Abstract
The nuclear factor of activated T cells (NFAT) is involved in the transcriptional induction of cytokine and other immunoregulatory genes during an immune response. Among four distinct NFAT family members identified to date, mRNAs of NFAT1, NFATc, and NFATx are expressed in the thymus. Here, we report the distribution of these three NFAT family members in human fetal thymocyte subsets and in peripheral mature T cells. We show that NFATx mRNA was expressed in all T lymphocyte subsets tested and was highest in CD4+CD8+ double positive (DP) thymocytes. Conversely, NFAT1 mRNA was preferentially expressed in the mature CD4+ single positive (SP) populations. NFATc mRNA was present at low levels in all subsets but strongly induced upon treatment with phorbol ester and calcium ionophore. Interestingly, we detected NFAT-DNA binding complexes in DP thymocytes, albeit at lower levels than in CD4 SP cells. Corresponding to the mRNA expression, we observed that NFATx was responsible for the NFAT-DNA binding in DP thymocytes. Moreover, this DNA binding was inhibited by cyclosporin A, indicating that NFATx nuclear translocation was regulated by the calcineurin phosphatase in DP thymocytes. For the CD4 SP populations, NFAT1 and NFATc, and to some extent NFATx, were responsible for the NFAT-DNA binding complexes. These results indicate that NFAT family members are differentially regulated during the development of T cells, and that NFATx may play a distinct role in calcineurin-dependent signaling in DP thymocytes.
When stimulated via the TCR/CD3 complex in the presence of appropriate costimulatory signals, T cells become activated and produce an array of cytokines (1, 2, 3). This activation can be partially mimicked by the combined action of phorbol ester (such as PMA) and calcium ionophore (such as A23187). These reagents, respectively, induce activation of the protein kinase C/p21ras-dependent pathway and mobilization of intracellular calcium (4, 5). The increase of intracellular calcium, in particular, leads to the activation of the calmodulin-dependent calcineurin phosphatase. Subsequently, calcineurin dephosphorylates and promotes the nuclear translocation of the nuclear factor of activated T cells (NFAT),3 which is thought to then regulate the induction of the IL-2 gene (5, 6, 7), as well as other immunoregulatory protein genes, including TNF-α, IL-4, IL-5, IL-3, granulocyte macrophage-CSF, CD40 ligand, and Fas ligand (8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18).
Recently, four distinct cDNAs encoding NFAT family proteins have been reported: NFAT1 (also known as NFATp), NFATc, NFAT3, and NFATx (also known as NFAT4/NFATc3) (19, 20, 21, 22, 23, 24). All of these NFAT family proteins are capable of binding to the distal NFAT-binding sequence in the IL-2 promoter, and activating transcription driven from the NFAT sites when they are overexpressed (19, 20, 21, 22, 23, 24). Their similar recognition sequence can be explained by the fact that they all contain a highly conserved DNA-binding domain, which is referred to as Rel similarity domain because of its weak similarity to the DNA-binding domain of the Rel family proteins (20, 21, 22, 25). In addition, transcription dependent on the distal NFAT site in the IL-2 promoter requires the participation of other nuclear components, which can be induced by the action of PMA. These components include the AP-1 (Fos/Jun) family of transcription factors, which bind cooperatively with NFAT proteins to composite NFAT/AP-1 sites (26, 27, 28). Another domain, known as the NFAT homology region (29), is found in the amino terminus of the NFAT proteins, and is important for the regulation of their nuclear translocation and interaction with calcineurin (30, 31, 32). As for the transcriptional activation function of NFAT proteins, transactivation domains have been found both in their amino-terminal and carboxyl-terminal regions (31, 33).4
Despite the similarities, each NFAT member contains specific characteristics that are expected to endow it with unique functions. A remarkable difference is the mRNA distribution of each NFAT family member among different human and murine tissues. mRNAs that encode NFAT1, NFATc, and NFATx are expressed in the spleen, the thymus, the peripheral blood leukocytes, and in various nonlymphoid tissues (19, 20, 21, 22, 23, 24). On the other hand, NFAT3 is strongly expressed outside of the immune system (21). Interestingly, within the immune system, NFATx is expressed at significantly higher levels in the thymus than in peripheral lymphoid tissues (peripheral leukocytes and spleen) (21, 22, 23, 24), whereas NFAT1 and NFATc mRNAs appear to be predominant in peripheral lymphoid tissues (19, 20, 21). Moreover, several splice variants for each NFAT protein have been reported (21, 24, 29, 34),4 increasing the complexity of NFAT regulation. With regard to NFATx, one of its isoforms, termed NFATx1, is expressed preferentially in the thymus and T cells, while the NFATx2 isoform is expressed preferentially in skeletal muscle.4
Within the thymus, T lymphocytes undergo a series of developmental events that culminate with the production and selection of T cells that are mature, self-tolerant, and self-MHC restricted. During this developmental program, signals via TCR are thought to play a central role (35, 36, 37). The stages of development can be followed by the surface expression of markers such as CD4 and CD8. Immature precursors of T cells, which are negative for TCR, are included in the CD4−CD8− double negative (DN) thymocytes and eventually differentiate into TCRαβlow double positive (DP) thymocytes. During the DP stage, TCR is expressed in the cell surface and its engagement in thymocytes can result in either functional differentiation (positive selection) or programmed cell death (negative selection), both of which are critical in the generation of mature TCRαβhigh CD4 single positive (SP) and CD8 SP T cells.
Notably, although DN and SP thymocytes are able to produce cytokines in response to stimulation, DP thymocytes show a striking loss of such ability (38, 39, 40, 41, 42). To elucidate the mechanism of this functional transition, the regulation of transcription factors, including NFAT, has been analyzed in murine thymocytes upon stimulation with either APC or phorbol esters combined with calcium ionophore. Stimulation-dependent NFAT-DNA binding was detected in DN thymocytes, as well as in CD4 SP and CD8 SP thymocytes; in contrast, NFAT-DNA binding was barely detected in DP thymocytes (39, 41, 42, 43). These results correlated the presence of NFAT binding with the ability of thymocyte subsets to produce cytokines. However, the members of the NFAT family involved in the NFAT-DNA binding activities in different thymocyte subsets have not been defined.
The aim of this study is to analyze the distribution of distinct NFAT family members (NFAT1, NFATc, and NFATx) in human fetal thymocytes, and to provide the basis for analysis of the functional differences among NFAT family members during T cell development. We show that the mRNA expression of the NFAT family members is differentially regulated in thymocytes and mature T cells. Furthermore, we analyzed NFAT-DNA binding in the nuclear extracts from thymocyte subsets by using Abs specific to each NFAT family protein, showing preferential NFATx-DNA binding activity in stimulated DP thymocytes.
Materials and Methods
Thymus and peripheral blood
Human fetal thymi, between 18 and 22 wk of gestational age, were obtained from Advanced Bioscience Resources (Alameda, CA), in compliance with regulations issued by the state and the federal government. Buffy coat cells from healthy donors were obtained from Stanford Blood Center (Stanford, CA). These materials were used with the approval of the committee for the Protection of Human Subjects at DNAX Research Institute (Palo Alto, CA).
Separation of peripheral T cells and thymocytes
Peripheral CD4 SP T cells and thymocyte populations were purified by using anti-CD4- and anti-CD8-coated immunomagnetic beads. To purify peripheral CD4 SP T cells, PBMC was isolated from buffy coats by density centrifugation over Ficoll-Hypaque (Sigma Chemical Co., St. Louis, MO), and was mixed with anti-CD4-coated Dynabeads (Dynal A.S., Oslo, Norway). The mixture was incubated while gently rotating for 1 h at 4°C, and CD4+ cells were recovered by a magnetic particle concentrator. The conjugated beads were removed from the cells using DETACHaBeads reagent (Dynal A.S.), according to the manufacturer’s recommendations.
Fresh human fetal thymi were gently disrupted on a sieve, and thymocytes were released into single cell suspension on ice. They were incubated with anti-CD8-coated Dynabeads (Dynal A.S.) as described above, and separated into CD8+ and CD8− fractions by using a magnetic-particle concentrator. To recover DP thymocytes, the CD8+ fraction was incubated with DETACHaBeads reagent, and released cells were further incubated with anti-CD4-coated Dynabeads, followed by their detaching. For CD4 SP thymocytes, a similar procedure was performed on the CD8− fraction. To obtain DN thymocytes, a CD8− fraction was further incubated with anti-CD4- and anti-CD8-coated Dynabeads to complete the removal of the other populations. Residual erythrocytes and dead cells were removed by density centrifugation over Ficoll-Hypaque. Resulting populations were routinely checked for their purity by a FACScan (Becton Dickinson, Mountain View, CA), using phycoerythrin (PE)-conjugated anti-CD4 and FITC-conjugated anti-CD8 mAbs (Becton Dickinson). The purity of each population was >95% (Fig. 1).
Cell stimulation
Isolated thymocytes and peripheral T cells were cultured for 3 h (or 6 h for analysis of cytokine mRNA expression) in DMEM supplemented with 10% heat-inactivated FCS, antibiotics, 1 mM l-glutamine, and 1 mM sodium pyruvate, in the presence of PMA (5 ng/ml) and A23187 (0.25 μM), at 37°C in a humidified atmosphere with 5% CO2 in the air. PMA and A23187 were purchased from Calbiochem (La Jolla, CA). Where indicated, 1 μg/ml of cyclosporin A (CsA; Sandoz, Basel, Switzerland) was added to the cell suspension from 10 min before stimulation.
Competitive reverse transcriptase (RT)-PCR
Total RNA from stimulated and unstimulated cells was prepared by using RNA-STAT60 (Tel-Test, Friendswood, TX), and was reverse transcribed by using oligo(dT) and Superscript II reverse transcriptase (Life Technologies, Grand Island, NY), according to the manufacturer’s instructions. Competitive PCR was performed by using multispecific internal competitor DNA fragments, which contain various cDNA sequences of interest and which yield different-sized PCR products than do endogenous cytokine cDNA, as described previously (44, 45, 46). A multispecific internal competitor DNA construct for cytokine genes was kindly provided by Dr. H. Groux (DNAX Research Institute, Palo Alto, CA), and the sequences of PCR primers specific to human IL-2, IL-4, and IFN-γ are described elsewhere (46, 47). The internal standard for NFAT genes was generated using a PCR-based construction, by connecting the following sequences of the specific primers: NFATx, sense: 5′-ACCAGCCCGGGAGACTTCAATAGA-3′, antisense: 5′-AAATACCTGCACAATCAATACTGG-3′; NFATc, sense: 5′-GCCGCAGCACCCCTACCAGT-3′, antisense: 5′-TTCTTCCTCCCGATGTCCGTCTCT-3′; NFAT1, sense: 5′-AGAAACTCGGCTCCAGAATCC-3′, antisense: 5′-TGGTTGCCCTCATGTTGTTTTT-3′. The sense-primer DNA sequences of NFATx, NFATc, and NFAT1 were connected in this order, followed by the reverse complementary sequences of antisense primers in the same order. A linker sequence of an irrelevant DNA (β-galactosidase) was included to yield approximately 350 bp of amplified products (100–170 bp shorter than the corresponding RT-PCR products from the cDNA of NFATx, NFATc, and NFAT1).
cDNA was adjusted to concentrations equal to those of a housekeeping gene (β-actin) by performing competitive PCR between β-actin cDNA and internal competitor DNA. These adjusted volumes of cDNA were then used to quantitate genes of interest using serial dilutions of competitors in each reaction in the presence of specific primers. PCR cycles were 30 s at 94°C, 30 s at 60°C, and 30 s at 72°C, with appropriate cycles for each amplification (see Figs. 2 and 3). PCR products were analyzed on a 2% NuSieve agarose gel (FMC BioProducts, Rockland, ME) and stained with SYBR Green I DNA-staining dye (Molecular Probes, Eugene, OR) for the estimation of relative intensity of the different bands on a STORM 860 image analyzer (Molecular Dynamics, Sunnyvale, CA). The ratios of amplified cDNA to amplified standards were plotted against the different dilutions of the standards with known concentrations, from which the relative amount of cDNA was deduced. The concentrations of cDNA were calculated and displayed as relative values to the amounts of β-actin cDNA.
Plasmid construction and expression of NFAT proteins
pME-NFATx and pME-NFATc are expression plasmids containing human NFATx and NFATc full-length cDNA, under the control of the SRα promoter in the pME18S mammalian expression vector (22). pME-P(ZR24) contains a truncated sequence of human NFAT1 cDNA, corresponding to the nucleotide residue 1345–2587 of the published human NFAT1 cDNA (29) (covering its Rel similarity domain and a part of the carboxyl-terminal domain). The original cDNA clone pZR24 was isolated from a Jurkat cell cDNA library (22), and the PCR-based construction of pME-P(ZR24) was conducted with this template and the following oligonucleotide primers: sense 5′-CTCGAGCCATGTGGCCCAAGCCGCTGG-3′ and antisense 5′-CTCGAGCCTATGAGTAGTGGATCACAGG-3′ (XhoI sites are underlined). The PCR product was ligated into the pME18S vector after cleavage by XhoI, and the construction was verified by DNA sequencing. Transfection of plasmids into COS7 cells by the DEAE-dextran method, and the subsequent preparation of cytosolic extracts were conducted as described previously (22, 48).
Preparation of nuclear extract
Nuclear extracts from human fetal thymocytes and peripheral T cells were prepared as described elsewhere (23, 49, 50), with modifications. In brief, 4 × 107 cells were washed twice with cold PBS and resuspended in buffer A (10 mM HEPES, pH 7.6, 15 mM KCl, 2 mM MgCl2, and 0.1 mM EDTA), followed by 10 min of incubation on ice. Then, an equal amount of buffer B (buffer A + 0.4% Nonidet P-40 (Calbiochem)) was added and mixed, and the resulting nuclei were pelleted by brief, low-speed centrifugation. The nuclear pellets were washed with buffer A, resuspended in 200 μl of buffer C (50 mM HEPES, pH 7.9, 50 mM KCl, 0.1 mM EDTA, and 10% glycerol) containing 0.3 M ammonium sulfate (pH 7.9), and rotated at 4°C for 30 min. The nuclear debris was pelleted by high-speed centrifugation at 70,000 rpm for 45 min. Then, the proteins in the supernatant were precipitated by the addition of an equal volume of 3.0 M ammonium sulfate, followed by centrifugation at 50,000 rpm for 20 min. The resulting nuclear proteins were resuspended in buffer C, and were stored at −80°C until use. All the buffers were supplemented with 1 mM DTT, 1 mM PMSF, 2 mg/ml pepstatin A, and 2 mg/ml leupeptin. Protein concentration was estimated by Bradford assay (Bio-Rad Laboratories, Hercules, CA). Nuclear extracts from PMA-stimulated HeLa cells were prepared by a procedure previously described (51) .
Anti-NFAT Abs
Antihuman NFATx antisera were generated by immunizing rabbits with bacterially produced recombinant proteins derived from human NFATx. An antiserum, termed α-SB, raised against a recombinant protein (amino acid residues 729–893 of the human NFATx protein (22)), was used in this study. The anti-NFATc mAb and the anti-NFAT1 antiserum were purchased from Upstate Biotechnology, Inc. (Lake Placid, NY) and Affinity Bioreagents, Inc. (Golden, CO), respectively.
Electrophoretic mobility shift assay (EMSA)
Protein-DNA binding reaction was conducted in the presence of 10 mM Tris (pH 7.9), 100 mM KCl, 1 mM EDTA, 0.25 mg/ml BSA, and 10% glycerol with 0.75 mg of poly(dI-dC) (Pharmacia Biotech Inc., Piscataway, NJ) in 15 μl of reaction mixture. A total of 2 μg of nuclear extracts were incubated with a radiolabeled double stranded oligonucleotide probe (1 ng) for 20 min at room temperature, along with cold oligonucleotide competitors (20 ng), if any. For supershift assays, an anti-NFAT Ab/antiserum, or a control antiserum, was then added to the reaction mixture, followed by an additional 20 min of incubation at room temperature. The resulting protein-DNA complexes were separated by electrophoresis on a nondenaturing 4% polyacrylamide gel supplemented with a 0.25× Tris-borate-EDTA (TBE) buffer, for 2 h at 120 V at room temperature. The double strand oligonucleotides used in this study are as follows: the distal NFAT site from the human IL-2 promoter, 5′-gatcGGAGGAAAAACTGTTTCATACAGAAGGCGT-3′; Sp-1 site, 5′-ggATTCGATCGGGGCGGGGCGAGC-3′; canonical AP-1 site, 5′-tcgaGCTATGACTCATCCG-3′ (only sense strands are shown; sequence overhangs are lowercased).
Results
Differential expression of NFAT family genes in thymocyte populations
We and other groups have shown that the mRNAs of NFAT1, NFATc, and NFATx are expressed in the thymus (20, 21, 22, 23, 24). We have also observed that NFATx mRNA is expressed more predominantly in the thymus than in peripheral blood leukocytes and spleen (22, 24). Conversely, NFAT1 appears more predominant in peripheral blood leukocytes and the spleen than in the thymus (21, 22, 52). These observations suggested different NFAT family members may be involved in T lymphocyte populations of different developmental stages. To analyze the expression of NFAT family members during T cell development, we isolated DN, DP, and CD4 SP thymocytes from human fetal thymi. Peripheral CD4 SP T cells, isolated from buffy coat cells, were also included for comparison. The isolated populations were more than 95% pure when analyzed on a FACS for their surface expression of CD4 and CD8 (Fig. 1). In addition to surface markers, we characterized the isolated populations by measuring their ability to produce cytokines in response to stimulation with phorbol ester (PMA) and calcium ionophore (A23187). The expression of the IL-2, IL-4, and IFN-γ mRNAs in each population was semiquantitatively measured by competitive RT-PCR, using specific primers and competitor-DNA fragments designed for these cytokine genes (Fig. 2 and Table I). As previously observed in murine thymocytes (39, 40, 41, 42), stimulation-dependent expression of IL-2, IL-4, and IFN-γ gene transcripts was observed in DN thymocytes, in CD4 SP thymocytes, and in peripheral CD4 SP T cells, when the cells were stimulated with PMA/A23187 for 6 h. In contrast, in DP thymocytes, the expression of all the cytokine mRNAs tested was below the level of detection under similar stimulation conditions. In addition, the cytokine-gene expression was significantly higher (10–20-fold) in peripheral CD4 SP T cells than in CD4 SP and DN thymocytes (Table I).
Cells . | Cytokine mRNA Expression (1000 × Molar Ratios to β-Actin mRNA)b . | . | . | . | . | . | |||||
---|---|---|---|---|---|---|---|---|---|---|---|
. | IL-2 (22 cycles) stimulation . | . | IL-4 (28 cycles) stimulation . | . | IFN-γ (24 cycles) stimulation . | . | |||||
. | 0 h . | 6 h . | 0 h . | 6 h . | 0 h . | 6 h . | |||||
DN thymocytes | ND | 843 | ND | 9.2 | ND | 70.7 | |||||
DP thymocytes | ND | ND | ND | ND | ND | ND | |||||
CD4SP thymocytes | ND | 1,588 | ND | 2.8 | ND | 41.3 | |||||
Peripheral CD4SP T cells | ND | 24,630 | ND | 56.3 | ND | 851.4 |
Cells . | Cytokine mRNA Expression (1000 × Molar Ratios to β-Actin mRNA)b . | . | . | . | . | . | |||||
---|---|---|---|---|---|---|---|---|---|---|---|
. | IL-2 (22 cycles) stimulation . | . | IL-4 (28 cycles) stimulation . | . | IFN-γ (24 cycles) stimulation . | . | |||||
. | 0 h . | 6 h . | 0 h . | 6 h . | 0 h . | 6 h . | |||||
DN thymocytes | ND | 843 | ND | 9.2 | ND | 70.7 | |||||
DP thymocytes | ND | ND | ND | ND | ND | ND | |||||
CD4SP thymocytes | ND | 1,588 | ND | 2.8 | ND | 41.3 | |||||
Peripheral CD4SP T cells | ND | 24,630 | ND | 56.3 | ND | 851.4 |
Cells were stimulated with PMA (5 ng/ml) and A23187 (0.25 μM) and harvested at 6 h of incubation. ND, not detected at the amplification with indicated cycles.
Cytokine mRNA expression was analyzed by competitive RT-PCR as described in Materials and Methods.
We then analyzed the mRNA levels of NFATx, NFATc, and NFAT1 in the same thymocyte subsets and peripheral CD4 SP T cells. For this competitive RT-PCR, we designed primer pairs specific for each NFAT family member as well as an internal competitor DNA fragment to be included in all the relevant PCR reactions (Fig. 3,A). The input cDNAs made from total RNAs were normalized to contain similar amounts of cDNAs encoding the housekeeping gene, β-actin (see Materials and Methods). This allowed us to get a quantitative measurement for the expression levels of the different NFAT mRNAs (Fig. 3,B). Interestingly, DP thymocytes were the highest producers of NFATx mRNA among the subsets examined (Fig. 3,B, upper panel). The most immature DN thymocytes, as well as mature CD4 SP thymocytes and peripheral CD4 SP T cells, expressed NFATx mRNA, but at levels that were one-half or one-third of the levels found in DP thymocytes. The expression of NFATx mRNA was either unaffected or slightly attenuated after stimulation with PMA/A23187. In contrast to NFATx, NFAT1 mRNA was expressed at higher levels in peripheral CD4 SP T cells and CD4 SP thymocytes than in the immature DN and DP thymocytes (Fig. 3,B, middle panel). In CD4 SP T cells, a decrease of NFAT1 mRNA expression was observed after stimulation for 3 h; such a change was not evident in other populations. NFATc mRNA, on the other hand, was induced in all the populations examined; NFATc was expressed weakly in unstimulated cells, and was induced severalfold after 3 h of stimulation (Fig. 3 B, bottom panel). Similar inducible expression of NFATc has been reported previously using several lymphoid cells and T cell lines (20, 22). Taken together, we found that mRNA expression of each NFAT family member was regulated differently during T cell development in the thymus.
Because the efficiency of PCR for each NFAT cDNA is not exactly equal, it is difficult to compare the levels of mRNAs across NFAT species. Nevertheless, control experiments using plasmid DNAs bearing cloned NFAT cDNAs indicated that the PCR efficiency did not vary more than twofold for the different NFAT members (data not shown). Thus, the general trend of these results also coincides with those obtained previously by Northern blots (21, 22). That is, our results show that NFATx was the most abundantly expressed NFAT species in all unstimulated thymocytes and that comparable levels of NFATx and NFAT1 mRNAs were seen in the periphery (Fig. 3,B). Also of interest is the fact that the levels of NFATc mRNA in stimulated DP thymocytes were comparable to that of NFATx (Fig. 3 B, lower panel).
Stimulation-dependent NFAT-DNA complex formation from the nuclear extract from total thymocytes
Since each thymocyte subset expressed different sets of NFAT family members, we next analyzed the NFAT proteins in thymocytes. We first examined conditions for isolating the NFAT-DNA complex in unfractionated total thymocytes. Human fetal thymocytes were stimulated with PMA/A23187 or were left unstimulated, and nuclear extracts were prepared from these cells using a modified procedure that included ammonium-sulfate fractionation (see Materials and Methods). EMSA was then performed by using a double stranded oligonucleotide encompassing the distal NFAT-binding site from the human IL-2 promoter as a probe. The NFAT-DNA binding complex was observed in the nuclear extract from 3-h stimulated total thymocytes, but not in unstimulated cells (Fig. 4, lanes 1 and 2). The amount of binding complex from total thymocytes was comparable to that formed by an equivalent quantity of nuclear extract from stimulated peripheral CD4 SP T cells (lane 7). In addition, we show that the complexes were quenched by a 20-fold excess amount of unlabeled oligonucleotides carrying either the NFAT or AP-1 site, but not by an irrelevant oligonucleotide carrying the Sp-1 site (lanes 3-5 and 8-10). Thus, specific NFAT-DNA binding complexes are inducible in total thymocytes, as well as in peripheral CD4 SP T cells.
NFAT family proteins in the NFAT-DNA complex from total thymocytes
To determine which NFAT family members are involved in the complexes from thymocytes, we tested Ab/antisera against NFAT proteins for their specificity to the different human NFAT proteins. As sources for the NFAT proteins, we used cytosolic fractions of COS7 cells that were transiently transfected with pME-NFATx1, pME-NFATc, or pME-P[ZR24], which are mammalian expression plasmids of human NFATx, NFATc, or a truncated form of NFAT1, respectively. The addition of an anti-NFATx antiserum (termed α-SB, raised against a truncated NFATx protein) to the binding reaction significantly altered (supershifted) the electrophoretic mobility of the NFATx-DNA binding complex (Fig. 5, lane 3), but not the NFATc (lane 7) or NFAT1 (lane 11) binding complexes. This indicated that this antiserum specifically recognized the human NFATx protein. Similarly, the addition of the anti-NFAT1 antiserum specifically supershifted the NFAT1-DNA binding complex (lanes 5, 9, and 13). The anti-NFATc Ab also specifically supershifted the NFATc-DNA complex (lanes 4, 8, and 12), although the retardation of the mobility was less than in the case of the anti-NFATx and anti-NFAT1 antisera. This is probably a reflection of the fact that the anti-NFATc mAb recognized a single epitope on NFATc, and that the predicted molecular size of NFATc is smaller than that of NFATx and NFAT1 (7, 20).
We then analyzed the NFAT proteins involved in the NFAT-DNA complexes from total thymocytes as well as peripheral CD4 SP T cells (Fig. 6). A small portion of the NFAT-DNA band of stimulated total thymocytes was supershifted by the anti-NFATx antiserum (lane 2). In stimulated peripheral CD4 SP T cells, supershift by the anti-NFATx antiserum was also observed, although it was weaker than in the thymocytes (lane 7). Parts of NFAT-DNA complexes from both stimulated total thymocytes and CD4 SP T cells were also supershifted by the addition of a saturating amount of anti-NFATc Ab (lanes 3 and 8), although the intensity of the resulting supershifted band was more evident in total thymocytes. In contrast, addition of the anti-NFAT1 antiserum resulted in more of the band being supershifted in stimulated CD4 SP T cells than in total thymocytes (lanes 4 and 9). These observations indicate that at least three NFAT1 family members were found in the NFAT complex in total thymocytes, and that NFATx-DNA binding is observed preferentially in thymocytes, while NFAT1 is preferentially observed in CD4 SP T cells.
NFATx is the major NFAT protein involved in the stimulation-dependent NFAT-DNA complex in DP thymocytes
The above observations prompted us to investigate which NFAT family member contributes to the formation of NFAT-DNA complexes in different developmental stages of T cells. To this end, we prepared nuclear extracts from isolated DP and CD4 SP thymocytes before and after 3 h of stimulation with PMA/A23187. The nuclear extract from stimulated CD4 SP thymocytes demonstrated significant DNA-binding activity to the NFAT oligonucleotide probe (Fig. 7 A, lane 7), as previously reported (39, 41, 42). In addition, and contrary to previous reports (39, 41, 42, 43), NFAT-DNA binding activity was also detected in the nuclear extract from stimulated DP thymocytes (lane 2). These NFAT-DNA binding complexes were sequence-specific and quenched by unlabeled NFAT and AP-1 oligonucleotides (lanes 3-5 and 8-10), as in the case of total thymocytes. The electrophoretic mobility of the NFAT-DNA complex from the DP thymocytes was slightly slower than in the case of the CD4 SP thymocytes (compare lanes 2 and 7), suggesting the involvement of distinct components of the NFAT complexes between these populations. EMSA analysis of DN thymocytes was not available due to insufficient yield of this population.
We then analyzed the NFAT-DNA complexes in nuclear extracts from DP thymocytes and CD4 SP thymocytes, using Ab/antisera specific to NFAT family members (Fig. 7 B). Remarkably, the band of NFAT-DNA complex from stimulated DP thymocytes was significantly supershifted only by the addition of the anti-NFATx antiserum (lane 2). In contrast, the complex from DP thymocytes was not supershifted by the anti-NFATc Ab or the anti-NFAT1 antiserum (lanes 3 and 4), indicating that NFATx was the protein that primarily contributed to the formation of the NFAT-DNA complex in DP thymocytes. On the other hand, the band of NFAT-DNA complex formed in stimulated CD4 SP thymocytes was supershifted mainly by the anti-NFATc Ab and the anti-NFAT1 antiserum, and very weakly supershifted by the anti-NFATx antiserum (lanes 7-9). Therefore, similar to the mRNA expression data shown above, these results demonstrate that different NFAT proteins were involved in the NFAT-DNA binding complexes from distinct thymocyte subsets.
NFAT-DNA binding activity in DP thymocytes is regulated by cyclosporin A-sensitive signaling
Because the NFAT-DNA binding activity in stimulated DP thymocytes contained only the NFATx protein, we examined whether the regulation of NFAT-DNA binding activity in DP thymocytes was similar to that in mature T cells (Fig. 8). We show that addition of CsA, which is known to inhibit the nuclear translocation of NFAT proteins (5, 7), to the DP thymocytes significantly inhibited appearance of stimulation-dependent NFAT-DNA binding in the nuclear extract from these cells (Fig. 8). Thus, NFAT proteins in both immature DP thymocytes and mature CD4 SP T cells are regulated by a similar CsA-sensitive signaling pathway.
Discussion
In this report, we have shown that mRNAs of NFAT family members are differentially expressed during T cell development in the human thymus. We also analyzed the stimulation-dependent NFAT-DNA binding complexes from different T lymphocyte subsets. In mature CD4 SP thymocytes and peripheral CD4 SP T cells, NFAT1 and NFATc proteins largely participated in the NFAT-DNA binding activities (Fig. 4). This finding is in agreement with a recent immunoprecipitation analysis of NFAT proteins in peripheral T cells, showing abundance of NFAT1 and NFATc proteins in these cells (53). In addition, we have shown that, contrary to previous reports (39, 41, 42, 43), there exists an NFAT-DNA binding activity that can be induced in DP thymocytes upon stimulation. Remarkably, this NFAT-DNA complex contained primarily the NFATx protein.
Concerning the difference among previous observations in DP thymocytes, we can list three main factors, which are nonexclusive, that might have contributed to our detection of NFAT-DNA binding activity in this subset. First, the nuclear extracts in this study were prepared by a procedure that included fractionation with ammonium sulfate (see Materials and Methods). By this fractionation, NFAT proteins were enriched in our nuclear extracts. Second, we characterized NFAT-DNA binding complexes by EMSA using the low ionic-strength buffer 0.25× TBE. When we performed similar assays with a different buffer, such as a Tris-glycine buffer, NFAT-DNA binding activity in DP thymocytes was poorly observed (data not shown). Third, it is also possible that one or more factors that could inhibit NFAT-DNA binding in DP thymocytes were removed during the ammonium-sulfate fractionation of the nuclear extract. Currently, however, we have no clear evidence regarding the presence of such inhibitors. Altogether, it is likely that our procedures raised the level of detection for a weak NFAT-binding activity in DP thymocytes. On this note, it has been reported that the affinity of the NFATx protein to the IL-2 NFAT site is lower than that of the other NFAT family proteins (21, 23).
Moreover, the NFATx-DNA binding in DP thymocytes was quenched by the presence of AP-1 oligonucleotides (Fig. 6 A), postulating that AP-1 or AP-1-related factors participated in this protein-DNA complex. AP-1-DNA binding has been reported to be remarkably reduced in the DP subset (41, 42), although it was detectable in other studies (43, 54, 55). Our procedures also increased the level for detection of AP-1-DNA binding (data not shown). Presumably, this also could have aided in our detection of NFAT-DNA binding.
Our data indicate that NFAT1, NFATx, and NFATc may have distinct roles during the development of T cells. With respect to NFAT1, its mRNA was shown to be expressed well in mature CD4 SP thymocytes as well as in peripheral CD4 SP T cells, but at very low levels in immature DN and DP thymocytes (Fig. 3). In agreement with the mRNA data, NFAT1-DNA binding was clearly detected in CD4 SP populations but not in DP thymocytes (Figs. 6 and 7 B). These data suggest that NFAT1 may not be involved in the development of immature thymocytes. This is consistent with recent observations in NFAT1-deficient mice, in which thymic development proceeded normally while splenic T cells from these mice showed an early impaired production of cytokines such as IL-4, granulocyte macrophage-CSF, and TNF-α upon ligation of their TCR (18, 56).
As for NFATx, its mRNA expression was observed in all subsets tested but showed a striking predominance in DP thymocytes (Fig. 3). In accordance with this, NFATx-DNA binding in response to the TCR-mimicking stimulation by PMA/A23187 was detected clearly in DP thymocytes but only weakly in CD4 SP populations (Fig. 7,B). No NFAT1- or NFATc-DNA binding was detected in DP thymocytes that had been stimulated for 3 h, suggesting a unique role for NFATx in these cells. However, DP thymocytes lack the ability to produce cytokines upon stimulation (Refs. 38–42 and Table I). Thus, although NFATx can induce transcription driven by NFAT sites when transiently overexpressed (21, 22, 24), it is evident that NFATx-DNA binding is not sufficient to activate cytokine expression in DP thymocytes. That is, additional transcriptional activation factors may be required, or alternatively, inactivation of inhibitory activities may be needed.
NFATx has been reported to bind the distal NFAT site in the IL-2 promoter and an NFAT site in the IL-4 promoter with weaker affinity than NFAT1 and NFATc (21, 23), suggesting that the DNA-binding specificity of NFATx is slightly different from that of NFAT1 and NFATc. Consequently, NFATx may act through unspecified alternative DNA sequences different from the NFAT sites that are currently known in cytokine gene promoters. NFATx may then regulate the expression of additional or a different pool of genes. On this point, recent studies have indicated that both p21ras-dependent and calcineurin-dependent pathways are involved in positive selection, a TCR-mediated process that occurs during the differentiation of DP thymocytes into MHC-restricted CD4 SP or CD8 SP mature thymocytes (reviewed in 57 . More specifically, CsA and FK506, inhibitors of calcineurin, have been shown to block positive selection but not negative selection (58, 59). We showed that the regulation of NFATx in DP thymocytes was dependent on a calcineurin-dependent pathway, since it was inhibited by the action of CsA (Fig. 8). We have also recently shown that nuclear translocation of NFATx protein is regulated through direct interaction with calcineurin (31). Thus, it is tempting to speculate that NFATx could be involved in TCR signaling-dependent and CsA-sensitive events in DP thymocytes, including positive selection. This may also include signaling through the pre-TCR complex, which is expressed in immature thymocytes before the TCR (60, 61), and seems to have a critical role in T lymphocyte differentiation.
We also observed that NFATx mRNA is abundantly expressed in unstimulated DN thymocytes (Fig. 3). The modest amounts of NFAT1 mRNA in DN, as in DP thymocytes, appear inadequate to produce detectable NFAT1-DNA binding activity. Consequently, it is likely that NFATx will also be the major NFAT-DNA binding species in DN thymocytes after stimulation. Thus, in contrast to NFAT1, NFATx, by virtue of its expression pattern, may play an important role throughout T cell development. Interestingly, DN thymocytes possess the potential to express several cytokines when stimulated (Fig. 2 and Table I). Hence, if NFAT complexes are required for this coordinate induction, then NFATx may be the major NFAT species regulating this early induction. Likewise, NFATx may play a role in the expression of NFATc in the thymus since NFATc mRNA induction is dependent on a calcium-regulated signaling pathway (Fig. 3).
It is also notable that NFATx is detected in mature CD4 SP T cells at the mRNA (Fig. 3) and at the protein level (53), albeit at lower levels than in DP thymocytes. These findings suggest that NFATx may play a role not only in the function of DN and DP thymocytes, but also in the function of mature T cells. Further investigation, such as studies using NFATx-deficient mice, may provide information to elucidate the functions of NFATx.
Finally, the regulation of NFATc in thymocyte subsets is also of interest. In agreement with previous observations (20, 22), NFATc mRNA was shown to be induced upon stimulation with PMA/A23187, not only in CD4 SP, but also in DN and DP thymocytes (Fig. 3). This suggests that NFATc may also be involved in the development of immature thymocytes, and that NFATc mainly affects gene expression at a later point in time, after its induction. However, NFATc regulation is more complex. Despite significant levels of NFATc mRNA expression (comparable to that of NFATx) in stimulated DP thymocytes, no NFATc-DNA binding was detected in these cells after stimulation for 3 h (Fig. 7 B) and 6 h (data not shown). Moreover, Lyakh, Ghosh, and Rice (53) recently reported that despite significant levels of NFATc protein in stimulated peripheral T cells, NFATc-DNA binding to the IL-2 NFAT binding site was poorly detected in these cells. Thus, it appears that there are additional factors regulating NFATc-DNA binding. Currently, we do not know whether these regulations are specific to the IL-2 NFAT binding site. Evidently, further work is required to understand this inhibition.
In summary, the present study describes for the first time the involvement of particular NFAT family members in distinct human thymocyte subsets. We have clearly shown that the major NFAT family members involved in the NFAT-DNA binding complexes differ dramatically between immature DP thymocytes and mature CD4 SP populations. Our observations provide a framework for further studies pertaining to calcineurin-regulated signaling pathways that impinge upon transcriptional regulation by NFAT proteins during T cell development in the thymus.
Acknowledgements
We thank Dr. Hyun-Jun Lee for critical comments and help throughout the study, and Dr. Shin-ichiro Imai for initial characterization of anti-NFATx antisera. We also thank Drs. Maria-Grazia Roncarolo and Hervé Groux for advice on the treatment of human fetal thymocytes and semiquantitative RT-PCR, and Drs. Dan Chen and René de Waal Malefyt for critical discussion. We are also grateful to Debra Ligget for oligonucleotide synthesis, and to Allison Helms for DNA sequencing. We thank Dr. Maribel Andonian and Gary Burget for graphical support, and Margaret Tanner Angelopoulos for critical reading of the manuscript.
Footnotes
DNAX Research Institute of Molecular and Cellular Biology is supported by Schering-Plough Corporation.
Abbreviations used in this paper: NFAT, nuclear factor of activated T cells; DN, double negative; DP, double positive; SP, single positive; PE, phycoerythrin; CsA, cyclosporin A; RT-PCR, reverse transcriptase-polymerase chain reaction; EMSA, electrophoretic mobility shift assay.
Imamura, R., E. S. Masuda, Y. Naito, S. Imai, T. Fujino, T. Takano, K. Arai, and N. Arai. Carboxy-terminal 15 amino acid sequence of NFAT×1 possibly created by tissue-specific splicing is conserved among NFAT family proteins and is essential for transactivation activity in T cells. Submitted for publication.