Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Prolonged eosinophil survival, i.e., reduced apoptosis, is implicated in the pathogenesis of chronic allergic inflammation (1). Eosinophil apoptosis, similar to that of other cells, is regulated by a variety of cytokines and cell surface receptors in a multistep process. The classic hematopoietins, interleukin (IL)-3, IL-5, and granulocyte macrophage-colony stimulating factor (GM-CSF), constitute a cytokine family well known to inhibit apoptosis in these cells (4). The signaling pathways for IL-5 and GM-CSF have been partially delineated and found to be dependent on phosphorylation of Lyn and Syk and activation of p21 Ras/Raf-1, Jak-2/Stat, and PI3 kinase pathways (7, 8). Although IL-5, produced by T cells and to a lesser extent eosinophils, is the most specific cytokine for eosinophils, GM-CSF is widely expressed by human myofibroblasts, endothelial cells, airway epithelial cells, and eosinophils themselves, suggesting a potentially important role as both an autocrine and paracrine factor (4, 9). Moreover, GM-CSF production has been shown to be greater in asthmatic airways when compared with controls, and correlates with both the frequency of nonapoptotic eosinophils and macrophages in bronchial biopsy specimens and the severity of asthma (2). These findings underscore the importance of GM-CSF and IL-5 on eosinophil apoptosis in vivo.

Aside from these well-recognized hematopoietins, the family of common  chain-sharing cytokines present in diseases presumed to have a Th2 lymphocyte involvement has also been investigated for its effects on eosinophil survival and functions. IL-2 has been shown to neither prolong nor reduce eosinophil survival. IL-4, a significant Th2 cytokine in allergic diseases, has been shown to induce eosinophil accumulation but, surprisingly, may enhance apoptosis in eosinophils from normal and atopic dermatitis subjects (12). IL-9 has been reported by others to inhibit eosinophil apoptosis via upregulation of IL-5 receptor mRNA (13). IL-13 modestly prolonged eosinophil survival and upregulated the expression of CD69 cell surface protein and mRNA expression (14). Of great interest to our investigation, IL-15 has been shown to prolong NK cell, intestinal intraepithelial lymphocyte, synovial T cell, and neutrophil survival (15- 18). Although effects on other cells have been investigated, the role of IL-15 in eosinophil survival has not yet been established. Notably, IL-15 mRNA is expressed in most tissues, with high levels observed in skeletal muscle, placenta, heart, spleen, liver, kidney, dermal layers of skin, and lung (19), in contrast to the expression IL-2 mRNA, the cytokine it most closely mimics, which is limited to activated T cells and to tissues that contain T cells. A role for IL-15 in allergic diseases has been proposed, as IL-15 induced allergen-specific human T cells to produce IL-5, and when combined with IL-18, induced production of GM-CSF from human NK cells (20, 21). Moreover, a patient with hypereosinophilic syndrome has recently been described as having elevated serum levels of IL-15 (22). From these data, we hypothesize that IL-15 produced at the site of allergic inflammation might play a role in recruitment, activation, and survival of eosinophils. In this study, we provide data to suggest that IL-15 regulates eosinophil apoptosis in vitro via autocrine production of GM-CSF. We also demonstrate that nuclear factor (NF)-B activation is required for IL-15 to exert this effect. Moreover, IL-15 is shown to be present in asthmatic airways. These findings suggest potential involvement of IL-15 in allergic diseases.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Eosinophil Donor Information

Peripheral blood was obtained by venipuncture from 21 different patients with allergic rhinitis or asthma. Subjects were between the ages of 18 and 40, and in otherwise good health. No oral, inhaled, or intranasal corticosteroids were allowed for the previous 4 wk, antihistamines for the previous 5-7 d (long-acting) or 24 h (short acting), and no nonsteroidal anti-inflammatory drugs, antilipid drugs, or ethanol for 48 h before donation. All subjects were nonsmokers, and subjects had not donated blood within the preceding 8 wk. Informed consent was obtained according to a protocol approved by the National Jewish Research and Medical Center IRB Committee.

Eosinophil Isolation and Culture

Eosinophils were purified using a negative immunomagnetic procedure. Briefly, heparinized whole blood was centrifuged (1,000 × g, 30 min) over a Percoll density gradient (density 1.090 g/ml; Pharmacia Biotech, Piscataway, NJ) to separate mononuclear cells from granulocytes. After removal of the mononuclear cell band, erythrocytes were lysed by incubation (2× for 30 s) with sterile, cold water. The remaining white blood cells were incubated with 200 µl of anti-CD16, 100 µl of anti-CD3, and anti-CD14-coated microbeads (Miltenyi Biotec, Auburn, CA) for 60 min and were then passed through steel mesh columns that had been previously washed with PEB (phosphate-buffered saline with 5 mM ethylenediamine tetraacetic acid and 0.5% human serum albumin). The cells in the eluent were stained and examined microscopically. Purity of eosinophils in all experiments was > 99%. After isolation, eosinophils were transferred to 5 ml polypropylene round-bottom tubes in 1-ml aliquots (0.5 million/ml) in Iscove's medium (GIBCO BRL, Grand Island, NY) with 20% fetal calf serum (Hyclone, Logan, UT). Eosinophils were variously stimulated with 100 U/ml of recombinant human tumor necrosis factor (TNF)- (Genzyme, Cambridge, MA), IL-2 (5 ng/ml), IL-4 (5 ng/ml), IL-15 (50 ng/ml) (Pepro Tech Inc., Rocky Hill, NJ), or GM-CSF (100 pg/ml). The concentration of cytokines was chosen from prior titrations (data not shown) of IL-2, IL-4, and IL-15 from 1-100 ng/ml, with or without TNF-, which showed that IL-2 and IL-4, with or without TNF-, did not appear to exhibit a dose-dependent response. The maximal effect of IL-15, with or without TNF-, was seen at 25 ng/ml. For blockade of NF-B activity, eosinophils were incubated with Bay 11-7082 (3 µm), an inhibitor of IB phosphorylation (CalBiochem, San Diego, CA) at 37°C for 30 min before the addition of cytokines. For antibody neutralization of cytokine activity, purified mouse monoclonal anti-human IL-3, IL-5, and GM-CSF antibodies (IgG₁) (Genzyme, Cambridge, MA) (50 µg/ml) were added immediately to unstimulated and stimulated eosinophils. Culture with isotype control antibody had no effect on eosinophils. Cultures were maintained at 37°C in a 5% CO₂ environment. Every effort was made to minimize exposure to lipopolysaccharide (LPS) during eosinophil isolation and culture. Human serum albumin and fetal calf serum were screened for LPS by limulus amebocyte lysate kit (Associates of Cape Cod, Woods Hole, MA): contamination with LPS was less than 25 and 6 pg/ml, respectively.

Assessment of Eosinophil Viability and Apoptosis

Eosinophils were gently vortexed and then removed from each tube at 24, 48, and 72 h. Viability of eosinophils was then assessed at these different time points by trypan blue exclusion. Those eosinophils that excluded trypan blue after a 5-min incubation were considered viable cells. At least 100 eosinophils were counted under the light microscope. Apoptosis was assessed by light microscopy following Kimura staining on the basis of cells exhibiting classic apoptotic morphology: nuclear and cytoplasmic condensation. Flow cytometric analysis of hypodiploid DNA based on propidium iodide staining in a FACSCaliber Flow cytometer (Becton Dickson, Franklin Lakes, NJ) was also used to verify apoptosis. All three methods were highly correlated, with nuclear morphology showing the earliest evidence of apoptosis, followed by DNA fragmentation and loss of trypan blue exclusion. All data are presented as mean ± SEM. Significant differences between groups were determined by one-way analysis of variance (ANOVA) (JMP software program; SAS Institute, Cary, NC). If the overall P value was less than 0.05, the Tukey's test was then used to confirm differences in individual groups. A P value of < 0.05 was considered significant in all cases.

Enzyme-Linked Immunosorbent Assay Measurements

For cytokine measurements, eosinophils were cultured at 2 × 10⁶/ml for 24 to 48 h. This increased cell density did not alter the effects of cytokines on viability. Cells were then pelleted by a 10-min centrifugation at 4°C and the supernatants carefully removed and immediately frozen at 70°C for analysis of cytokine production. Cytokine enzyme-linked immunosorbent assays (ELISAs) were performed according to the manufacturer's instructions (Pharmingen, San Diego, CA). OptEIA reagents were used for these ELISAs and sensitivity was ~ 7 pg/ml.

p65 Nuclear Translocation by Confocal Microscopy

Immunostaining was performed as described by Thomassen and colleagues (23). Briefly, eosinophils were resuspended at 1 × 10⁶ cells/ml. IL-15, TNF-, GM-CSF, and Bay 11-7082 were added as in the apoptosis studies described previously. After 15 min, samples were placed on ice, cytospun, and fixed in acetone for 30 s. Cells were then permeablized and blocked in 0.1% Triton plus 2% goat serum for 30 min. Samples were then incubated with 1 µg of rabbit polyclonal antibody against human p65 (Santa Cruz Biotech, Santa Cruz, CA) overnight at 4°C after which they were washed twice and chromotrope-2R (0.2%) added for 30 min to block nonspecific granular staining by secondary antibody as described by Dewson and colleagues (24). Samples were washed twice and 1 µg of Cy3 goat anti-rabbit and 10 µM DAPI added for 1 h, at which point they were washed twice, mounted, and imaged under oil using a Leica microscope (Leica Inc., Deerfield, IL) and Slidebook software (Denver, CO). Eosinophils were serially sectioned, and deconvolved images were obtained from equatorial sections. Isotype control antibody staining was performed and was negligible.

Immunohistochemical Staining of IL-15 in Endobronchial Biopsies

Endobronchial biopsies were obtained from subjects with asthma and nonatopic normal subjects. All asthmatic subjects met the American Thoracic Society criteria for asthma. Among the 13 subjects with asthma, 2 were classified as moderate and 11 as severe asthmatics. Moderate asthmatics had an FEV₁ of < 80% predicted, were on -agonists and inhaled corticosteroids or theophylline, and did not have a history of urgent health care utilization or oral corticosteroid use. Subjects with severe asthma were patients referred to our institute for severe, oral corticosteroid-dependent asthma with frequent hospitalization and emergency room visits. Normal control subjects (n = 5) had no history of respiratory diseases and were on no medications. The protocol was approved by the institutional review board, and all subjects gave informed consent. Briefly, specimens were fixed in acetone at 20°C overnight and then embedded in glycol methacrylate resin. Tissue blocks were stored at 20°C. Serial 2-µM sections were cut from well-oriented tissue blocks with a Reichert Ultracut E ultramicrotome (Leica).

Tissue sections were stained using goat antibody against human IL-15 (R&D Systems, Inc., Minneapolis, MN). A three-step indirect immunostaining method was used. Sections were treated with 0.3% hydrogen peroxide in 0.05 M Tris-buffered saline (TBS) (pH 7.6) for 30 min to inhibit endogenous peroxidase, followed by incubation with 1% normal goat serum for 30 min to block potential nonspecific binding sites. The slides were then incubated with primary antibody for 2 h at room temperature, followed by incubation with biotinylated rabbit anti-goat IgG for 1 h at room temperature. Thereafter, avidin-biotin-peroxidase complex (Vecter Labs, Burlingame, CA) was added to the slides for 45 min at room temperature. After rinsing the slides in TBS, 0.03% aminoethylcarbazole (AEC) in 0.03% hydrogen peroxide was used as a substrate to develop a peroxide-dependent red color reaction. Slides then were counterstained with Mayer's hematoxylin and covered with Crystalmount (Biomeda Corp., Foster City, CA). Appropriate control slides were similarly treated but with the primary antibody replaced by nonimmune serum or TBS. One slide from each tissue block was stained with hematoxylin-eosin for routine histology.

Analyses of IL-15-Positive Cells

A National Institutes of Health (NIH) scion image analysis program was used to measure the submucosa area, which included the subepithelial basement membrane (SBM), and deeper submucosa was defined as the area between the external side of the SBM and the internal side of the smooth muscle and/or mucous glands. Blood vessels, smooth muscle, and mucous glands were excluded. Nucleated cells positive for IL-15 were counted at ×400 magnification (measured area, 0.5 [range 0.3-0.7] mm²). The cell count was expressed as number of cells/mm². Comparisons between tissues from subjects with asthma and normal controls were made using the Wilcoxon Rank Sum test and group data expressed as medians ± interquartile range (IQR). The observer was blinded to the clinical state of the subjects.

Double-Immunofluorescent Staining

To identify the cellular sources of IL-15-positive cells in the submucosa of endobronchial biopsy tissue, double-immunofluorescent staining was performed in two subjects with severe asthma. Tissue sections were incubated with a mixture of a mouse monoclonal cell marker antibody and a goat polyclonal IL-15 antibody, followed by incubation with a mixture of FITC-conjugated anti-goat and rhodamine-conjugated anti-mouse secondary antibodies (Dako, Carpinteria, CA). The slides were mounted and observed with an Olympus BX51 immunofluorescence microscope. Control slides were similarly treated but with the primary antibodies replaced by nonimmune goat serum and mouse IgG.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Inhibition of Spontaneous Eosinophil Apoptosis

IL-15 was assessed for its ability to modulate eosinophil survival in vitro, and its effects were contrasted to those of the other common  chain receptor-sharing cytokines. Culture of control eosinophils was followed by a loss in viability determined by trypan blue exclusion in a time-dependent fashion (Figure 1A). In contrast to IL-2 and IL-4, which had only modest effects, IL-15 significantly enhanced eosinophil survival when compared with controls, and the effect was most pronounced at 72 h (59.5 ± 5.9% versus 23.2 ± 3.6%, P < 0.0001). Additionally, the effect of costimulation with TNF- on eosinophil survival was examined, as it is often present in the Th2 milieu and plays a pivotal role in regulation of apoptosis in other cell types. While TNF- alone appeared to have a modest effect on prolonging eosinophil survival, as previously reported (25), TNF- only slightly augmented the pro-survival effect seen from IL-15 alone (Figure 1B). Of note, when TNF- was combined with IL-4, a significant prosurvival effect (which was not seen with IL-4 alone) became evident (67.8 ± 3.8 versus 23.2 ± 3.6 for controls, P < 0.0001). Although the combination of TNF- and IL-2 appeared to enhance eosinophil viability (versus controls), the effect was not different than that seen from TNF- alone (P = 0.26).

View larger version (22K): [in this window] [in a new window]   Figure 1.   Survival of eosinophils cultured with IL-2, IL-4, and IL-15 either alone (A), or in combination with TNF- (B), assessed by trypan blue exclusion. (A) IL-15 significantly enhanced eosinophil survival over controls at 72 h, whereas IL-2 and IL-4 did not. (B) Although TNF- modestly enhanced eosinophil survival in the presence and absence of IL-15, when TNF- was combined with IL-4, a significant pro-survival effect over controls and TNF- alone was demonstrated. Although the combination of TNF- and IL-2 appeared to enhance eosinophil viability over controls, the effect was not different than that seen from TNF- alone (P = 0.26). *P < 0.001 versus control, P < 0.0001 versus TNF-. Data are represented as mean ± SEM values of 9-14 experiments with eosinophils from different subjects.

Loss of viability in these cultures was due to apoptosis as confirmed by DNA fragmentation assessed by flow cytometry and by examination of nuclear morphology (Figure 2). The number of cells exhibiting hypodiploid DNA was significantly reduced (P < 0.001, n = 5) in eosinophils cultured with IL-15 alone or in combination with TNF- (Figure 2, top panel). This was in contrast to IL-4 (shown) and IL-2 (not shown), which did not alter spontaneous eosinophil apoptosis when compared with controls. TNF- by itself modestly reduced the amount of hypodiploid DNA. As above, the combination of TNF- and IL-4 inhibited DNA fragmentation to the same degree as IL-15. By contrast, the effect of IL-2 in combination with TNF- was not significantly different than that seen with TNF- alone (data not shown). Similar results were obtained with Kimura staining of cells, showing significantly fewer pyknotic nuclei in eosinophils cultured with IL-15 alone or with TNF- and IL-4 with TNF- (Figure 2, bottom panel). DNA fragmentation and nuclear morphology results were highly correlated with viability (data not shown), demonstrating that IL-15 alone or in combination with TNF-, and IL-4 combined with TNF-, maintained eosinophil viability by inhibiting spontaneous apoptosis.

View larger version (16K): [in this window] [in a new window]   View larger version (129K): [in this window] [in a new window]   Figure 2.   Assessment of apoptosis at 72 h of culture by DNA fragmentation (top panel) and nuclear morphology (bottom panel). (A) Control; (B) TNF-; (C) IL-15; (D) TNF- and IL-15; (E) IL-4; (F) TNF- and IL-4. Top panel: DNA fragmentation was assessed by flow cytometric analysis following staining of DNA with propidium iodide. Apoptosis (percentage of eosinophils containing hypodiploid DNA, shown as % of cells) was reduced when eosinophils were cultured with IL-15, either alone or in combination with TNF-, and with the combination of IL-4 and TNF-. Bottom panel: Pyknotic nuclei as seen from Kimura staining of cells corresponded with the findings from DNA analysis. Results are representative of at least seven experiments for each condition.

We then investigated by ELISA (sensitivities of > 7 pg/ml) whether IL-15 and the other common  chain receptor- sharing cytokines stimulated the production of the classic hematopoietins, or TNF-, from the cultured eosinophils and hence prolonged eosinophil survival. GM-CSF, IL-5, and TNF- were not detected in supernatants collected at 24 and 48 h from any conditions of eosinophil culture (data not shown). In an alternative approach to this question, neutralizing antibodies against GM-CSF, IL-3, and IL-5 at 50 µg/ml were added to the culture immediately after the addition of IL-15. Whereas anti-IL-5 had little to no effect on cells stimulated with IL-15, with or without TNF- (Figure 3), anti-GM-CSF demonstrated a significant inhibitory effect throughout 72 h, reducing the survival of IL-15-stimulated eosinophils nearly to that seen in unstimulated eosinophils. These data suggest that IL-15 stimulated eosinophils to secrete small amounts of GM-CSF below the detectable level of ELISA but sufficient to prolong survival of the cells. The effect of anti-IL-3 was initially similar to that of anti-GM-CSF, suggesting an early autocrine effect of IL-3 but, over the complete time course, was not significantly different (P = 0.13) compared with eosinophils stimulated with IL-15. Anti-GM-CSF, and to a lesser extent anti-IL-3, also exhibited similar inhibition of the enhanced survival in response to combined stimulation with TNF- and IL-4, whereas anti-IL-5, again, had no effect (data not shown).

View larger version (26K): [in this window] [in a new window]   Figure 3.   Effects of neutralizing antibodies to GM-CSF, IL-3, and IL-5. (A) Eosinophils were cultured with IL-15 in the presence or absence of neutralizing antibodies. The pro-survival effect of IL-15 was completely abrogated when neutralizing antibody to GM-CSF was added to the culture. Neutralizing antibody to IL-5 did not alter the prosurvival effect of IL-15 Neutralizing antibody to IL-3 partially inhibited the pro-survival effect of IL-15. (B) A similar result was seen when these neutralizing antibodies were used in the presence of IL-15 in combination with TNF-. *P < 0.05 versus control; P < 0.01 versus IL-15; P = 0.02 versus TNF- plus IL-15. Data are represented as mean ± SEM values of five to seven experiments, with eosinophils from different subjects.

NF-B Activation

Recently, NF-B, a ubiquitous transcription factor, has been implicated in regulation of apoptosis and activation of granulocytes (16, 26) including eosinophils (29, 30). Furthermore, as IL-15 has been demonstrated to result in NF-B activation in neutrophils (26), we asked whether this transcription factor was involved in the observed effects of IL-15 in eosinophils. Accordingly, freshly isolated eosinophils were incubated with Bay 11-7082 at 3 µM for 30 min before the addition of cytokines. Although Bay 11- 7082 (an inhibitor of IB phosphorylation) at 3 µM alone did not alter survival of unstimulated eosinophils at 24, 48, or 72 h, it completely inhibited the prosurvival effect of IL-15 (Figure 4A). Furthermore, a similar result was seen when this inhibitor was used with TNF- and IL-15 (Figure 4B) and with TNF- and IL-4 (data not shown). Interestingly, the pro-survival effect of TNF- alone (Figure 4C) was converted to an enhanced apoptotic signal if NF-B was blocked by Bay 11-7082 (viability of 16.2 ± 5.3 versus 42 ± 8.2% for controls at 48 h, P = 0.02), confirming the findings of Ward and colleagues (28).

View larger version (25K): [in this window] [in a new window]   Figure 4.   Effects of the NF-B inhibitor, Bay 11-7082. (A) Bay 11-7082 at 3 µM alone did not alter the survival rate of eosinophils cultured in medium alone. However, Bay 11-7082 almost completely inhibited the pro-survival effect of IL-15. (B) A similar effect was obtained when Bay 11-7082 was added to eosinophils cultured with IL-15 in combination with TNF-. (C) A proapoptotic effect of TNF- alone was unmasked when NF-B activation was inhibited by Bay 11-7082. (D) Inhibition of NF-B activation by Bay 11-7082 also resulted in abrogation of pro-survival effect seen from exogenous GM-CSF. *P < 0.05 versus control; P < 0.05 versus IL-15; and P < 0.0001 versus TNF- plus IL-15. Data are represented as mean ± SEM values of seven to nine experiments, with eosinophils from different subjects.

As shown in Figure 5, eosinophils treated with Bay 11- 7082 in the presence and absence of cytokines clearly showed increased hypodiploid DNA and pyknotic nuclei, showing that the eosinophils had undergone apoptotic and not necrotic cell death. Therefore, it appeared that NF-B activation is involved in regulation of spontaneous eosinophil apoptosis by these cytokines. Further evidence for a link between the pro-survival effect of IL-15 and NF-B pathways was provided by confocal microscopy (Figure 6). Nuclear translocation of the p65 subunit, indicative of NF-B activation, was noticeably increased after 15-min stimulation with TNF-, IL-15, or the combination of TNF- and IL-15 (or TNF- and IL-4; data not shown) when compared with control eosinophils. Importantly, and as predicted, p65 translocation in stimulated eosinophils, under all conditions, was inhibited by Bay 11-7082. These data further suggest that NF-B plays a role in the reduction of spontaneous apoptosis by IL-15, with or without TNF-.

View larger version (22K): [in this window] [in a new window]   View larger version (129K): [in this window] [in a new window]   Figure 5.   Reduction of eosinophil viability by Bay 11-7082 was due to apoptosis. Eosinophils were cultured for 72 h under the following conditions: (A) control; (B) Bay 11-7082; (C) IL-15; (D) Bay 11-7082 and IL-15; (E) TNF- and IL-15; and (F) TNF-, IL-15, and Bay 11-7082. Left panel: DNA fragmentation was assessed by staining with propidium iodide and flow cytometric analysis as in Figure 2. Although IL-15, alone or in combination with TNF-, decreased the number of apoptotic eosinophils (demonstrating hypodiploid DNA), addition of Bay 11-7082 to the cultures resulted in increased apoptosis. Right panel: Pyknotic nuclei as seen from Kimura staining of cells corresponded with the findings from DNA analysis. Results are representative of at least three experiments for each condition.

View larger version (53K): [in this window] [in a new window]   Figure 6.   NF-B activation demonstrated by fluorescent confocal microscopy: p65 nuclear translocation in control eosinophils compared with those stimulated with TNF-, IL-15, or the combination of IL-15 and TNF- (15 min) and those pretreated with the NF-B inhibitor, Bay 11-7082, before stimulation. Left panel shows localization of p65 and right panel shows nuclei by DAPI staining. Results are representative of three experiments from different individuals.

Because autocrine production of GM-CSF was required for the pro-survival effect of IL-15 (and TNF- with IL-15 or IL-4), and GM-CSF has recently been shown to induce low levels of NF-B nuclear translocation in eosinophils (30), we asked whether NF-B can also act downstream of GM-CSF. Exogenous GM-CSF at 100 pg/ml yielded comparable, prolonged survival to IL-15 (50 ng/ml), and this pro-survival effect was also inhibited by Bay 11- 7082 (Figure 4D). However, we could detect no evidence for p65 translocation by confocal microscopy at 5, 10, 20, 30, or 60 min following stimulation with GM-CSF up to 250 pg/ml (data not shown). Thus, while we have demonstrated that NF-B appears to act upstream of GM-CSF production in IL-15-stimulated eosinophils, it is also possible that it is acting downstream of GM-CSF (see DISCUSSION).

Increased Expression of IL-15 in Airway Tissue of Asthmatic Individuals

To explore a potential role of IL-15 as a regulatory cytokine of eosinophil apoptosis in vivo, we measured IL-15 expression by immunohistochemistry in endobronchial biopsy tissues from nine subjects with moderate to severe asthma and four normal individuals. IL-15 expression in epithelial cells varied greatly among these subjects and did not reach statistical difference between subjects with asthma and control groups. Of note, however, subjects with asthma appeared to have a significantly higher number of IL-15-positive submucosal cells/mm² than control subjects (medians of 35 [19-79, IQR] versus 19 [4-28, IQR], respectively, P < 0.05) (Figures 7A and 7B). Further investigation of the inflammatory infiltrate identified CD68+ macrophages as the predominant source of IL-15 (Figures 7C and 7D), consistent with other inflammatory infiltrates (31). Hence, our results suggest that IL-15 is present in airways of subjects with asthma and may contribute to prolonging eosinophil survival in vivo.

View larger version (124K): [in this window] [in a new window]   Figure 7.   IL-15 expression in control (A) and asthmatic airway tissue (B). Representative photomicrographs of staining for IL-15 in endobronchial biopsy samples. The epithelial staining for IL-15 varied greatly among both control subjects and individuals with asthma, with no significant difference demonstrated between the two groups. However, the number of positively stained cells (black arrows) in submucosal areas of subjects with asthma was significantly higher than the control group, (medians of 35 [19-79, IQR] versus 19 [4-28, IQR] per mm2, respectively, P < 0.05 by Wilcoxon Rank Sum test). Measurements were performed by immunohistochemistry in tissues obtained from nine subjects with asthma and four normal controls. Double-immunofluorescent staining of tissue from a subject with asthma to identify submucosal cells positive for IL-15 demonstrated that IL-15 (C) was most strongly associated with CD68+ macrophages (D) (white arrows).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Similar to its pro-survival effect on NK cells, T cells, and neutrophils (16), IL-15 inhibited spontaneous apoptosis of human peripheral blood eosinophils in vitro. The mechanism appeared to involve autocrine production of GM-CSF, and to a lesser extent, IL-3, but not IL-5. Although we were unable to detect GM-CSF by ELISA, the amount produced was presumably sufficient to prolong eosinophil survival as evident by blockade with neutralizing antibody to GM-CSF. This phenomenon has been reported by Esnault and colleagues where peripheral blood eosinophils successfully transfected with GM-CSF mRNA showed significantly increased survival despite undetectable secretion of GM-CSF (32). Furthermore, these investigators showed that conditioned media from the transfected eosinophils prolonged survival of nontransfected eosinophils. One possibility is that secreted GM-CSF is readily bound to its receptor and therefore unavailable to the assay. Of interest, when employing exogenous GM-CSF, a concentration of 100 pg/ml was needed in our studies to achieve comparable survival as seen with IL-15. Thus, it is conceivable that endogenously produced GM-CSF has greater biologic potency or t_1/2 than recombinant protein. This may represent a unique property of GM-CSF, which appears to be the predominant pro-survival cytokine produced by eosinophils themselves (33).

We have demonstrated that activation of NF-B is a required downstream event in the inhibition of eosinophil apoptosis by IL-15. NF-B plays a pivotal role in apoptosis regulation of many cells (28, 37). The canonical form of NF-B is a heterodimer, which usually consists of two proteins, a p65 (RelA) subunit and a p50 subunit. Activation of NF-B is regulated by IB proteins, which, in response to stimuli, are phosphorylated, polyubiquitinated, and finally degraded by the proteasome. The released NF-B is then allowed to enter the nucleus, where it binds to specific sequences of target genes (40). Like IL-15, TNF- also activated NF-B, and in keeping with previous reports, had a modest effect on inhibiting eosinophil apoptosis. Of interest, when activation of NF-B was inhibited, the proapoptotic effect of TNF- on eosinophils was unmasked. This is consistent with findings from Ward and colleagues, and supports a role for NF-B as a pivotal regulator in eosinophil apoptosis (28).

We also investigated the effects of other common  chain receptor-sharing cytokines on eosinophil apoptosis. IL-2, an important cytokine for T cell survival, did not alter eosinophil survival. Similar to IL-2, IL-4 by itself neither prolonged nor enhanced spontaneous eosinophil survival. These findings are in contrast to a previous report by Wedi and colleagues, which showed that IL-4 alone enhanced spontaneous eosinophil apoptosis (34). We were unable to reproduce this result using the same concentrations of IL-4. This discrepancy might be due to the difference in the in vivo environment of eosinophils harvested for each study, i.e., normal donors by Wedi and colleagues and donors with allergies here. However, we found that when combined with TNF-, IL-4 significantly enhanced eosinophil viability, and this synergism is the focus of ongoing investigation. Similar to IL-15, the pro-survival effect of IL-4 combined with TNF- appeared to be mediated through the autocrine production of GM-CSF and dependent upon NF-B activation. Although the presence of the individual chains of each cytokine receptor, i.e., IL-2R, IL-2R, IL-4R, and IL-9R, has been established for eosinophils, IL-15R expression, which is required for IL-15 high-affinity binding (13, 41, 42), awaits availability of appropriate reagents. However, since IL-15 shares the receptors with IL-2, and in view of the differential effects of IL-15 and IL-2 on eosinophils, our results support the existence of IL-15R on human eosinophils and the importance of IL-15 as a member of this cytokine family that can by itself prolong eosinophil survival.

One set of target genes, which might be involved in this antiapoptotic pathway induced by IL-15, are those of the Bcl-2 family. Salmon and colleagues have shown that synovial T cells exposed to IL-15 in vitro showed upregulation of Bcl-2 and Bcl-X_L resulting in decreased spontaneous apoptosis of these cells (18). The roles of these Bcl-2 family proteins in eosinophil apoptosis have been studied by others, although the results from these studies have been inconclusive. Although it appears that GM-CSF can upregulate the expression of Bcl-X_L/S and Bcl-2 in eosinophils (43), we were unable to show that IL-15 modulated these members of Bcl-2 family proteins (data not shown), perhaps due to insensitivity of the assay. Alternatively, it is conceivable that IL-15 stimulation involves another yet to be identified anti-apoptotic protein in eosinophils such as A1, which, along with Bcl-X, are potential NF-B-dependent candidates. Whether the expression of A1 in eosinophils can be induced by prosurvival cytokines has not yet been studied and deserves further exploration.

We have shown that autocrine production of GM-CSF and NF-B activation are required for IL-15 to prolong eosinophil survival, and thus our data are concordant with earlier demonstrations of NF-B-mediated regulation of GM-CSF transcription (46, 47). From our data, it is, however, unclear whether GM-CSF itself activates NF-B as a downstream event in eosinophils. Although the inhibitor Bay 11-7082 inhibited the prosurvival effect of GM-CSF at 48 and 72 h, we were unable to demonstrate p65 translocation by confocal microscopy, perhaps because of sensitivity or timing of the assay. Further, we note that data using EMSA has shown very low levels of NF-B translocation in eosinophils stimulated with GM-CSF (10 ng/ml for 30 minutes) (30). In neutrophils, GM-CSF has been shown to activate the PI 3-kinase/Akt pathway, resulting in delayed apoptosis (48). Recent studies suggest that Akt is activated in blood eosinophils isolated from allergic donors (49) perhaps explaining why PI3 kinase inhibitors do not enhance eosinophil apoptosis (21) (and indeed have no effect on eosinophils stimulated with IL-15; data not shown). Alternatively, Akt can also be activated independently of PI3 kinase (50), suggesting an alternative pathway for Akt to promote eosinophil survival. Interestingly, Akt has recently been shown to activate NF-B via activation of IB degradation (51, 52). Therefore, we speculate that in addition to mediating IL-15-induced autocrine production of GM-CSF as our data indicates, NF-B may also be activated downstream following GM-CSF activation of Akt.

Lastly, the presence of IL-15 in vivo was explored in a pilot study of subjects with mostly severe asthma. Although the epithelial staining for IL-15 varied greatly among both control and asthmatic individuals, the submucosa of the asthmatic group had a significantly higher number of positively stained cells, predominantly CD68⁺ macrophages, compared with control subjects. Although many questions are raised by these preliminary data regarding both disease severity and the effects of medications on IL-15 expression, these data, along with initial reports associating IL-15 with Th2 inflammation (21, 22) suggest a role for IL-15 as an eosinophil pro-survival cytokine in asthmatic tissue. Collectively, we propose that IL-15 plays a role in allergic diseases by inhibiting eosinophil apoptosis and that its anti-apoptotic effects appear to be mediated through autocrine production of GM-CSF and ultimately by NF-B activation.