We report that NF-AT1 and NF-AT4 are expressed cytoplasmically in resting eosinophils, whereas NF-AT2 and NF-AT3 have not been seen. Likewise, NF-AT1 mRNA and NF-AT4 mRNA have been detected in resting eosinophils, and their levels can be significantly up-regulated by the Th2-associated cytokines IL-4 and IL-5. There is no detectable NF-AT protein expression in the nuclei of resting eosinophils. However NF-ATs appear in the nuclei of IL-4-, IL-5-, or ionomycin-stimulated eosinophils. Only NF-AT1 and NF-AT4, but not NF-AT2 and NF-AT3, have translocated into the nuclei in IL-4- or IL-5-stimulated eosinophils. These findings delineate a novel pathway in the cytokine network in which Th2 lymphocytes “control” eosinophils via the release of IL-4 and IL-5, and activation of NF-AT in eosinophils. The findings also suggest that a later feedback “talking” may exist between eosinophils and Th2 lymphocytes.
Nuclear factor of activated T cells is a family of transcription factors implicated in the control of cytokine and early immune response gene expression. NF-AT1-deficient mice are prone to developing a classically allergic phenotype characterized by eosinophilia and increased production of Th2-associated cytokines (1). Mice lacking NF-AT1 and NF-AT4 have allergic blepharitis, interstitial pneumonitis, and an increase in serum IgG1 and IgE levels, secondary to a selective increase in Th2-associated cytokines (2). NF-AT proteins are expressed not only in T lymphocytes but also in other cells including B lymphocytes, NK cells, mast cells, monocytes, thymus and spleen cells (3), and vascular smooth muscle cells (4). Thus far, it has not been determined whether NF-AT proteins are expressed in eosinophils. In the present study, we have investigated the expression of NF-AT proteins and the corresponding mRNA in peripheral eosinophils and the regulation of these proteins by IL-4 and IL-5.
Materials and Methods
Purification of eosinophils
Eosinophils were purified from healthy, nonallergic volunteers as described in detail elsewhere (5). Briefly, a Percoll gradient (1.082 g/ml; Pharmacia, Uppsala, Sweden) was used to enrich eosinophils; anti-CD16-coated MACS magnetic particles (Miltenyi Biotech, Bergisch Gladbach, Germany) were used to deplete neutrophils. Eosinophil purity was invariably ≥97% (lymphocytes <0.5%). The whole procedure was conducted at 4°C in a Ca2+-free and Mg2+-free medium.
The purified eosinophils were spun down on a slide, fixed, and immersed in 1% BSA blocking buffer for 10 min to avoid unspecific binding; next, primary Ab (20 μg/ml of either anti-NF-AT1 (mouse mAb 4G6-G5), anti-NF-AT2 (mouse mAb 7A6), anti-NF-AT3 (goat pAb3 C-20), or anti-NF-AT4 (goat pAb C-20); Santa Cruz Biotechnology, Santa Cruz, CA) was added. Eosinophils were then incubated overnight at 4°C, followed by the addition of secondary Ab, and were visualized by an alkaline phosphatase staining system (Dako, Glostrup, Denmark).
Real time quantitative RT-PCR assay
All real time quantitative RT-PCRs were performed as described elsewhere (6). Briefly, total RNA was reverse transcribed and submitted to real time quantitative PCR in an ABI Prism 7700 Sequence Detector System (Perkin-Elmer, Norwalk, CT). By using a SYBR Green PCR Core Reagents Kit (Perkin-Elmer), fluorescence signals were generated during each PCR cycle via the 5′ to 3′ endonuclease activity of AmpliTaq Gold to provide real time quantitative PCR information. The following sequences of the specific primers (Amersham Pharmacia Biotech, Little Chalfont, U.K.) were used: NF-AT1 sense, 5′-AGAAACTCGGCTCCAGAATCC-3′; NF-AT1 antisense, 5′-TGGTTGCCCTCATGTTGTTTTT-3′; NF-AT2 sense, 5′-GCCGCAGCACCCCTACCAGT-3′; NF-AT2 antisense, 5′-TTCTTCCTCCCGATGTCCGTCTCT-3′; NF-AT3 sense, 5′-GGTTTCCCGGCCAGTCCAGGTCTA-3′; NF-AT3 antisense, 5′-AAGGGGCGGGGAAGGAAGGAAACT-3′; NF-AT4 sense, 5′-ACCAGCCCGGGAGACTTCAATAGA-3′; NF-AT4 antisense, 5′-AAATACCTGCACAATCAATACTGG-3′.
The target cDNA was adjusted to amounts equal to those of the housekeeping gene (β-actin) according to the manufacturer’s instructions. PCRs were optimized under the following conditions: 15 s at 95°C and 60 s at 60°C with 40 cycles for each amplification. To express the results, two terms were used: ΔRn, representing the normalized reporter signal minus the baseline signal established in the first few cycles PCR, and CT (threshold cycle), representing the PCR cycle at which an increase in reporter fluorescence signal above a baseline can first be detected.
Electrophoretic mobility shift assay (EMSA)
Nuclear extracts were prepared as described by McCaffrey et al. and Aramburu et al. (7, 8). Briefly, 400 μl of ice-cold Dignam buffer A and then 25 μl of 10% Nonidet P-40 were added into the cells. The cells were vortexed and centrifuged (9000 rpm, 30 s, 4°C). Pelleted nuclei were lysed in 50 μl of Dignam buffer C and centrifuged (12,000 rpm, 10 min, 4°C); the resulting supernatants were diluted (1/1) with Dignam buffer D. Double-stranded synthetic oligonucleotide DNA probes were end-labeled with [γ-32P]ATP (5000 Ci/mmol) and T4 polynucleotide kinase (Amersham Pharmacia Biotech Inc.). The sequences of the oligonucleotide probes used (5′ to 3′, one strand) were as follows (9, 10): human IL-2 distal NF-AT site (NF-AT huIL-2), GGAGGAAAAACTGTTTCATACAGAAGG (binding sequences are underlined); consensus AP-1 site, CGCTTGATGACTCAGCCGGAA. EMSA reactions were performed at room temperature in a final volume of 25 μl. Nuclear extracts (3 μg of protein per reaction volume) were incubated for 15 min in binding buffer (8), followed by the addition of 0.5 ng 32P-labeled probes to react for 15 min; next, the samples were electrophoresed on nondenatured 5% polyacrylamide gels in 0.25 Tris-borate-EDTA (TBE) buffer. In some experiments, cold probes used as competitors were added at the beginning of the reaction to identical aliquots of the extracts (100-fold higher than labeled probes). The mAbs used for supershift (all at 10 μg/ml) were incubated with the nuclear extracts on ice for 30 min.
Results and Discussion
Expression of NF-AT in eosinophils
By immunocytochemistry assay, we have observed that NF-AT1 (Fig. 1,A) and NF-AT4 (Fig. 1,D) are cytoplasmically expressed in resting eosinophils. NF-AT2 (Fig. 1,B) and NF-AT3 (Fig. 1,C) have not been detected. Fig. 1, E and F, are the mouse and goat isotype Ab negative control, respectively. We have also observed a tendency for NF-AT1 and NF-AT4 to be expressed in the nuclei in IL-4- and IL-5-activated eosinophils. NF-ATs have also been detected in peripheral T lymphocytes (data not shown).
mRNA of the NF-AT family are expressed in eosinophils and regulated by IL-4 and IL-5
NF-AT1 and NF-AT4 mRNA have been detected in resting eosinophils; NF-AT2 and NF-AT3 mRNA have not (Fig. 2,A). There are ∼1.9 × 103 copies for NF-AT1 and ∼6.7 × 102 copies for NF-AT4 in the resting eosinophils, whereas no copies coding NF-AT2 and NF-AT3 have been found within 40 cycles. NF-AT1 mRNA have been significantly up-regulated in IL-4- and IL-5-stimulated (all at 10 ng/ml) eosinophils (Fig. 2,B). There are ∼6.4 × 103 copies and ∼5 × 103 copies for NF-AT1, respectively. NF-AT4 mRNA have also been significantly up-regulated in the IL-4- and IL-5-stimulated eosinophils (Fig. 2,C). There are ∼3.0 × 103 copies and ∼4.4 × 103 copies for NF-AT4, respectively. A linear relationship between CT and log starting quantity of the standard DNA template or target cDNA (NF-AT1 and NF-AT4) has been detected (Fig. 2 D). We have also observed that NF-AT1, NF-AT2, NF-AT3, and NF-AT4 have been expressed in purified T lymphocytes, and their expression can be up-regulated by IL-4 and IL-5, respectively (data not shown).
IL-4- and IL-5-induced activation of NF-AT in eosinophils
There is no detectable NF-AT in the nuclei of resting eosinophils. NF-AT huIL-2 binding proteins have been detected in the nuclei of IL-4- or IL-5-stimulated (1 h) eosinophils, which indicates that IL-4 and IL-5 can induce NF-AT activation, resulting a translocation of NF-AT in eosinophils (Fig. 3,A). Ionomycin (Ion) (0.1 μg/ml) also activates NF-AT to nuclear translocation. Cyclosporin A (0.1 μg/ml) is a potent inhibitor of activation of NF-AT (Fig. 3,A). NF-AT complexes in IL-4-, IL-5-, or Ion-stimulated eosinophils have been completely and specifically competed by a 100-fold molar excess of the cold NF-AT huIL-2 probe, whereas no competition is induced by the cold AP-1 site probe. A 100-fold molar excess of the cold AP-1 site probe can completely and specifically compete the 32P-labeled AP-1 probe (Fig. 3,B). It is not clear why the cold AP-1 site probe does not compete the NF-AT huIL-2 probe in our experiments, because the binding complex probably contains both NF-AT and AP-1 (8, 9). NF-AT translocation into the nuclei of IL-4- and IL-5-stimulated eosinophils has been detected within 15 min and lasts for ≥120 min (Fig. 4,A). In the current study, AP-1 was seen in freshly isolated eosinophils; increased activity of AP-1 in nuclei of eosinophils after stimulation with IL-4 or IL-5 was observed for 120 min (Fig. 4,B). NF-AT1 complex was supershifted by the mouse anti-NF-AT1 mAb (Fig. 4,C). Neither isotype mouse IgG nor mouse anti-NF-AT2 mAb can affect the relative mobility of the NF-AT complex. These results confirm the presence of NF-AT1, but not NF-AT2, in the nuclei of IL-4- or IL-5-activated eosinophils. Likewise, the presence of NF-AT4, but not NF-AT3, in the nuclei of IL-4- or IL-5-activated eosinophils has also been confirmed (data not shown). The mechanism by which IL-4 and IL-5 induce the appearance of NF-ATs in the nuclei of eosinophils is unknown, because the cytokines are not known to induce a Ca2+ influx, which is thought to be necessary for nuclear translocation of NF-AT. It will be of interest to further investigate this mechanism. In Fig. 4 C, it appears that the entire band is supershifted with anti-NF-AT1. This finding seems to be in contrast to the conclusions for NF-AT4. The reason for this phenomenon may be that the NF-AT4 content in eosinophils is so small that the NF-AT1 band overlaps the NF-AT4 band.
In mice lacking NF-AT1, resulting an in vivo model of allergic inflammation, an increased pleural accumulation of eosinophils and serum IgE levels and an abnormal pattern of IL-4 gene expression are seen (11, 12). The eosinophilia and high levels of serum IgE in these mice are apparently dependent upon the overproduction of IL-4 and IL-5 (1). Mice lacking NF-AT4 have impaired development of CD4 and CD8 single-positive thymocytes and peripheral T cells as well as hyperactivation of peripheral T cells (13). Our findings mentioned above suggest that the NF-AT family play a critical role in inducible gene transcription during the immune response, especially in the allergic response. Our results show that eosinophils, one of the key cells in allergic inflammation, may not only be a passive target for the deficiency, dysfunction, or inactivation of the NF-AT family in Th2 lymphocytes, but may also be an active regulator during the initiation, development, limitation, and termination of allergic inflammation.
Kiani et al. demonstrated that NF-AT1 is also involved in down-regulating the late phase of IL-4 gene transcription-inhibiting Th2 responses. This is evidence that NF-AT proteins regulate not only the initiation but also the termination of gene transcription (14). Our results are in support of this observation in light of the biological functions of eosinophils and Th2 lymphocytes. Based on our results, we might propose that NF-AT could be deficient or genetically inactivated in certain circumstances such as allergic conditions. A challenge such as allergen stimulation initiates a cascade of events of allergic inflammation, including overproduction of IL-4 and IL-5, activation of eosinophils, and eosinophilia. An over-release of IL-4 and IL-5 could subsequently activate NF-AT in eosinophils. The cascade of biological events in eosinophils after NF-AT activation would result in later feedback “talking” to Th2 lymphocytes to induce secondary activation of NF-AT in the Th2 lymphocytes (may also be other types of immune cells). This later feedback talking could lead to the termination of allergic disorders. A schematic view for this proposal is illustrated in Fig. 5. It will be quite interesting to investigate the exact mechanism of the secondary activation of NF-AT in Th2 lymphocytes during the later feedback talking on the pathway in the cytokine network between eosinophils and Th2 lymphocytes, especially in cells from allergic subjects.
We thank Lisbeth Abrahamsen and Ulla Minuva for their excellent technical assistance.
T.J., S.Q., and A.M. are supported by generous grants from the Danish Allergy Research Center and from the Alfred Benzons Foundation.
Abbreviations used in this paper: pAb, polyclonal Ab; EMSA, electrophoretic mobility shift assay; Ion, ionomycin; NF-AT huIL-2, distal NF-AT site of the human IL-2 promoter; CT, threshold cycle.