Calcium-dependent activation of interleukin-21 gene expression in T cells.

Interleukin (IL)-21 is a gamma(c)-dependent cytokine produced by activated T cells with important actions for T, B, and NK cells. The IL-21 gene is adjacent to the IL-2 gene, and like IL-2, IL-21 is strongly induced at the transcriptional level after T cell activation. Interestingly, however, in contrast to the IL-2 gene, a calcium ionophore alone was sufficient to induce IL-21 gene expression in preactivated T cells. Two DNase I hypersensitivity sites were found in the IL-21 gene, corresponding to nucleotide sequences that are conserved in humans and mice. One site is located at the IL-21 promoter region and conferred T cell receptor-mediated IL-21 gene transcription. TCR-induced IL-21 gene expression was inhibited by cyclosporin A and FK506. Correspondingly, the IL-21 5'-regulatory region contains three NFAT binding sites, and induction of IL-21 promoter activity was impaired when these sites were mutated or following treatment with cyclosporin A. Thus, our studies reveal that in contrast to IL-2, a calcium signal alone is sufficient to mediate induction of the IL-21 in preactivated T lymphocytes and that this induction appears to result from specific NFAT binding.

In addition to its effects on T cells, IL-21 can enhance the proliferative response of human and murine B cells stimulated in vitro with antibodies to CD40, but it inhibits B cell proliferation in response to anti-IgM ϩ IL-4 (3) and can augment B cell death (7,8). Within the B cell lineage, IL-21 critically regulates immunoglobulin (Ig) production and cooperates with IL-4 for the production of multiple antibody classes in vivo (9). IL-21R Ϫ/Ϫ mice have markedly diminished IgG1 but greatly elevated IgE levels in response to antigen (9), and correspondingly, IL-21 can inhibit antigen-induced IgE production (10). Mice lacking expression of both IL-21R and IL-4 exhibit a dysgammaglobulinemia with severely impaired IgG and IgE responses (9). IL-21 differentially influences B cell fate depending on the signaling context. Although IL-21 induces the death of resting B cells (7), it promotes differentiation of B cells activated in a T cell-dependent manner into memory and plasma cells, explaining how IL-21 can be proapoptotic for B cells in vitro yet critical for antigen-specific immunoglobulin production in vivo (8).
In T cells, IL-21 can cooperate with IL-15 or IL-7 to drive the expansion of T cells, particularly of CD8 ϩ T cells (11). This effect correlates with the ability of IL-21 to promote cytolytic activity and anti-tumor effects (11). IL-21 has also been reported to act on NK cells. One report indicated that it can increase NK cell maturation, whereas another suggested it can inhibit NK cell expansion (3,6). To better understand the basis for regulation of this critical cytokine, we now have studied the regulatory elements controlling expression of this gene.

MATERIALS AND METHODS
Mice and Cell Culture-C57BL/6 mice were obtained from the Jackson Laboratory. NFATc2 knock-out mice were a gift from L.H. Glimcher (Harvard School of Public Health, Boston, MA). Mice used in these experiments were 8 -10 weeks of age. All experiments were performed under protocols approved by the National Institutes of Health Animal Use and Care Committee and followed the NIH guidelines "Using Animals in Intramural Research." Single-cell suspensions from the spleen were prepared and cultured at 1 ϫ 10 6 /ml in RPMI 1640 medium containing 10% fetal bovine serum, 2 mM glutamine, 100 units/ml penicillin G, and 100 g/ml streptomycin. CD4 ϩ T cells were isolated from the spleens by positive selection with MACS purification (Miltenyi Biotech). These isolated CD4 ϩ T cells were activated by incubating in the presence of 2 g/ml of anti-CD3 and 2 g/ml of anti-CD28 mAbs (Pharmingen) for 2 days and expanded in the presence of 100 units/ml of IL-2 for another day.
5Ј-RACE-5Ј-RACE was performed using the GeneRacer kit (Invitrogen). Total cellular RNA isolated from activated CD4 ϩ T cells was dephosphorylated using calf intestinal alkaline phosphatase, and the 5Ј-"cap" of mRNA was removed using tobacco acid pyrophosphatase. An RNA oligonucleotide was then ligated to the full-length, decapped mRNAs, and reverse transcription was performed using oligo(dT) as a primer. The resultant cDNAs were PCR-amplified using the GeneRacer 5Ј-primer and a primer specific for the IL-21 gene (5Ј-TTCAGCTGT-TCAACAATGTC-3Ј). The GeneRacer 5Ј-nested primer and the same IL-21 primer were then used in a second, nested PCR.
Quantitative Real Time PCR-CD4 ϩ splenocytes isolated from C57BL/6 mice were activated with anti-CD3 and anti-CD28 for 2 days, * 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.
Plasmid Constructs-To generate the wild type (WT) mouse IL-21 promoter luciferase reporter construct, we subcloned the Il21 gene Ϫ1605 to ϩ75 region 5Ј to the luciferase gene between the BgiII and MluI sites in the polylinker of the pGL3-Basic luciferase reporter vector (Promega). Site-directed mutagenesis of the Il21 NFAT sites in this plasmid was performed using the QuikChange kit (Stratagene). Three mutagenic primers, 5Ј-CCATGTGTCTCTTTTTAAGGTCTAAGATGCA-GATG-3Ј, 5Ј-CGTACACCTAGCCAACTTAAAAGAAAAACGAGTTAC-3Ј, and 5Ј-CACTATACAAAGATTTAAGGGCTGCAATGGGAGGG-3Ј were used to introduce the indicated underlined 3-bp changes into the NFAT sites at Ϫ203, Ϫ121, and Ϫ76, respectively. Fragments spanning HS2 (ϩ1139 to ϩ1683) were generated by PCR and cloned in both orientations 5Ј and 3Ј to the SV40 promoter in pGL3-Promoter. Transient Transfections and Luciferase Assays-Transient transfections of normal mouse T cells were performed by electroporation, as described previously (13).
Electrophoretic Mobility Shift Assays (EMSAs)-CD4 ϩ splenocytes isolated from C57BL/6 mice were activated with anti-CD3 and anti-CD28 mAbs for 2 days, and these activated cells were washed and then cultured in the presence of IL-2 for 1 day. Nuclear extracts were prepared as described (14) from these untreated preactivated CD4 ϩ cells or cells that had been treated with anti-CD3 mAb for 1 h at 37°C. EMSAs were performed as described previously (13)  Boxed are potential binding sites for NFAT proteins in the HS1 region, and for AP1 and IRF1 proteins in the HS2 region. In A, the transcription start site of the murine IL-21 gene is underlined and indicated by the arrow. gels (29:1 acrylamide:bisacrylamide) containing glycerol and 0.5ϫ Tris borate-EDTA buffer. For supershifting assays, nuclear extracts were preincubated for 10 min with NFATc1 or NFATc2 mouse mAbs (Affinity Bioreagents, Inc., Golden, CO). Oligonucleotide sequences from the IL-21 promoter and IL-2 distal NFAT binding sites used as probes are shown in Fig. 5A.

RESULTS
To investigate the regulation of IL-21 expression, we first analyzed IL-21 mRNA levels in CD4 ϩ T cells. CD4 ϩ splenocytes isolated from C57BL/6 mice were activated with anti-CD3 and anti-CD28 mAbs for 2 days, washed, and then cultured in the presence of IL-2 for 1 day. These preactivated CD4 ϩ cells were then either left unstimulated or restimulated with various stimuli. IL-21 mRNA was increased ϳ10-fold by anti-CD3 but not by anti-CD28 (Fig. 1A). Co-stimulation with both anti-CD3 and anti-CD28 showed similar effects to those seen with anti-CD3 alone (Fig. 1A). The induction of IL-21 mRNA was primarily at the level of transcription because the transcription inhibitor, actinomycin D, markedly diminished the inducibility (Fig. 1A). This induction did not require de novo protein synthesis, because the protein synthesis inhibitor, cycloheximide, did not diminish and in fact increased IL-21 mRNA levels in response to anti-CD3 ϩ anti-CD28 (Fig. 1A). Cycloheximide is a protein synthesis inhibitor that superinduces the expression of many genes by preventing the degradation of otherwise labile mRNAs (15). Interestingly, IL-21 mRNA levels were increased by ionomycin alone, whereas phorbol 12-myristate 13-acetate had little effect by itself and did not augment the effect of ionomycin (Fig. 1A). This is in marked contrast to the regulation of the IL-2 gene in which both phorbol 12-myristate 13-acetate and ionomycin are required for its expression (16) (Fig. 1B). Consistent with the importance of a calcium signal, rapamycin had no effect, but cyclosporin A and FK506 markedly inhibited induction of IL-21 mRNA in response to ionomycin or anti-CD3ϩanti-CD28 (Fig. 1A); this indicates a role for the calcineurin/NFAT signaling pathway in IL-21 mRNA transcription.
To identify the regions of the IL-21 gene that mediate its induction, we performed DNase I hypersensitivity assays with preactivated CD4 ϩ T cells that were either unstimulated or restimulated for 1 h with anti-CD3 ϩ anti-CD28. Nuclei were then isolated and subjected to DNase I hypersensitivity assays. DNA was extracted, digested with NsiI, which cleaves the 6.5-kb 5Ј and 7.1-kb 3Ј to the transcription start site, and Southern blotted with the indicated probe. Two principal hypersensitive sites, HS1 (upstream of the previously reported 5Ј-end of the IL-21 cDNA) and HS2 (in intron 2), were identified (Fig. 2). Both HS1 and HS2 were detected in the unstimulated cells, and the intensities of these sites were modestly increased in restimulated cells. The nucleotide sequences of both the HS1 (Fig. 3A) and HS2 (Fig. 3B) regions are well conserved in humans and mice, consistent with the presence of important regulatory regions in these areas.
The transcriptional start site in the IL-21 gene was identified using 5Ј-RACE and mRNA from activated CD4 ϩ T cells (data not shown). Using an IL-21-specific primer (see "Materials and Methods"), we identified a single product containing 69 bases of 5Ј-untranslated mRNA; the 5Ј-nucleotide was assigned position ϩ1 as indicated in Fig. 3A. The full-length mRNA is thus 16 nucleotides longer than an mRNA previously reported (NCBI accession number NM_021782). These data confirm that there is not a cryptic upstream exon and that HS1 is located in the proximal 5Ј-regulatory region.
To test whether HS1 and HS2 correspond to TCR response elements, each HS region was separately subcloned into a luciferase reporter construct. When the HS1 region was used, potent inducibility was observed in response to ionomycin as well as anti-CD3 ϩ anti-CD28 (Fig. 4A), and this inducibility was abrogated when the transfected cells were treated with cyclosporin A 30 min before stimulation (Fig. 4A). In contrast, when the construct containing the HS2 region was cloned in either orientation and either 5Ј or 3Ј to the IL-21 or SV40 promoters, only minimal TCR inducibility was found in transfection experiments in normal mouse splenocytes (Fig. 4B). Thus, HS1 appeared to be more important for IL-21 gene regulation, and we have focused on HS1 in this study.
Analysis of HS1 revealed several potential consensus sequences for factor binding, including three consensus NFAT

FIG. 4. Cyclosporin inhibits IL-21 promoter activity.
A, normal murine T lymphocytes were transfected with a WT HS1 region reporter construct. Transfected cells were incubated with medium, anti-CD3 ϩ anti-CD28 (␣CD3 ϩ ␣CD28), CsA, or anti-CD3 ϩ anti-CD28 ϩ CsA. Data are shown as the mean Ϯ S.E. B, fragments spanning HS2 (ϩ1139 to ϩ1683) were generated by PCR and cloned in both orientations 5Ј and 3Ј to the SV40 promoter in pGL3-promoter. These constructs were transfected into normal murine T lymphocytes and evaluated for their TCR responsiveness.
binding sites that are conserved in humans and mice (Fig. 3A). We investigated factor binding to 32 P-labeled probes spanning these sites using EMSAs and nuclear extracts from unstimulated or anti-CD3-stimulated mouse T cells. When probes spanning the NFAT sites at either Ϫ203 or Ϫ76 (Fig. 5A) were assayed, only very faint complexes were detected with extracts from either unstimulated or stimulated cells (Fig. 5B). However, the probe spanning the site at Ϫ121 (Fig. 5A) generated a complex with extracts from unstimulated cells (Fig. 5, B, lane 3 and C, lane 1), which was modestly increased when cells were stimulated with anti-CD3 (Fig. 5, B, lane 4 and C, lane 2). An antibody to NFATc2 (also known as NFAT1 and NFATp) supershifted the complex to generate a slower mobility complex, with this effect being more evident in activated cells (Fig. 5C,  lanes 3 and 4). In contrast, an antibody to NFATc1 (also known as NFAT2 and NFATc) had no effect (lanes 5 and 6). The mutation of the NFAT site eliminated the formation of complexes from resting (lane 7) and stimulated (lane 8) cells.
As a control, we also investigated factor binding to a probe spanning the distal NFAT binding sites in the IL-2 gene. When the IL-2 NFAT probe was used, the formation of complexes was weaker than with the IL-21 probe. Some binding activity was detected in unstimulated cells (lane 9), which increased following stimulation with anti-CD3 (lane 10). The complex formed with the IL-2 NFAT probe was evaluated for the presence of NFATc1 and NFATc2 using antibodies to each protein. An antibody to NFATc2 supershifted the complex to generate a slower mobility complex (lanes 11 and 12), whereas the anti-NFATc1 antibody almost completely blocked the formation of complexes (lanes 13 and 14). Thus, both NFATc2 and NFATc1 proteins can bind to the IL-2 NFAT probe. In contrast, the IL-21 NFAT probe appeared to preferentially bind NFATc2. The fact that much of the complex formed with the IL-21 probe was not supershifted suggested that additional protein(s) contribute to the binding activity seen in unstimulated cells.
We next sought to confirm NFATc2 binding to the IL-21 promoter in vivo using chromatin immunoprecipitation assays. Despite in vitro binding activity in both resting and stimulated cells, we found a marked increase in NFATc2 binding in vivo after cellular stimulation with anti-CD3 (Fig. 5D). No binding of NFATc2 was observed to a control exon region of glyceraldehyde-3-phosphate dehydrogenase gene, indicating the specificity of the binding to IL-21 promoter. Thus, NFATc2 binding in vivo corresponded to the induction of Il21 gene expression.
To determine whether NFATc2 was uniquely essential for IL-21 induction, we next evaluated the ability of anti-CD3 and anti-CD28 to induce IL-21 mRNA expression in NFATc2-deficient mice. Interestingly, IL-21 mRNA induction was not diminished in these mice (Fig. 6). This is analogous to the normal induction of IL-2 mRNA and increased induction of IL-4 mRNA in these animals as well (Fig. 6) (17). The normal induction despite the absence of NFATc2 suggests the compensation by and action of other NFAT family proteins in vivo, as has been suggested for the IL-2 and IL-4 genes as well (17).
Despite the normal induction of IL-21 mRNA in NFATc2 knock-out mice, our data overall suggested that the NFAT binding sites were critical. Specifically, the inhibition by cyclosporin A of the induction of IL-21 mRNA (Fig. 1) and promoter activity (Fig. 4) by anti-CD3 and anti-CD28 was consistent with a critical role for the NFAT sites in the HS1 region. To

FIG. 5. NFAT bind to the IL-21 promoter in vitro and in vivo.
A, sequence of the Ϫ203, Ϫ121, Ϫ76 (spanning the Ϫ216/Ϫ182, Ϫ136/Ϫ101, Ϫ89/Ϫ55, respectively), and IL-2 NFAT probes. For the Ϫ121 probes, both wild type (WT) and mutant (mt) probes are shown. Mutated nucleotides are shown in lowercase, and hyphens indicate identity to the WT sequence. B, EMSA using the Ϫ203, Ϫ121, and Ϫ76 probes in A and nuclear extracts from untreated preactivated mouse T lymphocytes or from cells treated with anti-CD3 (␣CD3) (see "Material and Methods"), as indicated. C, EMSA using the WT and mutant Ϫ121 probes and the IL-2 NFAT probe in A and nuclear extracts from untreated mouse T lymphocytes (lanes 1, 3, 5, 7, 9, 11, and 13) or from anti-CD3-treated cells (lanes 2, 4, 6, 8, 10 , 12, and 14) in the presence of antibodies to NFATc2 (lanes 3, 4, 11, and 12) or NFATc1 ( lanes 5, 6, 13, and 14), as indicated. D, chromatin immunoprecipitation assays reveal inducible NFAT binding to the IL-21 promoter in vivo. Murine T lymphocytes were cultured in the absence or presence of anti-CD3 for 1 h, and then chromatin immunoprecipitation assays were performed with preimmune serum or anti-NFATc2. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a negative control. Data are shown as the mean Ϯ S.E. further define their role, we mutated each NFAT site individually and in combination and assayed the TCR responsiveness of each mutant construct in normal mouse T lymphocytes. Whereas the wild type construct showed ϳ15-fold TCR inducibility, selective mutation of the NFAT sites at Ϫ203 (mutant m1) and Ϫ76 (mutant m3) markedly decreased TCR inducibility (Fig. 7), even though these sites individually show little binding by EMSA (Fig. 5B). Selective mutation of the site at Ϫ121 most potently bound factors in vitro (mutant m2) or double mutation of any two of the sites (mutants m4, m5, and m6) even more potently decreased TCR inducibility, whereas simultaneous mutation of all three NFAT sites (mutant m7) abrogated TCR inducibility (Fig. 7). These results suggest that all of the NFAT sites are required for maximal TCR inducibility and are consistent with at least some factor binding to all three sites in vivo; this binding may be cooperative. By itself, the Ϫ121 NFAT site nevertheless appears to be the most important. DISCUSSION IL-21 is a type I cytokine that is made by activated T cells and exerts actions on B cells, T cells, and NK cells. Its receptor is most like the IL-2 receptor ␤ chain and the gene encoding IL-21 is adjacent to the Il2 gene. Similar induction patterns might have suggested similar regulation for these two genes, and our mutagenesis data indicate that the Il21 gene, analogous to the Il2 gene, is dependent on NFAT binding sites.
The activity of NFAT proteins is tightly regulated by the calcium/calmodulin-dependent phosphatase calcineurin, a primary target for inhibition by cyclosporin A and FK506 (19). Calcineurin controls the translocation of NFAT proteins from the cytoplasm to the nucleus of activated cells by interacting with an N-terminal regulatory domain conserved in the NFAT family.
In this study, we have described a TCR response element in the 5Ј-regulatory region of the Il21 gene. Several results support the hypothesis that NFAT is required for TCR-mediated transactivation of the Il21 gene. First, the critical IL-21 promoter region contains three NFAT binding sites, the nucleotide FIG. 6. Normal induction of IL-21 and IL-2 mRNA in NFATc2 ؊/؊ mice. Preactivated murine CD4 ϩ lymphocytes from WT or NFATc2 Ϫ/Ϫ mice were stimulated with immobilized anti-CD3 (␣CD3 ϩ ␣CD28) antibody for 4 h, and levels of IL-21, IL-2, and IL-4 mRNA were analyzed by real time PCR. Where indicated, CsA was added 30 min before TCR cross-linking. For each sample, the bar represents the ratio of IL-21, IL-2, or IL-4 mRNA to that of 18 S ribosomal mRNA. Data are shown as the mean Ϯ S.E. sequence of the IL-21 promoter is evolutionarily conserved, and a chromatin immunoprecipitation assay showed the in vivo binding activity of NFATc2 to this region. Second, site-specific mutagenesis of these NFAT sites significantly diminished IL-21 promoter activity. Third, CsA, an inhibitor of calcineurin that potently inhibits nuclear translocation of NFAT and NFATdependent transcription (19,40,41), markedly decreased TCRmediated transactivation of IL-21 promoter and IL-21 gene expression, implying that calcineurin regulates IL-21 gene transcription. Transient transfection experiments using luciferase reporter constructs showed that CsA could almost completely block anti-CD3 ϩ anti-CD28-mediated IL-21 promoter activity, and this specific role of the NFAT binding sites was confirmed using a construct in which all three NFAT sites were mutated. The fact that NFATc2 Ϫ/Ϫ mice had normal IL-21 gene induction is consistent with the potential utilization of more than one NFAT-binding protein in vivo.
Recently, Mehta et al. (42) reported that NFATc2 and T-bet contribute to Th2-specific regulation of IL-21 expression. They studied IL-21 expression in EL4 and D10 cells, and like us, they found that three NFAT binding sites were important for TCRstimulated IL-21 promoter activity. We demonstrate in primary T cells the importance of this region and moreover that the chromatin structure in the promoter region has an open conformation. However, our data differ from those of Mehta et al. (42) in that we find that IL-21 expression remains strong in NFATc2 knock-out T cells. The basis for this dissimilarity is unclear.
Despite the evidence in support of NFAT recruitment to the Il2 and Il4 genes (19,43) as well as the Il21 gene, the transcription of the Il21 gene in preactivated T cells is strikingly different from the genes encoding IL-2 and IL-4 in that it can be mediated by calcium ionophore alone without requiring protein synthesis or protein kinase C activation. Thus, our studies demonstrate that IL-21 induction by antigen in primary T cells is primarily dependent on a calcium signal and contrasts in this regard to these other activation-dependent cytokines that require both calcium and protein kinase C signals (e.g. IL-2 and IL-4) or to another activation-dependent protein whose expression only requires a protein kinase C signal (e.g. IL-2 receptor ␣ chain). The lack of a requirement for a protein kinase C signal for IL-21 expression suggests that AP1 family proteins are not important or are much less important for Il21 gene induction than for induction of the Il2 gene. It is interesting to speculate that the constitutive DNA binding activity seen even with untreated preactivated cell extracts (Fig. 5B) could contribute to the ability of ionomycin alone to activate the gene. In other words, the constitutive DNA binding activity in vitro may correspond to a factor that can cooperate with NFAT proteins in vivo, suggesting that the Il21 gene is poised to receive and respond to a calcium signal as a major component of its regulation. The characterization of such a factor is an interesting area for future investigation. In summary, we have demonstrated properties of regulation of the Il21 gene and have characterized a critical region containing NFAT binding sites that corresponds to a hypersensitive site, clarifying the molecular basis for the rapid induction of Il21 gene expression in T cells.