INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS REFERENCES

Several observations link eosinophils to the development and progression of inflammatory diseases such as allergy and asthma (1). In the case of asthma this is supported by findings of increased concentrations of eosinophils in peripheral blood and bronchoalveolar lavage fluid (BALF), as well as pulmonary deposition of injury-mediating eosinophilic compounds in affected individuals (2).

Two distinct families of polypeptides regulate the fate of eosinophils: the eosinophilopoietic cytokines and the chemo-kines. As for the cytokines, granulocyte-macrophage colony-stimulating factor (GM-CSF) is a potent regulator of eosinophil production and function, and promotes the survival of eosinophils by suppressing apoptosis (programmed cell death) (3, 4). Among the chemokines, both RANTES and eotaxin play important roles in controlling the migration of eosinophils to inflamed tissues (5), but whether they contribute to eosinophil survival is unclear. One way to reduce numbers of tissue eosinophils would be to block each of these factors with specific antagonists. An alternative approach is to use molecules that selectively modulate eosinophil apoptosis (6) rather than using glucocorticoids (7), since severe side effects of steroids limit their therapeutic potential.

We recently generated a GM-CSF analogue (E21R) that was devoid of agonistic activity and selectively antagonized GM-CSF-mediated functions in cell lines (8). Surprisingly, E21R induced apoptosis of certain proliferating hemopoietic cells (9). This prompted us to examine whether E21R could regulate the programmed cell death of eosinophils, and in particular to examine its effect on in vivo-activated and chemo-kine-activated eosinophils. We report here that E21R induced apoptosis of unstimulated eosinophils in the absence of GM-CSF. Furthermore, E21R induced apoptosis of eosinophils activated with either RANTES or eotaxin, whereas a RANTES-antagonist did not affect eosinophil survival. Our data indicate that by regulating eosinophil apoptosis, E21R might offer a useful therapeutic approach in diseases characterized by eosinophilia.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS REFERENCES

Purification and Culturing of Eosinophils

Peripheral blood was collected from nine healthy subjects with differential counts of less than 2% eosinophils, and from nine asthmatic patients with differential counts in which eosinophils ranged from 7 to 16%. Written consent was given by each subject, and the study was approved by ethics committees. The eosinophils were isolated after dextran sedimentation of erythrocytes followed by discontinuous metrizamide gradient separation (Metrizamide; Nycomed, Oslo, Norway), yielding a cell purity above 96%, or were subjected to density centrifugation (Histopaque, Sigma, UK) followed by a negative selection procedure using anti-CD16-coated microbeads and the magnetic activated cell sorting (MACS) system (Becton Dickinson, Cowley, UK) (10), yielding a cell purity above 99%. The eosinophils were cultured at 10⁶ cells/ml in RPMI medium supplemented with bicarbonate (0.23% wt/vol), fetal calf serum (FCS) (10% vol/vol), L-glutamine (1.7 mM), penicillin (10.5 µg/ml), and gentamicin (14 µg/ml).

Eosinophilopoietic Cytokines and Chemokines

In our cultures we used human GM-CSF (Genetics Institute, Cambridge, MA) and the GM-CSF analogue E21R (BresaGen, Adelaide, Australia). Recombinant E21R was produced in Escherichia coli as a protein fused to the N-terminus of GM-CSF. Inclusion bodies containing E21R were dissolved and refolded, and the correctly folded E21R was purified by ion-exchange chromatography on Q-Sepharose (Pharmacia, Uppsala, Sweden) and shown to be > 95% pure as judged by sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDS-PAGE).

We also used eotaxin and RANTES (100 µg/ml). To some cultures we added an antagonist to RANTES (100 µg/ml) that was made by truncating the N-terminal end of wild-type RANTES, as previously described (11).

Assessment of Eosinophilic Peroxidase and Eosinophil Cationic Protein Release

Purified eosinophils from four asthmatic subjects were stimulated with either GM-CSF (1 ng/ml), E21R (10 µg/ml), or with E21R (10 µg/ml) for 30 min followed by GM-CSF (1 ng/ml) for 30 min. We then added Sephadex G-15 beads (Sigma Chemical Co.) coated with serum for 30 min. Beads incubated with Hank's balanced salt solution (HBSS) were used as a negative control. Eosinophil peroxidase (EPO) release was measured in supernatants collected from the eosinophil cultures, using an o-phenylenediamine and H₂O₂ color change assay (12), and eosinophil cationic protein (ECP) release was measured with serum-coated beads and a radioimmunoassay (Pharmacia). EPO and ECP releases were each expressed as a percentage of the total cellular content of the two substances as measured after a freeze-thaw cycle.

Measurement of GM-CSF Production

Eosinophils (10⁶ cells/ml) from asthmatic subjects were cultured for 36 h before supernatants were assayed for concentrations of GM-CSF protein with an enzyme-linked immunosorbent assay (ELISA) kit according to the instructions of the manufacturer (R&D Systems, Minneapolis, MN; sensitivity > 0.5 pg/ml).

Determination of Eosinophil Apoptosis

Cells were harvested at selected time points, and the numbers of apoptotic cells were quantitated through the assay of reduced binding of propidium iodide to DNA, using flow cytometry (EPICS Profile II; Coulter Electronics, Hialeah, FL) as described previously (9). Chromosomal DNA was extracted from lysed cells with organic solvents and run on 1.2% agarose gels (9).

Statistics

Each measurement from every subject was done in triplicate, and the corresponding medians were used to calculate mean and SEM values for subjects. We used Pearson's correlation coefficient (r) for data in Table 2. Differences were evaluated with the Kruskal-Wallis test and Bonferroni's test, and were considered significant at p < 0.05. 

	RESULTS and DISCUSSION

E21R Antagonizes GM-CSF-Stimulated Eosinophil Release of EPO and ECP

We have previously shown that E21R acts as a specific GM-CSF receptor antagonist on human leukemic cells and neutrophils (8). To establish that E21R could also act as a GM-CSF antagonist on eosinophils, we first studied EPO and ECP release from eosinophils collected from asthmatic subjects and cultured with cytokines prior to stimulation with serum-coated beads. Figure 1 shows that E21R itself did not alter either EPO or ECP release. GM-CSF markedly stimulated the release of both EPO and ECP, but this was completely inhibited when E21R was added together with GM-CSF. Winqvist and colleagues showed that the mechanism by which serum-coated beads leads to eosinophil degranulation depends on complement (13). In addition, immunoglobulin complexes (such as IgA- and IgG-coated beads) were shown to be potent stimuli for eosinophil degranulation (14). The eosinophil degranulation by immobilized Ig may be mediated by G proteins and tyrosine kinases (15).

View larger version (14K): [in this window] [in a new window]   Figure 1.   GM-CSF-stimulated EPO and ECP release is inhibited by E21R. Cells were incubated in medium, medium with GM-CSF or E21R, or medium with a combination of GM-CSF and E21R. Eosinophil degranulation was stimulated by incubation of eosinophils with serum-coated Sephadex G-15 beads. Total cellular EPO and ECP release was 9,538 ± 845 and 24,845 ± 2,176 ng/106 cells (n = 4), respectively. The EPO or ECP release from eosinophils cultured with beads incubated in HBSS never exceeded 2.3%. Triplicate measurements were made with cells from each of four asthmatic subjects, and the corresponding median for each subject was used to calculate the mean and SEM as shown here. *p < 0.05.

Neither interleukin-3 (IL-3)-nor IL-5-stimulated EPO or ECP release were affected upon adding E21R (data not shown).

E21R Induces Apoptosis of Eosinophils from Healthy Subjects

We next examined the effect of E21R on eosinophil lifespan, using cells from healthy subjects. Figure 2A shows that E21R significantly accelerated apoptosis of eosinophils. By 15 h about 80% of eosinophils cultured with E21R (10 µg/ml) were apoptotic, whereas all cells cultured in medium alone or with GM-CSF (10 ng/ml) remained viable. Most eosinophils kept in medium alone died within 36 h, whereas all those in medium supplemented with GM-CSF survived throughout the observation period (Figure 2A), as previously demonstrated (3, 4). Dose-response experiments showed that 10 µg/ml of E21R induced apoptosis of most eosinophils by 15 h (Figure 2B). Conversely, all eosinophils remained viable when GM-CSF at 10 ng/ml or more was added to cultures supplemented with a fixed dose of E21R (10 µg/ml). These results show that the effect of E21R on eosinophil survival was selectively mediated through the GM-CSF receptor complex and not by nonspecific toxicity.

View larger version (16K): [in this window] [in a new window]   Figure 2.   The GM-CSF analogue E21R induces apoptosis of eosinophils. (A) Time course. Cells from healthy subjects were incubated in medium, medium with GM-CSF (10 ng/ml), or medium with E21R (10 µg/ml). Values are mean and SEM for five subjects. *p < 0.05 compared with medium alone. (B) Dose-response curves for E21R-induced cell death. Eosinophils were cultured for 15 h with increasing doses of E21R or with increasing doses of GM-CSF titrated against 10 µg/ml of E21R. Data are from the same subjects as in (A).

We next sought further evidence that E21R induced apoptosis of eosinophils. Figure 3 shows electron microscopic pictures of an eosinophil kept in medium alone (Figure 3A) or with E21R (Figure 3B, 10 µg/ml) for 15 h, the latter showing morphologic features consistent with apoptosis. DNA extracted from these E21R-treated eosinophils had been degrated into low-molecular-weight fragments, whereas DNA from eosinophils kept in medium alone or with GM-CSF (10 ng/ml) remained intact (Figure 3C). Eosinophils kept in medium alone for 36 h also showed morphologic signs of apoptosis (data not shown). These data are consistent with the notion that eosinophils deprived of hemopoietic survival factors undergo apoptosis (16), and suggest that E21R induces an active apoptotic process evidenced at early time points.

View larger version (40K): [in this window] [in a new window]   Figure 3.   E21R-treated eosinophils exhibit features of apoptosis. Eosinophils from a healthy subject were cultured for 15 h in medium (A) or medium with E21R (B). The E21R-treated eosinophil shows morphologic signs of apoptosis. (C ) DNA was extracted from eosinophils and displayed on a gel. Although DNA from E21R-treated cells showed fragmentation, the DNA isolated from untreated cells or cells treated with GM-CSF remained intact. Equal quantities of DNA were loaded onto each lane. A molecular marker in kilobases is shown at the top. One of five representative experiments is shown.

E21R Induces Apoptosis of Chemokine-Activated Eosinophils

The chemokines RANTES and eotaxin are potent activators of eosinophil function (5), but their role in eosinophil survival is not clear. Although both RANTES and eotaxin stimulated eosinophil migration in vitro (data not shown), Table 1 shows that neither RANTES nor eotaxin altered the survival of eosinophils after 24 h as compared with the survival of cells cultured in medium alone. A RANTES antagonist also failed to influence eosinophil survival, whereas it blocked RANTES-induced eosinophil migration (data not shown), and thus this RANTES antagonist did not behave like E21R. Moreover, E21R induced apoptosis when used alone or in the presence of RANTES and eotaxin (Table 1).

E21R Induces Apoptosis of Eosinophils from Asthmatic Patients

We initially found that eosinophils collected from asthmatic patients survived longer than unstimulated eosinophils from healthy subiects when cultured in medium alone (data not shown). This probably reflects former cells' endogeneous production of survival factors such as GM-CSF, since this cytokine was clearly present in supernatants following a 36-h culture of these cells (Table 2), whereas no GM-CSF was detectable in supernatants collected from eosinophils of healthy controls. Table 2 shows that E21R (10 µg/ml) induced apoptosis in all five cases studied, albeit to a variable degree. The fraction of apoptotic cells was in every case inversely related both to the blood differential count of eosinophils (r = 0.91, p < 0.05) and to eosinophil GM-CSF production (r = 0.94, p < 0.05). Although a fraction of eosinophils producing GM-CSF died when cultured with E21R at 10 µg/ml, virtually all eosinophils were apoptotic at higher concentrations of E21R, with the optimal concentration directly correlating to the amount of GM-CSF being produced (data not shown).

A density shift toward an increased number of hypodense eosinophils in asthmatic patients as compared with normal controls has been suggested as a factor contributing to the disease: hypodense eosinophils have a greater number of various surface receptors and show increased cytotoxicity (4), and immature, hypodense eosinophils enhance their viability by constitutive production of GM-CSF (17). However, no significant difference in cell survival at the 15-h time point was detected when comparing E21R-treated eosinophils from the 23%/24% metrizamide interface with those from the 18%/20% interface (data not shown).

E21R probably affects eosinophil EPO and ECP release and survival via two different mechanisms. Its inhibitory effect on EPO and ECP release most likely results from passive blockade of the GM-CSF receptor (8), since E21R itself had no impact on the release and it had to be present in a thousandfold excess over GM-CSF in order to inhibit EPO and ECP release. This concentration requirement is consistent with the differences in affinity between the GM-CSF analogue E21R and wild-type GM-CSF for the GM-CSF receptor (8).

E21R-induced apoptosis is likely to be the result of an active process, since recent studies of myeloid cells indicate that it involves both transcriptional activation and protein synthesis (9). The data in our study show that by 15 h, E21R induces apoptosis of nearly all eosinophils of healthy subjects. The mechanism by which it does this is not yet clear, but it is probably not the effect of merely antagonizing endogeneous GM-CSF activity, since unstimulated eosinophils do not have detectable endogeneous messenger RNA (mRNA)for GM-CSF or detectable GM-CSF protein, in contrast to activated eosinophils (18). Moreover, we have tested four different monoclonal antibodies raised against the specific GM-CSF receptor  subunit, and none of these antibodies replicated the apoptotic action of E21R (data not shown). As a further point, E21R-induced apoptosis requires serine/threonine protein phosphorylation (9), in contrast to the postulated mandatory role of tyrosine phosphorylation in the antiapoptotic effect mediated by GM-CSF (19).

E21R induced apoptosis of eosinophils cultured with either RANTES or eotaxin. The newly identified receptor CCR-3 mediates functional activation of eosinophils upon binding of either RANTES or eotaxin (20, 21), but its role in regulating eosinophil survival is unknown. Moreover, although the RANTES antagonist can bind to multiple chemokine receptors (11), it had no impact on eosinophil survival. Our findings therefore indicate: (1) that chemokines are not central in the regulation of eosinophil survival or apoptosis; and (2) that engagement of the GM-CSF receptor with E21R leads to eosinophil apoptosis even in the presence of eosinophil-active chemokines.

Xing and coworkers recently showed that overexpression of GM-CSF in vivo in the rat lung induced eosinophilia and fibrotic lesions, suggesting a key role for GM-CSF in certain respiratory disorders such as asthma (22). In addition to being found in eosinophils, mRNA for GM-CSF has been demonstrated in T cells and specimens from respiratory tissue in asthmatic patients (2, 6); however, the possible correlate to actual protein production locally in vivo is not known. Given the dual role of E21R in blocking GM-CSF and in inducing apoptosis of eosinophils, it would be of interest to test its role as a potential new therapeutic agent for asthma and other diseases associated with eosinophilia of different etiologies.