Emerging evidence suggests a role for eosinophils in immune regulation of T cells. Thus, we sought to determine whether human eosinophils may exert their effect via differential generation of Th1 and Th2 chemokines depending on cytokines in their microenvironment and, if so, to establish the conditions under which these chemokines are produced. Eosinophils cultured with TNF-α plus IL-4 had increased mRNA expression and protein secretion of the Th2-type chemokines, CCL17 (thymus and activation-regulated chemokine) and CCL22 (macrophage-derived chemokine). Conversely, the Th1-type chemokines, CXCL9 (monokine induced by IFN-γ) and CXCL10 (IFN-γ-inducible protein-10), were expressed after stimulation with TNF-α plus IFN-γ. Addition of TNF-α appeared to be essential for IFN-γ-induced release of Th1-type chemokines and significantly enhanced IL-4-induced Th2-type chemokines. Inhibition of NF-κB completely blocked the production of both Th1 and Th2 chemokines. Activation of NF-κB, STAT6, and STAT1 was induced in eosinophils by TNF-α, IL-4, and IFN-γ, respectively. However, there was no evidence for enhancement of these signaling events when eosinophils were stimulated with the combination of TNF-α plus IL-4 or TNF-α plus IFN-γ. Thus, independently activated signaling cascades appear to lead to activation of NF-κB, STAT1, and STAT6, which may then cooperate at the promoter level to increase gene transcription. Our data demonstrate that TNF-α is a vital component for eosinophil chemokine generation and that, depending on the cytokines present in their microenvironment, eosinophils can promote either a Th2 or a Th1 immune response, supporting an immunoregulatory role for eosinophils.

There is growing evidence that eosinophils can contribute to the adaptive immune response by producing cytokines such as IL-4 and IFN-γ (1, 2, 3, 4, 5), and, as we have recently demonstrated (6), by regulating Th1 and Th2 cytokine generation by CD4+ T cells. Eosinophils have been primarily associated with Th2-type immune responses as follows: IL-5 is important for eosinophil maturation, differentiation, and survival, whereas IL-13 contributes to eosinophil recruitment to target organs, via induction of eotaxins (7).

Many factors contribute to the initiation and regulation of inflammatory cells. The characteristics of an inflammatory response are largely determined by the type of chemokines in the microenvironment. It is generally accepted that Th2 cells are preferentially attracted by CCL17 (thymus and activation-regulated chemokine) and CCL22 (macrophage-derived chemokine) (8, 9), whereas Th1 cells are preferentially recruited by CXCL9 (monokine induced by IFN-γ) and CXCL10 (IFN-γ-inducible protein-10) (8, 9, 10). Regulation of the generation of these chemokine is also distinct. CCL17 and CCL22 are typically induced by IL-4 (or IL-13) via a STAT6-mediated pathway, whereas CXCL9 and CXCL10 are secreted in response to IFN-γ via a STAT1-mediated pathway (11). A critical role for STAT1 and STAT6 in differential trafficking of Th1 and Th2 cells to the lung has been demonstrated in mice (12, 13). In STAT1-deficient mice, adoptively transferred Ag-specific Th1 cells failed to home to the airway in response to local allergen challenge. This failure to home was most likely due to the defect in Ag-induced expression of CXCL9 and CXCL10 in the airway, because intranasal administration of CXCL10 restored trafficking of Th1 cells in STAT1 knockout mice (12). Similarly, STAT6-deficient mice failed to express CCL17 and CCL22 and did not support trafficking of Th2 cells (12, 13). These studies provide further in vivo evidence for differential homing of Th1 and Th2 cells and suggest a link between STAT1 and induction of Th1-associated chemokines in comparison with STAT6 and Th2-associated chemokines. Thus, existing evidence suggests that an environment rich in IL-4 and IL-13 invokes STAT6-mediated generation of CCL17/CCL22, which in turn recruit Th2-type cells. Conversely, a Th1 environment is supportive of CXCL9/CXCL10 and further recruitment of Th1-type cells.

We and others have demonstrated the presence of increased levels of Th2-associated chemokines, CCL17 and CCL22, in bronchoalveolar lavage fluid of atopic asthmatic subjects after airway Ag challenge (14, 15, 16). Of note, bronchoalveolar lavage fluid concentrations of CXCL9 and CXCL10 were also markedly increased. Because allergen challenge induces a striking airway eosinophilia that is associated with Th2-type cytokines, including IL-5 and IL-13, the presence of high concentrations of Th1-type chemokines was not expected. Based on these observations, we designed studies to determine whether human eosinophils are a source of T cell-associated chemokines, to investigate the conditions under which these chemokines are released from eosinophils, and to establish several of the intracellular mechanisms responsible for eosinophil synthesis of CCL17, CCL22, CXCL9, and CXCL10.

Peripheral blood for eosinophil purification was obtained from allergic volunteers (skin prick test positive with a history of seasonal or perennial allergic rhinitis) ranging in age from 18 to 58 years. None of the subjects had taken inhaled corticosteroids, or other medications that would interfere with the study results, or had evidence of a respiratory infection within the previous 4 wk. Informed consent was obtained from each subject before participation. The study was approved by the University of Wisconsin-Madison Center for Health Sciences Human Subjects Committee.

Eosinophils were purified, as previously described (17), from heparinized peripheral blood, with eosinophils comprising 2–10% of the leukocytes. Briefly, the granulocyte fraction was obtained after centrifugation of HBSS-diluted blood over Percoll (1.090 g/ml); RBCs were lysed; and neutrophils, T cells, and monocytes were depleted, respectively, by anti-CD16, anti-CD3, and anti-CD14 immunomagnetic beads (AutoMac system; Miltenyi Biotec). The resulting eosinophils were >99% pure and >97% viable.

Eosinophils (2 × 106 cells/ml) were cultured in medium (RPMI 1640 with 10% FCS and 1% penicillin/streptomycin) alone or stimulated with 10 ng/ml Th2 (IL-4), Th1 (IFN-γ) cytokines, or the eosinophil-priming cytokine TNF-α, separately or in various combinations for up to 72 h, as indicated. In selected experiments, eosinophils were pretreated for 1 h with the indicated concentration of pharmacological inhibitors of JNK (JNK inhibitor II or its inactive analog), p38 MAPK (SB203580 or the inactive analog SB202474), MEK (U0126 or inactive analog U0124), or NF-κB (BAY 11-7082), and then cultured with IL-4 plus TNF-α or IFN-γ plus TNF-α for additional 24 h. All the inhibitors were purchased from Calbiochem. The concentration for these pharmacological agents was based on pilot studies to determine the quantity that resulted in the greatest inhibition with the least amount of cytotoxicity. At the concentrations used, eosinophil viability at 24 h was >80%. Th1 or Th2 chemokines released into the cell-free culture supernatant fluids were analyzed using a sandwich ELISA, as previously described (18). The coating and biotinylated Abs for CXCL10, CCL17, and CCL22 analysis were purchased from R&D Systems. CXCL9 coating and biotinylated Abs were purchased from BD Biosciences. The assay sensitivity was <3 pg/ml for CXCL9 and CCL22, and <12 pg/ml for CXCL10 and CCL17.

Total RNA was extracted from 1 × 106 eosinophils using the RNeasy mini kit (Qiagen) and reverse transcribed using 400 U of Super Script III reverse transcriptase (Invitrogen Life Technologies) at 37°C for 60 min in the presence of random hexamer primers (Promega). Real-time quantitative PCR was performed in the Applied Biosystems 7500 sequence detector using human CCL17-, CCL22-, CXCL10-, or CXCL9-specific primers and TaqMan probes (Applied Biosystems). Based on its similar transcription efficiency to the chemokine target genes and its consistent expression among treatment groups, the housekeeping gene, β-glucuronidase (GUS),3 was chosen to normalize the samples. The thermal cycling conditions were as follows: a 2-min 50°C uracil-N-glycosylase incubation step and an initial denaturation at 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. The efficiency of the Taqman assay was determined by assaying serial 10-fold dilutions of target cDNA. All samples were amplified in duplicate, and the mean was obtained for further calculations. The data are expressed as fold change using the comparative cycle threshold (Ct) method in which ΔCt = Ct of the chemokine gene minus Ct of GUS; ΔΔCt = ΔCt of stimulated cells at the indicated time points minus ΔCt of unstimulated cells at 0 h; and fold change = 2−ΔΔCt. If no transcripts were detectable at 0 h, a Ct value of 40 cycles (corresponding to the highest number of cycles) was used to calculate ΔCt.

Purified blood eosinophils were cultured in medium (RPMI 1640 with 5% FCS and 1% penicillin/streptomycin) alone or stimulated with 10 ng/ml IFN-γ or TNF-α alone or in combination. After 0, 3, and 24 h incubation, cells were washed and incubated with PE-conjugated mAbs to CD69 (Beckman Coulter), IFN-γRα (BD BioSciences), and TNFRI and TNFRII (R&D Systems) for cell surface staining. At the 24-h incubation time point, intracellular staining was also performed with the same Abs. Before staining, eosinophils were fixed with 2% paraformaldehyde for 10 min at 37°C and permeabilized for 30 min on ice by adding 0.5 ml of ice-cold 90% methanol. For analysis, 10,000 events were collected using a BD Immunocytometry Systems FACScan II, and data analyses were performed using the CellQuest software package (BD BioSciences).

Freshly purified eosinophils (6 × 106 cells/tube) were rested in medium (RPMI 1640 containing 25 mM HEPES and 0.1% human serum albumin) at 37°C for 30 min, and then stimulated with 100 ng/ml TNF-α, IL-4, or IFN-γ alone, or TNF-α plus IL-4 or TNF-α plus IFN-γ for 30 min. Cells were washed and then sonicated in lysis buffer (137 mM NaCl, 20 mM HEPES, 1 mM EDTA (pH 7.4), 10 mM NaF, 20 mM β-glycerophosphate, 1 mM orthovanadate, 1% Sigma-Aldrich mammalian protease inhibitor mixture, 1% Triton-X-100, 0.1% SDS, and 0.25% deoxycholate). Insoluble material was removed by centrifugation at 14,000 rpm/4°C for 10 min. The soluble fraction was assayed for total protein, and immunoblotting was performed, as previously described (19). Briefly, samples were resolved by SDS-PAGE, and proteins were transferred to a polyvinylidene difluoride membrane and probed with Abs specific for STAT1 phosphorylated on tyrosine residue 701 (anti-phosphotyrosine-701 STAT1; Upstate Biotechnology), STAT6 phosphorylated on tyrosine residue 641 (anti-phosphotyrosine-641 STAT6; Cell Signaling Technology), or anti-IκBα (Santa Cruz Biotechnology). Detection of immunoreactive bands was performed using Super Signal West Dura or Femto chemiluminescent substrate (Pierce). Equal protein loading was assured by immunoblotting the same cell lysates with Abs to total ERK1 and ERK2 (Upstate Biotechnology), STAT1 (Santa Cruz Biotechnology), STAT6 (Cell Signaling Technology), or β1 integrin (Chemicon International).

Purified blood eosinophils were incubated with medium (RPMI 1640 with 5% FCS and 1% penicillin/streptomycin) alone or stimulated with 100 ng/ml TNF-α, IL-4, or IFN-γ alone or TNF-α in combination with IL-4 or IFN-γ. After 15 min, cells were immediately fixed with 2% paraformaldehyde for 10 min at 37°C and permeabilized for 30 min on ice by adding 0.5 ml of ice-cold 90% methanol. Cells were washed with PBS containing 2% albumin and 0.2% sodium azide, resuspended at 2 × 106 cells/ml, and incubated for 45 min in the dark at room temperature with appropriate Abs. Abs included PE-conjugated mAbs to phospho-STAT1 (pY701; BD Biosciences), phospho-STAT6 (Y641; BD Biosciences), IκBα (H-4; Santa Cruz Biotechnology), and mouse IgG isotype control (BD Biosciences). Cells were washed and resuspended in 0.25 ml of FACS buffer. For analysis, 10,000 events were collected using a BD Immunocytometry Systems FACScan II, and data analyses were performed using the CellQuest software package (BD Biosciences).

Data are presented as the median with interquartiles of 25 and 75% or the mean ± SEM (for normally distributed data). The Wilcoxon signed rank test (or Student’s paired t test for normally distributed data) was used to compare different treatments or time points. Chemokine mRNA expressions were compared by treatment using repeated-measures ANOVA models. A p value of <0.05 was considered significant. Statistical analyses were performed using the SigmaStat software package (Jandel Scientific Software).

IL-4 and IFN-γ induce Th2 and Th1 chemokines, respectively, in a variety of cell types. Thus, we compared the response of highly purified human eosinophils to these cytokines. IL-4 alone led to a small, but statistically significant production of the CCL17 (Fig. 1,A), whereas IFN-γ alone did not induce eosinophil secretion of any of the chemokines tested (Fig. 1,B). Because TNF-α can amplify chemokine generation in other cell types (20), eosinophils were stimulated in the presence of TNF-α plus either IL-4 or IFN-γ. With TNF-α alone, eosinophils produced a small amount of CCL22 (Fig. 1,A). Addition of TNF-α to IL-4 augmented eosinophil generation of CCL17 and CCL22 (Fig. 1,A), but this cytokine combination had no effect on Th1-type chemokines (Fig. 1,B). When TNF-α was added with IFN-γ, eosinophils produced high concentrations of CXCL9 and CXCL10 (Fig. 1,B). In fact, eosinophils produced 90–200 times more CXCL10 and 7–18 times more CXCL9 than CCL17 and CCL22, respectively. The pattern and magnitude of Th1 and Th2 chemokine generation in response to IFN-γ or IL-4 alone or in combination with TNF-α were similar in eosinophils obtained from normal nonatopic subjects (data not shown). In the presence of the combination of all three cytokines, there was a statistically significant decrease in IL-4/TNF-α-induced CCL22 by IFN-γ (Fig. 1,A) and IFN-γ/TNF-α-induced CXCL9 by IL-4 (Fig. 1,B). CCL17 and CXCL10 also tended to decrease, but this did not reach statistical significance. There was no significant chemokine generation following PHA stimulation, indicating a lack of contamination with lymphocytes (Fig. 1). Stimulation with LPS alone or in combination with either IL-4 or IFN-γ did not produce detectable levels of Th1- or Th2-type chemokines (data not shown). These data demonstrate that eosinophils are a source of both Th1- and Th2-type chemokines, the predominant generation of which is determined by the presence of TNF-α plus IFN-γ or IL-4, respectively.

FIGURE 1.

Th1- and Th2-type chemokine generation by purified human eosinophils following cytokine stimulation. Eosinophils were cultured for 72 h at 2 × 106 cells/ml in the presence of 10 ng/ml IL-4, IFN-γ, or TNF-α alone or in combination. Additional cultures of purified eosinophils contained PHA (10 μg/ml) to confirm the absence of lymphocyte contamination. Concentrations of Th2-type chemokines (A) and Th1-type chemokines (B) in cell-free culture supernatant fluids were measured by ELISA. Data are expressed as the mean ± SEM of 14 individual eosinophil donors. ND, levels of chemokine were not detectable.

FIGURE 1.

Th1- and Th2-type chemokine generation by purified human eosinophils following cytokine stimulation. Eosinophils were cultured for 72 h at 2 × 106 cells/ml in the presence of 10 ng/ml IL-4, IFN-γ, or TNF-α alone or in combination. Additional cultures of purified eosinophils contained PHA (10 μg/ml) to confirm the absence of lymphocyte contamination. Concentrations of Th2-type chemokines (A) and Th1-type chemokines (B) in cell-free culture supernatant fluids were measured by ELISA. Data are expressed as the mean ± SEM of 14 individual eosinophil donors. ND, levels of chemokine were not detectable.

Close modal

In initial studies with eosinophils prepared by CD16 depletion of a granulocyte fraction (97–98% pure eosinophils), CXCL9 and CXCL10 were induced in response to IFN-γ alone, which is consistent with a previous report demonstrating that IFN-γ was sufficient for eosinophil production of CXCL9 and CXCL10 (21). In contrast, when highly purified (>99%) eosinophils (CD16-, CD3-, and CD14-depleted granulocyte fraction) were used, the combination of IFN-γ plus TNF-α appeared to be essential for eosinophil production of Th1-associated chemokines (Fig. 1 B). To determine whether a small mononuclear cell contamination (up to 3%) of eosinophils that had been prepared by the CD16 depletion method could affect eosinophil chemokine generation in vitro, PBMCs (6 × 104/ml) were added to highly purified eosinophils (2 × 106/ml). In the presence of only 3% PBMCs, IFN-γ alone was sufficient to increase CXCL10 from a baseline concentration of 10 ± 32 pg/ml to 3040 ± 3530 pg/ml (mean ± SD, p < 0.001, n = 12), and to increase CXCL9 from 8 ± 22 pg/ml to 2740 ± 1000 pg/ml (p < 0.001, n = 12). In the absence of eosinophils, an equivalent number of PBMCs (6 × 104/ml) produced only 80 ± 26 pg/ml CXCL9 and 109 ± 79 pg/ml CXCL10. Collectively, these data suggest that a small number of mononuclear cells can augment eosinophil generation of CXCL9 and CXCL10.

Several approaches were used to determine that the synergistic effect of IFN-γ and TNF-α on eosinophil generation of Th1-type chemokines was due to enhanced cell survival. Eosinophil viability was determined by trypan blue exclusion 48 h after culture with medium, IFN-γ, TNF-α, or IFN-γ plus TNF-α. Eosinophil viability (n = 4 experiments) was enhanced by IFN-γ alone (76 ± 4%) and by IFN-γ plus TNF-α (65 ± 7%), but not by medium (11 ± 5%) or TNF-α alone (13 ± 8%). Viability in the presence of the combination of IFN-γ plus TNF-α was not greater than in the presence of IFN-γ alone (n = 4, p = 0.210). The synergistic effect of IFN-γ and TNF-α on Th1 chemokine generation was thus not due to a synergistic effect on eosinophil viability. Furthermore, addition of the prosurvival factor IL-5 had no significant effect on eosinophil generation of CXCL9 or CXCL10. IL-5 (100 pg/ml) was added at the initiation of eosinophil culture with IFN-γ alone or IFN-γ plus TNF-α. Both CXCL9 and CXCL10 remained below the level of detection in cultures treated with TNF-α alone or in the presence of IL-5, suggesting that the lack of Th1 chemokine production was not due to cell death. When cells were treated with IFN-γ plus TNF-α in the presence or absence of IL-5, respectively, levels of CXCL9 were 1750 ± 320 and 1830 ± 700 pg/ml (n = 3, p = 0.9), and levels of CXCL10 were 4500 ± 2180 and 6550 ± 4320 pg/ml (n = 3, p = 0.4). Eosinophil generation of comparable amounts of Th1 chemokines in the presence or absence of the prosurvival factor IL-5 suggests that the synergistic induction of Th1 chemokines is not due to enhanced eosinophil survival.

To determine the time course of Th1 (CXCL9, CXCL10) and Th2 (CCL17, CCL22) chemokine generation, highly purified eosinophils were cultured with TNF-α in combination with either IFN-γ or IL-4 for 0, 3, 20, 48, and 72 h (Fig. 2). Detectable levels of Th1 and Th2 chemokines were noted by 20 h and continued to increase for at least 72 h. Although the concentrations were dramatically different, the kinetics of Th1 and Th2 chemokine production was similar. The delay in generation of chemokines is consistent with eosinophil synthesis of new protein rather than a release of preformed and stored chemokines.

FIGURE 2.

Kinetics of Th1- and Th2-type chemokine generation by purified human eosinophils. Eosinophils were stimulated by TNF-α plus IL-4 (10 ng/ml) or IFN-γ (10 ng/ml), and cultured for 0, 3, 20, 48, and 72 h. Concentrations of Th2 chemokines (A) and Th1 chemokines (B) in cell-free culture supernatant fluids were measured by ELISA. Data are expressed as the mean ± SEM of four individual eosinophil donors.

FIGURE 2.

Kinetics of Th1- and Th2-type chemokine generation by purified human eosinophils. Eosinophils were stimulated by TNF-α plus IL-4 (10 ng/ml) or IFN-γ (10 ng/ml), and cultured for 0, 3, 20, 48, and 72 h. Concentrations of Th2 chemokines (A) and Th1 chemokines (B) in cell-free culture supernatant fluids were measured by ELISA. Data are expressed as the mean ± SEM of four individual eosinophil donors.

Close modal

Consistent with protein generation, CCL17 and CCL22 gene expression was induced by IL-4 alone and was modestly enhanced by TNF-α (Fig. 3,A). Conversely, the combination of TNF-α plus IFN-γ had a potent synergistic effect on CXCL9 and CXCL10 mRNA expression that was observed as early as 3 h after stimulation (Fig. 3,B). The high levels of CXCL9 and CXCL10 mRNA expression were sustained for at least 20 h (Fig. 3 B). In contrast to protein secretion, there were small, but statistically significant increases in CXCL9 and CXCL10 gene expression when eosinophils were incubated with IFN-γ alone. TNF-α alone did not induce significantly changes in levels of mRNA. Neither chemokine was detectable when eosinophils were incubated with medium alone (Ct >40 cycles, data not shown). These data provide further support for de novo protein synthesis of Th1 and Th2 chemokines and demonstrate the highly synergistic effect of TNF-α on IFN-γ-induced Th1 chemokine gene expression by highly purified human eosinophils.

FIGURE 3.

Gene expression of Th1- and Th2-type chemokines by purified human eosinophils following cytokine stimulation. Eosinophils were stimulated for 0, 3, 6, 9, or 20 h with A, 10 ng/ml IL-4 alone (□), TNF-α alone (), or TNF-α plus IL-4 (▪) for stimulation of CCL17 and CCL22 gene expression (n = 4 subjects), or B, 10 ng/ml IFN-γ alone (○), TNF-α alone (), or TNF-α plus IFN-γ (•) for stimulation of CXCL9 and CXCL10 gene expression (n = 7 subjects). Levels of mRNA were determined by real-time quantitative PCR. The housekeeping gene, GUS, was used to normalize the samples. The data are expressed as fold change (2−ΔΔCt), in which ΔCt = the Ct for chemokine gene minus the Ct of the housekeeping gene, GUS; ΔΔCt = the ΔCt of stimulated cells at different time points minus the ΔCt of unstimulated cells at 0 h. ∗, p < 0.05 vs TNF-α alone; , p < 0.05 vs IL-4 or IFN-γ alone.

FIGURE 3.

Gene expression of Th1- and Th2-type chemokines by purified human eosinophils following cytokine stimulation. Eosinophils were stimulated for 0, 3, 6, 9, or 20 h with A, 10 ng/ml IL-4 alone (□), TNF-α alone (), or TNF-α plus IL-4 (▪) for stimulation of CCL17 and CCL22 gene expression (n = 4 subjects), or B, 10 ng/ml IFN-γ alone (○), TNF-α alone (), or TNF-α plus IFN-γ (•) for stimulation of CXCL9 and CXCL10 gene expression (n = 7 subjects). Levels of mRNA were determined by real-time quantitative PCR. The housekeeping gene, GUS, was used to normalize the samples. The data are expressed as fold change (2−ΔΔCt), in which ΔCt = the Ct for chemokine gene minus the Ct of the housekeeping gene, GUS; ΔΔCt = the ΔCt of stimulated cells at different time points minus the ΔCt of unstimulated cells at 0 h. ∗, p < 0.05 vs TNF-α alone; , p < 0.05 vs IL-4 or IFN-γ alone.

Close modal

To evaluate the mechanisms for the highly synergistic effect of TNF-α on IFN-γ-induced Th1 chemokines, the ability of TNF-α to up-regulate IFN-γ receptors, as has been reported for other cell types (22, 23), was accessed. Eosinophil cell surface staining was examined at 0, 3, and 24 h after stimulation with TNF-α, IFN-γ, or IFN-γ plus TNF-α (Fig. 4, A–D). Cellular activation was confirmed by increased expression of CD69 within 3 h by TNF-α or IFN-γ alone (Fig. 4,A). CD69 was further increased at 24 h when eosinophils were stimulated with the combination of IFN-γ and TNF-α (Fig. 4,A). IFN-γRα was highly expressed on the eosinophil cell surface at baseline. In contrast to observations for other cell types (22, 23), eosinophil expression of IFN-γRα was significantly diminished when cells were cultured for 24 h with the combination of IFN-γ plus TNF-α (Fig. 4,B). In contrast to IFN-γRα, TNF-RII expression was only detectable at low levels before culture and was significantly increased at 24 h in the presence of IFN-γ or IFN-γ plus TNF-α (Fig. 4,C). TNF-RI, which is the primary TNF-α receptor on most cell types, was not detectable on the eosinophil cell surface and was not induced by IFN-γ or TNF-α alone or in combination (Fig. 4,D). The low levels of cell surface TNF receptors led us to determine whether the receptors are present intracellularly. Indeed, IFN-γRα, TNF-RI, and TNF-RII were detected in permeabilized eosinophils. Levels of intracellular IFN-γRα did not change in the presence of TNF-α or IFN-γ, whereas expression of both intracellular TNF-RI and TNF-RII was decreased when eosinophils were stimulated for 24 h with IFN-γ (Fig. 4, G and H). The simultaneous decrease in intracellular TNF-RII and increase in cell surface TNF-RII raise the possibility that IFN-γ induces the translocation of intracellular TNF-RII to the cell surface.

FIGURE 4.

Effect of TNF-α and IFN-γ on expression of their respective cell surface and intracellular receptors. A–D, Cell surface receptors were determined by flow cytometric analysis at 0, 3, and 24 h after incubation with medium (□), IFN-γ (light gray bars), TNF-α (dark gray bars), or the combination of IFN-γ and TNF-α (▪). E–H, 24 h after incubation, eosinophils were fixed, permeabilized, and stained for intracellular cytokine receptors. The data are expressed as median channel fluorescence (MCF) of eosinophils from six individual donors. ∗, p < 0.05 compared with medium; , p < 0.05 compared with IFN-γ alone; , p < 0.05 compared with TNF-α alone.

FIGURE 4.

Effect of TNF-α and IFN-γ on expression of their respective cell surface and intracellular receptors. A–D, Cell surface receptors were determined by flow cytometric analysis at 0, 3, and 24 h after incubation with medium (□), IFN-γ (light gray bars), TNF-α (dark gray bars), or the combination of IFN-γ and TNF-α (▪). E–H, 24 h after incubation, eosinophils were fixed, permeabilized, and stained for intracellular cytokine receptors. The data are expressed as median channel fluorescence (MCF) of eosinophils from six individual donors. ∗, p < 0.05 compared with medium; , p < 0.05 compared with IFN-γ alone; , p < 0.05 compared with TNF-α alone.

Close modal

Activation of STAT1, STAT6, and NF-κB is important in the signaling from IFN-γ, IL-4, and TNF-α receptors, respectively, on a variety of cell types. To determine whether these signal transduction pathways are induced upon stimulation of human eosinophil with IL-4 plus TNF-α or IFN-γ plus TNF-α, their activation was assessed by both intracellular flow cytometric analysis and immunoblotting. Using intracellular flow cytometric analysis, increased levels of phosphorylated STAT6 and STAT1 were detected in eosinophils stimulated with IL-4 and IFN-γ, respectively (Fig. 5,A). NF-κB activation was induced by TNF-α, as indicated by a decrease in the inhibitory protein, IκBα (Fig. 5,A). There was no evidence for a further increase in the activation of these phosphoproteins when eosinophils were cultured with IL-4 or IFN-γ in combination with TNF-α (Fig. 5,A). To our knowledge, this is the first report to demonstrate the utility of intracellular phosphoprotein analysis by flow cytometry to detect signal transduction in eosinophils. Compared with immunoblotting, intracellular flow cytometric analysis of eosinophil signaling pathways required 10–15 times fewer eosinophils, and the procedure could be completed in <5 h compared with several days. Immunoblot analysis (Fig. 5, B and C) confirmed the observations made using intracellular flow cytometry. Both methods demonstrated a similar pattern of IL-4-, IFN-γ-, or TNF-α-induced phosphorylation of STAT6 and STAT1, and degradation of IκBα, but no further increase in activation of these pathways was noted with the combination of IL-4 plus TNF-α or IFN-γ plus TNF-α. These data indicate an independent activation of signaling cascades that, in combination, may synergistically stimulate eosinophil generation of T cell-associated chemokines.

FIGURE 5.

Signaling cascades initiated by IL-4, TNF-α, and IFN-γ. A, Detection of phosphoproteins by intracellular flow cytometric analysis. Eosinophils were incubated with medium alone (dark gray), IL-4 alone (black, pSTAT6), IFN-γ alone (black, pSTAT1), TNF-α alone (black, IκBα), or the combination of IL-4 plus TNF-α or IFN-γ plus TNF-α (black). Isotype control Ab staining is indicated as light gray. Representative histograms of data from at least three subjects are shown for pSTAT6, pSTAT1, and IκBα degradation. B, Purified eosinophils were cultured with medium alone (lane 1), IL-4 (lane 2), TNF-α (lane 3), or both (lane 4). STAT6 activation was detected by immunoblotting with antiserum specific for phosphorylated STAT6, and NF-κB activation was detected by the proteolytic degradation of the inhibitory subunit IκBα. Total STAT6 and β1 integrin serve as the loading control. C, Purified eosinophils were cultured with medium alone (lane 1), IFN-γ (lane 2), TNF-α (lane 3), or both (lane 4). STAT1 activation was detected by immunoblotting with antiserum specific for phosphorylated STAT1, and NF-κB activation was detected by the proteolytic degradation of the inhibitory subunit IκBα. Total STAT1, ERK1, and ERK2 serve as the loading control. Data are representative of eosinophils from three subjects.

FIGURE 5.

Signaling cascades initiated by IL-4, TNF-α, and IFN-γ. A, Detection of phosphoproteins by intracellular flow cytometric analysis. Eosinophils were incubated with medium alone (dark gray), IL-4 alone (black, pSTAT6), IFN-γ alone (black, pSTAT1), TNF-α alone (black, IκBα), or the combination of IL-4 plus TNF-α or IFN-γ plus TNF-α (black). Isotype control Ab staining is indicated as light gray. Representative histograms of data from at least three subjects are shown for pSTAT6, pSTAT1, and IκBα degradation. B, Purified eosinophils were cultured with medium alone (lane 1), IL-4 (lane 2), TNF-α (lane 3), or both (lane 4). STAT6 activation was detected by immunoblotting with antiserum specific for phosphorylated STAT6, and NF-κB activation was detected by the proteolytic degradation of the inhibitory subunit IκBα. Total STAT6 and β1 integrin serve as the loading control. C, Purified eosinophils were cultured with medium alone (lane 1), IFN-γ (lane 2), TNF-α (lane 3), or both (lane 4). STAT1 activation was detected by immunoblotting with antiserum specific for phosphorylated STAT1, and NF-κB activation was detected by the proteolytic degradation of the inhibitory subunit IκBα. Total STAT1, ERK1, and ERK2 serve as the loading control. Data are representative of eosinophils from three subjects.

Close modal

Activation of the MAPK cascades ERK1/2, p38, and JNK are potential points of convergence in the NF-κB, STAT6, and STAT1 signaling networks. Thus, chemical inhibitors were used to determine the contribution of NF-κB and MAPK to the synergistic induction of chemokines by TNF-α plus IFN-γ or IL-4. Fig. 6 demonstrated that inhibition of cytokine-inducible IκBα phosphorylation with BAY 11-7082 to reduce NF-κB activation completely ablated eosinophil generations of Th1 and Th2 chemokines. In contrast, MAPK inhibitors led to partial and variable reduction in chemokine synthesis. Inhibition of MEK with U0126, p38 MAPK with SB203580, or JNK with JNK inhibitor II partially blocked CXCL9 and CXCL10 generation in response to the combination of TNF-α and IFN-γ. Th2 chemokines induced by IL-4 plus TNF-α had a different response profile to these inhibitors. CCL17 was partially blocked by MEK and JNK inhibitors, but was not responsive to the p38 MAPK inhibitor. Conversely, CCL22 was partially blocked by p38 MAPK and JNK inhibitors, but not by the MEK inhibitor. Collectively, these data indicate that MAPK contribute to, but are not critical for the synergistic effect of TNF-α plus IFN-γ or IL-4. Furthermore, it appears that whereas TNF-α and IFN-γ or IL-4 independently activate distinct signaling pathways, convergence between the pathways leads to NF-κB-dependent control of transcription or posttranscriptional events.

FIGURE 6.

Effect of NF-κB and MAPK pathway inhibitors on IL-4 plus TNF-α-induced Th2 chemokine and IFN-γ plus TNF-α-induced Th1 chemokine generation. Purified eosinophils were preincubated for 1 h with medium alone or the NF-κB inhibitor BAY 11-7082 (4 μM), the MEK inhibitor U 0126 or its inactive analog U 0124 (20 μM), the P38 MAPK inhibitor SB203580 or its inactive analog SB202474 (2 μM), or the JNK inhibitor II or its inactive analog (20 μM). Eosinophils were then cultured for 24 h with addition of TNF-α plus IL-4 or TNF-α plus IFN-γ (10 ng/ml each cytokine) to their corresponding wells. Concentrations of Th2 or Th1 chemokines in the corresponding culture supernatants were determined by ELISA. Cell viability determined by trypan blue exclusion at 24 h was >80% for all culture conditions. ∗, p < 0.05 compared with stimulants in the absence of inhibitor; #, p < 0.05 compared with corresponding inactive analog control. Data are expressed as the percentage change from IL-4 plus IFN-γ or TNF-α plus IFN-γ in the absence of inhibitor (mean ± SEM) from five individual eosinophil donors.

FIGURE 6.

Effect of NF-κB and MAPK pathway inhibitors on IL-4 plus TNF-α-induced Th2 chemokine and IFN-γ plus TNF-α-induced Th1 chemokine generation. Purified eosinophils were preincubated for 1 h with medium alone or the NF-κB inhibitor BAY 11-7082 (4 μM), the MEK inhibitor U 0126 or its inactive analog U 0124 (20 μM), the P38 MAPK inhibitor SB203580 or its inactive analog SB202474 (2 μM), or the JNK inhibitor II or its inactive analog (20 μM). Eosinophils were then cultured for 24 h with addition of TNF-α plus IL-4 or TNF-α plus IFN-γ (10 ng/ml each cytokine) to their corresponding wells. Concentrations of Th2 or Th1 chemokines in the corresponding culture supernatants were determined by ELISA. Cell viability determined by trypan blue exclusion at 24 h was >80% for all culture conditions. ∗, p < 0.05 compared with stimulants in the absence of inhibitor; #, p < 0.05 compared with corresponding inactive analog control. Data are expressed as the percentage change from IL-4 plus IFN-γ or TNF-α plus IFN-γ in the absence of inhibitor (mean ± SEM) from five individual eosinophil donors.

Close modal

We have established that human eosinophils produce both Th1 (CXCL9 and CXCL10)- and Th2 (CCL17 and CCL22)-associated chemokines, and that TNF-α plays a pivotal role in this process. Eosinophil secretion of CCL17 and CCL22 occurred in response to IL-4 and was significantly enhanced by the presence of TNF-α. Interestingly, CXCL9 and CXCL10 generation from highly purified eosinophils required stimulation by the combination of IFN-γ plus TNF-α. Given the association of eosinophils with Th2-type responses, their production of large quantities of CXCL9 and CXCL10 was somewhat unexpected. Nonetheless, our observations that eosinophils can produce both Th1 and Th2 chemokines are consistent with our previous in vivo findings that CXCL9, CXCL10, CCL17, and CCL22 are increased in the airway during eosinophilic airway inflammation in allergic asthma subjects (16).

To our knowledge, this is the first demonstration that eosinophils can synthesize and secrete Th2 chemokines. The kinetics of protein generation and gene expression indicate that CCL17 and CCL22 are actively synthesized and not stored by eosinophils. The amount of Th2 chemokines produced by eosinophils was modest and variable among blood donors. Whether donor-to-donor variability is due to polymorphisms of IL-4Rα, as has been observed for IL-4-induced chemokine generation by airway smooth muscle (24), is an interesting possibility that has not yet been investigated. Although the amount of CCL17 and CCL22 produced by human eosinophils was modest, levels were comparable to those published for human T cells (25). Regulation of eosinophil CCL17 and CCL22 was similar to what has been reported for some, but not all cell types. Eosinophils produced small amounts of Th2 chemokines in response to IL-4 alone, and in agreement with studies of airway smooth muscle (24), bronchial epithelial cell lines (26), and fibroblasts (27), TNF-α potently augmented IL-4-induced generation of CCL17 and CCL22. Preliminary data suggest that differential chemokine generation was not due to an effect on eosinophil survival. Addition of IL-5 prolonged eosinophil viability, but did not enhance their generation of Th2 chemokines in response to TNF-α plus IL-4. In addition, eosinophils from normal and atopic subjects had a similar chemokine response, indicating that eosinophils have the intrinsic ability to produce Th2 chemokines when appropriate stimuli are present. The effect of IFN-γ on Th2 chemokines varies depending on the cell type. In the current study, IFN-γ significantly decreased CCL22 and tended to reduce eosinophil generation of CCL17 in response to IL-4 plus TNF-α. A similar suppressive effect of IFN-γ on Th2 chemokines has been reported for airway smooth muscle, human monocytes, and monocyte-derived dendritic cells (24, 28). Conversely, a combination of IL-4, TNF-α, and IFN-γ was the most potent inducer of CCL17 by bronchial epithelial cell lines (26), and IFN-γ alone promoted CCL17 and CCL22 generation by primary human epidermal keratinocytes (29). Thus, regulation of CCL17 and CCL22 appears to be cell type dependent.

It has previously been established, in a mouse model of allergen-induced airway inflammation, that STAT6 is required for CCL17 and CCL22 (30). However, the effects of TNF-α on IL-4-induced activation of STAT6 and subsequent downstream signaling pathways leading to synthesis of CCL17 and CCL22 by human eosinophils have not been reported. By flow cytometric analysis and immunoblotting, we confirm that IL-4 induces phosphorylation of STAT-6 and that TNF-α leads to activation of NF-κB in human eosinophils. However, no evidence was found for enhanced activation of either factor when eosinophils were stimulated by the combination of IL-4 plus TNF-α. Use of pharmacological inhibitors demonstrated an apparent requirement for NF-κB in the production of both CCL17 and CCL22. In contrast, the MAPK inhibitors only partially blocked Th2 chemokines. Generation of both CCL17 and CCL22 was significantly, but not completely inhibited by JNK inhibitor II. Interestingly, CCL17 and CCL22 had differential sensitivity to MEK and p38 MAPK inhibitors. CCL22 was partially blocked by inhibition of p38 MAPK, but was not susceptible to a MEK inhibitor; CCL17 was partially blocked by inhibition of MEK, but was not susceptible to the p38 MAPK inhibitor. The findings in the current study suggest that TNF-α and IL-4 activate independent signaling cascades in eosinophils that in combination complement activation of CCL17 or CCL22, and that NF-κB is a point of convergence between the pathways.

TNF-α also played a pivotal role in eosinophil generation of CXCL9 and CXCL10. However, neither TNF-α nor IFN-γ alone was sufficient for Th1 chemokine secretion by eosinophils. The prominent synergy between IFN-γ and TNF-α for induction of CXCL9 and CXCL10 has been described for other cell types (20); however, the lack of protein secretion in response to IFN-γ alone is uncommon. Our observations contrast those of Dajotoy et al. (21). Although a synergistic effect was seen with IFN-γ plus TNF-α, in their study, IFN-γ alone was sufficient to induce generation of CXCL9 and CXCL10 by human blood eosinophils. We postulate that the requirement for TNF-α in our study is due to the lack of contaminating cell populations. When as few as 3% PBMCs were added to highly purified eosinophils, CXCL9 and CXCL10 could be induced by IFN-γ alone. Of note, when this small number of PBMCs was stimulated in the absence of eosinophils, the levels of chemokine secretion were very low. Thus, it is likely that when even very small numbers of mononuclear cells are present with eosinophils, they are stimulated by IFN-γ to release TNF-α, which then acts in concert with IFN-γ to induce eosinophil generation of CXCL9 and CXCL10. Likewise, in other cell types, IFN-γ-mediated induction of endogenous TNF-α and subsequent autocrine signaling (31) may explain the apparent lack of requirement for exogenous TNF-α.

Several observations support the notion that the synergistic induction of Th1 chemokines by TNF-α plus IFN-γ is not due to enhanced cell survival. First, although IFN-γ alone is a potent eosinophil survival factor, it does not induce eosinophil generation of Th1 chemokines. Second, addition of the prosurvival factor IL-5 to eosinophils cultured in the presence of TNF-α promoted cell viability, but chemokine generation was not detectable. Third, comparable levels of CXCL9 or CXCL10 were produced in response to the combination of IFN-γ plus TNF-α in the presence or absence of IL-5. Finally, the early (within 3 h) differences in Th1 chemokine gene expression in eosinophils stimulated with IFN-γ plus TNF-α compared with either cytokine alone are not likely to be due to apoptotic events.

The synergistic effect of IFN-γ and TNF-α for induction of several genes including HLA-DR and indoleamine dioxygenase has been attributed to TNF-α-mediated up-regulation of IFN-γ receptors within 4 h of cellular activation (22, 23). In our system, TNF-α had no effect on IFN-γRα, which was highly expressed on the cell surface of resting eosinophils. Conversely, cell surface TNFRII, which was only present at low levels on resting eosinophils, was significantly up-regulated 24 h after stimulation with IFN-γ. Interestingly, as cell surface TNFRII increased in response to IFN-γ, intracellular TNFRII decreased. These data raise the possibility that IFN-γ can induce the translocation of a preformed receptor to the cell surface. Nonetheless, the fact that enhancement of cell surface receptors was not detectable at 3 h suggests that increased expression of TNFRII may not be a critical factor for the synergistic effect of TNF-α and IFN-γ on subsequent chemokine secretion.

The synergistic action of IFN-γ and TNF-α on eosinophils involved independent STAT1 and NF-κB activation. Using Western blot or intracellular flow cytometric analysis, we demonstrated IFN-γ-induced STAT1 and TNF-α-induced NF-κB activation in human eosinophils. The level of STAT1 phosphorylation and IκBα degradation induced by their respective activator was not enhanced by the combination of IFN-γ and TNF-α. Using chemical inhibitors, we showed that induction of both Th1 chemokines was ablated in the presence of a NF-κB inhibitor and was partially suppressed by inhibitors of MEK, p38 MAPK, and JNK.

In addition to their induction of CXCL9 and CXCL10, IFN-γ and TNF-α can synergistically activate transcription of a number of genes, including indoleamine dioxygenase, ICAM-1, VCAM-1, MHC class I, and inducible NO synthase, which contain STAT1 and NF-κB binding sites in their promoter (20, 32). Several mechanisms have been proposed to explain this synergy, including the following: 1) enhanced affinity of transcription factors to their respective recognition sites (22); 2) physical interaction between IFN-γ- and TNF-α-induced transcription factors to a composite recognition site (33, 34); and 3) cross-talk between the two signaling pathways via activation of additional transcription factors such as IFN regulatory factor-1 or C/EBP-β (20, 32, 35). Further studies are needed to determine the prevailing scenario in human eosinophils. Of further note, CXCL9 and CXCL10 mRNA could not be detected by real-time quantitative PCR in resting eosinophils, despite attempts to enhance assay sensitivity by increasing the amount of target cDNA and the number of PCR cycles. The eosinophil’s lack of basal expression of CXCL9 and CXCL10 is unique in comparison with other cell types and may indicate a level of transcriptional repression that remains to be identified.

Generation of Th1 and Th2 chemokines provides further evidence for the eosinophil’s role in regulating T cells. Eosinophil production of CCL17 and CCL22 may serve as a positive feedback loop to enhance recruitment of Th2 cytokine-producing cells. The functional significance for eosinophil generation of Th1 chemokines is less clear. The expression of CXCR3 on human eosinophils (36) and the direct chemotactic effect of CXCL9 and CXCL10 on human eosinophils (37) raise the possibility that these Th1-type chemokines could also augment allergic inflammation. In contrast, CXCL9 and CXCL10 can bind to CCR3 (38, 39) and inhibit eotaxin-induced recruitment of eosinophils (40). Whether CXCL9 and CXCL10 enhance or dampen allergic inflammation may depend on the location and timing of their production. For example, in a mouse model, targeted overexpression of CXCL10 to the lung inhibited the development of Ag-induced airway eosinophilia (41). The authors suggested that this was most likely due to recruitment of Th1-type cells that shifted the balance away from a Th2 response. In contrast, systemic overexpression of CXCL10 augmented airway eosinophilia (42). In this case, CXCL10 was thought to directly recruit eosinophils to the airway. Finally, it is tempting to speculate that eosinophil generation of Th1-type chemokines may play a role in limiting viral infections in atopic patients. In addition to their antiviral effector functions, such as release of RNases (43), eosinophils could contribute to the recruitment of IFN-γ-producing Th1 cells.

In conclusion, this study demonstrated that, depending on the presence of TNF-α plus either IL-4 or IFN-γ in the microenvironment, human eosinophils can release Th2 as well as large amounts of Th1 chemokines, thus demonstrating their plasticity in regulating inflammatory immune responses.

We thank Julie Sedgwick for purification of blood eosinophils; Rose DeGrauw and Katie Clay for their technical assistance; Jacqueline Houtman for assistance with preparation of the manuscript; Mike Evans for assistance with statistical analysis; James Malter, Stephane Esnault, Zhong-Jian Shen, and Beatriz Quinchia-Rios for helpful discussion and technical advice; and Louis Rosenthal for critical review of the manuscript.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported in part by an institutional Specialized Center of Research grant (NIH HL56396) and a University of Wisconsin General Clinical Research Center grant (NIH M01RR03186).

3

Abbreviations used in this paper: GUS, β-glucuronidase; Ct, cycle threshold.

1
Alam, R., W. W. Busse.
2004
. The eosinophil: quo vadis?.
J. Allergy Clin. Immunol.
113
:
38
-42.
2
Adamko, D. J., S. O. Odemuyiwa, D. Vethanayagam, R. Moqbel.
2005
. The rise of the phoenix: the expanding role of the eosinophil in health and disease.
Allergy
60
:
13
-22.
3
Moqbel, R., P. Lacy.
2000
. New concepts in effector functions of eosinophil cytokines.
Clin. Exp. Allergy
30
:
1667
-1671.
4
Peruhype-Magalhaes, V., O. A. Martins-Filho, A. Prata, L. A. Silva, A. Rabello, A. Teixeira-Carvalho, R. M. Figueiredo, S. F. Guimaraes-Carvalho, T. C. Ferrari, R. Correa-Oliveira.
2005
. Immune response in human visceral leishmaniasis: analysis of the correlation between innate immunity cytokine profile and disease outcome.
Scand. J. Immunol.
62
:
487
-495.
5
Shinkai, K., M. Mohrs, R. M. Locksley.
2002
. Helper T cells regulate type-2 innate immunity in vivo.
Nature
420
:
825
-829.
6
Liu, L. Y., S. K. Mathur, J. B. Sedgwick, N. N. Jarjour, W. W. Busse, E. A. B. Kelly.
2006
. Human airway and peripheral blood eosinophils enhance Th1 and Th2 cytokine secretion.
Allergy
61
:
589
-598.
7
Rothenberg, M. E., S. P. Hogan.
2006
. The eosinophil.
Annu. Rev. Immunol.
24
:
147
-174.
8
Bisset, L. R., P. Schmid-Grendelmeier.
2005
. Chemokines and their receptors in the pathogenesis of allergic asthma: progress and perspective.
Curr. Opin. Pulm. Med.
11
:
35
-42.
9
D’Ambrosio, D., M. Mariani, P. Panina-Bordignon, F. Sinigaglia.
2001
. Chemokines and their receptors guiding T lymphocyte recruitment in lung inflammation.
Am. J. Respir. Crit. Care Med.
164
:
1266
-1275.
10
Elsner, J., S. E. Escher, U. Forssmann.
2004
. Chemokine receptor antagonists: a novel therapeutic approach in allergic diseases.
Allergy
59
:
1243
-1258.
11
Smit, J. J., N. W. Lukacs.
2006
. A closer look at chemokines and their role in asthmatic responses.
Eur. J. Pharmacol.
533
:
277
-288.
12
Mikhak, Z., C. M. Fleming, B. D. Medoff, S. Y. Thomas, A. M. Tager, G. S. Campanella, A. D. Luster.
2006
. STAT1 in peripheral tissue differentially regulates homing of antigen-specific Th1 and Th2 cells.
J. Immunol.
176
:
4959
-4967.
13
Mathew, A., J. A. MacLean, E. DeHaan, A. M. Tager, F. H. Green, A. D. Luster.
2001
. Signal transducer and activator of transcription 6 controls chemokine production and T helper cell type 2 cell trafficking in allergic pulmonary inflammation.
J. Exp. Med.
193
:
1087
-1096.
14
Bochner, B. S., S. A. Hudson, H. Q. Xiao, M. C. Liu.
2003
. Release of both CCR4-active and CXCR3-active chemokines during human allergic pulmonary late-phase reactions.
J. Allergy Clin. Immunol.
112
:
930
-934.
15
Pilette, C., J. N. Francis, S. J. Till, S. R. Durham.
2004
. CCR4 ligands are up-regulated in the airways of atopic asthmatics after segmental allergen challenge.
Eur. Respir. J.
23
:
876
-884.
16
Liu, L., N. N. Jarjour, W. W. Busse, E. A. Kelly.
2004
. Enhanced generation of helper T type 1 and 2 chemokines in allergen-induced asthma.
Am. J. Respir. Crit. Care Med.
169
:
1118
-1124.
17
Yamamoto, H., J. B. Sedgwick, W. W. Busse.
1998
. Differential regulation of eosinophil adhesion and transmigration by pulmonary microvascular endothelial cells.
J. Immunol.
161
:
971
-977.
18
Kelly, E. A., R. R. Rodriguez, W. W. Busse, N. N. Jarjour.
1997
. The effect of segmental bronchoprovocation with allergen on airway lymphocyte function.
Am. J. Respir. Crit. Care Med.
156
:
1421
-1428.
19
Bates, M. E., W. W. Busse, P. J. Bertics.
1998
. Interleukin 5 signals through Shc and Grb2 in human eosinophils.
Am. J. Respir. Cell Mol. Biol.
18
:
75
-83.
20
Gouwy, M., S. Struyf, P. Proost, J. Van Damme.
2005
. Synergy in cytokine and chemokine networks amplifies the inflammatory response.
Cytokine Growth Factor Rev.
16
:
561
-580.
21
Dajotoy, T., P. Andersson, A. Bjartell, C. G. Lofdahl, H. Tapper, A. Egesten.
2004
. Human eosinophils produce the T cell-attracting chemokines MIG and IP-10 upon stimulation with IFN-γ.
J. Leukocyte Biol.
76
:
685
-691.
22
Robinson, C. M., K. A. Shirey, J. M. Carlin.
2003
. Synergistic transcriptional activation of indoleamine dioxygenase by IFN-γ and tumor necrosis factor-α.
J. Interferon Cytokine Res.
23
:
413
-421.
23
Krakauer, T., J. J. Oppenheim.
1993
. IL-1 and tumor necrosis factor-α each up-regulate both the expression of IFN-γ receptors and enhance IFN-γ-induced HLA-DR expression on human monocytes and a human monocytic cell line (THP-1).
J. Immunol.
150
:
1205
-1211.
24
Faffe, D. S., T. Whitehead, P. E. Moore, S. Baraldo, L. Flynt, K. Bourgeois, R. A. Panettieri, S. A. Shore.
2003
. IL-13 and IL-4 promote TARC release in human airway smooth muscle cells: role of IL-4 receptor genotype.
Am. J. Physiol.
285
:
L907
-L914.
25
Wirnsberger, G., D. Hebenstreit, G. Posselt, J. Horejs-Hoeck, A. Duschl.
2006
. IL-4 induces expression of TARC/CCL17 via two STAT6 binding sites.
Eur. J. Immunol.
36
:
1882
-1891.
26
Sekiya, T., M. Miyamasu, M. Imanishi, H. Yamada, T. Nakajima, M. Yamaguchi, T. Fujisawa, R. Pawankar, Y. Sano, K. Ohta, et al
2000
. Inducible expression of a Th2-type CC chemokine thymus- and activation- regulated chemokine by human bronchial epithelial cells.
J. Immunol.
165
:
2205
-2213.
27
Kumagai, N., K. Fukuda, T. Nishida.
2000
. Synergistic effect of TNF-α and IL-4 on the expression of thymus- and activation-regulated chemokine in human corneal fibroblasts.
Biochem. Biophys. Res. Commun.
279
:
1
-5.
28
Bonecchi, R., S. Sozzani, J. T. Stine, W. Luini, G. D’Amico, P. Allavena, D. Chantry, A. Mantovani.
1998
. Divergent effects of interleukin-4 and interferon-γ on macrophage-derived chemokine production: an amplification circuit of polarized T helper 2 responses.
Blood
92
:
2668
-2671.
29
Horikawa, T., T. Nakayama, I. Hikita, H. Yamada, R. Fujisawa, T. Bito, S. Harada, A. Fukunaga, D. Chantry, P. W. Gray, et al
2002
. IFN-γ-inducible expression of thymus and activation-regulated chemokine/CCL17 and macrophage-derived chemokine/CCL22 in epidermal keratinocytes and their roles in atopic dermatitis.
Int. Immunol.
14
:
767
-773.
30
Fulkerson, P. C., N. Zimmermann, L. M. Hassman, F. D. Finkelman, M. E. Rothenberg.
2004
. Pulmonary chemokine expression is coordinately regulated by STAT1, STAT6, and IFN-γ.
J. Immunol.
173
:
7565
-7574.
31
Nguyen, V. T., E. N. Benveniste.
2002
. Critical role of tumor necrosis factor-α and NF-κB in interferon-γ-induced CD40 expression in microglia/macrophages.
J. Biol. Chem.
277
:
13796
-13803.
32
Paludan, S. R..
2000
. Synergistic action of pro-inflammatory agents: cellular and molecular aspects.
J. Leukocyte Biol.
67
:
18
-25.
33
Ganster, R. W., Z. Guo, L. Shao, D. A. Geller.
2005
. Differential effects of TNF-α and IFN-γ on gene transcription mediated by NF-κB-Stat1 interactions.
J. Interferon Cytokine Res.
25
:
707
-719.
34
Pine, R..
1997
. Convergence of TNFα and IFNγ signalling pathways through synergistic induction of IRF-1/ISGF-2 is mediated by a composite GAS/κB promoter element.
Nucleic Acids Res.
25
:
4346
-4354.
35
Agresti, C., A. Bernardo, N. Del Russo, G. Marziali, A. Battistini, F. Aloisi, G. Levi, E. M. Coccia.
1998
. Synergistic stimulation of MHC class I and IRF-1 gene expression by IFN-γ and TNF-α in oligodendrocytes.
Eur. J. Neurosci.
10
:
2975
-2983.
36
Liu, L. Y., N. N. Jarjour, W. W. Busse, E. A. Kelly.
2003
. Chemokine receptor expression on human eosinophils from peripheral blood and bronchoalveolar lavage fluid following segmental antigen challenge.
J. Allergy Clin. Immunol.
112
:
556
-562.
37
Jinquan, T., C. Jing, H. H. Jacobi, C. M. Reimert, A. Millner, S. Quan, J. B. Hansen, S. Dissing, H. J. Malling, P. S. Skov, L. K. Poulsen.
2000
. CXCR3 expression and activation of eosinophils: role of IFN-γ-inducible protein-10 and monokine induced by IFN-γ.
J. Immunol.
165
:
1548
-1556.
38
Xanthou, G., C. E. Duchesnes, T. J. Williams, J. E. Pease.
2003
. CCR3 functional responses are regulated by both CXCR3 and its ligands CXCL9, CXCL10 and CXCL11.
Eur. J. Immunol.
33
:
2241
-2250.
39
Loetscher, P., A. Pellegrino, J. H. Gong, I. Mattioli, M. Loetscher, G. Bardi, M. Baggiolini, I. Clark-Lewis.
2001
. The ligands of CXC chemokine receptor 3, I-TAC, Mig, and IP10, are natural antagonists for CCR3.
J. Biol. Chem.
276
:
2986
-2991.
40
Fulkerson, P. C., N. Zimmermann, E. B. Brandt, E. E. Muntel, M. P. Doepker, J. L. Kavanaugh, A. Mishra, D. P. Witte, H. Zhang, J. M. Farber, et al
2004
. Negative regulation of eosinophil recruitment to the lung by the chemokine monokine induced by IFN-γ (Mig, CXCL9).
Proc. Natl. Acad. Sci. USA
101
:
1987
-1992.
41
Wiley, R., K. Palmer, B. Gajewska, M. Stampfli, D. Alvarez, A. Coyle, J. Gutierrez-Ramos, M. Jordana.
2001
. Expression of the Th1 chemokine IFN-γ-inducible protein 10 in the airway alters mucosal allergic sensitization in mice.
J. Immunol.
166
:
2750
-2759.
42
Medoff, B. D., A. Sauty, A. M. Tager, J. A. MacLean, R. N. Smith, A. Mathew, J. H. Dufour, A. D. Luster.
2002
. IFN-γ-inducible protein 10 (CXCL10) contributes to airway hyperreactivity and airway inflammation in a mouse model of asthma.
J. Immunol.
168
:
5278
-5286.
43
Rosenberg, H. F., J. B. Domachowske.
2001
. Eosinophils, eosinophil ribonucleases, and their role in host defense against respiratory virus pathogens.
J. Leukocyte Biol.
70
:
691
-698.