFN-β signals through activation of JAK/STAT pathway, although PI3K and MAPK have also been involved. In this study, the role of PI3K, MEK1, and STAT1 in IFN-β-induced sIL-1Ra production is investigated in freshly isolated human blood monocytes. PI3K, but not MEK1 activation is essential for sIL-1Ra production in monocytes treated with IFN-β, as demonstrated by using the respective inhibitors of PI3K and MEK1, Ly294002 and PD98059. The use of cycloheximide and actinomycin D shows that sIL-1Ra was an immediate early gene induced by IFN-β and that PI3K was controlling sIL-1Ra gene transcription. Although both inhibitors of PI3K and MEK1 diminished the Ser727 phosphorylation of STAT1 induced by IFN-β, only Ly294002 inhibited sIL-1Ra production. Furthermore, the inhibition of STAT1-Ser727 phosphorylation by Ly294002 did not affect STAT1 translocation, suggesting that STAT1 was not involved in sIL-1Ra gene induction. This was confirmed in monocytes that were transfected with small interfering RNA specifically targeting STAT1. Indeed, monocytes in which effective STAT1 gene knockdown was achieved were fully responsive to IFN-β in terms of sIL-1Ra production. Taken together, the present data demonstrate that the induction of sIL-1Ra transcription and production by IFN-β in human monocytes involved PI3K, but not STAT1 activation.
Interleukin-1 receptor antagonist (IL-1Ra)3 is a member of the IL-1 family. Three protein forms of IL-1Ra resulting from the same gene have been described, two of them being intracellular and the third secreted (sIL-1Ra) (1). The function(s) of the intracellular forms of IL-1Ra is still elusive, but sIL-1Ra binds competitively to IL-1RI without inducing signal transduction, and thus inhibits IL-1α and IL-1β actions. Some stimuli induce sIL-1Ra in the absence of IL-1 production in human monocytes, including IL-3, IL-4, GM-CSF, leptin, and IFN-β (2, 3, 4, 5). Other stimuli, such as LPS and direct cellular contact with stimulated T cells, induce the production of both sIL-1Ra and IL-1β (6, 7, 8).
An imbalance between pro- and anti-inflammatory cytokines has been involved in the pathology of chronic immunoinflammatory diseases, such as multiple sclerosis and rheumatoid arthritis (9, 10). IFN-β has proved beneficial to patients with relapsing-remitting multiple sclerosis (11) and could be a potential therapy for rheumatoid arthritis (12, 13, 14). The therapeutic effects of IFN-β might be due to the restoration of the balance between pro- and anti-inflammatory cytokines (15, 16, 17). However, cellular mechanisms involved in cytokine production and targeted by IFN-β in chronic inflammatory diseases remain unclear (11). In human monocytes activated by proinflammatory stimuli, IFN-β displays opposite effects depending on the type of stimulus. Indeed, IFN-β inhibits TNF and IL-1β production in cell contact-mediated T lymphocyte signaling of monocytes (4, 18), in contrast with LPS-activated monocytes in which IFN-β enhances the production of both IL-1β and TNF (19). However, with the latter stimuli, the production of sIL-1Ra is enhanced upon addition of IFN-β, which as such potently induces the production of sIL-1Ra in monocytes (4). In contrast, IFN-β does not induce IL-1β protein or transcript (4, 19).
The type I IFN receptor complex (IFNAR) is expressed on most cell types and consists of two structurally related polypeptides, one of which binds the cytokine (IFNAR-2) and the other transduces the signal (IFNAR-1) (for review, see Ref.20). The canonical pathway of intracellular signaling used by IFN-β involves the activation of the two receptor-associated Janus protein tyrosine kinases JAK1 and Tyk2, which in turn activate by tyrosine phosphorylation members of the STAT family, STAT1, STAT2, and STAT3 (20, 21). This leads to the formation of transcriptional activator complexes, i.e., STAT1-STAT1, STAT1-STAT2, STAT1-STAT3, and STAT3-STAT3 (22, 23, 24). In addition, receptor ligation leads to the recruitment of downstream signaling elements, including STAT3 (25) and insulin receptor substrate (IRS) proteins to IFNAR-1 (26). Both IRS-1 and STAT3 have been shown to function as adapter proteins, linking IFNAR-1 to the p85 subunit of PI3K, resulting in enzyme activation (26). However, another study demonstrates a direct interaction between PI3K and IFNAR-1, but not with STAT3 (27). Other transduction pathways involving MAPK have been shown to modify the JAK-STAT pathway by interacting with IFNAR-1 (28).
Several intracellular pathways leading to sIL-1Ra production in human monocytes have been described that might depend on the stimulus or the cell differentiation/maturation stage. MAPK such as Raf-1 and ERK1/ERK2 have been involved in sIL-1Ra induction by LPS and leptin (29, 30). In contrast, serine/threonine phosphatases are involved in the induction of sIL-1Ra in monocytic cells activated by cellular contact with stimulated T cells (31), and STAT6 mediates the induction of sIL-1Ra by IL-4 (32). Furthermore, an LPS-inducible PI3K-dependent signaling pathway contributes to the elevated translation of sIL-1Ra in septic/LPS-adapted leukocytes, a pathway that does not affect the production of IL-1β (33). The present study addresses the question as to the signaling pathways involved in IFN-β induction of sIL-1Ra in human monocytes. PI3K, but not MEK1 activation is essential for sIL-1Ra production, as demonstrated by using pharmacological inhibitors of PI3K and MEK1. Furthermore, although PI3K is involved in Ser727 phosphorylation of STAT1, the latter factor, which is part of the canonical IFN-β signaling pathway, is not involved in sIL-1Ra expression, as shown by knocking down STAT1.
Materials and Methods
Materials
FCS, streptomycin, penicillin, l-glutamine, RPMI 1640, PBS free of Ca2+ and Mg2+, and TRIzol reagent (Invitrogen Life Technologies); Ficoll-Paque (Pharmacia Biotech); PMSF, neuraminidase, and polymyxin B sulfate (Sigma-Aldrich); and Complete Mini EDTA-free (Roche Diagnostics) were purchased from the designated suppliers. Human rIFN-β-1a (IFN-β) with a sp. act. of 3.97 × 108 IU/ml was a gift from ARES-Serono. IFN-γ (5 × 106 IU/ml) was obtained from Biogen Idec. Kinase inhibitors Ly294002, PD98059, and genistein were purchased from Calbiochem-Novabiochem. Other reagents were of analytical grade or better.
Monocytes
Cytokine production
Monocytes were activated with the indicated stimulus in RPMI 1640 medium supplemented with 10% heat-inactivated FCS, 50 μg/ml streptomycin, 50 U/ml penicillin, 2 mM l-glutamine, and 5 μg/ml polymyxin B sulfate (medium) in 96-well plates (5 × 104 cells/well/200 μl) and cultured for 24 h or as otherwise stated. The production of sIL-1Ra was measured in culture supernatants by commercially available enzyme immunoassay for sIL-1Ra (Quantikine; R&D Systems).
sIL-1Ra mRNA
Monocytes (5 × 106 cells/well/500 μl) were cultured in 24-well plates for 1 h with the indicated inhibitor and then for an additional 1 h with IFN-β (104 U/ml). Total RNA was isolated with TRIzol and analyzed by a commercially available RNase protection assay system kit with hck2 template set (BD Pharmingen), as previously described (34).
Western blot analysis
Monocytes were starved for 16 h at 37°C in RPMI 1640 medium supplemented with 1% heat-inactivated FCS in nonadherent conditions, i.e., in polypropylene tubes (Falcon; BD Biosciences). Cells were harvested and resuspended at 8 × 106 cells/ml in medium supplemented with 1% heat-inactivated FCS, and 500 μl was placed in 2-ml polypropylene tubes (Eppendorf) at 37°C. After 1 h, inhibitors were added or not for 45 min, and then cells were stimulated with IFN-β (104 U/ml). After the indicated time of incubation, the reaction was stopped by the addition of 800 μl of ice-cold PBS and centrifugation. Total cell lysate was prepared and subjected to Western blot analysis, as described previously (35). The blots were probed with anti-STAT1, anti-phospho Ser727 STAT1, and anti-phospho Tyr701 STAT1 (Upstate Biotechnology). Secondary HRP-conjugated goat anti-rabbit Abs were from DakoCytomation. Ab-bound proteins were detected by the Uptilight hrp Blot Chemiluminescent substrate (Uptima).
Preparation of nuclear extracts and EMSA
Cell nuclear extracts were prepared, as described previously (36). Protein concentrations were determined by the method of Bradford (37). Nuclear extracts were analyzed for STAT1-binding activity by EMSA. The binding reaction mixture contained 5 μg of protein, 2 μg of poly(dI-dC), and 10 μg of BSA in a final volume of 15 μl of extraction buffer containing 20% glycerol. Each reaction contained 5 × 104 cpm of unblunted dsSTAT1 oligonucleotide (5′-GTGCATTTCCCGTAAATCTTGTC-3′ and 5′-TGTAGACAAGATTTACGGGAAAT-3′) that was labeled by fill-in with DNA polymerase I large (Klenow) fragment in the presence of [α-32P]dCTP, as described (30). The reaction mixtures were incubated for 1 h at room temperature. Supershift was conducted by adding 100 μg/ml anti-STAT1 Ab (Upstate Biotechnology) 30 min before the end of the reaction. Free and bound DNA were separated by electrophoresis on a 4% nondenaturing polyacrylamide gel in 50 mM Tris, pH 8.0, containing 380 mM glycine and 2 mM EDTA. Gels were dried and subjected to autoradiography.
STAT1 knockdown
STAT1 was silenced by using the 4-for-Silencing kit provided by Qiagen. The four small interfering RNA (siRNA) sequences targeting human STAT1 were as follows: 1) 5′-GACCCAAUCCAGAUGUCUA-3′; 2) 5′-AAGTCATGGCTGCTGAGAATA-3′; 3) 5′-GUUCGGCAGCAGCUUAAAA-3′; and 4) 5′-GUCCUGAGUUGGCAGUUUU-3′. siRNAs were annealed according to manufacturer’s instructions, and then stored at −20°C before use. Monocytes were transfected with siRNA in Nucleofector device (Amaxa) by using Nucleofector human CD34 kit. Briefly, 5.6 μg of siRNA mixture (i.e., 1.4 μg of each STAT1-specific siRNA) was added to 3 × 106 monocytes that were previously washed in PBS, and resuspended in 100 μl of human CD34+ cell kit transfection solution. Cells were subjected to nucleofection using the U08 program. Control cells were either mock transfected or transfected with 6 μg of nonsilencing rhodamine-labeled siRNA (rho-siRNA). Transfected cells were immediately diluted in 2 ml of 37°C prewarmed RPMI 1640 complete medium and seeded into 12-well plates (2.1 ml/well). After 3 h, 30–40% of ρ-siRNA-transfected cells were labeled, as assessed by light and fluorescent microscopy. After 24 h, transfected cells were harvested, and their ability to produce sIL-1Ra upon IFN-β treatment was assessed, as described above. STAT1 knockdown was ascertained by quantitative real-time PCR and Western blot. Western blot was conducted, as described above, using anti-STAT1-specific Ab and anti-ERK1/2-specific Ab (Cell Signaling Technology) as a control. To ascertain that STAT1-knocked down cells were defective in STAT1-dependent responses, transfected cells were stimulated for 3 h with either IFN-β (1 × 104 U/ml) or IFN-γ (500 U/ml), and total mRNA was analyzed by real-time PCR for the expression of STAT1, sIL-1Ra, and the STAT1-dependent gene, FcγR1 (38, 39). Quantitative real-time PCR analysis (TaqMan quantitative ABI PRISM 7900 Detection System) was conducted after reverse transcription of mRNA prepared by RNeasy minikit (Qiagen). The expression level of mRNAs was normalized to the expression of a housekeeping gene (18S). STAT1, FcγR1, sIL-1Ra, and 18S probes were obtained from Applied Biosystems. All measurements were conducted in triplicates.
Statistics
When required, significance of differences between groups was evaluated using Student’s paired t test; p < 0.05 was considered to be significant.
Results
Activation of PI3K, but not of MEK1, is essential for optimal sIL-1Ra production in human monocytes stimulated by IFN-β
To determine whether MEK1 and PI3K were involved in sIL-1Ra production induced by IFN-β, monocytes were treated for 1 h with Ly294002 and PD98059, which inhibit PI3K and MEK1 activity, respectively, before stimulation with IFN-β. After 24-h incubation, the production of sIL-1Ra induced by 104 IU/ml IFN-β reached 9.8 ± 1.3 ng/ml (Fig. 1,A), no IL-1β being detectable (data not shown). Ly294002 inhibited the production of sIL-1Ra in a dose-dependent manner, reaching 60 ± 9% inhibition with 20 μM Ly294002, whereas PD98059 had no effect (Fig. 1,A). Confirming the inhibitor effects on protein production, Ly294002 inhibited sIL-1Ra mRNA level by 80% at the highest dose, whereas PD98059 had no effect (Fig. 1,B). These results did not depend on monocyte preparation. Indeed, as shown in Fig. 1,C, sIL-1Ra production was significantly inhibited (40 ± 20%) by 10 μM Ly294002, a suboptimal dose that was used to avoid cytotoxicity often observed at higher concentrations in long-term (24-h) cultures. Similarly, IFN-β-induced expression of sIL-1Ra transcript was inhibited by 58 ± 23% in the presence of 20 μM Ly294002 (Fig. 1 D). Together these results suggest that contrary to MEK1, PI3K was involved in the signaling pathway leading to sIL-1Ra production in monocytes stimulated by IFN-β.
PI3K is involved in the induction of sIL-1Ra gene transcription
Because the production of sIL-1Ra might be regulated at several levels, the involvement of PI3K in sIL-1Ra transcription or translation mechanisms was assessed. To ascertain that sIL-1Ra was an immediate early gene in IFN-β-treated monocytes, cycloheximide was used to interfere with the expression of a putative protein intermediate. The IFN-β induction of sIL-1Ra transcript was not inhibited, but enhanced, by cycloheximide (Fig. 2,A), demonstrating that sIL-1Ra was indeed an immediate early gene induced by IFN-β in human monocytes. Despite its enhancing effect on sIL-1Ra mRNA, cycloheximide effectively inhibited sIL-1Ra production in treated monocytes (Fig. 2,B). The inhibitory effect of the PI3K inhibitor Ly294002 was not affected by cycloheximide, the expression of sIL-1Ra transcript induced by IFN-β being inhibited 1.8-fold in the presence and absence of cycloheximide (Fig. 2,C). This further suggests that PI3K activation was directly involved in the induction of sIL-1Ra gene transcription or mRNA stabilization, and that this effect did not require protein neosynthesis. Levels of mRNA may be reduced by inhibiting transcription or by decreasing mRNA stability. To discriminate between effects of Ly294002 on transcription vs mRNA turnover, the stability sIL-1Ra transcript in the presence and absence of inhibitor was measured. As shown in Fig. 3,A, the transcription inhibitor, actinomycin D, abolished the induction of sIL-1Ra mRNA by IFN-β, thereby allowing the determination of mRNA stability in the absence of transcription. However, because sIL-1Ra mRNA reached a steady-state level between 12 and 15 h and remained stable at least until 24 h (data not shown) (4), the effect of Ly294002 on sIL-1Ra transcription or mRNA stability was conducted by adding actinomycin D 3 and 4 h after the addition of IFN-β and Ly294002, respectively, to avoid cytotoxic effect display by the inhibitor in long-term activation. In the absence of actinomycin D, sIL-1Ra mRNA levels enhanced as a function of incubation time with IFN-β (Fig. 3, B and C). The sIL-1Ra transcript levels were lower in the presence of Ly294002 throughout the experiment (Fig. 3, B and C). When actinomycin D was added to monocytes (i.e., 3 h after activation by IFN-β), the levels of sIL-1Ra transcript remained unchanged throughout the experiment regardless of the presence of Ly294002 (Fig. 3, B and C). This demonstrates that PI3K activation by IFN-β controlled sIL-1Ra gene transcription rather than contributing to sIL-1Ra mRNA stabilization.
IFN-β-induced phosphorylation of STAT1 on Ser727 involves PI3K activation
Because the canonical signaling pathway used by IFN-β occurs through the activation and transduction of STAT1, the effect of kinase inhibitors on STAT1 phosphorylation on both Ser727 and Tyr701 residues was examined. The phosphorylation of STAT1 on Ser727 was already observed after 15 min of monocyte treatment with IFN-β, reached a maximum at 30–60 min, and lasted for at least 2 h (Fig. 4,A). The phosphorylation of Tyr701 was observed after 15 min and was virtually undetectable after 2 h (Fig. 4,A). Treatment of monocytes with IFN-β in the presence of Ly294002 or PD98059 resulted in the inhibition of STAT1 phosphorylation on Ser727 by 22 and 63%, respectively (Fig. 4,B). The Tyr701 phosphorylation was significantly affected only by genistein, an unspecific tyrosine kinase inhibitor used as control, which also inhibited Ser727 phosphorylation of STAT1 (Fig. 4,B). Both Ly294002 and genistein inhibited sIL-1Ra production, in contrast with PD98059 that affected Ser727 phosphorylation without inhibiting the production of sIL-1Ra (Fig. 4, B and C). Although indirectly, this strongly suggests that the phosphorylation state of STAT1 was not related to the production of sIL-1Ra. Together these data imply that the production of sIL-1Ra induced by IFN-β in human monocytes depended on the activation of PI3K, but not of STAT1.
STAT1 nuclear translocation and binding to consensus DNA sequence are not inhibited by Ly294002
To further assess that Ly294002 inhibition of STAT1 Ser727 phosphorylation was not involved in sIL-1Ra inhibition, the effect of the inhibitor on STAT1 translocation and binding to IFN-γ activation site (GAS) consensus sequence was tested. As shown in Fig. 5, STAT1 translocation and binding to consensus sequence were inhibited in the presence of genistein, but not in the presence of Ly294002, confirming that the modulation of STAT1 Ser727 phosphorylation by PI3K was not important in STAT1 translocation. This demonstrates that PI3K did not inhibit STAT1 translocation and binding to GAS consensus sequence, further suggesting that STAT1 activation was not involved in the induction of sIL-1Ra expression mediated by PI3K in IFN-β-treated human monocytes.
STAT1 is not required for sIL-1Ra induction
To assess that STAT1 was not involved in the IFN-β-triggered signaling pathway leading to sIL-1Ra production, STAT1 was silenced in monocytes by using siRNA duplex interference. STAT1 mRNA expression level was silenced by 90 ± 1% in monocytes transfected with siRNA directed against STAT1 as compared with mock-transfected cells (Fig. 6). In monocytes transfected with the control siRNA, STAT1 mRNA expression level was slightly decreased by 23 ± 19% (Fig. 6,A). However, this decrease was not observed at the STAT1 protein levels. Indeed, the Western blot analysis showed that STAT1 protein expression was not affected in either mock-transfected cells or cells transfected with control siRNA, in contrast with cells transfected with siRNA directed against STAT1 (Fig. 6,B). As shown in Fig. 6,C, STAT1 knockdown did not affect sIL-1Ra production in monocytes activated by IFN-β, because similar levels of sIL-1Ra production were obtained in monocytes mock-transfected or transfected with control or STAT1-specific siRNA. To ascertain that STAT1 silencing affected genes, which expression depends on STAT1 activation, STAT1-knocked down monocytes were activated by either IFN-β or IFN-γ, and the expression of FcγR1 transcript was measured by real-time PCR. The activation of monocytes by either IFN-β or IFN-γ enhanced ∼10 times the basal expression of STAT1 transcript whether cells were transfected or not with STAT1 siRNA (Fig. 6,D). However, in monocytes knocked down for STAT1, the latter was inhibited by ∼80% independently of the stimulus. STAT1 silencing was effective, because of the expression of FcγR1 being significantly inhibited by 76 ± 3% and 38 ± 4% in STAT1 siRNA-transfected monocytes activated by IFN-β and IFN-γ, respectively, as compared with mock-transfected cells (Fig. 6,D). Noticeably, IFN-γ was 3 times more efficient than IFN-β in inducing FcγR1 transcript expression (Fig. 6,D). This demonstrates that STAT1 silencing affected the expression of STAT1-dependent genes, as exemplified by FcγR1. The difference in inhibition percentage between IFN-β- and IFN-γ-activated monocytes might be due to the premise that FcγR1 induction by IFN-γ depends on both STAT1 and PU.1 activation (39). IFN-γ slightly inhibited the basal expression of sIL-1Ra transcript (Fig. 6,D), but did not induce sIL-1Ra protein production (data not shown). sIL-1Ra mRNA was not affected by STAT1 knockdown in either IFN-β- or IFN-γ-activated monocytes (Fig. 6,D), confirming the results obtained at the protein level (Fig. 6 C). Together these results demonstrate that STAT1 activation was not required for the IFN-β induction of sIL-1Ra production in human monocytes.
Discussion
The major finding of this study is that in human monocytes treated by IFN-β, sIL-1Ra production is controlled by PI3K through a STAT1-independent pathway. Like PI3K, MEK1 activation contributes to Ser727 phosphorylation of STAT1. However, MEK1 does not regulate sIL-1Ra production. PI3K controls the induction of sIL-1Ra gene transcription, which is an immediate early gene triggered by IFN-β in human monocytes. This differs from data obtained on so-called septic leukocytes. Indeed, in the latter cells, PI3K regulates efficient translation of sIL-1Ra residual transcripts induced by LPS, without affecting its transcription (33). Thus, depending on the stage of monocyte maturation and/or type of stimulation, PI3K may regulate sIL-1Ra production at different levels, i.e., at the transcriptional level in monocytes and at the translational level in LPS-treated macrophages (33). The experiments conducted with actinomycin D in this study reveal that PI3K regulated the transcription of sIL-1Ra, but was not involved in the stabilization of sIL-1Ra mRNA, further confirming that different signal transduction pathways might be engaged depending on the cell type, i.e., monocytes or septic leukocytes, representing different pathophysiological conditions.
Experiments conducted with cycloheximide showed a superinduction of sIL-1Ra transcript. The superinduction of cytokine mRNA in monocytes-macrophages has been described previously (40, 41). However, in the latter studies, this phenomenon was described for proinflammatory cytokines, i.e., IL-1 and TNF; to our knowledge, this is the first time that this is observed for sIL-1Ra. There is no clear-cut explanation for this phenomenon, although it was hypothesized that superinduction of immediate early genes by protein synthesis inhibitors might occur via at least three ways, including: mRNA stabilization; activation of intracellular signaling cascades; and interference with transcriptional down-regulation (42).
It has been shown that PI3K may be cross-activated upon IFNAR-1 engagement by IFN-α (43), which triggers the activation of the IRS signaling system. Indeed, in hemopoietic cells, IFN-α triggers Tyr phosphorylation of IRS-1 and IRS-2 and their subsequent association with the Src homology 2-containing p85 regulatory subunit of PI3K (26). In Daudi cells, STAT3 rather than IRS fulfills the function of adapter protein between IFNAR-1 and PI3K (25), although experiments conducted in Tyk2-null cells demonstrate a direct interaction between PI3K and IFNAR-1 (27). From the present study, it appears that similarly to IFN-α, IFN-β triggers the activation of PI3K, leading to the production of sIL-1Ra, which pathway does not require STAT1 activation. The identity of the adapter protein (if any) coupling IFNAR-1 to PI3K in human monocytes stimulated by IFN-β remains to be determined. The possibility of STAT3 involvement was investigated by using JSI-124, a so-called STAT3-specific inhibitor (44). JSI-124 inhibited IFN-β-induced sIL-1Ra production, but failed to inhibit STAT3 phosphorylation (data not shown), suggesting that in human monocytes, the latter inhibitor affected another mechanism.
In accordance with previous studies (20, 23), IFN-β induces a sustained Ser727 phosphorylation of STAT1 that peaks at 30–60 min and persists up to 2 h. The PI3K inhibitor, Ly294002, inhibited both Ser727 phosphorylation of STAT1 and sIL-1Ra production, suggesting that both mechanisms are regulated by PI3K. However, the two events were not related. Indeed, although PD98059, a MEK1 inhibitor, and genistein, an unspecific tyrosine kinase inhibitor, inhibited Tyr701 and Ser727 phosphorylation of STAT1, only genistein inhibited sIL-1Ra production. Consequently, STAT1 was not involved in the IFN-β-triggered signaling pathway leading to sIL-1Ra production. This was also suggested by EMSA, in which Ly294002 did not inhibit STAT1 translocation to the nucleus and binding to GAS consensus sequence, in contrast with genistein. The inhibitory effect of genistein on STAT1 translocation and phosphorylation might be due to the inhibition of Tyk2, which in turn might hamper PI3K activation (45). This was reminiscent of previous studies that demonstrated that although Tyr701 phosphorylation is essential for nuclear translocation and DNA binding, additional phosphorylation of STAT1 at Ser727 generates maximal activation of transcription (46, 47). It was previously demonstrated that the MEK1 substrate, ERK2, is implicated in Ser727 phosphorylation of STAT1 (28, 48); in this study, we show that in addition to MEK1, PI3K also controls Ser727-STAT1 phosphorylation. The lack of involvement of STAT1 in sIL-1Ra induction by IFN-β was strongly established by the premise that sIL-1Ra production was fully induced in human monocytes that were knocked down for STAT1. Thus, in addition to JAK/STAT, IFN-β signals through other pathway(s) and particularly through PI3K-dependent pathways that lead to the expression of genes such as sIL-1Ra. These data are consistent with a recent study demonstrating that a PI3K-dependent, STAT1-independent signaling pathway regulates IFN-stimulated human monocyte adhesion (49). However, in the latter study, the evidence supported the PI3K-dependent, STAT1-independent adhesion of monocytes was obtained with bone marrow-derived macrophages isolated from STAT1−/− mice. In the present study, this is demonstrated directly in human monocytes that were knocked down for STAT1.
It was recently demonstrated that the engagement of TLR4 in mouse macrophages induced the phosphorylation of STAT1 in both Ser727 and Tyr701 residues (50). However, in contrast with the present results, the phosphorylation of Ser727 precedes that of Tyr701, which occurs only after 4-h stimulation by LPS. In the latter condition, PI3K markedly contributed to Tyr701-STAT1, but not to Ser727-STAT1 phosphorylation. Because part of TLR4 signaling occurs through the induction of an autocrine loop of IFN-β, it is possible that differences in signaling exist between monocytes and macrophages and/or between human and murine cells. However, the results obtained in the present study in freshly isolated monocytes are in accordance with previous observations in human monocyte-derived macrophages. Indeed, PI3K, but not the MAPK, ERK1/ERK2, proved to be involved in the induction of the anti-inflammatory cytokine IL-10 and the repression of TNF (51, 52). Together with our results, the latter studies demonstrate that PI3K plays an important part in the induction of anti-inflammatory cytokines such as IL-10 and sIL-1Ra in cells of the monocytic lineage. This contrasts with the requirement of PI3K in the chemokine-dependent migration of neutrophils and macrophages, i.e., a proinflammatory process, and should be taken into account when considering PI3K as a pharmaceutical target in inflammation (53).
In conclusion, this study demonstrates that PI3K displays a key regulatory function in sIL-1Ra production in human monocytes by regulating its transcription. Because sIL-1Ra is an important anti-inflammatory molecule, insight into the signaling mechanisms that control its production might contribute to the development of agents specifically interfering with inflammatory processes.
Disclosures
The authors have no financial conflict of interest.
Acknowledgements
We are indebted to Dr. W. Schlegel for discussion and scientific advice, and thank R. Rehm for skillful reading of the manuscript.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by Grant 3200-068286.02 from the Swiss National Science Foundation and a grant from the Swiss Society for Multiple Sclerosis. The authors have no financial conflict of interest.
Abbreviations used in this paper: IL-1Ra, IL-1R antagonist; GAS, IFN-γ activation site; IFNAR, type I IFN receptor complex; IRS, insulin receptor substrate; sIL-1Ra, secreted IL-1Ra; siRNA, small interfering RNA.