A murine interleukin-4 antagonistic mutant protein completely inhibits interleukin-4-induced cell proliferation, differentiation, and signal transduction.

We characterize here a highly efficient antagonist for interleukin-4 (IL-4) in the mouse system. In this double mutant of the murine IL-4 protein, both glutamine 116 and tyrosine 119 were substituted by aspartic acid residues. This variant (QY) bound with similar affinity to the IL-4 receptor α subunit as wild type IL-4 without inducing cellular responses. In contrast, QY completely inhibited in a dose-dependent manner the IL-4-induced proliferation of lipopolysaccharide-stimulated murine splenic B-cells, of the murine T cell line CTLL-2, and of the murine pre-B-cell line BA/F3. QY also inhibited the IL-4-stimulated up-regulation of CD23 expression by lipopolysaccharide-stimulated murine splenic B-cells and abolished tyrosine phosphorylation of the transcription factor Stat6 and the tyrosine kinase Jak3 in IL-4-stimulated BA/F3 cells. Selective inhibition of IL-4 may be beneficial in T-helper cell type 2-dominated diseases, like type I hypersensitivity reactions or helminthic infections. The QY mutant could be an attractive tool to study in vivo the therapeutic potential of IL-4 antagonists in mouse systems.

Interleukin-4 (IL-4) 1 is a pleiotropic cytokine derived from T-cells, thymocytes, and mast cells that has multiple effects on many cell types (1). Its functions include the differentiation of T helper cells to a T H 2 phenotype and the induction of sterile ⑀ transcripts in B-cells, which is a required step for class switching to IgE type antibodies (2,3). Among the clinically important features of IL-4 are the coordination of immune responses against helminthic macroparasites and its central role in the sensitization process of type I allergic diseases (4,5).
Site-directed mutagenesis has led to the discovery of two regions in the human IL-4 molecule that are important for interaction with the receptor chains (for review see Ref. 12). Substitution of glutamic acid 9 (Glu 9 ) or arginine 88 (Arg 88 ) leads to a near complete loss of binding to IL-4R␣ (13). A second important region of the human IL-4 molecule is defined by mutations of arginine 121, tyrosine 124, and serine 125, close to the C terminus of the protein. Mutations at these positions do not interfere with IL-4R␣ binding but can severely impair cellular responses, because they result in loss of interaction with ␥c (14). The most efficient single site mutant of this type is Y124D, which does not induce T-cell proliferation but is a partial agonist for up-regulation of CD23 on B-cells (15). In contrast, a double mutant where both Arg 121 and Tyr 124 have been replaced by aspartic acid (RY) has no agonistic activity in any assay employed so far (16). RY is therefore a perfect high affinity antagonist for human IL-4, as well as for IL-13, which also uses IL-4R␣ for signal transduction (16,17).
IL-4 antagonists may provide an effective way for the therapy of T H 2-dominated diseases. The major obstacle for testing in vivo biological tolerance and therapeutic effects of IL-4 antagonists is the species specificity of IL-4. There is approximately 60% DNA sequence homology (18,19) and no crossreactivity between human and mouse IL-4 (20). Due to the lack of receptor binding, human IL-4 antagonists cannot be effective in mice. This has prompted us to develop an efficient antagonist for IL-4 in the mouse system.

MATERIALS AND METHODS
Animals, Cells, and Viruses-BALB/c mice 6 -8 weeks of age were purchased from Charles River/Wiga. The Spodoptera frugiperda insect cell line Sf9 was cultured in Insect Xpress medium (Serva, Heidelberg, Germany) without adjuvants at 27°C in a 2-liter spinner flask aerated with 100% oxygen. Autographa californica nuclear polyhedrosis virus DNA (BaculoGold DNA) was from Pharmingen (Hamburg, Germany).
Production of IL-4 Wild Type and Double Mutant Proteins-Recombinant IL-4 proteins were produced in Sf9 cells following Baculovirus infection. The pEP-B-splice mIL-4 plasmid containing the cDNA for mouse IL-4 was kindly provided by Dr. Werner Mü ller (University of Cologne, Germany) (22). A fragment encoding the cDNA of mouse IL-4 was amplified by polymerase chain reaction with the synthetic 5Ј primer 5Ј-CGCGGATCCCATATCCACGGA-3Ј and the 3Ј primer 5Ј-GC-CGGATCCTACGAGTAATCC-3Ј, comprising a BamHI site and the first (for the 5Ј primer) and, respectively, last (for the 3Ј primer) four codons of wild type IL-4. The QY mutant gene was constructed using the 3Ј primer 5Ј-GCCGGATCCTACGAGTCATCCATATCCATGATGC-3Ј, resulting in substitution of both glutamine 116 and tyrosine 119 by aspartic acid. The constructs were sequenced and cloned into the Bacu-* This work was supported by the Deutsche Forschungsgemeinschaft (Du 220/2-1), by the Jubilä umsstiftung der Industrie und Handelskammer Wü rzburg-Schweinfurt, and by the Senator Kurt und Inge Schuster Stiftung. 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. Protein Purification-Sf9 supernatant was extensively dialyzed against 10 mM Tris/HCl, pH 8.0, followed by cation exchange chromatography on a CM-Sepharose fast flow column (Pharmacia, Freiburg, Germany) equilibrated with 10 mM Tris/HCl, pH 8.0, at 4°C. Bound protein was eluted with a 0 -1 M NaCl gradient. The material was applied to a reversed phase high pressure liquid chromatography column and eluted by a gradient of acetonitrile from 30 to 50%. IL-4 and QY mutant were identified by immunoblotting using rat anti-mouse IL-4 antibody (BVD4 -1D11; Pharmingen). The yield of purified recombinant protein was approximately 1 mg/liter Sf9 supernatant. Freezedried protein was stored at Ϫ20°C and dissolved in H 2 O before use. Protein was determined by the BCA method using bovine serum albumin as standard.
The biotinylated IL-4-BP was immobilized to a streptavidin-coated sensor chip CM5 (Pharmacia Biosensor AB, Uppsala, Sweden) at a density of 80 -100 pg/mm 2 in a BIAcore 2000 system. The association and dissociation of IL-4 and QY in HBS buffer (10 mM Hepes, pH 7.4, 150 mM NaCl, 3.4 mM EDTA, 0.005% Surfactant P-20) were analyzed at a flow rate of 50 l/min and 25°C. A set of sensograms was recorded at six different analytic concentrations. The data were evaluated using the BIA evaluation 2.1 software provided with the system as described (23). The rate constants evaluated from 15 independent measurements were used to calculate the mean values Ϯ standard deviation.
Proliferation Assays-Cytokine-induced proliferation of cells was determined by [ 3 H]TdR incorporation as described (24). CTLL-2 cells were used at a density of 5 ϫ 10 4 /ml and BA/F3 cells at 5 ϫ 10 5 /ml. T-celldepleted splenic B-cells were prepared from BALB/c mice and tested following standard procedures (25). Briefly, splenic cell suspensions were incubated with rat anti-mouse Thy1.2 antibody (30H12; Pharmingen) for 30 min, followed by incubation at 37°C for 45 min with Low-Tox-H rabbit complement (Cedar Lane, Hornby, Ontario, Canada). Viable and resting B-cells were separated on the 70 -65% interphase of a Percoll gradient (Pharmacia) and cultured at a density of 10 6 /ml with 5 g/ml Salmonella typhimurium LPS (Sigma, Deisen-hofen, Germany) plus the indicated concentrations of cytokines for 48 h. Determination of CD23 Expression-Splenic B-cells (10 6 /ml) were cultured in a 96-well plate with 5 g/ml LPS and IL-4 or QY as indicated. Cells were centrifuged after 40 h, and supernatant was withdrawn. After blocking with 100 l of phosphate-buffered saline, 3% bovine serum albumin for 1 h, cells were stained in opaque microtiter plates with 1 g/well rat anti-mouse CD23 (B3B4; Pharmingen) at 4°C for 1 h. Following incubation with goat anti-rat IgG-horseradish peroxidase conjugate for 45 min, detection was performed with an enhanced chemoluminescene based method (26) using a Microlumat LB 96-P luminometer (Berthold, Wildbad, Germany). ED 50 , ID 50 , and K i were calculated as described above.

RESULTS AND DISCUSSION
Previous experiments have shown that the receptor binding affinity of IL-4 is mainly determined by the high affinity interaction with IL-4R␣. Therefore, the kinetics of interaction between mouse IL-4 proteins and the mouse IL-4R␣ chain were studied by means of a BIAcore 2000 system employing a recombinant extracellular domain, IL-4-BP, immobilized to a biosensor matrix. The rate constants for the dissociation of the complex between IL-4-BP and IL-4 or QY were indistinguishable. The k off for both proteins was about 2 ϫ 10 Ϫ3 s Ϫ1 (Fig. 1 and Table I). The association rate constants were similar, but consistently a 50% lower on rate was found for QY in comparison with IL-4. The dissociation equilibrium constant K d calculated as k off /k on was about 400 pM for IL-4 and 800 pM for the QY variant.
The effects of the wild type and QY mutant protein on cell proliferation were determined with highly purified splenic Bcells co-cultured with LPS, the T-cell line CTLL-2, and the pre-B-cell line BA/F3. The biological activity of IL-4 produced by Sf9 cells was identical to commercially available IL-4 produced by E. coli (Fig. 2A). IL-4 induced proliferation of splenic B-cells, CTLL-2 cells, and BA/F3 cells (Fig. 2, A-C). The ED 50 values are summarized in Table II. The QY mutant by itself had no detectable activity in proliferation assays (Fig. 2, A and  B) but inhibited IL-4-stimulated proliferation of all three cell types in a dose-dependent fashion (Fig. 2, D-F). The concentrations required for half-maximal inhibition of IL-4-induced proliferation are given in Table II. As expected, in more sensi-  tive cells a higher dose of inhibitor was needed to block IL-4induced responses. The low affinity IgE receptor, CD23, is an IL-4-inducible B-cell differentiation marker. Half-maximal CD23 expression was induced by 19 pM IL-4 (Fig. 3A). Similiar to the proliferation assays, there was no detectable activity of the QY variant. QY prevented the IL-4-induced CD23 expression with halfmaximal inhibition reached at about 95-fold excess of the mutant (Fig. 3B).
The signal transduction of cytokines involves the activation of various tyrosine kinases and rapid phosphorylation of their substrates. IL-4 stimulates phosphorylation and activation of the transcription factor Stat6 and the tyrosine kinase Jak3 (27). To determine whether the QY mutant could inhibit IL-4 induced signaling, we have measured Stat6 and Jak3 phosphorylation in BA/F3 cells. IL-4 at 1 nM concentration induced tyrosine phosphorylation of both proteins, whereas the same amount of the QY mutant did not stimulate phosphorylation (Fig. 4). A 500-fold excess of the QY variant completely inhibited the IL-4-induced response (Fig. 4).
Antagonistic properties of human IL-4 variants are caused by the deletion of a hydrophobic patch on the surface of the IL-4 molecule and by the introduction of an electrostatic mismatch (28). No local instability is introduced in the protein structure, because main chains structures of antagonistic mutants are identical compared with wild type (29).
No structural data on mouse IL-4 are available, but some mutation studies have been reported. A deletion mutant of murine IL-4 lacking residues 118 -120 has impaired activity in a proliferation assay with the T-cell line CT4R but retains high affinity binding to IL-4R␣ (30). Such a deletion, however, may alter the protein structure, and inhibition of wild type protein was not tested. In the same study, replacement of Gln 116 with alanine had only minor effects on receptor binding and biological activity (30). A murine IL-4 variant with a single site mutation of tyrosine 119 (Y119D) has been mentioned in the literature, but in one report no data were shown (31), and in another one the same mutant induced agonistic effects similar to those of wild type IL-4 in differentiation assays, like MHC II up-regulation on B-cells (32). This is to be expected for the Y119D mutant, because murine and human ␥c interact in a structurally different way with IL-4. Human IL-4 can productively interact both with human and murine ␥c, but human ␥c is most severely affected by mutations in Tyr 124 (13), and murine ␥c is most severely affected by mutations in Arg 121 (33). For this reason, and in consideration of the partial agonist activities of human Y124D (13, 15), we decided to create the QY   double mutant, as a murine analog to the complete human IL-4 antagonist RY (16). As shown here, the QY mutant was by itself inactive and completely antagonistic for wild type IL-4, in all assays performed.
Like IL-4, IL-13 can also direct IgE class switching in human B-cells (34, 35) but apparently not in the mouse, because IL-4 knockout mice are unable to produce IgE (36,37). Anti-IL-4 antibodies or soluble IL-4 receptors may not be sufficient to treat for example type I hypersensitivity in humans, because neither would interfere with IL-13. Furthermore, both reagents can not only inhibit but also enhance effects of IL-4 by acting as carrier molecules that increase the lifetime of IL-4 in the serum (38). Antagonistic IL-4 mutants will inhibit both IL-4-and IL-13-mediated responses and cannot increase agonistic IL-4 effects. Because cytokine mutants are structurally nearly identical to the wild type protein, they should provide good immunological tolerance as well. The QY variant should allow to study such questions in vivo.