Follicular CD4+ Th (Tfh) cells provide B cell help in germinal center reactions that support class switching, somatic hypermutation, and the generation of high-affinity Abs. In this article, we show that deficiency in NFAT1 and NFAT2 in CD4+ T cells leads to impaired germinal center reactions upon viral infection because of reduced Tfh cell differentiation and defective expression of proteins involved in T/B interactions and B cell help, including ICOS, PD-1, and SLAM family receptors. Genome-wide chromatin immunoprecipitation data suggest that NFAT proteins likely directly participate in regulation of genes important for Tfh cell differentiation and function. NFAT proteins are important TCR and Ca2+-dependent regulators of T cell biology, and in this article we demonstrate a major positive role of NFAT family members in Tfh differentiation.
Nuclear factor of activated T cells (NFAT) transcription factors are key regulators of T cell activation (1) and exhaustion (2). NFAT1, NFAT2, and NFAT4 are expressed in cells of the immune system and have important roles in T cell development and function (3). Although NFAT family members make similar contacts with DNA (4), they show distinct expression patterns and functions, as judged by the nonoverlapping phenotypes of mice deficient in individual NFAT family members (5, 6, and reviewed in Ref. 7).
Follicular CD4+ Th (Tfh) cells are essential for mediating B cell help and inducing germinal center (GC) responses required for most high-affinity Ab responses. Tfh cells have been characterized by their expression of chemokine (C-X-C motif) receptor 5 (CXCR5) and the lineage-defining transcription factor B-cell CLL/lymphoma 6 (Bcl6) (8–10). Bcl6 does not act alone, and other transcription factors have also been identified as key regulators of Tfh differentiation, including STATs, Maf, BATF, IRF4, ASCL2 (reviewed in Refs. 11, 12), and LEF-1 and TCF-1 (13, 14). NFAT2 is highly expressed in Tfh cells (15); however, the roles played by NFAT family members in Tfh cells are not well understood. Tfh cells have enhanced calcium signaling with NFAT nuclear translocation compared with Th1 cells, suggesting that NFATs may have preferential responsibilities in Tfh cells compared with Th1 cells (16). In this article, we investigated the role of NFAT1 and NFAT2 in the generation of Tfh cells and GC responses to acute viral infection.
Materials and Methods
Mice and infections
SMARTA [lymphocytic choriomeningitis virus (LCMV) gp66-77-IAb–specific] CD45.1+ mice (17) were crossed with Nfat1−/−, Nfat2fl/fl CD4Cre, or Nfat1−/− Nfat2fl/fl CD4Cre mice (2). All mice were maintained in specific pathogen-free barrier facilities and used according to protocols approved by the La Jolla Institute for Allergy and Immunology animal care and use committees. Mice were infected i.p. with 2 × 105 PFU LCMV Armstrong 5 strain. The adoptive transfer experiments using SMARTA cells were performed as previously described (17), unless otherwise stated. Postinfection, splenocytes were harvested and cells were stained with Abs against cell-surface markers as previously described (17). LCMV-specific serum IgG was quantified by ELISA as previously described (18). ELISA data were analyzed by area under the curve (AUC). AUC analysis better accounts for both the quantity and the quality of the IgG, because it accounts for the shape of the curve. AUC total peak area above baseline calculations (GraphPad Prism 6.0) was done for each individual sample, log transformed.
Isolation and culture of T cells
Spleens and lymph nodes were harvested from 6- to 8-wk-old mice. Naive CD4+ cells were purified using EasySep kit (Stem Cell) and activated with anti-CD3 and anti-CD28 as previously described (2).
RNA sequencing and chromatin immunoprecipitation followed by next generation sequencing data analyses
Previously published human RNA sequencing (RNA-seq) data from Weinstein et al. (19) were used. The mouse homologs of genes differentially expressed in Tfh versus non-Tfh cells were identified. NFAT1 chromatin immunoprecipitation followed by next generation sequencing (ChIP-seq) data for memory-like CD8+ T cells were from our previous report (2).
Statistical analyses were done using Prism 6.0 (GraphPad). The p values were calculated using two-tailed unpaired Student t tests with a 95% confidence interval.
Results and Discussion
Reduced Tfh cells and GCs in NFAT1,2-deficient mice
We investigated the role of two NFAT family members, NFAT1 and NFAT2, in the generation of Tfh cells using single- and double-deficient Nfat1−/−, Nfat2fl/fl Cd4-Cre mice (2). The frequency of total B cells was not affected in these mice (data not shown). The mice were infected with LCMV Armstrong strain, which causes an acute viral infection, and the frequency of GC B cells and Tfh cells were determined in spleens 8 d postinfection. Single NFAT1 or NFAT2 deficiency did not significantly affect the generation of GL7+ Fas+ GC B cells compared with wild type (WT) control mice. However, deficiency of both NFAT1 and NFAT2 led to an almost complete absence of GC B cells (Fig. 1A, 1B). LCMV-specific IgG responses were severely decreased in NFAT1,2-deficient mice, commensurate with the failure to develop GCs (Fig. 1C, Supplemental Fig. 1A).
We then sought to determine whether lack of GC responses was due to impaired Tfh cell generation. Similar frequencies of total CD4 T cells were observed between WT and NFAT-deficient animals at day 8 after acute viral infection, and CD44 expression was comparable (data not shown). In contrast, a substantial decrease in the frequency of CXCR5+ Bcl6+ GC Tfh cells was observed in NFAT1,2-deficient compared with WT mice (Fig. 1D, 1E). Moreover, expression of key receptors associated with Tfh cell function, such as ICOS, Ly108, PD-1, and CXCR5, were all reduced in NFAT1,2-deficient CD4+ T cells (Fig. 1F, 1G). Both Th1 and Tfh cells develop in response to acute viral infections, and KLRG1 expression is associated with highly polarized Th1 cells in some contexts (20, 21). The increased KLRG1 expression on NFAT-deficient cells suggests that, in the absence of NFATs, the cells become even more polarized toward Th1 (Fig. 1F, 1G). In summary, our results suggest that NFAT family members have an important role in the generation or maintenance of Tfh cells in vivo upon viral infection, without which GC and antiviral Ab responses are defective.
T cell–intrinsic defect in Tfh cell generation in NFAT1,2-deficient mice
To assess CD4 T cell–intrinsic activities of NFAT1 and NFAT2, we generated Nfat1−/−, Nfat2fl/fl Cd4-Cre CD45.1+ SMARTA TCR transgenic mice. Naive SMARTA CD45.1+ CD4+ T cells were transferred into congenic mice that were then infected with LCMV. Expansion of the adoptively transferred cells and expression of the activation marker CD44 was similar between WT and NFAT1,2-deficient SMARTA cells 8 d postinfection (Supplemental Fig. 1B–D, 1J), indicating the capacity of NFAT-deficient CD4 T cells to mount an Ag-specific response. Mice receiving NFAT1,2-deficient SMARTA CD4 T cells developed fewer GC B cells compared with WT controls (Fig. 2A, 2B), correlating with a decrease in the generation of GC Tfh (CXCR5+ Bcl6+) (Fig. 2C, 2D, Supplemental Fig. 1H), as well as all CXCR5-expressing SMARTA cells (Supplemental Fig. 1E–G). Expression of the PD-1 and ICOS receptors associated with Tfh cell functions was also affected, in this case predominantly by NFAT2 (Fig. 2E, 2F). More severe GC defects had been observed in germline knockout mice (Fig. 1), suggesting a possibility for non-CD4 T cell–intrinsic effects. However, the SMARTA transfer experiment results were also consistent with the presence of endogenous WT T cell responses capable of providing T cell help to B cells. In summary, our results demonstrate an essential role of NFAT family members in the generation or maintenance of Tfh cells in vivo upon viral infection.
Among the non-Tfh SMARTA cells, a dramatic increase in KLRG1+ cells was observed for NFAT2- and NFAT1,2-deficient cells (Fig. 2G–I, Supplemental Fig. 1L). This observation is consistent with these cells being an unusual form of highly polarized Th1 effector cells previously identified in mice with Mycobacterium tuberculosis infection (20). Moreover, we observed no significant differences in Tbet expression between WT and NFAT1,2-deficient cells (Fig. 2I, 2J). Thus, our results indicate that NFAT expression prevents the generation of KLRG1+ effector CD4 T cells in response to infection and specifically regulates Tfh cell generation.
Defective induction of key receptors in NFAT1,2-deficient mice accounts for the defect in GC formation
To examine potential mechanisms by which NFATs control Tfh differentiation and function, we activated WT and NFAT1,2-deficient CD4 T cells with anti-CD3 and anti-CD28 in vitro and characterized expression of genes of interest. Rapid induction of CD69 and CD25 by NFAT1,2-deficient CD4 T cells was comparable with that of WT controls at early times (3, 6 h), although expression of CD69 and CD25 was not maximally maintained in the absence of NFAT1 and NFAT2 (Supplemental Fig. 2A). In contrast, induction of ICOS protein expression was severely impaired on NFAT1,2-deficient naive CD4+ T cells compared with their WT counterparts at all time points examined (Fig. 3A). ICOS expression is critical for Tfh differentiation (17, 22), and therefore loss of ICOS expression is likely one major cause of Tfh loss in the absence of NFATs. Expression of CD40L, PD-1, and SLAM was also significantly impaired on in vitro–activated NFAT1,2-deficient CD4 T cells (Supplemental Fig. 2); each receptor has roles in Tfh help to B cells (23–25). Thus, our results indicate that NFAT family members have a major role in the induction of receptors involved in Tfh cell biology.
Given the early defects in upregulation of ICOS on NFAT1,2-deficient CD4 T cells, we asked whether early stages of Tfh differentiation in vivo depend on NFAT transcription factors. Pilot experiments revealed NFAT1,2-deficient SMARTA CD4 T cells exhibited a delay in accumulation compared with WT SMARTA CD4 T cells in vivo by day 3 in response to an acute viral infection (Supplemental Fig. 1J). Thus, in subsequent experiments, 4-fold more NFAT1,2-deficient SMARTA cells were transferred to partially compensate for the delayed expansion (Supplemental Fig. 1J). Nascent Tfh cells can be identified as CXCR5+ Bcl6+ CD4 T cells at day 3 postinfection (Fig. 3B, 3C). NFAT1,2-deficient cells had severely diminished frequencies of CXCR5+ Bcl6+ (or CXCR5+PD-1+) nascent Tfh cells in comparison with WT CD4 T cells (Fig. 3B, 3C, Supplemental Fig. 1K). The Tfh deficiency was not simply a reflection of a generalized T cell activation defect, because the frequency of early Th1 cells (CD25hi Bcl6−) was elevated (Fig. 3D, 3E) and mean fluorescence intensities of the activation marker CD25 were high (Fig. 3D and data not shown). Thus, our results demonstrate that NFAT family members are selectively important for early Tfh differentiation, most likely via regulation of Icos and other Tfh-associated differentiation genes.
To identify direct roles of NFAT in the regulation of Tfh differentiation and function, we examined binding of NFAT1 to genes that influence Tfh differentiation or are significantly expressed on Tfh cells (Fig. 4A). ChIP-Seq analysis from memory-like CD8+ T cells stimulated in vitro with PMA and ionomycin (2) shows NFAT1 binding sites in the proximal promoters and/or within the transcribed regions of the Icos, Il2ra, and IL6ra genes, which encode receptors that have prominent roles in early Tfh differentiation (17, 26, 27). These ChIP-seq associations, in conjunction with the ICOS expression defects observed in NFAT1,2-deficient CD4 T cells and the early Tfh differentiation defects of NFAT1,2-deficient CD4 T cells in vivo (Fig. 3), suggest that NFAT transcription factors might regulate Tfh differentiation upstream of Bcl6, as also recently observed for TCF-1 and LEF-1 transcription factors (13, 14). NFAT1 binding sites are also in the vicinity of the Cxcr5 and Slamf1 genes. NFAT1 did not bind to the Bcl6 locus in CD8 T cells (data not shown), but whether NFAT proteins bind to Bcl6 in CD4 T cells is unknown. Whether NFAT transcription factors cooperate with LEF-1, TCF-1, or other transcription factors to regulate Bcl6 expression and Tfh differentiation remains to be determined.
When comparing global gene expression patterns of genes differentially upregulated in human Tfh versus non-Tfh cells (19) (Fig. 4B), 40% of Tfh-associated upregulated genes possessed at least one NFAT1 binding site within 10 kb of their transcription start sites (Fig. 4B, dark blue dots). Furthermore, 47% of genes downregulated in Tfh compared with non-Tfh cells also possessed an NFAT1 binding site within 10 kb of their transcription start sites (Fig. 4B, dark red dots). Thus, these bioinformatics analyses suggest that NFAT family members extensively regulate Tfh-associated genes.
Overall, our results highlight important roles for NFAT family members in Tfh biology. NFAT expression was required for appropriate expression of ICOS and CD40L, which are essential for Tfh differentiation, and NFATs likely regulate additional essential components of Tfh differentiation based on bioinformatics analysis of NFAT ChIP-seq data in conjunction with Tfh gene expression profiles. It was recently reported that Ca2+ signaling was also necessary for Tfh IL-21 expression (16), suggesting that NFATs are required for many functions of Tfh cells during an antiviral response.
A previous study reported that NFAT2-deficient mice have enhanced GCs in the context of an immunization model; this was associated with a loss of follicular regulatory T cells through regulation of CXCR5 expression, but no direct impact on Tfh cells was ascribed to NFAT2 (15). In contrast, in the experiments reported in this article, NFAT2 positively regulates Tfh differentiation. More strikingly, acute infection of NFAT1,2-deficient mice demonstrated that NFAT1 and NFAT2 cooperate for the generation of Tfh cells during antiviral responses. One explanation for the discrepancies could be that immunization and viral infection elicit different in vivo responses. In fact, we have not observed follicular regulatory T cell differentiation by SMARTA CD4 T cells in LCMV infections (data not shown).
Consistent with the findings presented in this report, mice deficient in both Stim1 and Stim2, endoplasmic reticulum calcium sensors that control the gating of the store-operated calcium released–activated calcium channel ORAI1 upstream of NFAT activation (28), displayed decreased Ab responses and decreased CD40L expression in CD4+ T cells upon LCMV infection (29). Moreover, GC Tfh cells exhibit extensive calcium signaling in GCs (30, 31) and were also observed to have enhanced NFAT nuclear localization (16), in accord with our work. It has been previously observed that Tfh differentiation is associated with TCRs capable of greater MHC class II:peptide complex binding capacity (TCR dwell time) compared with other Th subsets (32, 33). Given that NFAT activity is TCR activation dependent, we infer from the Tfh defects associated with NFAT1,2 deficiency that NFAT1 and NFAT2 synergy might occur optimally in vivo in the context of extended TCR dwell times. Overall, our results have significant implications for therapeutic approaches to modulate B cell function by modulating the activity of NFAT transcription factors.
We thank Ryan Hastie for help in maintaining the mouse colony, and C. Kim, K. Gunst, and L. Nosworthy at the La Jolla Institute Flow Cytometry Facility for help with cell sorting experiments.
This work was supported by National Institutes of Health Grants R01AI 109842 (to A.R.), AI40127 (to A.R.), R01AI072543 (to S.C.), and U19 AI109976 (to S.C.); a postdoctoral fellowship from the Jane Coffin Childs Memorial Fund (to G.J.M.); and a postdoctoral fellowship from the Pew Latin American Fellows Program in the Biomedical Sciences (to R.M.P.).
The online version of this article contains supplemental material.
The authors have no financial conflicts of interest.