Characterization and Immunotherapeutic Implications for a Novel Antibody Targeting Interleukin (IL)-13 Receptor α2*

Background: Antibodies specific for tumor-associated antigens (TAAs) have emerged as valuable research, diagnostic, and therapeutic agents. Results: A novel antibody against TAA IL13Rα2 has been generated and characterized. Conclusion: The antibody possesses a high specificity and affinity for IL13Rα2 and competes with IL-13 for binding to IL13Rα2. Significance: Future studies testing the therapeutic and diagnostic properties of this antibody in IL13Rα2-expressing tumors are now possible. The high affinity interleukin-13 receptor α2 (IL13Rα2) is selectively expressed at a high frequency by glioblastoma multiforme (GBM) as well as several other tumor types. One approach for targeting this tumor-specific receptor utilizes the cognate ligand, IL-13, conjugated to cytotoxic molecules. However, this approach lacks specificity because the lower affinity receptor for IL-13, IL13Rα1, is widely expressed by normal tissues. Here, we aimed to develop and characterize a novel monoclonal antibody (mAb) specific to IL13Rα2 for the therapeutic purpose of targeting IL13Rα2-expressing tumors. Hybridoma cell lines were generated and compared for binding affinities to recombinant human IL13Rα2 (rhIL13Rα2). Clone 47 demonstrated binding to the native conformation of IL13Rα2 and was therefore chosen for further studies. Clone 47 bound specifically and with high affinity (KD = 1.39 × 10−9 m) to rhIL13Rα2 but not to rhIL13Rα1 or murine IL13Rα2. Furthermore, clone 47 specifically recognized wild-type IL13Rα2 expressed on the surface of CHO and HEK cells as well as several glioma cell lines. Competitive binding assays revealed that clone 47 also significantly inhibited the interaction between human soluble IL-13 and IL13Rα2 receptor. Moreover, we found that N-linked glycosylation of IL13Rα2 contributes in part to the interaction of the antibody to IL13Rα2. In vivo, the IL13Rα2 mAb improved the survival of nude mice intracranially implanted with a human U251 glioma xenograft. Collectively, these data warrant further investigation of this novel IL13Rα2 mAb with an emphasis on translational implications for therapeutic use.

Proteins expressed by tumor cells but not by normal cells are attractive molecules for the selective delivery of cytotoxic mol-ecules. Accordingly, interleukin-13 receptor ␣2 (IL13R␣2), 2 the high affinity receptor for interleukin-13 (IL-13), is a promising candidate. IL13R␣2 is expressed at a high frequency in the aggressive and incurable form of primary brain tumor known as glioblastoma multiforme (GBM) (1)(2)(3) as well as by other solid tumors (4). In contrast, normal tissues express little to no IL13R␣2 with the exception of the testes (6). Notably, IL13R␣1, a different receptor with low affinity for IL-13, is expressed ubiquitously by many tissues (7)(8)(9), making it a poor candidate for selective targeting of tumor-specific immunotherapeutic applications.
Until recently, IL13R␣2 was thought to act as a decoy receptor for IL-13 (10). However, recent studies have challenged that theory based on studies demonstrating that upon binding of IL-13 to IL13R␣2 downstream signaling occurs in specialized pulmonary macrophages (11) as well as in pancreatic ductal (12) and ovarian carcinoma cells (13). Moreover, overexpression of IL13R␣2 in GBM but not in normal brain tissue (14,15) uniquely positions this receptor as a candidate for targeting tumor cells. GBM is a highly infiltrative tumor often making complete surgical removal impossible. Moreover, GBM is highly resistant to radiation and chemotherapy (16), warranting further development of novel and targeted therapies for the treatment of patients. Several studies have investigated the therapeutic properties of an IL-13 fusion protein conjugated to a recombinant cytotoxin derived from Pseudomonas exotoxin A (IL-13PE) that induces apoptosis in IL13R␣2-expressing glioma cells in vitro, in preclinical animal models, and in patients tested in clinical trials (17)(18)(19)(20)(21)(22). However, such agents lack a high specificity of interaction with IL13R␣2 due to the alternative binding of ubiquitously expressed IL13R␣1. Therefore, developing highly selective antibody fragments that can be combined with toxins for specificity to IL13R␣2-expressing cells is considered to be a promising pursuit.
Previous work has investigated a phage display library approach for selecting small antibody fragments specific to human IL13R␣2 followed by their evaluation in vitro and in vivo (23). Despite the high specificity of interaction with IL13R␣2, conjugation with toxins has failed to increase cytotoxicity in IL13R␣2-expressing glioma and renal cell carcinoma cell lines when compared with the effects of IL-13PE38. The low affinity of generated antibody fragments is the most reasonable explanation for the lack of success. Antibody fragments derived from phage display libraries are known to be lower in affinity and avidity than antibodies generated by conventional hybridoma technology (24). Modifications of those small antibody fragments are often required to enhance their affinity and avidity to targeted proteins. In recent years, monoclonal antibodies have shown increasing success as targeted anticancer and diagnostic agents (25,26), and a further search for high affinity reagents with restricted specificity to tumor-associated antigens is in progress. Historically, the hybridoma cell line specific to the antigen IL13R␣2, however, has been unavailable to the scientific community. Thus, the goal of the present study was to discover, develop, and characterize a high affinity antibody that specifically recognizes IL13R␣2 expressed on the surface of cancer cells. Here, we demonstrate the generation of an antibody possessing the properties critical for immunotherapeutic targeting of IL13R␣2-expressing tumors in vivo and potentially suitable for various other applications.

EXPERIMENTAL PROCEDURES
Materials-Lipofectamine 2000 and the pEF6/Myc-His vector were obtained from Invitrogen. mAbs to IL13R␣2 (clones YY-23Z and B-D13) and the IsoStrip mouse monoclonal antibody isotyping kit were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The mAb to IL13R␣2 (clone 83807) and recombinant human and mouse IL13R␣2hFc and IL13R␣1hFc chimeras were purchased from R&D Systems (Minneapolis, MN). Biotinylated horse anti-mouse antibodies and the Elite kit were obtained from Vector Laboratories (Burlingame, CA). 3,3Ј-Diaminobenzidine substrate was purchased from Dako (Carpinteria, CA). Goat anti-mouse antibody conjugated with peroxidase was purchased from Chemicon International (Temicula, CA), and Pngase F was purchased from New England Biolabs (Ipswich, MA). The QuikChange Lightning site-directed mutagenesis kit was purchased from Agilent Technologies, Inc. (Santa Clara, CA), and the RNeasy Plus kit was received from Qiagen (Valencia, CA). The cDNA iScript kit, 7.5% Tris-HCl gel, and ImmunStar WesternC developing reagent and protein marker were purchased from Bio-Rad. The human IL-13 ELISA kit was purchased from eBioscience (San Diego, CA). GBM12 and GBM43 were kindly provided by Dr. David C. James (University of California-San Francisco), and the cDNA encoding human wild-type IL13R␣2 was obtained from Dr. Waldemar Debinski (Wake Forest University).
Immunization-To obtain monoclonal antibodies with specificity to native IL13R␣2, the human recombinant IL13R␣2hFc fusion was used for immunization of animals and in all screening assays. Two 6-week-old female BALB/c mice were immunized with intraperitoneal injection of 10 g of rhIL13R␣2hFc protein in complete Freund's adjuvant followed by intraperitoneal injection of 10 g of rhIL13R␣2hFc protein in incomplete Freund's adjuvant at a 2-week interval for 2 months. Two weeks after the last intraperitoneal injection and 3 days before the fusion, a boost was performed by the combination of intravenous and intraperitoneal injection of 10 g of antigen without Freund's adjuvant. The fusion of mouse spleen cells with the mouse myeloma cell line X63.Ag8.653 subclone P3O1 was performed by using a procedure described by Köhler and Milstein (27). Hybridoma supernatants were assayed for the presence of IL13R␣2 antibodies using the enzyme-linked immunosorbent assay (ELISA). Selected populations were cloned, and supernatants were assayed to identify the clones with strongest binding.
Generation of CHO Cell Line Expressing Human IL13R␣2-The cDNA encoding human wild-type IL13R␣2 was amplified with the following primer pair: forward, 5Ј-GCTTGGTA-CCGAATGGCTTTCGTTTGCTTGGC-3Ј and reverse, 5Ј-GTTTTTGTTCGAATGTATCACAGAAAAATTCTGG-3Ј. The purified PCR product was restricted with KpnI and BstBI enzymes, agarose gel-purified, and subsequently cloned into the pEF6/Myc-His vector in a reading frame with Myc and His 6 tags. CHO cells were plated at 80% confluence and transfected with a plasmid encoding the IL13R␣2 using Lipofectamine 2000. The following day, 4 g/ml blasticidin was added for selection of cells that had stably incorporated and expressed the IL13R␣2 transcript. A stable population of cells was further subcloned in 96-well plates at a density of one cell/well. Ten days later, single clones were screened by flow cytometry for cell surface expression of IL13R␣2 using an antibody to IL13R␣2 (clone B-D13). The clone with the highest level of IL13R␣2 expression was selected and expanded for subsequent screening of hybridomas secreting IL13R␣2 antibodies.
ELISA-96-well plates were coated with 50 l of human or mouse recombinant IL13R␣2hFc or IL13R␣1hFc or human control IgG at a concentration of 1 g/ml overnight at 4°C. Following washes with TBS-Tween 20 buffer and blocking with 1% nonfat dry milk, 50 l of purified antibodies, serum, or hybridoma supernatants at various dilutions were applied to the plate and incubated for 1 h at room temperature. Bound antibodies were detected with goat anti-mouse antibodies conjugated to alkaline phosphatase following the development with alkaline phosphatase substrate. Plates were read at A 405 using a UniRead 800 plate reader (BioTek).
PCR-To determine the expression of IL13R␣2 in various glioma cells and astrocytes, total RNA was generated from the cell pellets using the RNeasy Plus kit. 200 ng of total RNA was then converted into cDNA using the cDNA iScript kit. The cDNA was further amplified for IL13R␣2 and GAPDH for 30 cycles using IL13R␣2 and GAPDH primers and visualized on a 1% agarose gel.
Surface Plasmon Resonance-The affinity and rates of interaction between IL13R␣2 (clone 47) mAb, commercially available IL13R␣2 mAbs (clones 83807 and B-D13), and target (rhIL13R␣2) were measured with a Biacore 3000 biosensor through surface plasmon resonance (SPR). The mAbs were immobilized (covalently) to the dextran matrix of the sensor chip (CM5) using the amino coupling kit. The carboxyl groups on the sensor surfaces were activated with an injection of a solution containing 0.2 M N-ethyl-NЈ-(3-diethylamino-propyl)carbodiimide and 0.05 M N-hydroxysuccinimide. The immobilization procedure was completed by the injection of 1 M ethanolamine hydrochloride to block the remaining ester groups. All steps of the immobilization process were carried out at a flow rate of 10 l/min. The control surface was prepared similarly with the exception that running buffer was injected rather than mAbs. Binding reactions were performed at 25°C in HBS-P buffer (20 mM HEPES, pH 7.4, 150 mM NaCl, and 0.005% (v/v) surfactant P20) using a flow rate of 20 l/min. Target (rhIL13R␣2) was added at various concentrations in the flow during the binding phase. The amount of protein bound to the sensor chip was monitored by the change in refractive index (represented by response units (RU)). The instrument was programmed to perform a series of binding measurements with increasing concentrations of target over the same surface. Triplicate injections of each concentration of target were performed. Sensorgrams (plots of changes in RU on the surface as a function of time) were analyzed using BIAevaluation v4.1. Affinity constants were estimated by curve fitting using a 1:1 binding model.
Data Preparation and Kinetic Analysis-The estimation of kinetic parameters was performed by repetitive injections of a range of target concentrations over the immobilized mAbs. Data were prepared by the method of "double referencing." This method utilizes parallel injections of each target sample over a control dextran surface as well as running buffer injections over both the immobilized mAbs and control dextran surfaces. Subtraction of these sensorgrams yielded the control; this was subtracted from the experimental sensorgram. Each data set (consisting of sensorgrams of increasing target concentrations over the same level of immobilized mAbs) was analyzed using various kinetic models. The BIAevaluation v 4.1 software was then used for data analysis. Affinity constants were estimated by curve fitting using a 1:1 binding model. Sensorgram association and dissociation curves were fit locally or globally. The rate of complex formation during the sample injection is described by an equation of the following type: dR/dt ϭ k a C (R max Ϫ R) Ϫ k d R (for a 1:1 interaction) where R is the SPR signal in RU, C is the concentration of analyte, R max is the maximum analyte binding capacity in RU, and dR/dt is the rate of change of SPR signal. The early binding phase (300 s) was used to determine the association constant (k a ) between mAb and target. The dissociation phase (k d ) was measured using the rate of decline in RU on introduction of free buffer at the end of target injections. Data were simultaneously fit by the software program (global fitting algorithm), and the dissociation constant (K D ) of the complexes was determined as the ratio k a /k d . For quantitative analysis, three independent replicates were performed for each sample. Data are expressed as mean Ϯ S.E.
Competitive Binding Assay-For the competitive binding plate assay, a 96-well plate was coated with 50 l of affinitypurified hrIL13R␣2hFc at 1 g/ml in carbonate buffer, pH 9.6 and stored overnight at 4°C. After washing with PBS contain-ing 0.05% Tween 20, mAbs to IL13R␣2 (10 g/ml) or control mIgG were added for 30 min at room temperature. After washing, 50 l of purified rhIL-13 in PBS and 0.1% BSA at 10 ng/ml was added for a 1-h incubation at room temperature and assayed for bound rhIL-13 using detection reagents from a human IL-13 ELISA kit. Separately, HEK cells expressing wildtype or 4-amino acid mutants in the IL13R␣2 sequence were pretreated with either rhIL-13 or mAb IL13R␣2 (clone 47) at 2 g/ml for 30 min on ice followed by a 1-h incubation with IL13R␣2 (clone 47) mAb or rhIL-13 at 100 ng/ml, respectively. Binding of rhIL-13 to IL13R␣2 alone or in presence of the competitor was detected with human IL-13 mAb-FITC. Binding of IL13R␣2 (clone 47) mAb to rhIL13R␣2 alone or in presence of the competitor was detected with anti-mouse antibody conjugated to Alexa Fluor 649 and analyzed by flow cytometry.
Mutagenesis of IL13R␣-Previously, Tyr 207 , Asp 271 , Tyr 315 , and Asp 318 of the human IL13R␣2 were identified as residues crucial for the interaction with human IL-13 (28). To determine whether those residues were important for binding of the IL13R␣2 (clone 47) mAb to IL13R␣2, the Tyr 207 , Asp 271 , Tyr 315 , and Asp 318 were mutated to Ala separately or at the same time (4-amino acid mutant) using the QuikChange Lightning site-directed mutagenesis kit according to the manufacturer's recommendations. Sequencing of selected clones was performed in house and confirmed the presence of the selected mutation. HEK cells were transfected with wild-type or mutated variants of IL13R␣2 cDNA in the pEF6 Myc-His vector using Lipofectamine Plus transfection reagent. 48 h after transfection, the cells were collected and analyzed for binding to IL13R␣2 (clone 47) mAb via flow cytometry.
Western Blot-The rhIL13R␣2 was applied to a 7.5% Tris-HCl gel (Bio-Rad) at 200 ng/lane and resolved under reducing conditions. After the transfer of proteins to a PVDF membrane (Bio-Rad) and blocking with 2% nonfat dry milk, the membrane was stained with anti-IL13R␣2 mAb (clones YY-23Z and B-D13) at 2 g/ml or with supernatant collected from hybridoma clones (diluted 10 times), followed by goat antimouse antibody conjugated to peroxidase. ImmunStar West-ernC was used to develop the reaction. Images were captured using a Bio-Rad ChemiDoc imaging system.
Immunohistochemistry-The GBM tissues were collected in accordance with a protocol approved by the Institutional Review Board at the University of Chicago. Flash frozen brain tumor tissues were cut to a thickness of 10 m. Tissue sections were fixed with Ϫ20°C methanol and stained for human IL13R␣2 using mouse IL13R␣2 (clone 47) mAb at a concentration of 3 g/ml or isotype control mIgG1. The bound antibodies were detected with biotinylated horse anti-mouse antibodies (1:100). The antigen-antibody binding was detected by the Elite kit with 3,3Ј-diaminobenzidine substrate. Slides were analyzed using the CRI Panoramic Scan Whole Slide Scanner and Panoramic Viewer software.
Animal Study-All animals were maintained and cared for in accordance with the Institutional Animal Care and Use Committee protocol and according to National Institutes of Health guidelines. The animals used in the experiments were 6 -7 week-old male athymic nu/nu mice. Mice were anesthetized with an intraperitoneal injection of ketamine hydrochloride/ xylazine (25 mg/ml/2.5 mg/ml) mixture. To establish intracranial tumors, a midline cranial incision was made, and a rightsided burr hole was placed 2 mm lateral to the sagittal suture and ϳ2 mm superior to . Animals were positioned in a stereotactic frame, and a Hamilton needle was inserted through the burr hole and advanced 3 mm. Intracranial penetration was followed by (i) injection of 2.5 ϫ 10 4 U251 glioma cells in 2.5 l of sterile PBS in combination with 200 ng of mIgG or IL13R␣2 (clone 47) mAb or (ii) 3 days postintracranial injection of glioma cells with PBS or 10 g of IL13R␣2 (clone 47 or B-D13) mAb as described previously (29). All mice were monitored for survival. Three animals from each group were sacrificed at day 17, and brains were harvested and frozen for sectioning, hematoxylin and eosin (H&E) staining, and microscopic analysis.
Statistics-The differences between groups were evaluated by Student's t test or one-way analysis of variance with post hoc comparison Tukey's test or Dunnett's test. For the in vivo survival data, a Kaplan-Meier survival analysis was used, and statistical analysis was performed using a log rank test. p Ͻ 0.05 was considered statistically significant.

Characterization of Antigen and Screening of Hybridoma
Clones Secreting Anti-IL13R␣2 Antibodies-The primary goal of this study was to generate a high affinity monoclonal antibody suitable for targeting of the IL13R␣2 expressed on the surface of tumor cells. We therefore immunized mice and screened the resulting hybridoma clones for reactivity against the antigen, rhIL13R␣2, in its native conformation. A platebound ELISA utilizing a hybridoma clone against rhIL13R␣2, YY-23Z, was established for the detection of rhIL13R␣2. The concentration of rhIL13R␣2 absorbed to the plastic at 1 g/ml was found to be suitable for the detection of antibody binding (Fig. 1A). Next, the rhIL13R␣2hFc was characterized for its "nativity" by utilizing a pair of commercially available antibodies recognizing only the native (found on the cell surface) and denatured (using Western blotting under reducing conditions) forms of IL13R␣2 and for its binding properties to rhIL13R␣2 in ELISA with antibody clones B-D13 and YY-23Z, respectively. Both clones B-D13 and YY-23Z were able to recognize the rhIL13R␣2hFc in a plate-bound ELISA (Fig. 1B). Denaturation of antigen at 95°C for 5 min in the presence of ␤-mercaptoethanol completely abolished the ability of the antibody clone B-D13 to recognize antigen by ELISA, whereas the YY-23Z clone retained the ability to bind the denatured antigen. Thus, the rhIL13R␣2hFc absorbed to the plastic of ELISA plates containing both native and denatured forms of the protein. Analysis of serum from animals immunized with a fusion of rhIL13R␣2 and hFc revealed the presence of antibodies against both rhIL13R␣2 and human Fc fragment (data not shown). To select antibodies specific for the IL13R␣2 portion of the fusion, human IgG was included as an additional negative control for the screening of hybridoma populations. Of the 39 screened primary populations, only 15 populations were specific to IL13R␣2, and four were reactive with human IgG. Finally, five clones strongly reacting with native IL13R␣2 were further expanded and recloned. The two clones recognizing only denatured antigen were selected from the separate immunization set with rhIL13R␣2hFc chimera (data not shown). Supernatants from selected clones were compared for their ability to bind hrIL13R␣2 in a plate-bound ELISA (Fig. 1C) and by Western blotting (Fig. 1D). Fig. 1C shows that clone 47 strongly binds to the antigen in plate-bound ELISA but not by Western blotting, indicating the ability of clone 47 to recognize a native conformation of the antigen. Therefore, clone 47 was selected for all subsequent studies and analysis for its properties. Clone 47 was found to be of the IgG1 isotype possessing a chain (data not shown).
Specificity of Binding for the IL13R␣2 (Clone 47) mAb to Recombinant Human IL13R␣2 and IL13R␣2 Expressed at the Cell Surface-We investigated the binding properties of the IL13R␣2 (clone 47) mAb to rhIL13R␣2 versus the commercially available clones 83807 and B-D13 in a plate-bound ELISA. Fig.  2A shows strong and specific binding of clone 47 to rhIL13R␣2 when compared with clones 83807 and B-D13. Clone 47 reaches the plateau of binding at the low concentration of 0.05 g/ml. None of the antibodies showed binding to human IgG utilized as an additional negative control in these experiments (data not shown).
To further verify the specificity of interaction for clone 47 with human IL13R␣2, a clonal line of CHO cells expressing the full size wild-type human IL13R␣2 (clone 6) was generated. Binding of the antibody to control CHO cells transfected with an empty vector was compared with that of CHO cells express-ing IL13R␣2. Again, the IL13R␣2 (clone 47) mAb demonstrated strong and specific binding to IL13R␣2 expressed on the cell surface but not to control CHO cells, indicating that this antibody specifically recognizes a native conformation of the IL13R␣2 (Fig. 2B). Clone 47 demonstrated the strongest affinity for IL13R␣2 at the lowest tested concentration of 0.25 g/ml. Notably, other selected hybridoma clones demonstrated similar specificity of interaction with IL13R␣2 expressed on the cell surface of CHO cells but not with control CHO cells (data not shown). Data obtained in a plate-bound ELISA also revealed that clone 47 does not interact with the low affinity receptor for IL-13, the IL13R␣1 (Fig. 2C), or mouse recombinant IL13R␣2, further validating the specificity of interaction between clone 47 and IL13R␣2 (Fig. 2D). Clones 83807 and B-D13 did not show binding to mouse rIL13R␣2 in agreement with current understanding of the cross-reactivity of these antibodies with mouse IL13R␣2.
We next characterized the binding capacity of clone 47 with various glioma cell lines, the patient-derived glioma lines GBM12 and GBM43, and normal human astrocytes. Increased expression of the IL13R␣2 gene relative to normal brain tissue is reported in 44 -47% of human GBM resected specimens (3) and in up to 82% (14 of 17) primary cell cultures derived from GBM and normal brain explants (2) . Fig. 3, A and B, show the flow charts of the comparative staining of glioma cells, human astrocytes, and HEK cells expressing recombinant human IL13R␣2 on the cell surface with the IL13R␣2 (clones 47, 83807, and B-D13) mAb. Fig. 3, A and B, reveal (i) various levels of IL13R␣2 expression on the cell surface and (ii) superior binding of the clone 47 versus clones B-D13 (1.2-4.6-fold difference between the cell lines) and 83807 to the surface of analyzed cell lines. Interestingly, we observed a near complete absence of the binding of clone 83807 to glioma cell lines in contrast to HEK cells expressing IL13R␣2. No binding of clone 47 was detected with normal human astrocytes, confirming the specificity of interaction of clone 47 with human glioma cells expressing IL13R␣2. The expression of IL13R␣2 mRNA in these cells generally correlates with the level of IL13R␣2 expression on the cell surface. Moreover, cells expressing low to no mRNA expression for IL13R␣2, including U118 and primary human astrocytes, demonstrated low to no expression for IL13R␣2 on the cell surface (Fig. 3B). In additional experiments, N10 glioma cells were incubated with either the IL13R␣2 (clone 47) mAb at 1 g/ml or the IL13R␣2 (clone 47) mAb preincubated with a 10ϫ excess of rhIL13R␣2 (supplemental Fig. 1A) and analyzed by flow cytometry. A significant ablation of interaction between the IL13R␣2 (clone 47) mAb in the presence of a 10ϫ excess of rhIL13R␣2 was found when compared with clone 47 alone. Similarly, preincubation of N10 cells with either a 10ϫ excess of rhIL-13 or IL13R␣2 (clone 47) mAb almost completely blocked the interaction between the antibody or rhIL-13 and N10 cells (supplemental Fig. 1B), indicating a specificity of recognition between IL13R␣2 expressed on the surface of glioma cells and clone 47 (supplemental Fig. 1).
To verify that the IL13R␣2 (clone 47) mAb possessed the ability to bind IL13R␣2 on the surface of glioma cells in situ, intracranial glioma xenografts of U251 cells expressing green fluorescent protein (GFP) were established in nude mice. Three weeks later, animals were sacrificed, and cells were obtained and placed into in vitro culture conditions. After 48 h, the cells were collected and stained with control mIgG or IL13R␣2 (clone 47) mAb. Cultured GFP-expressing U251 cells served as a positive control (data not shown). GFP-positive U251 cells represented ϳ56% of the total cells (Fig. 3C, panel a), and 96% of the cells were reactive with the IL13R␣2 (clone 47) mAb (Fig. 3C, panel c), whereas GFP-negative cells did not interact with the antibody (Fig. 3C, panel b). These data further confirm that the IL13R␣2 (clone 47) mAb specifically recognizes glioma cells expressing IL13R␣2 in mouse xeno- grafts and is not reactive with other cells from the mouse brain.
Affinity Studies-Surface plasmon resonance was used to determine the affinity and rate of interaction between the IL13R␣2 (clone 47) mAb and rhIL13R␣2. All measurements were done in comparison with two commercial antibodies against IL13R␣2, clones 83807 and B-D13. Fig. 4 shows the sensorgrams for each antibody. The measurements are summarized in Table 1. Fig. 4A shows that clone 47 demonstrates a prolonged and stable association with rhIL13R␣2 measured over a 30-min time frame, whereas clones 83807 (Fig. 4B) and B-D13 (Fig. 4C) dissociate relatively quickly. The affinity of binding for the IL13R␣2 (clone 47) mAb to rhIL13R␣2 was calculated at 1.39 ϫ 10 Ϫ9 M. This value exceeded the affinity of the commercially available antibody clones 83807 and B-D13 to rhIL13R␣2 by 75ϫ and 33ϫ, respectively. Clone 47 demonstrated the highest binding affinity (R max ) to rhIL13R␣2 at 390 RU when compared with 250 and 8 -16 RU for clones 83807 and B-D13, respectively. These data indicate that the IL13R␣2 (clone 47) mAb possesses properties superior to clones 83807 and B-D13 as well as demonstrates a higher affinity toward rhIL13R␣2.
A Novel Monoclonal Antibody Competes with rhIL-13 for Binding to IL13R␣2-To determine whether the IL13R␣2 (clone 47) mAb possesses inhibitory properties, competitive binding assays utilizing a rhIL13R␣2hFc chimera and HEK cells transiently expressing the human IL13R␣2 were performed. The competitive binding assay was set up in a plate-bound ELISA format. The rhIL13R␣2hFc absorbed to the plate served as the target antigen. To determine whether the IL13R␣2 mAb specifically inhibits the binding of IL-13 to rhIL13R␣2, plates were preincubated with a 100ϫ excess of mIgG, the IL13R␣2 (clone 47) mAb, or other IL13R␣2 mAb clones, including 83807, YY-23Z, and B-D13, followed by incubation with rhIL-13. Fig. 5A shows that the IL13R␣2 (clone 47) mAb significantly abolished the binding of rhIL-13 to rhIL13R␣2, whereas the IL13R␣2 mAb clones B-D13 and 83807 competed for binding of human IL-13 significantly less.
To further verify the inhibitory properties of the IL13R␣2 (clone 47) mAb, HEK 293T cells were transfected with a construct encoding wild-type or a 4-amino acid mutant form of IL13R␣2 cDNA in which Tyr 207 , Asp 271 , Tyr 315 , and Asp 318 residues were substituted to Ala. Previously, these residues of the human IL13R␣2 were identified as amino acids required for the interaction with the cognate ligand, IL-13. The presence of all four mutations in one molecule has been shown to result in near complete ablation of the binding of IL-13 to the mutated form of IL13R␣2 (28). After 48 h, the cells were pretreated with a 20ϫ excess of rhIL-13 or the IL13R␣2 (clone 47) mAb followed by incubation of the IL13R␣2 (clone 47) mAb or rhIL-13, respectively. Fig. 5B shows an ϳ50% binding inhibition of IL13R␣2 (clone 47) mAb by a 20ϫ excess of rhIL-13 to wildtype (WT) IL13R␣2 but not to a 4-amino acid mutant form of IL13R␣2. A 20ϫ excess of antibody abolished the binding of rhIL-13 to IL13R␣2 when expressed on the cell surface by 80%, which is similar to the result observed in plate ELISA. The residual binding of IL-13 to the 4-amino acid mutant form of IL13R␣2 was further decreased by an excess of the IL13R␣2 (clone 47) mAb (Fig. 5C). Collectively, these data suggest that the IL13R␣2 (clone 47) mAb specifically competes with rhIL-13 for the binding site on IL13R␣2. Also, these data suggest that the IL13R␣2 (clone 47) mAb and IL-13 have a significant overlap in their recognition site of the IL13R␣2 molecule. with rhIL13R␣2 as visualized by SPR in a Biacore 3000 are shown. The rhIL13R␣2 was injected at concentrations ranging from 1 to 100 nM (lower to upper curves) at a constant flow rate of 20 l/min over immobilized antibodies and over a control dextran surface (these values were subtracted from the signal). The association and dissociation phases were monitored for 300 s by following the change in SPR signal (colored curves) given in RU. Black curves represent the fit of the data to a one-site binding model. For derived kinetic parameters, see Table 1. Lower panels show residuals from a one-site binding model, indicating an excellent fit.

TABLE 1 Kinetics of monoclonal antibodies binding to the human recombinant IL13R␣2
The estimation of kinetic parameters was performed as described under "Experimental Procedures." The dissociation constant (K D ) of the complexes was determined as the ratio k a /k d . For quantitative analysis, three independent replicates were performed for each sample. Data are expressed as mean Ϯ S.E. These data demonstrate that the affinity of IL13R␣2 (clone 47) mAb to recombinant IL13R␣2 exceeds the affinity of commercially available mAb clones 83807 and B-D13 by 75ϫ and 33ϫ, respectively. Binding of the IL13R␣2 (clone 47) mAb to wild-type and mutant forms of IL13R␣2 was analyzed by flow cytometry. The IL13R␣2 mAbs 83807 and B-D13 were used as reference antibodies to exclude a possible influence of variations in the level of expression of wild-type or mutated variants of IL13R␣2 on the surface of HEK cells (Fig. 6A). Data were calculated as a ratio of IL13R␣2 (clone 47) binding to IL13R␣2 when compared with both antibody clones 83807 and B-D13. Fig. 6A demonstrates that the binding of IL13R␣2 (clone 47) mAb was not significantly affected by either the individual mutations or the 4amino acid mutant form of IL13R␣2 when compared with wildtype receptor. In contrast, binding of IL-13 to the 4-amino acid mutant form of IL13R␣2 was nearly abolished (Fig. 6B). These data indicate that the Tyr 207 , Asp 271 , Tyr 315 , and Asp 318 residues are not crucial for the interaction of IL13R␣2 (clone 47) mAb with IL13R␣2 but are necessary for binding to IL-13.
N-Linked Glycosylation Affects the Affinity of the IL13R␣2 mAb for IL13R␣2-N-Linked glycosylation has previously been demonstrated to be important for efficient binding of IL-13 to the cognate receptor, IL13R␣2 (30). Taking into consideration the significant overlap in epitope recognition between the IL13R␣2 (clone 47) mAb and IL-13, we hypothesized that N-linked glycosylation of the IL13R␣2 might also contribute to binding of the IL13R␣2 (clone 47) mAb. To test this hypothesis, rhIL13R␣2hFc was treated with Pngase F to remove N-linked glycosylation from the protein. The binding of the IL13R␣2 (clone 47) mAb to control and deglycosylated target protein was investigated. Treatment of rhIL13R␣2 with Pngase F was performed under native conditions (in the absence of SDS) to avoid denaturation of the rhIL13R␣2 affecting the binding of antibodies. Additional mAbs to IL13R␣2 (clones 83807, B-D13, and YY23Z) and rhIL-13 were included in the assay to demonstrate the specificity of binding. In a plate-bound ELISA, binding of the IL13R␣2 (clone 47) mAb to Pngase F-treated IL13R␣2 was decreased by 35% when compared with untreated protein (n ϭ 4; p Ͻ 0.001). The binding of the IL13R␣2 (clone 83807) was reduced by 80% when compared with untreated protein and completely absent for the IL13R␣2 mAbs B-D13 and YY-23Z, respectively (n ϭ 4; p Ͻ 0.001) (Fig. 7A). Binding of rhIL-13 with Pngase F-treated rhIL13R␣2 was also significantly diminished in agreement with previously published reports (30). To verify that Pngase F treatment resulted in deglycosylation of the protein, control and Pngase F-treated rhIL13R␣2hFc protein was resolved by Western blot. Fig. 7B shows that Pngase F-treated protein has a lower molecular weight, confirming the removal of N-linked glycans from the IL13R␣2 molecule. Binding of the IL13R␣2 (clone 47) mAb to Pngase F-treated U251 glioma and HEK 293 cells expressing wild-type IL13R␣2 was also decreased by ϳ30% (n ϭ 3; p Ͻ 0.05) when compared with control untreated cells (Fig. 7C).
Immunohistochemistry-We evaluated the ability of the IL13R␣2 (clone 47) mAb to detect IL13R␣2 in fresh frozen tissues. Flash frozen human GBM samples or the U251 glioma flank xenograft was stained with either isotype control mIgG1 or the IL13R␣2 (clone 47) mAb. Fig. 8 shows positive (brown) staining in the two human GBM samples albeit with different frequency of positive cells in the sample as well as a U251 glioma cell-based glioma xenograft. Positive staining was detected in two of the three GBM samples analyzed, which is consistent    N-linked glycosylation on the binding of IL13R␣2 to recombinant IL13R␣2. A, binding of IL13R␣2 to control and Pngase F-treated rhIL13R␣2. Plates were coated with hrIL13R␣2 at 1 g/ml and treated with native buffer or with 1 milliunit/well Pngase F in native buffer for 3 h at 37°C. An ELISA for binding of the IL13R␣2 (clone 47) mAb in comparison with antibody clones B-D13, 83807, and YY-23Z and rhIL-13 was performed, and the data of one representative experiment from three independent experiments are shown. A paired t test was used to evaluate the difference between control and Pngase F-treated groups (n ϭ 4). *, p Ͻ 0.5; **, p Ͻ 0.01; ***, p Ͻ 0.001. B, a Western blot shows the lower molecular weight of Pngase F-treated rhIL13R␣2 due to removal of N-linked glycosylation adducts from the molecule. C, flow cytometry shows the binding of IL13R␣2 mAbs to IL13R␣2-expressing U251 and HEK293 cells treated with 1 milliunit of Pngase F for 1 h at 37°C. The data are representative of three independent experiments. A paired t test was used to evaluate the difference between control and Pngase F-treated groups. *, p Ͻ 0.5. MFI, mean fluorescence intensity. Error bars represent S.D.
with the literature suggesting that fewer than 50% of primary GBM express IL13R␣2 (3). These data are also consistent with the ability of this antibody to recognize the native form of IL13R␣2 expressed on the cell surface and in ELISA applications as well as the compromised ability of this mAb to detect denatured antigen by Western blotting.
The Novel IL13R␣2 mAb Prolongs the Survival of Animals with an Intracranial Glioma Xenograft-We next sought to determine the potential therapeutic properties of the IL13R␣2 (clone 47) mAb in an orthotopic mouse model of human glioma. U251 glioma cells were intracranially injected into the brain of nude mice alone, in the presence of control mIgG, or with the IL13R␣2 (clone 47) mAb. Fig. 9A shows that animals in the control PBS (n ϭ 15) and mIgG (n ϭ 16) groups demonstrated a similar median survival of 27 and 25 days, respectively. In contrast, the survival of animals co-injected with the IL13R␣2 (clone 47) mAb (n ϭ 13) was significantly increased to a median of 34 days (p ϭ 0.0001; mIgG versus the IL13R␣2 mAb group). Analysis of H&E staining of the glioma xenografts from brains collected on day 17 revealed a similar pattern of glioma cell distribution in the brain of control groups. In contrast, the tumor mass in the group of animals co-injected with IL13R␣2 mAb was significantly decreased in size (Fig. 9B). Independently, U251 cells were inoculated in the brain of mice and 3 days later injected through the same burr hole with either PBS or the IL13R␣2 (clone 47 or B-D13) mAb as described previ-ously (29). Interestingly, the mice injected with clone 47 demonstrated improvement in median survival when compared with PBS and clone B-D13 groups (35 days versus 27 and 23 days, respectively; n ϭ 7; p Ͼ 0.05) (supplemental Fig. 2), similar to what was found in the co-injection experiment (Fig. 9A). Nevertheless, all animals ultimately succumbed to the disease. These data indicate that the IL13R␣2 (clone 47) mAb may possess the ability to promote tumor rejection of IL13R␣2-expressing U251 glioma cells in the mouse brain. This preliminary finding provides support for further investigation of this antibody for therapeutic purposes in various models and various experimental settings of IL13R␣2-expressing glioma and other malignancies.

DISCUSSION
Recent work has demonstrated that monoclonal antibodies have emerged as valuable research and diagnostic tools as well as therapeutic agents. Monoclonal antibodies specific for tumor-associated antigens have significant advantages over systemic chemotherapies due to the ability to specifically target cancer cells while avoiding interaction with untransformed tissue. Therefore, the search for novel "magic bullets" continues to grow, confirmed by a global market for therapeutic antibodies worth $48 billion as of 2010. Therapeutic antibodies are products of traditional hybridoma technology or screening of libraries for antibody fragments and their subsequent engineering into humanized fragments or full size molecules. Prior to this study, the hybridoma cell line secreting a high affinity antibody to the tumor-specific antigen IL13R␣2 was unavailable to the scientific community. Here, we describe the generation and characterization of a novel high affinity antibody to the tumorspecific antigen IL13R␣2 and discuss its potential use in different applications.
The specificity of interaction of newly discovered antibodies to human IL13R␣2 was analyzed by ELISA using the rhIL13R␣2hFc fusion protein, recombinant human IL13R␣2 expressed on the surface of CHO and HEK cells, and several glioma cell lines expressing IL13R␣2 at various levels by flow cytometry. Our novel antibody demonstrated a specificity of  interaction to human IL13R␣2 and did not cross-react with human IL13R␣1 or mouse IL13R␣2. Moreover, the specificity of binding to IL13R␣2 was confirmed in competitive binding assays using rhIL13R␣2hFc fusion protein by ELISA or by flow cytometry for detection of IL13R␣2 expressed on the surface of HEK cells. In these assays, IL13R␣2 (clone 47) mAb competed with recombinant human IL-13 for its epitope and was able to block ϳ80% of the binding between IL-13 and IL13R␣2. Conversely, human recombinant IL-13 was able to block ϳ50% of antibody binding to IL13R␣2. Similarly, a significant decrease in the binding of IL13R␣2 (clone 47) mAb to N10 glioma cells was observed when rhIL13R2hFc chimera and rhIL-13 were used as competitors. The binding of rhIL-13 to N10 cells was also abolished by IL13R␣2 (clone 47) mAb. These data indicate that the two molecules have significant overlap in their recognition sites for IL13R␣2.
IL-13 is a small 10-kDa molecule (31), whereas an antibody is ϳ15 times greater in molecular mass. The ability of rhIL-13 to compete with an antibody for a binding site suggests that the inhibitory property of the antibody is likely due to the specific interaction with amino acid residues contributing to the binding of IL-13 to the cognate receptor rather than to steric hindrance, which can also prevent the interaction of IL-13 with its receptor. Previously, Tyr 207 , Asp 271 , Tyr 315 , and Asp 318 were identified as critical residues of IL13R␣2 necessary for interaction with IL-13 (28). Indeed, in our assays, the binding of IL-13 to a mutant IL13R␣2 carrying a combination of all 4 amino acid mutations to alanine was significantly abolished when compared with the wild-type receptor in agreement with previously published work (28). However, binding of the IL13R␣2 mAb to either the individual or the 4-amino acid mutant form of IL13R␣2 was not significantly affected. These findings indicate that Tyr 207 , Asp 271 , Tyr 315 , and Asp 318 residues are not critical for the recognition of IL13R␣2 by the IL13R␣2 mAb. The human IL13R␣2 and murine IL13R␣2 are structurally conserved and share 59% amino acid identity (32). Moreover, Tyr 207 , Asp 271 , Tyr 315 , and Asp 318 residues are conserved in human and murine IL13R␣2. Absence of binding of the IL13R␣2 mAb to murine IL13R␣2hFc fusion further supports the hypothesis that these amino acid residues contribute to the binding of IL-13 to IL13R␣2 and are not critical for the interaction of this antibody with the receptor.
To further characterize the interaction of IL13R␣2 with our novel antibody, the affinity of the IL13R␣2 mAb was measured and compared with the binding properties of two commercially available antibodies using the surface plasmon resonance method. The affinity of the IL13R␣2 mAb was determined to be equal to 1.39 ϫ 10 Ϫ9 M, greatly exceeding the affinity of comparable commercially available antibodies by up to 75ϫ. In agreement with the affinity studies, the IL13R␣2 mAb (clone 47) demonstrated superiority to two commercial antibodies binding to the IL13R␣2 expressed on the surface of various glioma cells and in ELISA. Although many properties of antibodies, including the affinity and avidity, in vivo stability, rate of clearance and internalization, tumor penetration, and retention, should be considered prior to specific usage, it has been reported that higher affinity antibodies are better for immunotherapeutic tumor-targeting applications (33). The single chain antibody fragment (scFv) MR1Ϫ1 against epidermal growth factor receptor variant III demonstrates about 15ϫ higher affinity than the parental scFvMR1 and also showed on average a 244% higher tumor uptake than that for the scFvMR1 (34). It is likely that the high affinity properties of our IL13R␣2 mAb will be advantageous for applications utilizing antibodies or associated derivatives for targeting tumor cells expressing IL13R␣2.
Previous work has identified the N-linked glycosylation of IL13R␣2 as a necessary requirement for efficient binding to IL-13 (30). Taking into consideration that our IL13R␣2 mAb inhibits ϳ80% of IL-13 binding to the cognate receptor, IL13R␣2, it is reasonable to suggest that the binding of this antibody with the deglycosylated form of IL13R␣2 could also be affected. The IL13R␣2 molecule has four potential sites of N-linked glycosylation. The binding of the antibody to rhIL13R␣2 or to IL13R␣2 expressed on the surface of HEK or U251 cells treated with Pngase F was decreased by 35 and 30%, respectively, when compared with non-treated control. A partial change in binding activity for the clone 47 when compared with clones 83807 and B-D13 suggests that removal of carbohydrate adducts from IL13R␣2 with Pngase F causes conformational changes of the receptor, indirectly affecting the binding of both IL-13 (30) and the IL13R␣2 mAb to IL13R␣2. This also supports the hypothesis that the antibody binds directly to the IL13R␣2 amino acid backbone rather than due to interaction with post-translational carbohydrate moieties. Supporting this hypothesis, several studies have previously demonstrated that the conformational profile and structural rigidity of proteins depends on N-linked glycosylation (22,(35)(36)(37)(38).
To identify whether the IL13R␣2 mAb possesses possible therapeutic properties, we performed an in vivo study whereby glioma cells and the IL13R␣2 (clone 47) mAb were intracranially co-injected into brain, or antibody was injected into established tumor-bearing mice. Interestingly, the IL13R␣2 mAb was able to delay tumor progression and improve survival of animals with intracranial U251 glioma xenografts most significantly in the co-injected model, demonstrating a trend in the improvement of median survival in animals with established glioma. However, further optimization will be necessary to validate the therapeutic effect of this antibody in various experimental settings. Although the underlying mechanism for this antitumor effect remains unclear, the result suggests a possible therapeutic applicability for this antibody alone or as a therapeutic carrier with regard to the future treatment of IL13R␣2expressing glial and other lineage tumors. Several antibodies have been shown to mediate a cytotoxic effect in tumors through Fc-mediated activation of complement (39). Antibody-dependent cell-mediated cytotoxicity-induced activation of effector cells can also contribute to the cytotoxic effect of antibodies against targeted cells (40,41). Anti-IL13R␣2 derived from the sera of animals challenged with D5 melanoma cells expressing human IL13R␣2 demonstrates the ability to inhibit cellular growth in vitro (4). However, further detailed studies are required to delineate the properties of this antibody in vivo using additional models of IL13R␣2-expressing tumors.
The IL13R␣2 has been found to be expressed in several types of human cancers, including glioblastoma; medulloblastoma; Kaposi sarcoma; and head and neck, ovarian, pancreatic, and kidney cancers (2,(43)(44)(45)(46). Although the role of IL13R␣2 in some cancers is not yet defined, recent reports have demonstrated that IL13R␣2 contributes to the invasive phenotype of ovarian and pancreatic cancers (5,13). Moreover, Minn et al. (42) have suggested a relationship between IL13R␣2 expression and breast cancer metastasis to the lung. These studies confirm the importance of further evaluation of the therapeutic properties of clone 47 in additional experimental models of IL13R␣2overexpressing tumors. Additionally, Fichtner-Feigl et al. (11) demonstrated that the interaction of IL-13 with IL13R␣2 upregulates TGF-␤1, mediating fibrosis in a bleomycin-induced model of lung fibrosis. In light of this finding, it would be interesting to determine whether our novel antibody also has the ability to attenuate TGF-␤1-induced pulmonary fibrosis. However, the absence of cross-reactivity with mouse IL13R␣2 will hamper the study of this effect in mouse models, although it may be useful to evaluate other anti-mouse IL13R␣2 antibodies with inhibitory properties, translating those findings into an understanding of the potential implications of clone 47.
In conclusion, we report the generation of a novel antibody specific to human IL13R␣2. The antibody possesses a high affinity for IL13R␣2 and competes with IL-13 for the binding site on IL13R␣2. The antibody recognizes antigen expressed on the cell surface of glioma cells as well as other IL13R␣2-expressing cells, indicating the potential suitability for targeting IL13R␣2-expressing tumor cells in vivo. These properties combined with the existence of our hybridoma cell line also warrants further engineering of this antibody into smaller antibody fragments for genetic fusion with therapeutic proteins. Moreover, the novel antibody should be tested in various applications, including diagnostic imaging, delivery of antibody radionuclide conjugates, bioassays for the detection of IL13R␣2, and as a carrier for therapeutic agents in various types of IL13R␣2overexpressing tumors. Finally, we have confirmed here that our antibody has applicability in flow cytometry, ELISA, and immunocytochemistry, applications that are critical to study the biological regulation of IL13R␣2 in various contexts.