|Home | About | Journals | Submit | Contact Us | Français|
Th2 cells can be subdivided into subpopulations depending on the level of a cytokine and the subsets of cytokines they produce. We have recently identified the ETS family transcription factor PU.1 as regulating heterogeneity in Th2 populations. To define additional factors that might contribute to Th2 heterogeneity, we examined the PU.1 interacting protein IFN-regulatory factor (IRF)-4. When Th2 cells are separated based on levels of IL-10 secretion, IRF4 expression segregates into the subset of Th2 cells expressing high levels of IL-10. Infection of total Th2 cells, and IL-10 non-secreting cells, with retrovirus expressing IRF4, resulted in increased IL-4 and IL-10 expression, no change in IL-5 or IL-13 production and decreased Il9 transcription. Transfection of an IRF4-specific siRNA into Th2 cells decreases IL-10 production. IRF4 directly binds the Il10 gene as evidenced by ChIP assay, and regulates Il10 control elements in a reporter assay. IRF4 interacts with PU.1, and in PU.1-deficient T cells there was an increase in IRF4 binding to the Il10 gene, and in the ability of IRF4 to induce IL-10 production compared to wild type cells and Il10 promoter activity in a reporter assay. Further heterogeneity of IRF4 expression was observed in Th2 cells analyzed for expression of multiple Th2 cytokines. Thus, IRF4 promotes the expression of a subset of Th2 cytokines and contributes to Th2 heterogeneity.
CD4+ T helper (Th) cells play a critical role in modulating both innate and adaptive immune responses. Cytokines direct the differentiation of precursor CD4+ T-helper cells (Thp) into committed Th phenotypes, T-helper-1 (Th1) cells, Th17 cells and Th2 cells which are defined by their function and the array of secreted cytokines (1-3). Th2 cells mediate humoral immune responses by secreting IL-4, IL-5, IL-9, IL-13, and other cytokines including the anti-inflammatory cytokine IL-10. The differentiation of Th2 cells in vivo upon infection or in vitro generates a heterogenous population of Th2 cells that express and secrete various combinations of Th2-type cytokines at different levels (4-6). While Th2 heterogeneity is well documented, the origins are still unclear. Guo et al. reported that the Il4 gene in IL-4-producing and non-producing cells has differential levels of CpG methylation and histone acetylation. These modifications translate in an open DNA conformation and gene activation in IL-4-producing cells (7). We reported that PU.1 plays a critical role in establishing these phenotypes by preventing GATA-3 binding to the IL-4 locus resulting in an IL-4 low Th2 population (8). Knock-down of PU.1 expression resulted in increased homogeneity of Th2 cells (8). In addition to PU.1, it is likely that additional factors are involved in the establishment of Th2 subpopulations.
IRF4 was first identified as a protein recruited by the transcription factor PU.1 to an Ig κ 3′ enhancer in a cooperative manner (9-12) and contributes to the differentiation of Th2, Th17 and some functions of Treg cells (13-15). IRF4 also has distinct effects on cytokine production in naïve versus memory T cell populations (16). In Th2 cells, IRF4 binds to a site adjacent to a NFAT binding element and cooperates with NFATc1, NFATc2 and c-maf in the transcription of IL-4 (17, 18). These interactions are likely important as the activation of T cells triggers a rapid induction of IRF4 (19, 20) and IRF4-deficient T cells secrete decreased levels of Th2 cytokines (13). Ectopic IRF4 has been shown to increase the level of IL-10 secreted by Jurkat cells, but the mechanism by which IRF4 controls the expression of Th2-cytokines in differentiated Th2 cells has not been examined. In this report we determine that IRF4 is a regulator of a subset of Th2 cytokines, and contributes to Th2 heterogeneity.
Wild-type C57BL/6 and BALB/c mice (Harlan Bioscience, Indianapolis, IN) were used for Th1 and Th2 differentiation. Conditional mutant Sfpi1 mice on the C57BL/6 background were previously described (21) and were mated to lck-Cre expressing mice (noted as Sfpi1lck-/-). Mice were maintained in pathogen-free conditions and all studies were approved by the Indiana University School of Medicine Animal Care and Use Committee.
CD4+ T cells were purified from spleen and lymph nodes by positive selection using magnetic beads (Miltenyi Biotech, Auburn, CA) with purity greater than 97% by FACS analysis. Cells were activated with 2 μg/ml plate-bound anti-CD3 (145-2C11) + 1 μg/ml anti-CD28, in addition to 10 ng/ml IL-4 and 10 μg/ml anti-IFNγ (R4/6A2 or XMG) for Th2 differentiation, or 10 ng/ml IL-12 + 10 μg/ml anti-IL-4 (11B11) for Th1 differentiation. After 3 days of incubation, cells were expanded for a total of 5 or 6 days. Differentiated cells were restimulated with 2 μg/ml anti-CD3 at a concentration of 106 cells/ml for real-time PCR and ELISA as previously described (8, 22). Statistics were performed using a t-test with SPSS software.
Total cell extracts were prepared by lysing Th2 cells with lysis buffer (10% glycerol,1% Igepal, 50 mM tris-pH 7.4, 150 mM NaCl, 1 mM EDTA-pH 8) for 15 min on ice before centrifugation at 14,000 rpm for 15 min at 4°C. Nuclear and cytoplasmic proteins were prepared from differentiated Th2 cells using Nuclear and Cytoplasmic Extraction Reagents from Pierce Biotechnology. Protein extracts (25 μg) from naïve T, Th1, Th2 and Phoenix cells were separated on 4-12% SDS-PAGE and transferred onto Nytran membranes (Schleicher and Schuell Bioscience). Immunoprecipitation (IP) was performed as previously described (8). Nuclear cell lysates (1 mg Th2 extract) were incubated with control antibody, anti-IRF4, anti-PU.1 or anti-GATA-3 conjugated with protein G beads (Santa Cruz Biotechnology, Santa Cruz CA) overnight at 4°C before precipitation with protein G beads. Immunocomplexes were separated by SDS-PAGE and immunoblots were probed with the precipitating antibodies.
Th2 nuclear lysate (250 μg) was incubated with double stranded biotinylated oligonucleotides as described (23). The sequences for IL-10 promoter oligonucleotide are biotin-TGAGGTCTGAAGAAAATCAGCCCTCTCGGG and the reverse complement. For the competition assay the competitor oligonucleotide was incubated with the nuclear protein for 15min at RT before the addition of the biotinylated IL-10 promoter oligonucleotide. The IRF4 consensus binding site was deleted in the IRF4 mutant competitor oligonucleotide and the sequence is TGAGGTCTATCAGCCC TCTCGGG and the reverse complement. The sequence for the λB IRF4 competitor and GATA-3 oligonucleotide were previously described (8, 24). Protein-DNA complexes were separated by SDS-PAGE and immunoblots were probed with anti-IRF4.
The retroviral vector MIEG-hCD4 was previously described (8). The coding region for human IRF4 cDNA was amplified by PCR and cloned into MIEG-hCD4. The Phoenix GP packaging cells line was transiently transfected with 15 μg of purified plasmid by calcium phosphate precipitation. CD4+ T cells cultured under Th1 or Th2 conditions for two days were transduced by centrifugation at 1,800 rpm, 20°C for 2 hrs with 1.5 ml of retroviral supernatant containing 8 μg/ml of polybrene, 100 U/ml of human IL-2 and cytokines and antibodies for either Th1 or Th2 differentiation. Transduced Th2 were stained with anti-human CD4-PE from BD and purified by sorting before restimulation for real-time PCR or ELISA.
WT CD4+ T cells were cultured under Th2 condition for 5 days as described above before Amaxa nucleofection with scrambled siRNA or IRF4 specific siRNA, a pool of two different siRNA constructs (siRNA2 and siRNA3) described by Brüstle et al. (14). After 4h, treated cells were restimulated with 5 μg/ml, anti-CD3, 1 μg/ml anti-CD28, 10 ng/ml of IL-4 and 10 μg/ml XMG for 40h. Cells were then harvested and either analyzed by RT-PCR or restimulated with 2 μg/ml anti-CD3 for ELISA.
Cells were restimulated with 4 μg/mL α-CD3 for 3 hours prior to 3 hours treatment with 3 μM monensin. Intracellular cytokine staining was performed using fluorochrome conjugated antibodies (BD Pharmingen) and α-IL-4 PECy7 (eBioscience). IRF4 staining was performed using α-IRF4 (Santa Cruz Biotechnology) and the secondary Donkey anti-goat antibody (Jackson Immunoresearch) conjugated with Cy5 (Cyanine 5). Stained cells were analyzed with an LSRII instrument.
CD4+ T cells were differentiated into Th2 cells for 7 to 10 days as described above. The day before sorting, the differentiated cells were harvested counted and plated at 107 cells/ml in RPMI-1640 for 6 hours. After this resting period the cells were restimulated for 6hrs with 10ng/ml PMA and 1μg/ml ionomycin. Cells were harvested and washed with MACS buffer. The IL-10 secreting and non-secreting populations were labeled using the mouse IL-10 Secretion Assay (Miltenyi Biotec) before sorting with FACSVantage SE or FACSAria from Becton Dickinson. The cells were rested for 1 or 2 days before restimulation, analysis or transduction.
IL-10 high and -low Th2 (1 × 106) cells were sorted as described above and co-cultured with either 0.5 × 106 splenic CD11c+ cells purified by positive selection using magnetic beads (Miltenyi Biotech) or 1 × 106 total splenocytes from Balb/c mice. Cells supplemented with RPMI-1640 medium were activated with 5 μg/ml anti-CD3, and 4 μg/ml LPS. DC co-culture was harvested after 3 days for IFNγ and IL-6 ELISA and total splenocyte co-culture was harvested after 5 days for IgG1 ELISA.
CD4 T cells cultured under Th2 conditions for five days were fixed with formaldehyde to cross-link protein-DNA complexes and 5 × 106 cells were used per ChIP reaction. The ChIP assay was performed essentially as described (25) except that cells were resuspended in nuclear lysis buffer (50nM Tris pH 8.0, 10mM EDTA pH 8.0, 1% SDS, and protease inhibitor) at 4°C for 15 minutes before ultrasonication and the addition of one LiCl wash with the LiCl buffer (0.25 M LiCl, 1% Igepal, 1% sodium deoxycholate,1mM EDTA pH 8.0 and 10 mM Tris pH 8.0), and two TE buffer washes (1mM EDTA and 10 mM Tris pH 8.0). Real-time PCR was done with 2 μl (1.7 × 105 cells) of immunoprecipitated DNA for 30 cycles. DNA was then analyzed using qPCR. To calculate percent input, ChIP results for the specific antibody were determined using a standard curve of input DNA from the same cells. Control IgG ChIP results are subtracted from specific antibody ChIP results and results are shown as the mean specific binding of replicates ± SD. Primers for Il10 and Il4 ChIP have been previously described (8, 28). Primers for the Il9 promoter were CAG TCT ACC AGC ATC TTC CAG TCT AGC and GTG GGC ACT GGG TAT CAG TTT GAT GTC.
EL4 cells were cultured for three days before transfection of 106 cells by Amaxa nucleofection or by electroporation with expression plasmid, reporter vector containing the regulatory sequence and β-galactosidase in DMEM. Cells were immediately transferred to 6 well plates. After 24h the cells were harvested, washed with PBS and cultured in the presence or absence of 0.2 μg/ml of Ionomycin and 20 ng/ml of PMA for 24h. Harvested cells were lyzed with reporter lysis buffer (Promega) and the luciferase activity was measured for each sample and divided by the protein concentration and the β-galactosidase activity of the sample. The Il4 and Il10 reporters were described previously (28). The NFAT reporter is a trimer of a distal Il2 promoter-binding site and was generously provided by Gerald Crabtree.
Having established a role for PU.1 in regulating Th2 heterogeneity, we wanted to determine if other factors required for Th2 differentiation also contribute to the expression of subsets of Th2 cytokines. We separated Th2 cultures into IL-4 high and low, and IL-10-high and -low populations and examined the expression of Th2-associated transcription factors in each population. As previously described, PU.1 segregated into the IL-4-low population (8). IL-10-high cells expressed 15-fold more Il10 mRNA and secreted 50-fold more IL-10 than IL-10-low cells (Fig. 1A-C). Il4 expression was 2 fold higher in IL-10-high than in IL-10-low cells. The IL-10-high and low phenotypes were stable as there was still a significant difference in IL-10 production from these cells after an additional week in culture (Fig. 1B). In contrast to the enrichment for IL-4 and IL-10 production, the level of Il9 mRNA is 9-fold higher in IL-10-low than in IL-10-high cells (Fig. 1D). To determine if IL-10-high and -low cells had different functional characteristics in culture, we cultured CD11c+ splenocytes or total splenocytes in the absence or presence of either IL-10-high or -low Th2 cells. We observed that IL-10-low cells stimulated increased IL-6 and IFNγ production from LPS-stimulated DCs, while IL-10-high cells has less of an effect on IL-6 production and did not affect IFNγ production (Fig. 1E). Conversely, IL-10-high cells, but not IL-10-low cells were able to induce class switching to IgG1 in LPS-stimulated splenocyte cultures (Fig. 1F). These data suggest that IL-10-high and -low populations in Th2 cultures represent distinct states that have the potential for separate biological functions in vivo.
To determine if there were transcription factors that were associated with differences in IL-10 expression, we screened IL-10-high and -low cells for expression of a number of factors associated with regulating the Th2 phenotype, including Sfpi1, Maf, Bcl6, Zfpm1, Zbtb32, Runx1 and Irf4. We observed that Irf4 expression was enriched 5-fold in the IL-10 high population (Fig. 1G). To confirm that IRF4 protein was also differentially expressed between IL-10-high and -low cells we performed intracellular staining for IL-10 and IRF4. There was an 8-fold difference in expression of IRF4 in cells gated for IL-10-high or -IL-10 low expression (Fig. 1H). Since IRF4 expression is induced upon anti-CD3 stimulation of naïve CD4+ T cells, we compared the expression of IRF4 in naïve cells activated with anti-CD3 for 6 hours with the expression in IL-10-low and high cells. While IRF4 was induced following activation, IRF4 expression was still higher in IL-10-low cells, suggesting that IL-10-low cells are not activated but undifferentiated cells (Fig. 1H). To further show specificity for the expression of Irf4, we also examined the expression of related factors Irf1, Irf2 and Irf8. Irf1 and Irf2 were not differentially expressed between IL-10-high and -low cells (Fig. 1G). Irf8 mRNA was barely detectable in either subset in the absence of stimulation with anti-CD3 and expression was not different following activation of IL-10-high or -low cells (data not shown). These studies demonstrate a specific correlation of IRF4 expression with IL-10 production.
IRF4 was reported to be important for Th2 development since IRF4-deficient T cells secrete decreased levels of Th2 cytokines while IFNγ level increase (13). However the role of IRF4 in differentiated Th2 cells has not been clearly defined. The requirement for IRF4 in Th2 development suggested that ectopic expression of IRF4 into Th2 cells during differentiation would increase the level of secreted Th2 cytokines. To test this, we generated an IRF4-expressing bicistronic retroviral vector (Fig. 2A) and transduced differentiating Th2 cells on the second day of a five-day culture period. Cells sorted for hCD4 expression were restimulated with anti-CD3 and evaluated for cytokine production using ELISA. Ectopic expression of IRF4 in Th2 cells increased production of IL-10 and IL-4 by 8-fold and 4-fold, respectively, with no significant effect on IL-5 and IL-13 (Fig. 2B).
Since transduction of IRF4 increases specific cytokines in differentiating Th2 cells, we next tested whether it would alter the phenotype of IL-10-low cells isolated from differentiated Th2 populations. To assess the change in IL-10-low phenotype, we sorted IL-10-low cells from cells cultured under Th2 conditions for 10 days as described in Methods. IL-10-low cells were transduced with IRF4 RV and cultured for two days before selection and stimulation with anti-CD3 to assess the level of cytokine production. Transduction of IRF4 in IL-10 low cells enhanced the production of IL-10 and IL-4 by 6-fold and 4-fold, respectively, with no significant effect on IL-5, but a decrease in Il9 mRNA (Fig. 2C). These results demonstrate that IRF4 specifically increases IL-4 and IL-10 production from Th2 cells but does not induce other Th2 cytokines.
To further show that IRF4 is required for IL-10 expression in Th2 cultures, we differentiated Th2 cultures in vitro and transfected cells with IRF4-specific or scrambled siRNA. Since Irf4 is expressed at high levels in effector Th2 cells, the IRF4-specific siRNA resulted in only a partial reduction of Irf4 expression (Fig. 2D). However, there was a commensurate decrease in IL-10 production from cells transfected with IRF4-specific siRNA (Fig. 2D).
Th1 cells also secrete IL-10, though at levels much lower than those produced by Th2 cells and with fewer IL-10+ cells in the Th1 population than Th2 cultures, though most IL-10+ cells are also IFNγ+ (Fig. 3A and data not shown) (26, 27). As IRF4 is also expressed in Th1 cells (Fig. 3B), we wanted to determine if IRF4 expression also segregated with IL-10 expression in Th1 cultures. Intracellular staining for IL-10 and IRF4 in Th1 cells was similar to the pattern in Th2 cells where higher IRF4 was observed in IL-10-high cells (Fig. 3C). To further demonstrate that IRF4 functioned in Th1 cells to increase IL-10 production, we transduced differentiating Th1 cell with control or IRF4-expressing retroviruses as done with Th2 cells in Figure 2B and observed that IRF4 increased IL-10 production from Th1 cells (Fig. 3D).
Together, these data demonstrate that IRF4 contributes to IL-10 production in T helper subsets.
To determine if IRF4 is directly regulating Il10 and Il4, we tested the direct binding of IRF4 to the Il10 locus (Fig. 4A) using DAPA and chromatin immunoprecipitation. We first used a biotinylated oligonucleotide corresponding to an IRF4 consensus binding site in the Il10 promoter. Strepavidin-agarose was able to precipitate IRF4 in the presence but not the absence of oligonucleotide using a DAPA protocol (Fig. 4B). To demonstrate specificity for this interaction IRF4 was competed with an oligonucleotide from the Igλ enhancer that contains an IRF-4 binding site, but not with the Il10 promoter oligonucleotide with the IRF-4 binding site deleted or with a GATA-3 consensus oligonucleotide (Fig. 4B). To test this interaction in vivo we used chromatin immunoprecipitation and observed the Il10 promoter, the Il10 CNS element (28), and as controls the Il4 promoter and enhancer were enriched in IRF4 immunoprecipitates compared to control antibody precipitates (Fig. 4C). In contrast, IRF4 had minimal binding to the Il9 promoter (Fig. 4C).
IRF4 was previously shown to transactivate a reporter gene with the Il4 promoter (17), which we also observed in this study (Fig. 4D). To determine if IRF4 could transactivate gene expression from the Il10 regulatory elements we used a luciferase reporter containing either the Il10 promoter region or the Il10 CNS3 region, both of which contain IRF4 binding sites (Fig. 4A)(28). Upon transfection of EL4 cells with either of the Il10 reporters and IRF4-expressing or control pCEP4 vectors, the transactivation was measured by assessing luciferase activity. Co-transfection of IRF4 induced increased luciferase activity from promoter and CNS3 reporters (Fig. 4D). Stimulation of cells with PMA + ionomycin increased basal reporter activity and IRF4 co-transfection was able to further increase reporter activity (Fig. 4D). As a negative control, IRF4 did not activate an NFAT reporter plasmid (Fig. 4D). Thus, IRF4 binds and directly transactivates Il10 regulatory elements.
PU.1 was demonstrated to interact with GATA-3 using recombinant protein binding assays (29) and in Th2 cell co-immunoprecipitates (8). PU.1 is also known to interact with IRF4 in B cells (10, 12, 24, 30, 31). To investigate the interaction of PU.1 and IRF4 in Th2 cells, we used anti-PU.1, anti-IRF4 or anti-GATA-3 to immunoprecipitate complexes from Th2 nuclear lysate. PU.1, IRF4, and a small amount of GATA-3 were immunoprecipitated from Th2 cells with anti-IRF4 and anti-PU.1 (Fig. 5A). Immunoprecipitation with anti-GATA-3 confirmed the interaction between PU.1 and GATA-3, though little IRF4 was precipitated with this complex (Fig. 5A). Thus, although PU.1 interacts with IRF4 and GATA-3, these data suggest that PU.1-IRF4 and PU.1-GATA-3 are largely separate complexes, as little IRF4 was associated with GATA-3. Since IRF4 expression segregates in IL-10-high cells, these interactions would only be meaningful if PU.1 were also expressed in the same cells. To test this, we examined expression of Sfpi1, encoding PU.1, in IL-10-high and -low cells and observed that Sfpi1 mRNA was enriched in the IL-10-high population in a pattern similar to Irf4 expression (Fig. 5B). Since IL-10-high cells are comprised of both IL-4-high and -low cells, this does not contradict previous data showing a segregation of PU.1 expression in IL-4-high and -low populations.
To determine if the association of IRF4 and PU.1 had functional consequences, we used mice carrying a conditional allele of the Sfpi1 gene, crossed to lck-Cre transgenic mice (denoted as Sfpi1lck-/-). We then compared the function of IRF4 in wild type and Sfpi1lck-/- Th2 cells. Chromatin immunoprecipitation demonstrated that IRF4 binding to the Il10 promoter is greater in Sfpi1lck-/- Th2 cells than in C57BL/6 Th2 cells (Fig. 5C). A similar trend of IRF4 binding was observed in Il4 promoter (data not shown). The level of Il4 DNase hypersensitivity site VA (32, 33) in the IRF4 precipitates was also greater in Sfpi1lck-/- Th2 cells than in WT Th2 cells though binding to Il10 CNS3 was only modestly affected by PU.1-deficiency. Th2 cultures from Sfpi1lck-/- mice produced slightly more IL-10 than WT cultures, supporting observations in our previous report (8). Concomitant with increased IRF4 binding to the Il10 locus, transduced IRF4 induced more IL-10 in Sfpi1lck-/- Th2 cells than in WT Th2 cells. We did observe that IRF4 transduction had a less robust increase in IL-10 production in C57BL/6 background cells, compared to Balb/c Th2 cells (Fig. 5D vs. Fig. 2B) although IRF4 expression as assessed by flow cytometry was similar between Balb/c and C57BL/6 cells, and between C57BL/6 WT or Sfpi1lck-/- cells (data not shown). To further demonstrate the ability of PU.1 to interfere with gene activation by IRF4 we repeated the reporter assay in Figure 4D with the addition of a condition where PU.1- and IRF4-expressing vectors were co-transfected. Results demonstrate that PU.1 also interferes with the IRF4-dependent induction of the Il4 and Il10 promoter reporter vectors (Fig. 5E). Overall, these results suggest that PU.1 interactions limit the ability of IRF4 to transactivate Th2 cytokines.
The analysis in Fig. 1 is based on two-state separation of cells into IL-10-high and -low cells. However, IL-10-high and -low cells can be further divided based on co-expression of other cytokines. Since IRF4 induced production of IL-4 and IL-10, we examined the expression of IRF4 in populations of IL-4- and IL-10-positive cells by intracellular staining. IRF4 expression was highest in Th2 cells that were double-positive for IL-4 and IL-10, and lowest in cells that did not secrete either cytokine (Fig. 6). Interestingly, expression was intermediate but similar in IL-4- and IL-10-single positive cells. Results are shown for Balb/c cells and similar patterns are observed for C57BL/6 cells. Thus while IRF4 promotes IL-4 and IL-10 production, other factors also contribute to the decision of a cell to make one or both cytokines.
IRF4 plays an important role in the development of Th2 and Th17 cells (14). In Irf4-/- mice, development of Th2 cells is decreased, suggesting that it plays a role in the differentiation process (14). However, a role in differentiation does not preclude involvement in the regulation of specific cytokines in differentiated Th2 cells. In this report we demonstrate that IRF4 contributes to the heterogeneity of Th2 populations by increasing production of IL-4 and IL-10 and decreasing expression of Il9, while having no effects on IL-5 or IL-13. IRF4 expression segregates in Th2 cells between IL-10-high and IL-10-low cells, directly binds to and transactivates the Il10 gene, and ectopic expression of IRF4 can increase IL-10 production from IL-10-low cells. Thus, IRF4 is an instructive factor in establishing Th2 heterogeneity.
IL-10 is a regulatory cytokine produced by a number of cells including Th2 cells and Il10 regulation in each cell type may be distinct. IL-10 plays a critical role in controlling inflammation in vivo by selectively suppressing the expression of pro-inflammatory cytokines. In this report we show that IL-10-high and -low cells have differing effects on co-cultured cells in vitro, and it is likely that functions differ in vivo as well. In various cell types, the molecular mechanism regulating the expression of IL-10 involves binding of IRF1 and STAT3 to the promoter region of the Il10 gene locus (26, 34), and the regulation of the level of IL-10 mRNA by Sp1 and Sp3 (35). In Th2 cells, GATA-3 remodels the Il10 locus (34), and Jun family proteins bind the CNS3 region to induce Il10 transcription (28). Moreover, IL-10 production requires repeated stimulation to be completely imprinted within the Th2 population (36). In this report, we also show that IRF4 contributes to Il10 expression in Th2 and Th1 cells. The IL-12-dependent production of IL-10 in Th1 cells has been documented in human and mouse cells (26, 37, 38). Clearly, additional factors contribute to the difference in IL-10 production by Th1 and Th2 cells, including that GATA-3 is not expressed in Th1 cells, c-jun and JunB bind to the Il10 CNS3 in Th2 but not Th1 cells (28), and that Ets-1 is a negative regulator of IL-10 production in Th1 cells (39). We demonstrate that in Th2 cells IRF4 binds directly to the Il10 locus and is able to transactivate Il10 regulatory element reporter plasmids. Our results parallel the recent description of a role for IRF4 in Treg cells where Il10 was one of the prominently regulated genes (15). The fact that IRF4 binding sites exist in the Il10 promoter and CNS3 regions, and that it binds to the same regulatory element as Jun containing complexes suggests that these factors may cooperate. We did not observe strong interactions of IRF4 with GATA-3, suggesting that while these factors both contribute to Il10 expression, direct interactions are not required.
There are interactions of IRF4 with PU.1, which we previously demonstrated decreased IL-10 production (8). In the absence of PU.1, IRF4 had greater potential to bind Il10, transactivate the Il10 promoter and increase IL-10 production, suggesting that at least in part, PU.1 is able to modulate Il10 through the ability to interfere with IRF4 activity. It was surprising that we did not see increased IL-10 in Sfpi1lck-/- control-transduced cells (Fig. 5D), and this may be a result of the culture conditions required for retroviral transduction. The effects of PU.1-deficiency resulting in increased Th2 cytokine production are most dramatic when TCR stimulation is limiting (HCC, SLN and MHK, submitted). However, transduction of cells with limiting TCR stimulation was not efficient, and as a result of using optimal stimulation conditions for the retroviral transduction we did not observe altered IL-10 production in the absence of IRF4 transduction. The role of PU.1 must also be placed in the context of interactions with multiple transcription factors (Fig. 5A) and with the heterogeneity of the Th2 population where PU.1 is differentially expressed in subpopulations of cells. As previously noted, PU.1 is expressed highly in IL-4-low cells, and in this report is also expressed in IL-10-high cells. This might suggest that PU.1 is most highly expressed in an IL-4-low/IL-10-high population. However, it has thus far been difficult to sort cells stained for two cytokines preventing a more thorough analysis of this issue. As PU.1 expression is only present in a subpopulation of Th2, the effects of deficiency on IRF4 function might be obscured in a bulk Th2 population, unless IRF4 is overexpressed (Fig. 5D). Moreover, as ectopic PU.1 expression decreased IL-5 and IL-13 production, as well as IL-4 and IL-10, it is clear that interactions with IRF4 only account for a portion of the observed function of PU.1 in Th2 cells.
The reciprocal regulation of Il10 and Il9 in Th2 cells is striking and distinct from the IL-9 and IL-10-producing cells present in cultures primed with TGFβ and IL-4 (40, 41). While transduction of IRF4 in Th2 cells decreases Il9 expression, we observed only minimal IRF4 binding to the Il9 promoter in a ChIP assay. This suggested that the effects of IRF4 could be indirect, through the induction of IL-10. Indeed, neutralizing IL-10 in IL-10-high cells, that express low levels of Il9, modestly increased Il9 mRNA (data not shown). However, it is not clear which IL-10 activated pathways might be responsible for this regulation. It is also not clear why a cell would be specialized to express only one of these cytokines. IL-9 is a pleiotropic cytokine involved in the pathologic and physiologic evolution of asthma by recruiting eosinophils and lymphocytes to the lung, inducing mucus hypersecretion, mast cells hyperplasia in concert with IL-4, IL-5 and IL-13 (42), while IL-10 is a suppressive cytokine that may modulate many of these processes. It is possible that secretion of IL-9 by Th2 cells would only be effective if target cells did not receive a conflicting signal generated by IL-10. In this manner, Th2 heterogeneity may reflect functional specialization of cell types within the inflammatory microenvironment.
Increasing evidence suggests that the establishment of Th2 heterogeneity is not stochastic, but rather instructive, based on the expression of specific factors. As such, a growing list of transcription factors has specific effects on individual Th2 cytokines. IRF4 is induced following T cell activation, and expression is further increased following Th2 differentiation. Importantly, the level of IRF4 in IL-10-low cells is increased compared to that in recently activated T cells. As we have shown, IRF4 activates IL-10 and IL-4 production, while decreasing IL-9 and having little effect on IL-5 or IL-13. C-maf regulates IL-4 production but is not required for production of other Th2 cytokines (43). Similarly, Pias1 increases IL-13 production without affecting IL-4 or IL-5 expression (44). We have shown that PU.1 decreases expression of many Th2 cytokines, but increases expression of CCL22, a chemokine associated with Th2 inflammation (8). BOB.1/OBF.1 regulates PU.1 expression in Th2 cells and also affects the potential for Th2 cytokine production (45). Moreover, the expression levels of each of these factors, and other factors that contribute to Th2 cytokine production exist in gradients that correlate with cytokine producing phenotypes (Fig. 6). The similar level of expression of IRF4 in both IL-4- and IL-10-single positive cells supports the idea that other positive- or negative-acting factors overlay on the IRF4 gradient to generate the specific patterns of cytokine secretion. The mosaic of transcription factor gradients ultimately results in the heterogeneity observed in cytokine production from individual cells.
In this report, we have identified IRF4 as a regulator of Th2 heterogeneity by enhancing or decreasing the production of specific cytokines. IRF4 function, like GATA-3 as described in our previous report (8), is limited by the expression of PU.1 in Th2 cells, which binds IRF4 and decreases binding to target genes including Il10. Future work will examine how these factors interact to generate the population phenotype and what signals determine the expression of each factor within individual cells.
We thank Dr. Q Yu for help with primer design.
1This work was supported by U.S. Public Health Service Award AI057459 (to M.H.K.) from the National Institutes of Health.