Simultaneous Antagonism of Interleukin-5, Granulocyte-Macrophage Colony-Stimulating Factor, and Interleukin-3 Stimulation of Human Eosinophils by Targetting the Common Cytokine Binding Site of Their Receptors

HUMAN INTERLEUKIN-5 (IL-5), IL-3, and granulocyte-macrophage colony-stimulating factor (GM-CSF) are cytokines involved in hematopoiesis and inflammation.1 All 3 cytokines stimulate eosinophil production, function, and survival2-6 and therefore have the ability to influence inflammatory diseases, such as asthma, atopic dermatitis, and allergic rhinitis, in which the eosinophil plays a major effector role.7 IL-5, being the eosinophil-specific cytokine, has received most of the initial attention, with IL-5 mRNA and protein levels noted to be elevated in lung tissue and bronchoalveolar lavage fluid from symptomatic asthma patients.8-10 Correlations between IL-5 levels and allergen challenge and disease activity have also been seen. However, it is becoming apparent that not only IL-5, but also GM-CSF and IL-3 play a role in eosinophil production and activation in asthma, because there is evidence of both GM-CSF and IL-3 production at sites of allergic inflammation.11-17 The concomitant expression of these cytokines probably contributes to the total number of infiltrating eosinophils as well as to the degree of eosinophil activation. In addition, each of these cytokines may be responsible for the different phases and compartmentalization of the eosinophil infiltrate. Recent kinetic data from patients undergoing antigen challenge showed that IL-5 levels increased between days 2 and 7 postchallenge, whereas GM-CSF peaked at day 2 and remained elevated through to day 16. Furthermore, detection of GM-CSF extended beyond the site of allergen challenge.18 

IL-5, GM-CSF, and IL-3 stimulate eosinophils and other cells by binding to cell surface receptors that comprise a ligand-specific α chain and a β chain that is shared by the 3 receptors (β_c).1 Binding to each receptor α chain is the initial step in receptor activation; however, engagement of either α chain alone is not sufficient for activation to occur. Recruitment of β_c by each ligand:α chain complex follows, a step that has 2 major functional consequences. Firstly, it allows the binding of IL-5, GM-CSF, and IL-3 to become essentially irreversible. Secondly, it leads to full receptor activation. β_c is the major signaling component of these receptors, and its engagement leads to the activation of JAK-2, STAT-5, and other signaling molecules,19 culminating in the full plethora of cellular activities commonly associated with either IL-5, GM-CSF, or IL-3 stimulation, such as eosinophil adherence, priming for degranulation and cytotoxicity, and prolongation of viability.

To block or antagonize the activity of eosinophil-activating cytokines in vivo, 3 major approaches are being tried. One of them uses antibodies to the implicated cytokines. For example, antibodies to human IL-5 are being used in an animal model of allergen-induced asthma20,21 and have been shown to have a relatively long-lasting effect in preventing eosinophil influx into the airways and bronchial hyperresponsiveness. A second approach relies on IL-5 or GM-CSF mutants that can bind to their respective α chains with wild-type affinity but that have lost or shown reduced ability to interact with human β_c. IL-5 mutants such as E13Q, E13K, and E13R and the human GM-CSF mutant E21R directly antagonize the functional activation of eosinophils by IL-5 or GM-CSF, respectively.22,23 However, at least in the case of E13K, eosinophil survival is not antagonized, and, in fact, this mutant is able to support eosinophil survival.24 A third approach involves the use of soluble receptor α chains that can sequester circulating cytokines.25 However, this carries the risk of a cytokine:receptor α chain complex potentially interacting with surface-expressed β_c and triggering receptor activation. The common theme among these approaches is that they tackle a single receptor system involving either IL-5, GM-CSF, or IL-3, leaving the other 2 eosinophil-acting cytokines unaffected. Although the concomitant administration of IL-5 and GM-CSF antagonists may be considered, this could be clinically impracticable.

An alternative approach to blocking eosinophil-activating cytokines involves targeting β_c. Although β_c does not directly bind IL-5, GM-CSF, or IL-3 alone, it does associate with these cytokines complexed to the appropriate receptor α chain. Recent studies have identified the major binding sites of β_c for the IL-5:IL-5Rα, GM-CSF:GM-CSFRα, and IL-3:IL-3Rα complexes. Significantly, these sites are used by all 3 complexes and comprise the predicted B-C loop and F-G loops in the membrane proximal domain of β_c.26-28 Thus, targetting β_c is not only desirable, but also feasible, with the potential to allow the simultaneous inhibition of IL-5, GM-CSF, and IL-3 action by a single agent.

We show here the development of the first simultaneous antagonist of IL-5, GM-CSF, and IL-3 in the form of monoclonal antibody (MoAb) BION-1 directed against the major cytokine binding region of β_c. BION-1 blocked the production and activation of human eosinophils stimulated by IL-5, GM-CSF, or IL-3 by inhibiting the high-affinity binding of all 3 cytokines to eosinophils and preventing receptor heterodimerization and β_c phosphorylation. The results demonstrate the feasibility of blocking the 3 eosinophilopoietic cytokines with a single compound and support the notion of simultaneously inhibiting more than 1 cytokine by targetting the shared signaling subunit in their receptors.

Domain 4 of β_c was expressed with an N-terminal Flag-epitope in the form of an activated β_c mutant, ΔQP.29 An extracellular deletion removes domains 1 to 3 in this construct, but the transmembrane and cytoplasmic regions are retained. This was cloned into the eukaryotic expression vector pcDNA3 (Invitrogen, San Diego, CA).

Recombinant human IL-3 and GM-CSF were produced in Escherichia coli as described.22,30 Recombinant human IL-5 was purified from E coli by Bresagen (Adelaide, South Australia). Tumor necrosis factor α (TNFα) was a gift from Dr J. Gamble (Hanson Centre for Cancer Research, Adelaide, Australia). COS cells were transfected with receptor cDNA as described previously.26CHO β_c and CHO ΔQP cells stably expressing either full-length β_c or ΔQP, respectively, were generated by electroporation.22 TF1.8 cells were a gift from Dr J. Tavernier (University of Gent, Gent, Belgium). MO7e cells, a human megakaryoblastic cell line, were from Dr P. Crozier (Auckland, New Zealand). Human eosinophils were purified from the peripheral blood of normal or slightly eosinophilic volunteers via sedimentation through dextran and centrifugation through a discontinuous density gradient of hypertonic metrizamide, as previously described,31 and were greater than 95% pure. Human neutrophils and monocytes were purified from peripheral blood as described previously,2 32 with greater than 95% purity.

BALB/c mice were immunized intraperitonally with 1 × 10⁷ COS cells transfected with β_c or ΔQP expression constructs. The immunizations were repeated 4 times at 2-week intervals. Four weeks after the final immunization, a mouse was boosted with 2 × 10⁶ COS transfectants intravenously. Three days later, splenocytes were harvested and fused with NS-1 myeloma cells as previously described.33 Hybridoma supernatants were screened on CHO β_c or CHO ΔQP cells by flow cytometry, with untransfected CHO cells as a control. All antibodies were from single hybridoma clones as selected by a limiting dilution method. Using this procedure, several anti-β_cMoAbs were generated; 8B8, 8E4,34 and 1C135 are nonfunctional MoAbs raised against full-length β_c and have been used as control anti-β_c MoAbs throughout these studies. MoAbs were purified from ascites fluid or hybridoma supernatant by a protein A sepharose column. The isotypes of MoAbs were tested with a Mouse MoAb Isotyping Kit (Boehringer Mannheim, Mannheim, Germany). Fab fragments were generated using a Fab Preparation kit (Pierce, Rockford, IL) following the supplied protocol.

Freshly purified neutrophils, eosinophils, monocytes, or CHO and COS cell transfectants (5 × 10⁵) were incubated with 50 μL of hybridoma supernatant or 0.25 μg of purified MoAb for 45 to 60 minutes at 4°C. Cells were washed twice and then incubated with fluorescein isothiocyanate (FITC)-conjugated rabbit antimouse Ig (Silenus, Hawthorn, Victoria, Australia) for another 30 to 45 minutes. Cell were then washed and fixed before analyzing their fluorescence intensity on an EPICS-Profile II Flow Cytometer (Coulter Electronics, Hialeah, FL).

IL-3 and GM-CSF were radio-iodinated using the iodine monochloride method.36,¹²⁵I–IL-5 was purchased from Dupont NEN (North Sydney, New South Wales, Australia). MoAb BION-1 was radio-iodinated using the chloramine T method.37 Binding assays were performed as previously described.38 Briefly, 1 to 2 × 10⁶ cells were preincubated with either BION-1 whole IgG, Fab fragments, or control MoAbs or with a range of concentrations of IL-3 or TNFα for 1 hour at 4°C. Radio-labeled ligand was then added and incubated for a further 2 hours before the cells were separated from free label by centrifugation through fetal calf serum (FCS). Counts associated with the resulting cell pellets were determined by counting on a γ-counter (Cobra Auto Gamma; Packard Instruments Co, Meridien, CT). Nonspecific binding was determined for each ligand by determining binding in the presence of a 200-fold excess of unlabeled ligand.

Alanine substitutions in the B-C and F-G loops of domain 4 of the β_c have been described previously.26,27 The cDNAs for wild-type β_c and each of the β_cmutants in the B-C and F-G loops were introduced into COS cells by electroporation,26 and binding assays were performed on the transfected cells 48 hours posttransfection. The anti-β_cMoAb, BION-1, and 1C1 were labeled using the chloramine T method,37 and binding assays were performed essentially as described previously.27 

TF-1.8 cells were grown in the presence of 2 ng/mL of GM-CSF. The cells were starved for 24 hours before setting up proliferation assays as described previously.33 From dose-response curves, the half-maximal proliferation dosage of IL-5 (0.3 ng/mL), GM-CSF (0.03 ng/mL), IL-3 (0.1 ng/mL), or erythropoietin (EPO; 5 ng/mL) was chosen to perform proliferation experiments in the presence of a range of concentrations of MoAbs. The ³H-Thymidine incorporation of each sample was determined by liquid scintillation.

Ethics approval was obtained to collect bone marrow cells from healthy adult donors. The mononuclear cells were isolated after dextran sedimentation and density gradient centrifugation. The cells (50,000/mL) were cultured in semisolid methylcellulose medium for 14 days in humidified conditions supplied with 5% CO₂. Colonies of more than 40 cells were scored after staining as previously described.39 

The maximal dose of IL-5 required to support eosinophil survival after 36 hours was determined. Eosinophils were then cultured with 1 nmol/L of IL-5, GM-CSF, or IL-3 plus anti-β_c MoAbs for 36 hours. The viability of eosinophils was quantitated by propidium iodide staining and flow cytometry analysis as described.40 

Eosinophils were treated with a range of concentrations of cytokines for 3 hours either in the presence or absence of anti-β_cMoAbs. CD69 expression was measured by immunofluorescence using an anti-CD69 monoclonal MoAb coupled to phycoerythrin (PE; Becton Dickinson Immunocytochemistry Systems, San Jose, CA).

MO7e cells were surface labeled with ¹²⁵I using the lactoperoxidase method, as described previously.41 The labeled cells were preincubated with MoAbs BION-1, control anti-β_c MoAb 1C1 (0.5 mg/mL), or medium alone for 1 minutes before being stimulated or not with IL-3 (6 nmol/L) for 5 minutes. Cells were lysed in lysis buffer consisting of 137 mmol/L NaCl, 10 mmol/L Tris-HCl (pH 7.4), 10% glycerol, and 1% Nonindet P-40 with protease and phosphatase inhibitors (10 μg/mL leupeptin, 2 mmol/L phenylmethlysulphonyl fluoride, 10 μg/mL aprotonin,and 2 mmol/L sodium vanadate) for 30 minutes at 4°C, followed by centrifugation of the lysate at 10,000g for 15 minutes to remove cellular debris. The lysate was precleared with mouse-Ig–coupled Sepharose beads for 18 hours at 4°C and incubated with anti–IL-3Rα, anti-β_c MoAb beads for 2 hours at 4°C. The beads were washed 6 times with lysis buffer and immunoprecipited proteins were separated on 7.5% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) under reducing conditions. The gel was transferred onto nitrocellulose and the immunoprecipited proteins were detected by a PhosphorImager (Molecular Dynamics, Sunnyvale, CA). The blot was then probed with an antiphosphotyrosine MoAb, 3-365-10 (Boehringer Mannheim, Frankfurt, Germany), as described previously.35 

We have previously shown by site-directed mutagenesis that domain 4 of β_c and, more specifically, the putative B-C and F-G loops are involved in high-affinity binding and receptor activation by IL-5, GM-CSF, and IL-3.26-28 To ascertain the feasibility of targetting this region of β_c for developing molecules that can simultaneously block IL-5, GM-CSF, and IL-3 stimulation of eosinophils, we have attempted to raise functional MoAbs against this domain. However, previous attempts using cells expressing the full-length β_c as the immunogen failed to produce any MoAb to domain 4 of β_c, and immunization with chemically synthesized peptides encompassing either the predicted B-C and F-G loops or the whole domain 4 led to MoAb that recognized the immunogen but failed to bind to native β_c. We have now succeeded in obtaining a blocking MoAb by immunizing mice with COS cells overexpressing a cDNA encoding only the extracellular domain 4 of β_c (ΔQP). This MoAb, termed BION-1, immunoprecipitated full-length β_c as well as ΔQP from transfected COS cells and recognized native β_c in purified eosinophils, neutrophils, and monocytes as judged by flow cytometry (Fig 1).

Given that domain 4 of β_c is crucial for the high-affinity binding of IL-5, GM-CSF, and IL-3, we examined whether BION-1 was able to affect this binding. Initially, we found that BION-1, either as a whole IgG or as a Fab fragment, inhibited, in a dose-dependent manner, the binding of ¹²⁵I–IL-5,¹²⁵I–GM-CSF, and ¹²⁵I–IL-3 to the human erythroleukemic cell line TF-1.8 (data not shown). Significantly, BION-1 blocked the binding of IL-5, GM-CSF, and IL-3 to primary purified human eosinophils (Fig 2). In contrast, other anti-β_c MoAb or IgG₁ MoAb controls failed to do so. For each cytokine, we used the lowest practicable concentration of radioligand, thereby allowing binding to high-affinity sites with minimal binding to low-affinity sites.

To demonstrate that the inhibitory effect of BION-1 was solely due to the blocking of high-affinity receptors, saturation binding studies were performed on TF-1 cells and peripheral blood mononuclear cells (PBMNC) in the presence or absence of BION-1 (Fig 3). Binding was performed with GM-CSF over a wide range of concentrations, ie, 10 pmol/L to 10 nmol/L. Scatchard transformation of the binding data obtained shows that BION-1 but not anti-β_c antibody 8E4 completely blocked GM-CSF binding to high-affinity binding sites but not to low-affinity sites (Fig 3). These results, together with those in Fig 2, suggest that BION-1 recognizes the same binding site on β_c used by IL-5, GM-CSF, and IL-3 or binds to an epitope in close proximity to it.

In an attempt to define the region in β_c recognized by BION-1, we used several mutants of β_c and examined whether substitutions of individual aminoacids in the predicted B-C loop or F-G loop impaired BION-1 binding. To perform these experiments, we transfected COS cells with wild-type and mutant β_c and measured the affinity of ¹²⁵I–BION-1 and¹²⁵I-1C1 (as binding control). The results showed that β_c mutants carrying either substitutions M363A/R364A, E366A, or R418A were not recognized by BION-1 (Table 1), indicating that these residues form part of the BION-1 epitope.

To address this issue further, we performed a reciprocal experiment in which increasing concentrations of IL-3 were used to compete for¹²⁵I–BION-1 binding. As can be seen in Fig 4, the addition of IL-3 to TF1.8 cells expressing endogenous IL-3 receptor α and β_c chains prevented, in a dose-dependent manner, the binding of¹²⁵I–BION-1 to β_c. In contrast, TNFα did not prevent the binding of ¹²⁵I–BION-1 to β_c. These results emphasize the close proximity of the epitopes on β_c that the cytokines and BION-1 bind.

To ascertain whether the inhibition of IL-5, GM-CSF, and IL-3 binding by BION-1 was translated into inhibition of IL-5, GM-CSF, and IL-3 stimulation, we initially used the factor-dependent TF-1.8 cell line that proliferates in response to either IL-5, GM-CSF, IL-3, or EPO. We found that BION-1 inhibited the stimulation of TF-1.8 cell proliferation by IL-5, GM-CSF, and IL-3 in a dose-dependent manner, with an ED₅₀ of approximately 0.2 to 2 mmol/L, whereas a control anti-β_c MoAb , 8E4, had little or no effect (Table 2). The Fab fragment of BION-1 behaved similarly with virtually identical ED₅₀ values to the BION-1 antibody (data not shown). Additionally, the stimulating ability of EPO was not inhibited by BION-1, showing the specificity of BION-1 for the IL-5/GM-CSF/IL-3 receptors system (data not shown).

Because eosinophils are believed to be the major effector cells in asthma and they respond to IL-5, GM-CSF, and IL-3, we examined BION-1 for its ability to block eosinophil production and survival in response to these 3 cytokines. We found that BION-1, but not another anti-β_c MoAb, 8E4, inhibited the ability of IL-5, GM-CSF, and IL-3 to stimulate the formation of eosinophil colonies from human bone marrow cells (Table 3). In addition, BION-1 inhibited the prosurvival activity of IL-5, GM-CSF, and IL-3 on purified peripheral blood human eosinophils (Fig 5). Because these cytokines are essential for maintaining eosinophil viability, blocking of β_c by MoAb BION-1 promoted eosinophil cell death to levels similar to those observed in the absence of cytokines (Fig 5).

Eosinophil functional activity can be activated by IL-5, GM-CSF, and IL-3 as well as by TNFα, with the latter using p55 and p75 of the TNFα receptor but not β_c. A sign of eosinophil activation is the upregulation of the CD69 surface antigen, a marker of eosinophil activation in asthma42 and in vitro (Fig 6A). We found that BION-1 inhibited the upregulation of eosinophil CD69 induced by IL-5, GM-CSF, and IL-3 (Fig 6B), whereas another anti-β_c MoAb failed to do so. The blocking effect of BION-1 was found to be specific for the IL-5, GM-CSF, and IL-3 receptors, because the stimulating activity of TNFα was not inhibited (Fig 6B).

We have previously shown that IL-3, GM-CSF, and IL-5 induce dimerization of the respective α chains with β_c, a phenomenon that leads to receptor activation as measured by phosphorylation of β_c on tyrosine residues.35 43 We used this system to show that preincubation of Mo7e cells with BION-1 blocks receptor dimerization and tyrosine phosphorylation of β_c, whereas anti-β_c MoAb 1C1 was unable to prevent receptor dimerization and activation (Fig 7). This result suggests that, although IL-5, GM-CSF, and IL-3 direct binding to β_c is not detectable, a direct interaction is obligatory for receptor dimerization and subsequent cellular activation.

We show here that targetting the common β subunit of the human IL-5, IL-3, and GM-CSF receptors allows for the simultaneous blocking of the binding and stimulating activities of all 3 eosinophilopoietic cytokines. The MoAb BION-1 and its Fab fragment thus represent initial reagents with which to inhibit eosinophil production, survival, and activation in vitro and in vivo. Furthermore, the ability of BION-1 to directly bind to the cytokine-interacting region of β_coffers a new approach to identifying small molecule antagonists of β_c.

IL-5, GM-CSF, and IL-3 are the only cytokines so far discovered that stimulate the production of eosinophils from the bone marrow. In addition, these cytokines stimulate the adhesion capacity of eosinophils, their cytotoxic activity, and their survival. Because of these properties, IL-5, GM-CSF, and IL-3 have been implicated in chronic inflammatory conditions mediated by eosinophils, of which allergic inflammation such as asthma is one of the best studied. In bronchial asthma, eosinophils have long been recognized to play a central role, as judged by their accumulation in the bronchoalveolar lavage fluid and the presence in the sputum of asthmatics of eosinophil cationic proteins believed to contribute to local tissue damage.7 Of the different eosinophil regulators, IL-5 has been a major focus of attention, largely due to its specificity for the eosinophil lineage. Elevated IL-5 mRNA and protein levels have been observed in biopsies of bronchial mucosa in asthma.8,9Furthermore, recruitment of eosinophils is associated with elevated IL-5 in the airways of asthma patients.10 Given the involvement of IL-5 in allergic inflammation, new therapeutic approaches have involved the use of anti–IL-5 antibodies20,21 and the development of specific IL-5 antagonists.23 24 

BION-1 may offer an alternative approach by blocking IL-5 through its interaction with β_c, with the added advantage of also blocking GM-CSF and IL-3. It is becoming clear that not only IL-5, but also other eosinophilopoietic (GM-CSF and IL-3)11,14,15,44-46 as well as noneosinophilopoietic cytokines (IL-4)47-49 are elevated in asthma. Furthermore, the production of IL-5, GM-CSF, and IL-3 by CD4 cells is believed to prolong eosinophil survival in vivo.12,13 In the case of GM-CSF, for example, transgene expression allows the development of allergic airway inflammation.50,51 The extent to which these cytokines contribute to the lung inflammation is not clear. Their concomitant production may accentuate local eosinophilia or, alternatively, induce temporal and site-specific phases of eosinophil recruitment and activation.18 In either case, the simultaneous antagonism of the 3 eosinophilopoietic cytokines may more profoundly downregulate eosinophil-dependent inflammation.

The use of BION-1 or similar molecules that can simultaneously inhibit IL-5, GM-CSF, and IL-3 would be advantageous not only in asthma, but also in other eosinophil-dependent inflammatory diseases such as chronic hyperplastic sinusitis52 and nasal polyposis,16 in which all 3 cytokines have been found to be elevated. Furthermore, BION-1 and functionally analogous molecules may also be antagonists of choice in certain leukemias that produce and respond to GM-CSF and IL-3. A novel model of chronic myeloid leukemia has recently implicated bcr/abl in the production of excess GM-CSF and IL-3 and the development of disease in mice.53 The degree of β_c inhibition required would have to be carefully ascertained in each case, because the complete absence of β_c can lead, at least in mice, to lung disease.54 

BION-1 was able to greatly decrease eosinophil colony formation by IL-5, GM-CSF, and IL-3. This illustrates the importance of β_c for eosinophil production and is consistent with β_c gene knockout experiments showing a major reduction in eosinophil numbers in the peripheral blood and bone marrow of these mice.54 In addition, BION-1 profoundly inhibited eosinophil viability down to levels observed in the absence of cytokines, a property similar to that seen by some glucocorticoids such as fluticasone 17-propionate.55 This ability of BION-1 to inhibit IL-5–, IL-3–, and GM-CSF–mediated eosinophil viability suggests its use as an adjuvant of steroid therapy, its use to reduce the dosage of concomitant steroid therapy, and its use in steroid-resistant asthma.

The membrane proximal domain 4 of β_c was chosen as the immunogen to generate BION-1, because this domain contains the major sites supporting cytokine high-affinity binding and receptor activation.26-28 We and others have previously shown that the B-C and F-G loops in domain 4 of β_c support high-affinity binding to IL-5, IL-3, and GM-CSF and are important for receptor activation. Consistent with this and with the inhibition of BION-1 and IL-3 for each other’s binding, we found that residues M363 R364, E366 in the predicted B-C loop, and R418 in the predicted F-G loop form part of the BION-1 epitope. Interestingly, these residues do not appear to be directly involved in mediating high-affinity binding; however, they may be close in the 3-dimensional structure to Y365, H367, I368, and Y421, the mutation of which has a profound effect on cytokine binding and activation.26-28 Thus, these data suggest a more complex cytokine binding site than that found by mutagenesis and identify new target amino acids in β_c for the development of novel antagonists. Furthermore, these data, together with the ability of BION-1 to prevent IL-3–mediated dimerization of the IL-3 receptor α chain and β_c, support the concept that receptor dimerization and subsquent activation require ligand contacting both receptor subunits. However, we cannot rule out that, given the size of BION-1 as a Fab molecule of 50,000 MW, inhibition of receptor dimerization is the result of steric hindrance.

The development of BION-1 may have 2 immediate uses. Firstly, it may be used as a therapeutic in its own right after humanization and affinity maturation to improve its ED₅₀. The use of MoAbs in clinical medicine is rapidly gaining momentum to the point that they now encompass 27% of biotechnology therapeutics in development and several anticytokine antibodies such as anti-TNFα56 are already showing their efficacy in clinical trials. In relation to IL-5, animal models show that anti–IL-5 antibodies are proving to be efficacious in asthma.31 Secondly, BION-1 may be used as the basis for a novel screening assay for antagonists of β_c. The observation that IL-3 could reciprocally inhibit the binding of ¹²⁵I-labeled BION-1 to β_c (Fig4) suggests that compounds with the same specificity may be obtained. The major advance lies in the fact that, unlike IL-5, GM-CSF, and IL-3, which require prior binding to each α subunit before interacting with β_c, BION-1 can directly bind to β_c, thereby facilitating a cell-free solid-phase assay on which natural or engineered products may be screened for inhibition of binding of labeled BION-1 to immobilized β_c or domain 4 of β_c. The approach described here may also be extended to other receptors systems, such as the common subunit of the IL-4 and IL-13, which mediates the allergen-induced asthma triggered by IL-4 and IL-13.57 58