The ability of IFN-β to induce IL-10 production from innate immune cells is important for its anti-inflammatory properties and is believed to contribute to its therapeutic value in treating multiple sclerosis patients. In this study, we identified that IFN-β stimulates IL-10 production by activating the JAK1- and PI3K-signaling pathways. JAK1 activity was required for IFN-β to activate PI3K and Akt1 that resulted in repression of glycogen synthase kinase 3 (GSK3)-β activity. IFN-β–mediated suppression of GSK3-β promoted IL-10, because IL-10 production by IFN-β–stimulated dendritic cells (DC) expressing an active GSK3-β knockin was severely reduced, whereas pharmacological or genetic inhibition of GSK3-β augmented IL-10 production. IFN-β increased the phosphorylated levels of CREB and STAT3 but only CREB levels were affected by PI3K. Also, a knockdown in CREB, but not STAT3, affected the capacity of IFN-β to induce IL-10 from DC. IL-10 production by IFN-β–stimulated DC was shown to suppress IFN-γ and IL-17 production by myelin oligodendrocyte glycoprotein-specific CD4+ T cells, and this IL-10–dependent anti-inflammatory effect was enhanced by directly targeting GSK3 in DC. These findings highlight how IFN-β induces IL-10 production and the importance that IL-10 plays in its anti-inflammatory properties, as well as identify a therapeutic target that could be used to increase the IL-10–dependent anti-inflammatory properties of IFN-β.
Type I IFNs (IFN-α subtypes, β, ε, κ, ω, δ, and τ) are a group of cytokines that exhibit a diverse number of biological effects. Originally identified as having antiviral properties, type I IFNs were subsequently shown to possess antiproliferative, antitumor, antiangiogenic, and anti-inflammatory effects (1, 2). Type I IFNs were also demonstrated to exhibit a multitude of immunomodulatory effects on cells of the innate and adaptive immune systems (1, 3). The biological properties of type I IFNs are largely due to their ability to transduce cellular signals from the type I IFNR complex to the nucleus via the activation of JAKs and STATs (4). Specifically, binding of type I IFNs to the IFNR complex, consisting of two transmembrane chains (IFNAR-1 and IFNAR-2), results in the activation of two receptor-associated tyrosine kinases of the JAK family TYK2 and JAK1 (5). In turn, activated JAK1 and TYK2 phosphorylate and activate downstream members of the STAT family, including STAT1, STAT2, and STAT3 (3, 6). The subsequent homo- or heterodimerization of STATs results in their ability to translocate into the nucleus where they bind IFN-γ–activated site regulatory elements and IFN-stimulated response element sites that initiate transcription of IFN-stimulated genes (3, 7).
Signaling via the type I IFNR complex results in the activation of multiple intracellular-signaling pathways, including activation of the PI3K pathway (8, 9). PI3K is a heterodimeric molecule consisting of an 85-kDa regulatory subunit that contains Src homology 2-dependent domains capable of binding phospho-tyrosine motifs (YXXM or YXXW) and a 110-kDa catalytic domain (10). PI3K possesses serine and lipid kinase activity that allows it to generate phosphatidylinositol 3, 4, 5 triphosphate that mediates the recruitment of signaling molecules with pleckstrin homology domains, as well as phosphorylate downstream-signaling molecules on serine residues (10). Several laboratories reported mechanisms that are involved in the ligand-dependent/independent recruitment of PI3K to the type I IFNR complex and its subsequent activation (9, 11, 12). Although these studies suggested that the underlying cellular mechanisms responsible for the recruitment and activation of PI3K differ depending upon cell types, collectively these studies highlighted the importance that the PI3K pathway plays in mediating several of the biological properties of type I IFNs (12, 13). The ability of type I IFNs to activate PI3K was shown to be important for the induction of IFN-inducible genes, activation of NF-κB, regulation of apoptosis and cell survival, and mRNA translation via its capacity to regulate the mammalian target of rapamycin/p70 S6 kinase pathway (12–15).
IFN-β was shown to exhibit anti-inflammatory properties. Studies by Chang et al. (16) showed that IFN-β production from LPS-stimulated macrophages promoted the induction of the anti-inflammatory cytokine IL-10 and that the levels of IL-10 were dependent upon the ability of IFN-β to activate the type I IFNR complex. In the absence of type I IFN production or type I IFN signaling, LPS-stimulated mouse macrophages produced increased proinflammatory cytokines (16). IFN-β was also reported to augment the levels of IL-10 from LPS-stimulated human PBMCs that, in turn, suppressed the production of the proinflammatory cytokine IL-12 via an IL-10–dependent mechanism (17). Studies using human monocytes extended these findings by identifying that type I IFNs inhibit IL-12 p40 production at the transcriptional level by decreasing the level of PU.1-binding activity at the upstream Ets site within the IL-12 p40 promoter (18). Within a clinical setting, type I IFNs have been used for the treatment of relapsing–remitting multiple sclerosis (MS) (19) and were shown to increase the serum levels of IL-10 in MS patients (20). Although the precise therapeutic mechanism of action of IFN-β in MS treatment is unclear, the ability of IFN-β to augment IL-10 is believed to play a fundamental role in its therapeutic value (21). Despite the numerous reports linking IFN-β to the induction of IL-10, the cellular mechanisms regulating this process are unknown.
In the current study, we identified that IFN-β stimulates IL-10 production by activating the JAK1- and PI3K-signaling pathways. JAK1 activity was required for IFN-β to activate PI3K and Akt1, which resulted in repression of glycogen synthase kinase 3 (GSK3)-β activity and subsequently promoted IL-10 production. Moreover, we found that IFN-β increased the phosphorylated levels of CREB and STAT3 but only a knockdown in CREB levels reduced IFN-β–stimulated IL-10 production. IL-10 production by IFN-β–stimulated dendritic cells (DC) suppressed IFN-γ and IL-17 production by myelin oligodendrocyte glycoprotein (MOG) -specific CD4+ T cells; this IL-10–dependent anti-inflammatory effect was enhanced by directly targeting GSK3 in DC. These findings identify how IFN-β induces IL-10 production and the important role that IL-10 plays in its anti-inflammatory properties, as well as identify a therapeutic target that could be used to increase the IL-10–dependent anti-inflammatory properties of IFN-β.
Materials and Methods
Media and reagents
Cells were cultured in RPMI 1640 medium supplemented with 10% FBS (R10), 50 μM 2-ME, 1 mM sodium pyruvate, 2 mM l-glutamine, 20 mM HEPES, 50 U/ml penicillin, and 50 μg/ml streptomycin (RPMI complete). Ultra-pure LPS from Escherichia coli22
Preparation of human DC, Akt1 knockout DC, GSK3-β (S9A) DC, and JAK1-deficient mouse embryonic fibroblasts
PBMCs were obtained from healthy donors as per protocols approved by the University of Louisville, Institutional Review Board, Human Subjects Protection Program, study #503.05. Monocytes were isolated by negative selection using the human monocyte isolation kit II from Miltenyi Biotec. The purity of monocytes was routinely >90%, as determined by flow cytometry using an FITC-labeled anti-CD14 Ab. Human monocyte-derived DC were generated by culturing negatively selected human monocytes (2 × 106/ml) in low-attachment tissue-culture plates containing RPMI complete media, IL-4 (5 ng/ml), and GM-CSF (10 ng/ml) for 7 d. On day 3, 10 ml media was collected, centrifuged, and resuspended in 10 ml fresh RPMI complete containing IL-4 (5 ng/ml) and GM-CSF (10 ng/ml). On day 7, nonadherent cells were collected and positively selected for CD1c expression using the Miltenyi Biotec CD1c cell isolation kit. The purity of isolated DC was typically >95%, as determined by flow cytometry using an FITC-labeled CD1a or FITC-labeled CD1c Ab. Mouse DC were generated from the bone marrow of C57BL/6 mice by culturing bone marrow cells in the presence of 20 ng/ml rGM-CSF for 10 d in RPMI complete, as previously described (23). Analysis of the resulting cell population by flow cytometry demonstrated that >85% of the cells stained positive for CD11c expression. Wild-type (WT) and JAK1-deficient mouse embryonic fibroblasts (MEFs) were obtained from Dr. Robert D. Schreiber (Washington University, St. Louis, MO). Akt1 knockout mice were obtained from Dr. Morris J. Birnbaum and were characterized previously (24). GSK3 (S21A/S9A) knockin mice were generated by McManus et al. (25).
Measurement of GSK3-β kinase activity
DC (4 × 106) were plated in six-well plates and stimulated with IFN-β for 30 or 60 min. Cell lysates were prepared using immunoprecipitation-compatible lysis buffer, and protein concentrations were determined using the bicinchoninic acid assay (Pierce). GSK3-β in cell lysates (300 μg protein per sample) was immunoprecipitated using GSK3-β–specific Ab (G6414; Sigma-Aldrich) and EZview Red Protein G Affinity Gel Beads for 24 h at 4°C. The immunoprecipitates were subjected to an in vitro kinase assay using 5 μl the GSK3-specific substrate peptide and 2.5 μl γ-[32P]ATP (∼3000 Ci/mmole, 10 mCi/ml) for 30 min at 37°C. Twenty microliters of the assay mixture was spotted on a P81 cellulose phosphate square and then washed sequentially in phosphoric acid solution, ethanol, and acetone. The cellulose phosphate squares were dried, and cpm were collected using a liquid scintillation counter.
Transfection, Western blot, and IL-10 production
DC were transfected with nontargeting control siRNA, siRNA–GSK3-β, siRNA-CREB, siRNA-STAT3, pcDNA3–GSK3-β (S9A), pcDNA3–GSK3-β (K85A), or pcDNA3 (empty vector control) using Lipofectamine RNAiMAX or Lipofectamine LTX, following the manufacturer’s protocol (Invitrogen). Cell lysates were prepared as previously described (26–28). Images were acquired using the Kodak Image Station 4000MM system (Eastman Kodak, New Haven, CT). For siRNA studies, the levels of total GSK3-β, CREB, STAT3, and β-actin were assessed by Western blot on day 3. Control cells were pretreated for 2 h with 0.01% DMSO (organic solvent control for SB216763 or LY294002). For siRNA studies, transfected DC were stimulated with IFN-β (1000 U/ml) 3 d posttransfection, and cell-free supernatants were assayed for IL-10 levels by ELISA.
Generation of MOG33–55-specific CD4+ T cells
C57BL/6 mice (males, 6–8 wk of age) were immunized s.c. with 200 μg MOG35–55 peptide (New England Peptide) in CFA. Spleens were isolated 2 wk after immunization, and CD4+ T cells were purified using the CD4++ T cells isolated from immunized mice were harvested, plated at 2 × 106/well, and cultured with 5 × 105/well gamma-irradiated splenic cells in cell-free supernatants isolated from nonstimulated DC, IFN-β–stimulated DC, or DC stimulated with IFN-β (500 U/ml) in the presence of GSK3 inhibition. For the generation of cell-free supernatants, 2 × 105 DC/well, 4 × 105 DC/well, or 2 × 106 DC/well were cultured in a final volume of 200 μl and stimulated with 500 U/ml IFN-β for 20 h. Nonstimulated DC cultures included SB216763 (10 μM) to ensure that no carryover effect was involved in affecting the cytokine production by MOG33–55-specific CD4+ T cells. Cultures were stimulated for 72 h with or without 10 μg/ml MOG33–55 peptide. Neutralizing mouse anti–IL-10 Ab (clone JES5-2A5) or rat IgG1 isotype control (IC) Ab were preincubated with supernatants for 1 h at 37°C and used at 5 μg/ml.
RNA extraction and first-strand cDNA synthesis were performed using the 5 Prime PerfectPure RNA Cultured Cell Kit and High-Capacity cDNA Archive kit (Applied Biosystems). Real-time PCR was performed using an ABI 7500 system. GAPDH mRNA levels were determined for each time point and used as the endogenous control. Fold increase was calculated according to the comparative cycle threshold method (29).
Cells were plated at a density of 2.5 × 105 cells/well in a 96-well flat-bottom plate. The intracellular phospho–GSK3-β (S9) levels were detected by flow cytometry, as previously described (30
Statistical significance between groups was evaluated by ANOVA and the Tukey multiple-comparison test using the InStat program (GraphPad, San Diego, CA). Differences between groups were considered significant at p values <0.05.
IFN-β augments IL-10 production from LPS-stimulated human DC
IFN-β was shown to mediate anti-inflammatory effects on the immune system (16–18, 31). Chang et al. (16) demonstrated that the production of IFN-β by LPS-stimulated bone marrow-derived macrophages played a fundamental role in their ability to produce IL-10. Thus, we initially determined whether IFN-β mediated a similar effect in human DC. DC stimulated with LPS in the presence of an anti–IFN-β Ab produced significantly (p < 0.01) less IL-10 compared with cells stimulated with LPS in the presence of an IC Ab (Fig. 1A). DC stimulated with LPS in the presence of an IC exhibited >2-fold increase in the mean fluorescent intensity levels of IL-10 compared with DC stimulated with LPS in the presence of an anti–IFN-β Ab (Fig. 1B). Next, we determined whether IFN-β was capable of inducing IL-10 from DC in the absence of LPS. Compared with untreated cells, the addition of IFN-β significantly (p < 0.01) enhanced IL-10 production at all concentrations tested (Fig. 1C). No detectable effect was observed on the levels of other proinflammatory cytokines produced by IFN-β–stimulated human DC, including TNF, IL-12 p40, IL-1β, or IL-23 (Fig. 1D–G). These results demonstrated that the LPS-mediated production of IFN-β, as well as IFN-β alone, can induce IL-10 production from DC.
JAK1-dependent activation of the PI3K pathway
Binding of type I IFNs to the IFNR complex results in the activation of two receptor-associated tyrosine kinases: TYK2 and JAK1 (5, 9, 31). The activation of these tyrosine kinases results in the initiation of a signaling cascade involving STAT1, STAT2, and STAT3 (3, 6). Because JAK1 activity was shown to be critically involved in many aspects of type I IFN signaling (32, 33), we initially wanted to determine whether JAK1 was involved in the ability of IFN-β to induce IL-10 from DC. As shown in Fig. 2A, inhibition of JAK in IFN-β–stimulated human DC reduced the levels of phosphorylated JAK1. In addition, inhibition of JAK significantly (p < 0.01) decreased the levels of IL-10 produced by IFN-β–stimulated human DC (Fig. 2B). Because our findings demonstrated an important role for JAK1 in regulating IL-10, we wanted to identify the downstream targets by which JAK1 was affecting the ability of IFN-β to induce IL-10 from DC. For these studies, we initially focused on the STAT and PI3K/Akt pathways, because of their importance in regulating IL-10, and the findings by several groups demonstrating that JAK1 can affect both of these signaling pathways (34–37). As shown in Fig. 2C, JAK inhibition potently decreased the levels of serine or tyrosine phosphorylation of STAT1, STAT2, and STAT3 in human DC stimulated with IFN-β. Furthermore, the phosphorylated levels of Akt in IFN-β–stimulated JAK1-deficient MEFs (Fig. 2D) or IFN-β–stimulated human DC (Fig. 2E) pretreated with a JAK inhibitor were also attenuated compared with control cells stimulated with IFN-β. Because activated Akt was shown to phosphorylate and inactivate the constitutively active serine/threonine kinase GSK3-β, we next determined whether JAK1 affected the levels of phosphorylated GSK3-β. Similar to the effects of JAK1 inhibition on the levels of phosphorylated Akt, IFN-β stimulation of JAK1-deficient MEFs or human DC treated with a JAK inhibitor did not induce any discernible increase in the levels of GSK3-β (S9) compared with nonstimulated control cells (Fig. 2D, 2E).
We next investigated whether the IFN-β–mediated phosphorylation of Akt and GSK3-β were dependent or independent of PI3K activity. For this, human DC were stimulated with IFN-β in the presence or absence of the PI3K inhibitor LY294002, and the levels of phosphorylated Akt and phosphorylated GSK3-β were determined by Western blot. As shown in Fig. 3A and 3B, IFN-β–stimulated DC exhibited increased phosphorylated levels of Akt and GSK3-β compared with nonstimulated cells. Inhibition of PI3K abrogated the ability of IFN-β to induce the phosphorylation of Akt and GSK3-β (Fig. 3A, 3B). We also examined the effects of PI3K inhibition on the levels of phosphorylated GSK3-β in IFN-β–stimulated DC at the single-cell level by flow cytometry. Stimulation of cells with IFN-β increased the percentage of DC expressing GSK3-β from 0.63% in the nonstimulated control cells to >31% in IFN-β–stimulated DC (Fig. 3C). Furthermore, inhibition of PI3K abrogated the ability of IFN-β to increase the levels of GSK3-β phosphorylation (Fig. 3C).
Our current data demonstrated that activation of the PI3K pathway in IFN-β–stimulated cells was dependent upon JAK1 (Fig. 2A, 2B). Because previous reports documented the importance of the PI3K pathway in regulating the JAK/STAT pathway (8, 11, 38), we next investigated whether the ability of IFN-β to activate the PI3K pathway affected the activation/phosphorylation of STATs. Compared with nonstimulated cells, IFN-β–stimulated human DC exhibited increased levels of phosphorylated STAT1, STAT2, and STAT3 (Fig. 3D), but inhibition of PI3K had no observable effect on the phosphorylated levels of STAT1, STAT2, or STAT3. These findings are in contrast with those observed upon JAK inhibition in that the inhibition of JAK abrogated the capacity of IFN-β to activate STAT1, STAT2, or STAT3 (Fig. 2C). Taken together, these results demonstrated that JAK1 is involved in the activation of the PI3K and STAT pathways in IFN-β–stimulated DC, but PI3K does not affect the ability of IFN-β to activate STAT1, STAT2, or STAT3.
IFN-β–induced IL-10 production by DC is dependent on PI3K activity
The PI3K pathway was shown to play a pivotal role in the regulation of the host inflammatory response (39–42). Because our data showed that inhibition of PI3K decreased the IFN-β–induced phosphorylation of two downstream targets of the PI3K pathway Akt and GSK3-β (Fig. 3A–C), we next investigated the functional role that this pathway was playing in the ability of IFN-β to induce IL-10 from DC. As shown in Fig. 4A, inhibition of PI3K significantly (p < 0.001) decreased the levels of IL-10 produced by IFN-β–stimulated human DC. Additionally, IFN-β stimulation increased the mean fluorescent intensity levels of human DC expressing IL-10 compared with nonstimulated DC (Fig. 4B). In sharp contrast, PI3K inhibition potently reduced the capacity of IFN-β–stimulated DC to produce IL-10 (Fig. 4B).
Because the current study showed that inhibition of PI3K negatively affected the ability of IFN-β to stimulate IL-10 production from DC and resulted in the loss of phosphorylated Akt, we next determined whether Akt1 played a functional role in the ability of IFN-β–stimulated DC to produce IL-10. As shown in Fig. 4C, Akt1-deficient mouse DC (mDC) stimulated with IFN-β produced significantly less (p < 0.001) IL-10 compared with WT mDC stimulated with IFN-β. No significant differences (p > 0.05) were observed between the levels of IL-10 produced by nonstimulated Akt1-deficient mDC or IFN-β–stimulated Akt1-deficient mDC (Fig. 4C). However, direct inhibition of GSK3 in Akt1-deficient mDC significantly (p < 0.01) augmented the levels of IL-10 produced by Akt1-deficient cells stimulated with IFN-β (Fig. 4C). Taken together, these data demonstrated that Akt1 is required for IFN-β to induce IL-10 from mDC.
To this point, our findings demonstrated that the activation of the PI3K/Akt pathway was involved in the ability of IFN-β to stimulate IL-10 production from DC. Moreover, inhibition of GSK3 largely rescued the defective IL-10 production by IFN-β–stimulated mDC lacking Akt1 (Fig. 4C). Thus, we next determined whether the ability of the PI3K pathway to mediate the phosphorylation and inactivation of GSK3-β (S9) affected the levels of IL-10 produced by IFN-β–stimulated DC. An initial assessment of the IL-10 mRNA and protein levels demonstrated that the inhibition of GSK3 significantly (p < 0.01) enhanced the steady-state levels of IL-10 mRNA and the amount of IL-10 protein produced by IFN-β–stimulated DC compared with cells treated with IFN-β alone (Fig. 4D, 4E). GSK3 exists as two major isoforms (GSK3-α and GSK3-β) (43). However, as shown in Fig. 3B using a dual-specific Ab capable of detecting phosphorylated GSK3-α (S21) and GSK3-β (S9), only phosphorylated levels of GSK3-β (S9) were detected in IFN-β–stimulated DC. In addition, a knockdown in the cellular levels of GSK3-β protein significantly (p < 0.01) increased IL-10 production by human DC stimulated with IFN-β compared with control cells stimulated with IFN-β (Fig. 4F). To more directly determine whether GSK3-β activity was regulating IL-10 production by IFN-β–stimulated DC, we next used a kinase dead and constitutively active form of GSK3. IFN-β stimulation of DC expressing a kinase dead mutant of GSK3-β (K85A) also significantly (p < 0.001) enhanced the levels of IL-10 compared with empty vector-transfected cells stimulated with IFN-β (Fig. 4G). In contrast, the ectopic expression of a constitutively active GSK3-β (S9A) mutant severely attenuated (p < 0.001) the IL-10 produced by DC compared with empty vector-transfected cells stimulated with IFN-β (Fig. 4G). Because ectopic-expression studies can give variable results due to differences in construct-expression levels, we also determined how an inability of IFN-β to induce the phosphorylation and inactivation of GSK3 affected IL-10 production in mDC expressing a constitutively active knockin of GSK3 (25). IFN-β–stimulated mDC expressing a constitutively active knockin of GSK3-β (S9A) produced significantly (p < 0.001) less IL-10 compared with WT control DC stimulated with IFN-β (Fig. 4H). Taken together, these results demonstrated that the repression of GSK3-β activity in IFN-β–stimulated DC plays an important role in their ability to produce IL-10.
IFN-β induces partial inactivation of GSK3-β activity
The data from the current study demonstrated that IFN-β stimulation of DC enhanced the phosphorylated levels of GSK3-β (S9) and that preventing the phospho-inactivation of GSK3-β (S9) suppressed the ability of IFN-β to induce IL-10 production by DC. Our findings obtained from the pharmacological inhibition of GSK3, siRNA-mediated knockdown of GSK3-β, as well as the ectopic expression of a kinase dead GSK3-β demonstrated that these techniques augmented the levels of IL-10 produced by IFN-β–stimulated DC compared with DC stimulated with IFN-β alone. Taken together, these findings suggested that the differences in IL-10 production among these groups were likely due to the relative inactivation of GSK3-β activity by IFN-β. To test this possibility, we initially monitored the activity of GSK3 via its ability to phosphorylate the GSK3 substrate glycogen synthase (GS) (44, 45) in IFN-β–stimulated human DC. Compared with nonstimulated DC, DC stimulated in the presence of IFN-β exhibited a 42% reduction in the phosphorylated levels of GS (Fig. 5A, 5B). However, stimulation of DC with IFN-β in the presence of the GSK3 inhibitor SB216763 resulted in a near 100% reduction in the phosphorylated levels of GS compared with nonstimulated controls (Fig. 5A, 5B). Next, we used a GSK3-β kinase activity assay kit to directly measure GSK3-β kinase activity by monitoring the ability of immunoprecipitated GSK3-β from IFN-β–stimulated human DC to incorporate γ-[32P]-ATP into a GSK3-specific substrate peptide. Compared with nonstimulated DC, IFN-β–stimulated DC exhibited a significant (p < 0.05) reduction in GSK3-β kinase activity at all time points tested (Fig. 5C). In addition, blockade of PI3K using LY294002 resulted in abrogation of the capacity of IFN-β to reduce GSK3-β activity (Fig. 5C). GSK3-β activity in IFN-β–stimulated DC pretreated with a GSK3-specific inhibitor was similar to the level observed in the IC Ab group (Fig. 5C). Thus, although IFN-β–stimulated DC exhibited a reduction in GSK3 activity, IFN-β stimulation alone did not completely abolish GSK3 activity, as was noted with the GSK3 inhibitor SB216763. Therefore, our data demonstrating that the IL-10 levels produced by IFN-β–stimulated DC expressing a constitutively active GSK3 knockin were severely reduced, whereas direct GSK3 inhibition augmented IL-10, highlight the importance of GSK3 activity in controlling IL-10 production by IFN-β–stimulated DC. Moreover, these data directly showed that stimulation of DC with IFN-β reduced but did not abrogate GSK3 activity via the PI3K pathway.
GSK3-β regulates IFN-β–induced IL-10 by the transcription factor CREB
We next determined how GSK3 was controlling the levels of IL-10 produced by IFN-β–stimulated DC. The transcriptional activity of CREB and STAT3 can be regulated by GSK3 (46, 47), and CREB and STAT3 were reported to affect the transcriptional regulation of the IL-10 gene (34, 48). However, based on our current data that showed PI3K inhibition did not affect total or phosphorylated STAT3 levels (Fig. 3D), we initially assessed whether IFN-β–mediated inhibition of GSK3 affected CREB. Compared with nonstimulated control human DC, neither the GSK3 inhibitor SB216763 nor PI3K inhibitor LY294002 substantially affected the levels of nuclear phosphorylated CREB (S133) in nonstimulated cells (Fig. 5D). In contrast, the nuclear levels of CREB (S133) were significantly (p < 0.001) enhanced in IFN-β–stimulated DC compared with nonstimulated control cells (Fig. 5D). Moreover, cells pretreated with the GSK3 inhibitor SB216763 significantly (p < 0.01) enhanced the nuclear levels of CREB (S133), whereas blockade of PI3K activity abrogated (p < 0.001) this increase in CREB levels compared with DC stimulated with IFN-β alone (Fig. 5D).
To determine whether the increased levels of CREB played a functional role in the ability of IFN-β to induce IL-10, CREB was knocked down in human DC (Fig. 5E). CREB-specific siRNA reduced the cellular levels of CREB by >90% compared with nontransfected cells or cells transfected with control pools of siRNA (Fig. 5E). The knockdown in CREB levels potently (p < 0.01) reduced the levels of IL-10 produced by IFN-β–stimulated DC compared with control siRNA-transfected DC stimulated with IFN-β (Fig. 5E). Moreover, CREB knockdown significantly (p < 0.001) reduced the ability of GSK3-inhibited DC to produce IL-10 in the presence of IFN-β compared with GSK3-inhibited control DC stimulated with IFN-β (Fig. 5E). Although we did not observe any discernible effect on the activity of STAT3 in PI3K-inhibited DC (Fig. 3D), stimulation of cells with IFN-β increased the phosphorylation of STAT3 (Fig. 3D). Therefore, our findings concerning that only CREB activity was affected by PI3K does not rule out the possibility that STAT3 affects IL-10 production independently of PI3K. Therefore, we also assessed whether STAT3 knockdown affected the levels of IL-10 produced by IFN-β–stimulated DC. As shown in Fig. 5F, a knockdown in the cellular levels of STAT3 did not significantly (p > 0.05) affect the levels of IL-10 produced by IFN-β–stimulated DC. These findings demonstrate that the ability of IFN-β to augment the nuclear levels of CREB via the suppression of GSK3 activity plays a fundamental role in the ability of IFN-β to promote IL-10 production from DC.
The IL-10–dependent suppressive capacity of IFN-β–stimulated mDC on IL-17 and IFN-γ responses by MOG33–55-specific CD4+ T cells is augmented by direct inhibition of GSK3 in DC
IL-10 is an anti-inflammatory cytokine that can suppress IFN-γ and IL-17 production by CD4+ T cells. IFN-β–treated mDC produced IL-10 that could be significantly increased (p < 0.01) upon GSK3 inhibition (Fig. 6A, 6B), and this is consistent with our present findings using human DC. We next determined whether the ability of IFN-β to stimulate mDC to produce IL-10 could suppress IFN-γ and IL-17 production by MOG33–55-specific CD4+ T cells. MOG33–55-specific CD4+ T cells cultured in supernatants derived from IFN-β–stimulated mDC produced significantly (p < 0.05) less IFN-γ and IL-17 compared with MOG33–55-specific CD4+ T cells cultured with supernatants from nonstimulated control mDC (Fig. 6C–F). Moreover, the addition of an anti–IL-10–neutralizing Ab significantly (p < 0.01) reduced the ability of supernatants derived from IFN-β–stimulated mDC to suppress IFN-γ and IL-17 production (Fig. 6E, 6F). We next wanted to define how limiting the number of IFN-β–stimulated DC affected IL-17 and IFN-γ production by MOG33–55-specific CD4+ T cells and whether the suppressive effects of IFN-β–treated DC on CD4+ T cells could be enhanced by GSK3 inhibition. At a DC/CD4+ T cell ratio of 1:5, IFN-β–stimulated DC significantly (p < 0.05) suppressed the levels of IFN-γ produced by MOG33–55-specific CD4+ T cells compared with control cells treated with medium alone (Fig. 6G). At a DC/CD4+ T cell ratio of 1:5 or 1:10, supernatants derived from IFN-β–stimulated DC pretreated with GSK3 inhibitor exhibited a significant (p < 0.001) reduction in the levels of IFN-γ and IL-17 compared with supernatant from DC treated with IFN-β alone (Fig. 6G, 6H). Moreover, the enhanced suppressive effects of GSK3-inhibited DC stimulated with IFN-β were largely IL-10 dependent (Fig. 6G, 6H). Collectively, the findings from the present study identified and characterized the cell-signaling pathway by which IFN-β induced IL-10 production from DCs and also demonstrated an important role for the IL-10–dependent anti-inflammatory properties of IFN-β (Fig. 7).
Past studies showed that type I IFNs exert a multitude of anti-inflammatory properties (31). Wang et al. (17) reported that IFN-β exhibited a differential effect on the levels of pro- and anti-inflammatory cytokine production by LPS-stimulated human PBMCs in which the levels of IL-12 were reduced, whereas IL-10 production was increased. IFNAR-deficient mice lacking the ability to respond to type I IFNs were also documented to exhibit a potent proinflammatory cytokine response upon LPS administration (16). In sharp contrast, the analysis of anti-inflammatory cytokine production in these mice revealed a major defect in their ability to produce IL-10. These findings by Chang et al. (16) further showed that the elevated inflammatory response in IFNAR-deficient mice was the direct result of their defective IL-10 production. A similar cytokine phenotype, in which there was a distinct regulation of pro- and anti-inflammatory cytokines, was reported with GSK3 (36, 41). Several groups showed that GSK3 plays a central role in determining the magnitude and nature of the inflammatory response via its ability to suppress proinflammatory cytokine production while concurrently augmenting the production of IL-10 (36, 41). The findings from the current study demonstrated that the ability of IFN-β to suppress GSK3-β was a critical intracellular signaling event required for IFN-β to induce IL-10 from DC. Although it is not known how IFN-β differentially affects the levels of IL-10 and IL-12, it is tempting to speculate that GSK3 is involved in this effect. Because GSK3 can negatively affect multiple mechanisms regulating IL-12, including the induction of IL-10, the displacement of NF-κB from CBP, and the nuclear enhancement of CREB levels, it will be interesting to decipher whether these cellular processes play a role in the ability of IFN-β to affect IL-12 levels.
Our present findings demonstrated that JAK1 activity was required for the activation of STAT1, STAT2, and STAT3, as well as the activation of the PI3K pathway. In contrast, no effects were observed in the ability of the PI3K pathway to influence the phosphorylated levels of JAK1, STAT1, STAT2, or STAT3. Thus, it seems that the signaling pathway mediated via JAK1 is critical for the downstream activation of the STAT and PI3K pathways in IFN-β–stimulated DC. The ability of JAK1 to mediate activation of the PI3K/Akt pathway was shown to play a critical role in the ability of type I IFNs to induce IL-10 production from human DC via their capacity to repress GSK3 activity (Fig. 7). Although the focus of the current study was to delineate the functional role that the PI3K pathway played in regulating IL-10 production from innate immune cells stimulated with IFN-β, the capacity of IFN-β to activate STAT3 was also reported to have a fundamental role in type I IFN-mediated IL-10 production. Studies by Ziegler-Heitbrock et al. (34) demonstrated that STAT3 played a major role in the capacity of IFN-α–stimulated monocytes to produce IL-10. In contrast, our current findings showed that siRNA-mediated knockdown of STAT3 levels did not affect the ability of IFN-β–stimulated DC to produce IL-10. However, our laboratory demonstrated that a knockdown of the cellular levels of STAT3 reduced the IL-10 levels produced by IFN-α–stimulated DC (H. Wang and M. Martin, unpublished observations). Although the reason for the ability of STAT3 to differentially influence IL-10 production from IFN-α–stimulated DC, but not IFN-β–stimulated DC, is not clear, several studies identified that the intracellular signaling pathways activated by these type I IFNs can diverge (12, 33). Future studies will be needed to delineate the intracellular signaling pathways mediated by type I IFNs and how IFN-α and IFN-β differ with regard to their STAT3 dependency for IL-10 production from DC.
IL-10 is an anti-inflammatory cytokine that plays a pivotal role in controlling excessive inflammation and maintaining immune homeostasis. This is especially evident in IL-10 knockout mice that exhibit highly elevated inflammatory responses to microbial stimulation and develop inflammatory bowel disease (49). IL-10 was shown to suppress innate and adaptive inflammatory responses, and these studies highlighted the potential importance of this cytokine in treating inflammatory diseases. Indeed, because of the reported anti-inflammatory properties of IL-10 on Th1, Th2, and Th17 CD4+ T cells (50–52), identifying the molecular mechanisms controlling IL-10 production has been an area of intense investigation. Our present findings identified and characterized the cellular mechanism by which IFN-β stimulated IL-10 production from DC and demonstrated a functional role for the ability of IFN-β to induce IL-10 (Fig. 7). Specifically, the capacity of IFN-β to repress GSK-3 and increase the levels of IL-10 production by DC mediated IL-10–dependent bystander suppression effects on MOG-specific CD4+ T cell responses, because the inhibitory effects of cell-free supernatants derived from IFN-β–stimulated DC on Th1 and Th17 cytokine production could be reduced by the addition of an IL-10–neutralizing Ab. In addition to our current findings showing IFN-β stimulation of DC induces IL-10, previous studies showed that IFN-β can suppress the levels of Th1- and Th17-promoting cytokines produced by LPS-stimulated DC that, in turn, negatively impacts Th1/Th17 differentiation (53, 54). Moreover, it was shown that IFN-β can directly stimulate DC to produce IL-27 that inhibits the differentiation of Th17 cells (54). Although IFN-β seems to negatively or positively affect pro- and anti-inflammatory cytokines produced by DC that, in turn, impact T cell differentiation and T cell effector responses, IFN-β also mediates anti-inflammatory properties by acting directly on CD4+ T cells. It was recently demonstrated that IFN-β added to CD4+ T cell cultures inhibited the polarization of Th17 cells but increased the ability of CD4+ T cells to produce IL-10 (54). Taken together, these studies highlight several of the immunoregulatory properties exerted by IFN-β on cells of the innate and adaptive immune compartments and aid in understanding the broad impact that IFN-β has on the inflammatory response.
In conclusion, the current study identified the underlying cellular mechanisms by which IFN-β induced the production of IL-10 via its ability to activate the PI3K pathway and repress GSK3-β activity. This process played a critical role in augmenting the nuclear levels of the transcription factor CREB that, in turn, augmented the transcriptional regulation of the anti-inflammatory cytokine IL-10. Because IFN-β treatment was shown to increase the serum levels of IL-10 present in MS patients, and this effect is believed to play an important role in the therapeutic value of type I IFNs, our results indicate that GSK3 could serve as a therapeutic target for modulating the ability of IFN-β to induce IL-10.
We thank Dr. Birnbaum (University of Pennsylvania) for providing Akt1-deficient cells and Dr. Robert D. Schreiber (Washington University) for providing JAK1-deficient MEFs.
Disclosures The authors have no financial conflicts of interest.
This work was supported by Grants R01DE017680 and R01DE017921 from the National Institute of Dental Research (to M.M.).
H.W. and M.M. designed experiments, performed research, analyzed data, and wrote the manuscript; J.B., C.A.G., Y.T., M.R.B., T.G., and D.F.K. performed research and analyzed data. All authors approved the manuscript.
Abbreviations used in this article:
- dendritic cell
- glycogen synthase
- glycogen synthase kinase 3
- isotype control
- mouse dendritic cell
- mouse embryonic fibroblast
- myelin oligodendrocyte glycoprotein
- multiple sclerosis
- small interfering RNA
- Received May 4, 2010.
- Accepted November 4, 2010.