Accumulation of NFAT mediates IL-2 expression in memory, but not naïve, CD4⁺ T cells

Abstract

In contrast to naïve CD4⁺ T cells, memory CD4⁺ T cells rapidly express high levels of effector cytokines in response to antigen stimulation. The molecular mechanism for this specific behavior is not well understood. The nuclear factor of activated T cells (NFAT) family of transcription factors plays an important role in the transcription of many cytokine genes. Here we show that memory CD4⁺ T cells rapidly induce NFAT-mediated transcription upon T cell receptor ligation whereas NFAT activation in naïve CD4⁺ T cells requires longer periods of stimulation. The difference in kinetics correlates with the low levels of NFATc1 and NFATc2 proteins present in naïve CD4⁺ T cells and their high levels in memory CD4⁺ T cells. Accordingly, IL-2 expression requires NFAT activation only in memory CD4⁺ T cells whereas it is NFAT-independent in naïve CD4⁺ T cells. Thus, the accumulation of NFATc1 and NFATc2 in memory CD4⁺ T cells represents a previously uncharacterized regulatory mechanism for the induction of early gene expression after antigen stimulation.

Memory CD4⁺ T cells behave like effector cells instead of naïve CD4⁺ T cells because they are able to rapidly express and secrete high levels of effector cytokines in response to antigen stimulation (1, 2). Memory CD4⁺ T cells generated from effector T helper (Th) 2 cells produce high levels of Th2 cytokines upon activation whereas memory CD4⁺ T cells derived from effector Th1 cells produce Th1 cytokines (3). It is still an open question why a given antigen rapidly induces the expression of cytokine genes in memory but not naïve CD4⁺ T cells.

Nuclear factor of activated T cells (NFAT) represents a family of transcription factors that are characterized by the presence of a calcineurin-binding domain (NFATc family) and a Rel-homology DNA binding domain (4, 5). The four calcineurin-dependent NFATc family members (NFATc1, NFATc2, NFATc3, and NFATc4) are widely expressed in different tissues, although NFATc4 appears to be excluded from lymphoid cells (4). Different isoforms for each of the members have been identified as a result of alternative splicing, and, although their specific functions remain unclear, they seem to have a restricted tissue distribution (6, 7). NFATc1 isoforms are differentially expressed in CD4⁺ T cells with the NFATc1C isoform being predominant in nonstimulated CD4⁺ T cells while the NFATc1A isoform is expressed in effector CD4⁺ T cells (8).

It has been described that NFAT transcription factors reside in the cytoplasm in a hyperphosphorylated state and translocate to the nucleus in response to calcium signals. The accumulation of NFAT in the nucleus is the result of nuclear import promoted by calcineurin-mediated dephosphorylation (9) and nuclear export through phosphorylation mediated by several kinases (4). In the nucleus, NFAT family members together with other nuclear transcription factors mediate transcription of specific genes including cytokine and cytokine receptor genes. NFAT was first identified as a transcription factor that regulates the activity of the IL-2 gene promoter in the human Jurkat T cell line (10). However, IL-2 production in naïve CD4⁺ T from NFATc2 or NFATc3 single deficient mice is apparently normal (11–13). NFATc1-deficient mice die during embryogenesis, but NFATc1-deficient CD4⁺ T cells from RAG2^−/−NFATc1^−/− chimeras also show normal IL-2 production (14, 15). In contrast, T cells from NFATc1/c2 double deficient RAG2^−/− chimera mice produce less IL-2, but these mice suffer from pathological plasma cell infiltrates in several organs and their T cells display an activated, memory-like phenotype (16). IL-2 production by activated total splenocytes from mice double deficient for NFATc2/c3 is reduced, but these mice develop spontaneous allergic blepharitis and interstitial pneumonitis, and T cells also display a memory phenotype (17). However, it remains unclear whether the reduced IL-2 production in NFAT double deficient mice is a direct cause of the NFAT deficiency in Th cells or an indirect result of the phenotype in those mice.

Here we show that naïve CD4⁺ T cells contain very low levels of NFATc1 and NFATc2 proteins compared with memory CD4⁺ T cells. Consequently, inhibition of NFAT activation interferes with IL-2 expression only in memory CD4⁺ T cells, but not in naïve CD4⁺ T cells. We propose that accumulation of NFAT transcription factors in memory CD4⁺ T cells is one mechanism that can facilitate an efficient cytokine gene expression as seen in a memory immune response.

Results

Accumulation and Rapid Activation of NFAT in Memory CD4⁺ T Cells.

We have previously shown that NFAT-mediated transcription in naïve CD4⁺ T cells requires long periods of time (60–72 h) after stimulation, whereas high NFAT transcriptional activity is rapidly (6–12 h) induced in effector Th2 cells (18). Despite their resting stage, memory CD4⁺ T cells produce large amounts of cytokines soon after antigen stimulation when compared with naïve CD4⁺ T cells. To examine the kinetics of NFAT-mediated transcription in the memory CD4⁺ population, naïve (CD44^low) and memory-like (CD44^high) CD4⁺ T cells were isolated from NFAT-luciferase reporter transgenic mice (18, 19) and activated with anti-CD3 and anti-CD28 mAbs for different periods of time. As early as 36 h we could detect NFAT transcriptional activity in activated memory CD4⁺ T cells, but not in naïve CD4⁺ T cells (Fig. 1A). We examined whether the increased NFAT-transcriptional activity observed in memory CD4⁺ T cells correlates with increased NFAT DNA binding in an EMSA. Upon T cell receptor (TCR) ligation, high levels of NFAT DNA-binding activity were rapidly induced in memory CD4⁺ T cells but was almost undetectable in naïve CD4⁺ T cells (Fig. 1B). In contrast, NF-κB DNA binding was similar in activated naïve and memory CD4⁺ T cells (Fig. 1B). Thus, the kinetics of NFAT activation in response to T cell activation is more rapid in memory CD4⁺ T cells than in naïve CD4⁺ T cells.

Fig. 1.

Increased levels of NFATc1 and NFATc2 in memory CD4⁺ T cells. (A) A total of 5 × 10⁵ naïve (CD44^low) or memory (CD44^high) CD4⁺ T cells purified from NFAT-luciferase transgenic mice were activated with anti-CD3 and anti-CD28 mAbs. Luciferase activity was measured at the indicated periods of time. (B) Naïve and memory CD4⁺ T cells were isolated and activated as in A for 4 h. Nuclear extracts were examined by EMSA using oligos containing consensus NFAT or NF-κB DNA binding sequences. (C) The localizations of NFATc1, NFATc2, and NFATc3 (red) in freshly isolated naïve and memory CD4⁺ T cells were examined by immunostaining and confocal microscopy. Nuclei were stained with YOYO (green). Yellow color indicates nuclear localization of NFAT. Each Inset shows one representative cell in higher magnification. Stainings with secondary antisera are shown as control. (D) Whole-cell extracts of freshly isolated naïve and memory CD4⁺ T cells were used for Western blot analysis. The blot was probed consecutively with antibodies for NFATc1, NFATc2, NFATc3, and GATA3 as well as STAT1 as loading control. (E) Relative mRNA levels of NFATc1 and NFATc2 in unstimulated naïve and memory CD4⁺ T cells were examined by quantitative real-time RT-PCR. Representative experiments of two (D and E) or three (A–C) are shown.

To test whether the rapid activation of NFAT-mediated transcription observed in memory CD4⁺ T cells was the result of nuclear accumulation of NFAT before antigen stimulation, freshly isolated naïve and memory CD4⁺ T cells were examined by confocal microscopy for the localization of NFATc1, NFATc2, and NFATc3. None of the three NFAT family members was detected in the nucleus of unstimulated naïve or memory CD4⁺ T cells (Fig. 1C). Interestingly, we observed an accumulation of NFATc1 and NFATc2 proteins in the cytoplasm of memory CD4⁺ T cells compared with the almost undetectable amounts present in naïve CD4⁺ T cells (Fig. 1C). Although most memory cells have substantial amounts of both NFATc1 and NFATc2, only a small percentage of naïve cells contain detectable levels of these transcription factors [supporting information (SI) Fig. 5]. The difference in NFATc3 levels was less remarkable. To confirm that NFATc1 and NFATc2 protein levels were indeed higher in memory CD4⁺ T cells than in naïve CD4⁺ T cells before antigen stimulation, we measured NFAT levels in freshly isolated naïve and memory CD4⁺ T cells by Western blot analysis. Substantial levels of NFATc1 and NFATc2 proteins were present in memory CD4⁺ T cells, but they were almost absent in naïve CD4⁺ T cells (Fig. 1D). We also observed a preferential accumulation of the NFATc1A isoform in memory CD4⁺ T cells (Fig. 1D). There was no difference in protein levels of NFATc3 or the unrelated transcription factor GATA3 between naïve and memory CD4⁺ T cells (Fig. 1D). Although the differences were less remarkable, the levels of NFATc1 and NFATc2 mRNA were also higher in memory CD4⁺ T cells than in naïve CD4⁺ T cells (Fig. 1E), indicating that the up-regulation of these transcription factors in memory cells is in part due to increased gene expression. Thus, the rapid activation of NFAT in memory CD4⁺ T cells correlates with the presence of preexistent NFATc1 and NFATc2 proteins in memory CD4⁺ T cells.

NFATc1 and NFATc2 Expression Is Reprogrammed During the Generation of Antigen-Specific Memory CD4⁺ T Cells.

We examined whether NFATc1 and NFATc2 expression is up-regulated during the generation of memory from activated naïve CD4⁺ T cells. Naïve CD4⁺ T cells isolated from pigeon cytochrome c (cyt c) AND TCR transgenic mice (20) were activated with anti-CD3 and anti-CD28 mAbs, and NFAT protein levels were determined by Western blot analysis. NFATc1A and NFATc1B, as well as NFATc2, were almost undetectable in unstimulated naïve CD4⁺ T cells (Fig. 2A), and NFATc1C could be detected only after long exposures (data not shown). However, after 1 day of stimulation NFATc1 and NFATc2 levels were up-regulated and remained high during the differentiation into effector cells (Fig. 2A). An accumulation of NFATc1 and NFATc2 was also observed in both Th1 and Th2 effector cells differentiated with cyt c peptide and antigen-presenting cells in the presence of polarizing cytokines (Fig. 2B). Interestingly, NFATc1A levels were slightly increased in Th2 compared with Th1 effector cells (Fig. 2B). High levels of NFATc2 were also present in Th1 and Th2 effector cells (Fig. 2B).

Fig. 2.

Reprogramming of NFATc1 and NFATc2 expression during the differentiation of antigen-specific naïve CD4⁺ T cells into memory CD4⁺ T cells. (A) Antigen-specific CD4⁺ T cells were isolated from cyt c TCR (AND) transgenic mice and stimulated with anti-CD3 and anti-CD28 mAbs for the indicated periods of time. Whole-cell lysates were subjected to Western blot analysis for NFATc1 and NFATc2. TRAF5 is shown as loading control. The multiple bands observed in the NFAT blots represent phosphorylated NFAT isoforms. (B) Naïve CD4⁺ T cells from AND mice were activated with cyt c peptide and antigen-presenting cells for 4 days under polarizing Th1 and Th2 conditions. Whole-cell extracts of Th1 and Th2 effector cells were used to examine NFATc1 and NFATc2 expression by Western blot analysis. Actin levels were examined as loading control. (C) AND effector Th1 and Th2 cells were transferred into γ-irradiated and bone marrow-reconstituted host mice. After 4 weeks, antigen-specific memory CD4⁺ T cells generated in vivo were isolated, and whole-cell extracts were used to examine NFATc1 and NFATc2 expression by Western blot analysis. Actin is shown as loading control. Representative experiments of three are shown.

To examine NFAT levels in antigen-specific memory CD4⁺ Th1 and Th2 cells, naïve CD4⁺ T cells from AND mice were antigen-stimulated and differentiated into Th1 and Th2 effectors for 4 days in vitro. The cells were then transferred into host mice and reisolated as resting antigen-specific memory Th1 and Th2 cells 4 weeks later. NFAT protein levels were examined by Western blot analysis. In agreement with the results obtained with the polyclonal populations of naïve and memory CD4⁺ T cells (Fig. 1D), NFATc2 levels were greatly up-regulated in in vivo generated antigen-specific memory Th1 and Th2 compared with naïve CD4⁺ T cells (Fig. 2C). NFATc1B levels were also higher in Th1 and Th2 memory CD4⁺ T cells (Fig. 2C). Increased NFATc1A levels were observed in Th2 relative to Th1 memory CD4⁺ T cells while NFATc1C levels were elevated in Th1 compared with Th2 memory CD4⁺ T cells (Fig. 2C). Together, these results indicate that the expression of NFATc1 and NFATc2 is reprogrammed during the development from naïve to memory CD4⁺ T cells to maintain high levels of both transcription factors in memory cells.

NFAT Is Required for IL-2 Production in Memory, but Not Naïve, CD4⁺ T Cells.

Although NFAT was initially identified as a transcription factor that regulates IL-2 gene transcription, the low levels of NFATc1 and NFATc2 present in naïve CD4⁺ T cells before stimulation and the almost undetectable NFAT DNA-binding activity early upon activation questioned whether NFAT significantly contributes to IL-2 production in these cells. Kinetics analysis of IL-2, NFATc1, and NFATc2 expression in naïve CD4⁺ T cells upon activation suggests that induction of IL-2 expression (>20-fold by 8 h) occurs before an up-regulation of these transcription factors (<3-fold by 8 h) (SI Fig. 6). We therefore examined the contribution of NFAT to IL-2 expression in naïve and memory CD4⁺ T cells using transgenic mice that express a truncated, dominant-negative mutant of NFAT (dnNFAT) specifically in T cells (21). The dnNFAT mutant inhibits transcriptional activity mediated by all calcineurin-dependant NFATs by competing for calcineurin binding, but it does not inhibit transcription mediated by AP-1 and NF-κB (22). Expression of dnNFAT in CD4⁺ T cells from these mice partially inhibits NFAT DNA binding and NFAT-mediated transcription (21). These mice do not show any pathological phenotype, and the distribution of CD4⁺ and CD8⁺ T cells as well as naïve and memory CD4⁺ T cells is similar to WT mice (22). The expression of the dnNFAT transgene is also comparable between naïve and memory CD4⁺ T cells (SI Fig. 7A). IL-2 production of activated naïve and memory CD4⁺ T cells from dnNFAT transgenic mice or WT littermates was analyzed by ELISA. IL-2 levels were comparable between naïve CD4⁺ T cells from WT mice and from dnNFAT transgenic mice (Fig. 3A). In contrast, dnNFAT memory CD4⁺ T cells produced considerably lower levels of IL-2 than WT memory CD4⁺ T cells (Fig. 3A). Similar results could be observed after 48 h of activation (SI Fig. 7B).

Fig. 3.

Requirement of NFAT for IL-2 production in memory CD4⁺ T cells. (A) A total of 5 × 10⁵ naïve and memory CD4⁺ T cells were isolated from WT and dnNFAT transgenic mice and activated with anti-CD3 and anti-CD28 mAbs for 24 h. IL-2 levels in the cell supernatant were measured by ELISA. (B) Naïve and memory CD4⁺ T cells were isolated from WT and dnNFAT transgenic (Tg) mice and activated with anti-CD3 and anti-CD28 mAbs for 24 h. IL-2 mRNA levels were examined by a ribonuclease protection assay using total RNA. L32 and GAPDH mRNA levels are shown as loading controls. (C) A total of 5 × 10⁵ CD25-negative memory (CD44^high) CD4⁺ T cells were isolated from dnNFAT mice and WT littermates by cell sorting. Cells were stimulated with anti-CD3 and anti-CD28 mAbs for 24 h. IL-2 levels in the supernatant were analyzed by ELISA. (D) A total of 5 × 10⁵ memory CD4⁺ T cells were isolated from WT and dnNFAT transgenic mice and activated as in A for 48 h. IL-4 and IFNγ levels were examined by ELISA. (E) Effector Th1 and Th2 CD4⁺ T cells from AND and dnNFAT×AND mice were adoptively transferred into irradiated hosts. After 5 weeks they were purified as memory Th1 and Th2 cells and activated with anti-TCR(Vβ3) and anti-CD28 mAbs for 24 h. Supernatants were analyzed for IL-2 by ELISA. (F) A total of 1 × 10⁶ purified naïve CD4⁺ T cells from AND and dnNFAT×AND mice were activated with anti-CD3 and anti-CD28 mAbs for 24 h. IL-2 production was analyzed by ELISA. One representative experiment of two (B and C) or three (A and D–F) is shown. ELISA data represent the mean ± SE of triplicate determinations.

To confirm that the decreased IL-2 production in dnNFAT memory cells was due to reduced gene expression instead of higher consumption, we examined IL-2 mRNA levels from activated naïve and memory CD4⁺ T cells from WT and dnNFAT transgenic mice by ribonuclease protection assay. IL-2 mRNA levels were similar in naïve CD4⁺ T cells from dnNFAT transgenic mice compared with WT naïve CD4⁺ T cells (Fig. 3B). In contrast, lower levels of IL-2 mRNA were detected in dnNFAT memory cells compared with WT memory cells (Fig. 3B). Because NFAT has been involved in T regulatory cells' suppressive activity (23, 24) we confirmed that the requirement of NFAT for IL-2 expression in memory CD4⁺ T cells was not dependent on T regulatory function by excluding CD4⁺CD25⁺ T cells from the naïve and memory populations. Nevertheless, IL-2 production was still reduced in activated memory, but not naïve, CD4⁺ T cells from dnNFAT mice (Fig. 3C), supporting that NFAT is directly required for IL-2 gene expression in memory, but not naïve, CD4⁺ T cells.

In contrast to naïve CD4⁺ T cells, memory cells produce effector cytokines such as IL-4 and IFNγ early upon antigen stimulation. Thus, we also examined the production of IL-4 and IFNγ in naïve and memory CD4⁺ T cells from WT and dnNFAT mice upon activation for 48 h by ELISA. No IL-4 or IFNγ could be detected in naïve CD4⁺ T cells from WT and dnNFAT mice (data not shown). The production of IL-4 was substantially reduced in memory CD4⁺ T cells from dnNFAT mice compared with WT mice (Fig. 3D). Although the levels of IFNγ as determined by ELISA were even higher in dnNFAT memory cells (Fig. 3D), the IFNγ mRNA levels upon 24 h were slightly decreased (SI Fig. 7C). Proliferation was also comparable between memory CD4⁺ T cells from WT and dnNFAT mice (SI Fig. 7D). Thus, memory CD4⁺ T cells from the dnNFAT mice can still respond to TCR stimulation, and the compromised IL-2 production is not due to an overall deficiency in cytokine production or proliferation.

To show the requirement of NFAT for IL-2 expression in antigen-specific memory CD4⁺ T cells, naïve CD4⁺ T cells from dnNFAT×AND double transgenic and AND littermate control mice were differentiated into effector Th1 and Th2 cells in vitro in the presence of IL-4 or IL-12, respectively. We have previously shown that inhibition of NFAT activity in dnNFAT CD4⁺ T cells blocks IL6-driven Th2 differentiation but not IL4-driven Th2 differentiation (22) because exogenous IL-4 can directly induce IL-4 gene expression through activation of GATA3 and Stat6 without the need of NFAT (25). Accordingly, the presence of the dnNFAT transgene did not interfere with the differentiation of the antigen-specific CD4⁺ T cells into Th2 effector cells (SI Fig. 8). Likewise, inhibition of NFAT did not affect the differentiation of antigen-specific CD4⁺ T cells into Th1 cells in the presence of exogenous IL-12 (SI Fig. 8) because this cytokine acts primarily through Stat4 (26). Th1 and Th2 effector populations were then transferred into irradiated hosts and isolated as antigen-specific memory CD4⁺ Th1 and Th2 cells 5 weeks later. Similar recovery rates of AND and dnNFAT×AND memory CD4⁺ T cells were obtained (data not shown). Isolated memory CD4⁺ T cells were stimulated, and IL-2 production was measured by ELISA. IL-2 production was severely impaired in both memory Th1 and Th2 cells from dnNFAT×AND mice (Fig. 3E). Inhibition of NFAT did not decrease IL-2 production in antigen-specific naïve CD4⁺ T cells (Fig. 3F). Thus, NFAT is required for IL-2 expression only in antigen-specific memory, but not naïve, CD4⁺ T cells.

Inducible Inhibition of NFAT Interferes with IL-2 Production of Memory CD4⁺ T Cells.

To show that NFAT was required for IL-2 production upon antigen stimulation in memory CD4⁺ T cells instead for the generation of memory cells capable of producing IL-2, we generated inducible dnNFAT transgenic mice (Tet-dnNFAT) where the dnNFAT transgene is expressed in T cells only in the presence of tetracycline or doxycycline (dox). Western blot analysis showed that the dnNFAT transgene was present only in CD4⁺ T cells from dox-treated Tet-dnNFAT mice, but not in CD4⁺ T cells from untreated Tet-dnNFAT mice or WT littermates (Fig. 4A). The inducible levels of dnNFAT were sufficient to inhibit NFAT DNA binding in activated CD4⁺ T cells from dox-treated Tet-dnNFAT as tested by EMSA (Fig. 4B). Only a slight decrease was observed in activated CD4⁺ T cells from untreated Tet-dnNFAT mice (Fig. 4B), and the dox treatment itself did not have any effect on WT CD4⁺ T cells (data not shown).

Fig. 4.

Inducible inhibition of NFAT impairs IL-2 production in memory CD4⁺ T cells. (A) Whole-cell lysates of freshly isolated CD4⁺ T cells from dox-treated and untreated Tet-dnNFAT transgenic mice and WT littermates were subjected to Western blot analysis with anti-Flag mAb for determining dnNFAT expression. Actin is shown as loading control. n.s., nonspecific. (B) Total CD4⁺ T cells from dox-treated and untreated Tet-dnNFAT mice as well as WT littermates were activated with anti-CD3 and anti-CD28 mAbs for 24 h. Nuclear extracts were analyzed by EMSA for NFAT DNA binding. (C) A total of 1 × 10⁵ naïve (CD44^low) and memory (CD44^high) CD4⁺ T cells from dox-treated WT and Tet-dnNFAT mice were isolated and activated with anti-CD3 and anti-CD28 mAbs for 24 h in the presence of dox (100 ng/ml). IL-2 levels were measured by ELISA. The mean ± SD of IL-2 production of WT and Tet-dnNFAT CD4⁺ T cells relative to the production of WT CD4⁺ T cells (set as 100%) of four independent experiments is shown. ∗, P < 0.05 by one-sample t test.

We then investigated whether induced expression of dnNFAT affects IL-2 production in memory CD4⁺ T cells. The distributions of CD4 and CD8 T cell populations as well as naïve (CD44^low) and memory (CD44^high) CD4⁺ T cells were not affected in dox-treated Tet-dnNFAT mice compared with untreated Tet-dnNFAT or WT littermates (SI Fig. 9 A and B). Naïve and memory CD4⁺ T cells were isolated from WT and dox-treated Tet-dnNFAT mice and activated with anti-CD3 and anti-CD28 mAbs. The production of IL-2 by memory CD4⁺ T cells from dox-treated Tet-dnNFAT mice was consistently lower than the IL-2 production by WT memory CD4⁺ T cells (Fig. 4C). The degree of reduction of IL-2 varied among experiments (30–70%) because the inducible expression of dnNFAT in response to the dox treatment was largely variable, not only between independent experiments but also among individual mice within a given experiment. In contrast to memory CD4⁺ T cells, there was no significant difference in the IL-2 production between naïve CD4⁺ T cells from WT and Tet-dnNFAT mice (Fig. 4C). We did not observe impaired IL-2 production by memory CD4⁺ T cells from untreated Tet-dnNFAT mice when compared with WT memory CD4⁺ T cells (SI Fig. 9C). These results further demonstrate that NFAT is required for induction of IL-2 production by TCR signals in memory CD4⁺ T cells.

Discussion

Early studies using Jurkat T cells and T cell clones have shown a requirement of NFAT for IL-2 expression (27). The calcineurin inhibitor cyclosporin A also blocks IL-2 production by primary CD4⁺ T cells, but there is increasing evidence that calcineurin has several other cellular targets in addition to NFAT (e.g., MEF2, cofilin), and cyclosporin A treatment also affects NF-κB and AP-1 activity (28–31). However, the normal production of IL-2 in CD4⁺ T cells from mice deficient in NFATc1, NFATc2, or NFATc3 questioned the relative contribution of NFAT to IL-2 expression (11–13). Accordingly, in this study we show that naïve CD4⁺ T cells contain minimal levels of NFAT and that NFAT is dispensable for IL-2 production in these cells. CD4⁺ T cells from mice deficient for both NFATc2 and NFATc3 or NFATc1 and NFATc2 produce less IL-2, but the CD4⁺ T cells in these mice display a memory-like phenotype, and the production of a large number of cytokines is also impaired (16, 17). Because we show here that NFATc1 and NFATc2 accumulate in memory CD4⁺ T cells and that NFAT is required for IL-2 production in memory, it is possible that the reduction in IL-2 expression in NFATc1/c2 and NFATc2/c3 double deficient mice is due to the fact that these cells have an effector/memory-like phenotype.

Because the relative contribution of NFAT to IL-2 gene expression correlates with the protein levels of preexistent NFATc1 and NFATc2, this could explain why IL-2 expression in T cell lines that normally contain high NFAT protein levels requires NFAT (27). Similarly, we have previously shown that expression of the dnNFAT mutant interferes with IL-2 production in activated thymocytes that contain high levels of NFAT (21). We have also found that inhibition of NFAT interferes with IL-2 production in effector Th1 and Th2 cells according with the high levels of NFATc1 and NFATc2 present in these cells (data not shown).

Only the mechanisms that regulate the expression of the NFATc1 family member have been studied, probably because NFAT was originally identified as a preexisting cytoplasmic transcription factor (32). Two promoters have been identified in the NFATc1 gene that differentially regulate the expression of three NFATc1 isoforms using specific polyA signals (8, 33, 34). NFATc1C was shown to be the predominant form present in unstimulated CD4⁺ T cells whereas NFATc1A is present in effector cells (8). Here we demonstrate that NFATc1A is also highly expressed in memory CD4⁺ Th2 cells before stimulation whereas NFATc1C expression is elevated in memory Th1 cells, suggesting that these two isoforms may contribute specifically to the expression of Th2 cytokines or Th1 cytokines in memory cells, respectively. We have previously demonstrated that the presence of IL-6 during the activation of naïve CD4⁺ T cells induces NFATc2 gene and protein expression (22). Here we also show that NFATc2 protein levels are greatly up-regulated in response to T cell activation and remain high in effector and memory CD4⁺ T cells. Transcription factors of the C/EBP or STAT families could be responsible for the up-regulation of NFATc2 expression in effector and memory CD4⁺ T cells because sequence analysis revealed several potential binding sites within the NFATc2 promoter (unpublished data). However, whether these transcription factors mediate the NFATc2 up-regulation in activated CD4⁺ T cells and memory CD4⁺ T cells awaits further analysis.

The main feature of memory CD4⁺ T cells is the rapid expression of effector cytokines upon antigen stimulation, but the molecular differences to naïve CD4⁺ T cells are still not entirely clear (3). The strength of early phosphorylation events is reduced in memory CD4⁺ T cells compared with naïve CD4⁺ T cells, which may be because of a reduced expression of the adapter protein SLP-76 (35). However, downstream signaling events like MAPK signaling have been shown to be similar between naïve and memory CD4⁺ T cells (35), indicating that those are not responsible for the observed differences in cytokine expression. Similarly, calcium flux induced by TCR cross-linking or calcium ionophores is even reduced in memory CD4⁺ T cells compared with naïve CD4⁺ T cells (36). Another explanation for the strong response of memory CD4⁺ T cells upon antigen stimulation is demethylation at the loci of several important effector molecules like IL-4 and IFNγ, which occurs at the effector stage of naïve cells and persists when these cells become rested again (37). Here we demonstrate that accumulation of NFATc1 and NFATc2 can be an additional mechanism that contributes to the rapid effector function of memory CD4⁺ T cells. The accumulation of NFAT proteins could have a synergistic effect together with open cytokine gene loci for facilitating the rapid response of memory CD4⁺ T cells. Increased expression of specific transcription factors may therefore be an important mechanism of antigen-experienced immune cells to provide protection against recurrent infections.

Materials and Methods

Mice.

Two founder lines of the Tet-dnNFAT mice were generated by using the rtTA-M2 gene (38) downstream of the mouse proximal lck promoter while the Flag-tagged dnNFAT mutant (21) was under control of the Tet-responsive element of the pBi-EGFP vector (Clontech, Mountain View, CA). Both DNA fragments were coinjected into fertile oocytes for the generation of double transgenic mice. To induce dnNFAT expression, Tet-dnNFAT mice were fed with dox food (6 g/kg; Bio-Serv, Frenchtown, NJ) for 4 days before the experiment. NFAT-luciferase reporter transgenic mice contain the luciferase gene downstream of three copies of the NFAT binding site from the IL-2 promoter (18, 19). dnNFAT transgenic mice (21, 22) and cyt c TCR transgenic (AND) mice (20) have been described previously. All mice were backcrossed to B10.BR WT mice (The Jackson Laboratory, Bar Harbor, ME) for >10 generations. Procedures that involved mice were approved by institutional guidelines for animal care.

Cell Preparation and Activation.

CD4⁺ T cells were isolated by negative selection as previously described (22, 39) and stained with anti-CD4-Alexa Fluor 647 (BD Pharmingen, San Jose, CA) and anti-CD44-phycoerythrin (Caltag, Carlsbad, CA) mAbs. Naïve (CD4⁺CD44^low) and memory (CD4⁺CD44^high) T cells were purified (>98%) by cell sorting (FACS-Aria; Becton Dickinson) and gating in the low forward light scatter and low side light scatter lymphocyte population. The CD44^high CD4⁺ T cell population was negative for the activation markers CD25 and CD69. CD4⁺ T cells were activated with plate-bound anti-CD3 (2C11) (5 μg/ml) and soluble anti-CD28 (1 μg/ml) (BD Pharmingen) mAbs in Bruff's medium. Antigen-specific effector cells were generated by activating purified naïve CD4⁺ T cells from AND mice with pigeon cyt c peptide (5 μM) in the presence of mitomycin C-treated (50 μg/ml) DCEK-ICAM cells under polarizing Th1 [IL-12 (10 ng/ml) and anti-IL-4 (10 μg/ml)] or Th2 [IL-4 (15 ng/ml) and anti-IFNγ (10 μg/ml)] conditions as previously described (40–42). For antigen-specific memory cells, 1–2.5 × 10⁷ Th1 or Th2 effector cells were adoptively transferred into B10.BR host mice that had been previously thymectomized, lethally irradiated (950 rads), and reconstituted with T cell-depleted syngeneic bone marrow (AtxBM) (40). After 4 weeks, resting CD4⁺ memory mTh1 or mTh2 cells were purified by cell sorting using the TCR Vβ3-antibody.

EMSA.

Nuclear extracts were prepared as previously described (43, 44). Binding reactions were performed by using 2 μg of nuclear proteins and [³²P]dCTP end-labeled double-stranded oligonucleotide probes containing an NFAT binding site from the proximal IL-4 promoter (5′-gtaataaaattttccaatgtaaa-3′) (45) or a NF-κB consensus site from the κ intronic enhancer (46).

Additional Details.

For more details see SI Materials and Methods.

Abbreviations

NFAT

nuclear factor of activated T cells

cyt c

cytochrome c

dox

doxycycline

Th

T helper

TCR

T cell receptor.