Targeting IL-13Rα2-positive cancer with a novel recombinant immunotoxin composed of a single-chain antibody and mutated Pseudomonas exotoxin

Introduction

Targeting cell-surface receptors, antigens, or intracellular molecules has become an attractive strategy for development of novel cancer therapy. These strategies include gene therapy, antibody therapy, small molecules, and immunotoxin therapies (1–4). Immunotoxins are anticancer therapeutic chimeric agents that consist of a targeting moiety and a cytotoxic moiety and show enhanced specificity for killing target cells, thus minimizing the systemic and local toxicity to normal tissues. Several clinical trials have been completed with immunotoxins and cytotoxins that show survival benefit for patients with glioblastoma multiforme, T-cell leukemia, Hodgkin's disease, and hairy cell leukemia (1–3).

Interleukin-13 (IL-13) is one of the type 2 T helper cell-derived cytokines that binds to two receptor chains, IL-13Rα1 and IL-13Rα2. IL-13Rα1 chain is a low-affinity binding IL-13 receptor, and with IL-4Rα, it forms a high-affinity receptor complex that is involved in IL-13-induced signal transduction through either JAK-STAT or phosphatidylinositol 3-kinase (4). IL-13Rα2 chain binds IL-13 with high affinity and internalizes after binding to ligand without the involvement of other chains. Although this chain was shown to be nonsignaling receptor through JAK-STAT pathway, it is recently reported that this receptor is involved in activation of AP-1 pathway leading to up-regulation of transforming growth factor-β1 (5). In the past several years, high levels of IL-13Rα2 expression have been characterized in various malignant tumor cell lines and tissues derived from human malignant glioma, head and neck cancer, Kaposi's sarcoma, ovarian cancer, and renal cell carcinoma, whereas normal cells or tissues derived from same organs show very low or undetectable expression levels (3, 6–9). Although UniGene cluster EST database (Hs.336046) show higher expression of IL-13Rα2 gene in some tissues (e.g., adrenal gland, connective tissue, lung, pancreas, kidney, and brain), these studies did not examine IL-13Rα2 protein expression in tissues. We and others have reported very low level expression of IL-13Rα2 protein in corresponding normal tissues from which tumors arise (e.g., ovary, brain, head and neck, and kidney; refs. 3, 6–9). For example, ∼80% glioblastoma tumors express high level of IL-13Rα2, whereas corresponding normal brain tissues do not show detectable level of this receptor (6, 10).

To target IL-13R-positive cancer, IL-13 cytotoxin (IL-13-PE38), composed of IL-13 and a mutated form of Pseudomonas exotoxin, has been generated (11). IL-13-PE38 is highly cytotoxic to IL-13R-positive cancer cells in vitro and in vivo (3, 6–9, 11), and this cytotoxin has been evaluated in several phase I/II clinical trials in patients with glioblastoma multiforme (3, 12). These studies have shown that i.t. and peritumoral infusion of IL-13 cytotoxin is well tolerated and seems to show clinical benefit to patients with recurrent disease (13). Based on these results, a phase III clinical study was initiated in which two to three catheters were placed peritumorally and IL-13 cytotoxin at a concentration of 0.5 μg/mL was infused. This multicenter trial was completed for patient accrual in December 2005. The results for safety and overall survival when compared with standard Gliadel treatment are being carefully evaluated. We also generated an IL-13E13K-PE fusion protein, which bound IL-13Rα2-positive tumor cells with higher affinity compared with wild-type IL-13-PE38 (14). Although it was suggested that IL-13E13K-PE displays lower cytotoxicity to normal cells lacking IL-13Rα2 chain, both IL-13E13K-PE38 and wild-type IL-13-PE38 showed similar efficacy and safety profile in vitro and in vivo (14, 15). The reason both cytotoxins showed similar activity is due to similar binding to IL-13Rα1, which is expressed ubiquitously. To create a highly selective immunotoxin, an antibody to IL-13Rα2 fused to toxin was considered to be important.

The phage display antibody library technology has been found to be a useful strategy to isolate antigen-specific antibody fragments because of broad specificity and simple procedure avoiding immunization. In recent years, through phage display antibody library technology, a variety of antibody fragments with desired specificity have been isolated independent of immunogenicity of the antigen. This has been accomplished by using genes with rearranged complementarity determining region 3 or genes from immunized individuals. In addition, specificity in synthetic phage display antibody libraries is greater than that found in a panel of hybridomas generated from an immunized animal (16). The smaller size of antibody fragments facilitates for the modification to increase their binding affinity or the conjugation to toxin to produce immunotoxins.

In this study, we have isolated genes for specific antibodies to IL-13Rα2 from human single-chain Fv (scFv) antibody phage library and then created an immunotoxin construct by selecting two high-affinity clones of scFv and fused to gene for PE. This chimeric construct was expressed and purified from Escherichia coli and fusion protein was tested in IL-13Rα2-positive tumors in vitro and in vivo.

Materials and Methods

Cell Culture and Reagents

PM-RCC human renal cell carcinoma and U251 human glioblastoma multiforme cell lines were cultured as described previously (14). Recombinant IL-13 was purchased from Pepro Tech. Recombinant extracellular domain of IL-13Rα2 (ECDα2) and IL-13-PE38 were produced, characterized, and purified in the laboratory as described previously (14, 17). The Griffin.1 library was kindly provided by Dr. S. Ellis (Center for Protein Engineering, MRC Centre).

Selection of IL-13Rα2 ECD-Specific Phage Clones

The Griffin.1 library is a scFv phagemid library made from synthetic v-gene segments. This library was derived by recloning VH and VL from human synthetic Fab Lox library vectors into the phagemid vector pHEN2. Phage particles expressing scFv were prepared as described previously (18). Briefly, E. coli TG1 (Stratagene) were electroporated and grown in 10 mL 2× YT medium (16 g bacto-tryptone/10 g bacto-yeast extract/5 g NaCl/L in H₂O) plus 2% glucose by shaking for 1 h at 37°C. Ampicillin (100 μg/mL) and M13KO7 helper phage (4 × 10¹⁰ plaque-forming units) were then added, and the culture was grown for 1 more hour. Cells were pelleted and resuspended in 10 mL 2× YT medium containing 100 μg/mL ampicillin and 50 μg/mL kanamycin. The culture was then incubated overnight in shaking incubator at 37°C. Phage were purified from the culture supernatant by precipitation three times with polyethylene glycol, resuspended in PBS, titered, and stored at -80°C until panning. Panning of the synthetic scFv library was done in Maxisorp immunotubes (Nunc) coated with recombinant ECDα2 at 4°C overnight. After washing and blocking with 2% skim milk, phage (1 × 10¹²) colony-forming unit in 4 mL PBS containing 2% milk was incubated. This was mixed end-over-end at room temperature for 0.5 h and tubes were washed 10 times with 0.1% Tween-20 and 10 times with PBS for the first round (20 times of both Tween 20 and PBS washes in each subsequent round). Then, bound clones were eluted by 100 mmol/L triethylamine and neutralized. The eluted phage was infected into log-phase TG1 bacterial culture and colony-forming units were measured. The selection of positive clones was repeated for another three rounds of panning.

Phage and scFv ELISA

Immunoplates (96 wells) were coated overnight at 4°C with 200 μL/well of 10 μg/mL recombinant ECDα2. Plates were then washed three times with 0.05% Tween-20 in PBS and blocked with PBS containing 2% milk. After washing, 10 μL PEG-precipitated phage from the stored aliquot of phage from each round of selection and 90 μL PBS with 2% milk were added and incubated for 90 min at room temperature. Plates were then washed and incubated with horseradish peroxidase–conjugated anti-M13 antibody. Positive clones were detected by incubating with substrate solution (3,3′,5,5′-tetramethylbenzidine/peroxidase) as described by manufacturer's instructions.

Construction of Plasmids

IL-13Rα2(scFv)-PE38 chimeric gene was constructed by fusing sequence of IL-13Rα2 scFv antibodies with PE38. Seven clones for IL-13Rα2 were sequenced and clones 4 and 7 were selected for construction of immunotoxins. Restriction enzyme site was inserted by PCR using sense primer 5′-CAGCCCGGCCCATATGCAGGTGCAGCTGGTGCAGTC-3′ (for both clones 4 and 7) and antisense primer 5′-ATGATGATAAGCTTTCGCACGTTTGATTTCCAG-3′ (for clone 4) and 5′-ATGATGATAAGCTTTCGCACCTAGCACGGTCAG-3′ for the construct of clone 7 to incorporate NdeI and HindIII restriction enzyme site at the 5′ and 3′ termini, respectively. After digestion of PCR products by NdeI and HindIII, each product was inserted along with PE38 sequence in pET24a expression vector (Novagen). The sequences of resulting constructs were verified by ABI Prism 310 genetic analyzer (Perkin-Elmer).

Protein Expression and Purification of Immunotoxins

Both constructs (clones 4 and 7) for IL-13Rα2(scFv)-PE38 plasmid were transformed into E. coli BL21 (DE3) cells. The bacterial cultures were induced with 1 mmol/L isopropyl β-0-1-thiogalactopyranoside (IPTG) at A₆₀₀ = 1.2 for 6 h at 37°C. Proteins were obtained as inclusion bodies and, after washing, solubilized in 100 mmol/L Tris containing 7 mol/L guanidine HCl and 10 mg/mL dithioerythritol and refolded by rapid dilution. The refolding solution was dialyzed against 50 mmol/L Tris-HCl (pH 7.5). The proteins were purified by fast-protein liquid chromatography using Q-Sepharose (two times) and mono Q columns (Amersham Pharmacia). The purity at each step was verified by SDS-PAGE and Western blotting. The purity (>99%) of the final recombinant protein was verified by SDS-PAGE.

Cytotoxicity Assay

The in vitro cytotoxic activity of anti–IL-13Rα2(scFv)-PE38 was measured by the inhibition of cell viability or protein synthesis (14). Briefly, 1 × 10⁴ cells were cultured in leucine-free medium with varying concentrations of anti–IL-13Rα2(scFv)-PE38 for 22 h at 37°C. [³H]leucine (1 μCi; NEN Research Products) was added to each well and incubated for an additional 4 h. Cell viability was measured by Cell Counting Kit-8 (Dojindo). All assays were done in quadruplicate, and the concentration of anti–IL-13Rα2(scFv)-PE38 at which 50% inhibition of protein synthesis occurred was calculated (IC₅₀). Statistics for the IC₅₀ were determined by Student's t test.

Radio-Receptor Binding Assay

Recombinant human IL-13 was labeled with ¹²⁵I (Amersham) by the Iodo-Gen iodination reagent (Pierce) according to the manufacturer's instructions. The IL-13 equilibrium binding studies were done using the previously described method (15). Briefly, 1 × 10⁶ cells in 100 μL binding buffer (RPMI 1640 containing 0.2% human serum albumin and 10 mmol/L HEPES) were incubated for 2 h with 200 pmol/L [¹²⁵I]IL-13 (specific activity, 20.5 μCi/μg) with or without 200 nmol/L unlabeled anti–IL-13Rα2(scFv4)-PE38, anti–IL-13Rα2(scFv7)-PE38, IL-13, or ECDα2 at 4°C. Cell-bound [¹²⁵I]IL-13 was separated from unbound by centrifugation through a cushion of phthalate oils. Radioactivities in pelleted cells were counted by a gamma counter.

Animal Studies

Athymic nude mice 6 weeks old (∼20 g body weight) were obtained from the Frederick Cancer Center Animal Facilities (National Cancer Institute). Human malignant brain tumor xenografts were established in the nude mice by s.c. injection of U251 cells into flank. Palpable tumors developed within 5 days (tumor size, ∼25 mm² or ∼65 mm³). The mice then received injection of excipient (0.2% HSA in PBS) or chimeric immunotoxins anti–IL-13Rα2(scFv4)-PE38, anti–IL-13Rα2(scFv7)-PE38, or IL-13-PE38 by either i.p. (500 μL using a 27-gauge needle) or i.t. (30 μL using microinjection syringe) routes.

Serum Chemistry and Organ Histology

For serum chemistry and organ toxicity, serum and tissue samples were obtained 1 day after last injection of immunotoxins. Organs from these experimental animals were fixed with 10% formalin. Tissue sections (5 μm) prepared from paraffin-embedded blocks were stained with H&E.

Results

Selection of Clones with Binding Activities to ECDα2

To obtain a diverse set of antibody fragments against ECDα2, the panning system using recombinant ECDα2 protein was chosen. Three rounds of panning were done. The number of phage captured on ECDα2 increased ∼500-fold with the third round of panning (data not shown). Next, we panned with successively decreasing concentration of ECDα2 to select phage with the highest affinity for IL-13Rα2. Single colonies from the titration of captured phage were used to rescue seven individual phage clones (Fig. 1A ). These clones were analyzed by phage ELISA. All clones exhibited positive reactivity for binding to ECDα2. None of the clone showed nonspecific binding to bovine serum albumin coated as a negative control (Fig. 1B). All seven ELISA-positive phage were used to prepare phagemids. After sequencing, two clones (4 and 7) were selected showing light and heavy chains. The other five clones were lacking either light or heavy chain (data not shown). The only difference between clones 4 and 7 was in the DNA sequence of the epitope recognition portion of the molecule.

Expression and Purification of Anti-IL-13Rα2(scFv)-PE38

To determine whether anti-IL-13Rα2 monoclonal antibodies (mAb; scFv) could target a cytotoxic agent to IL-13R-expressing cells, we constructed immunotoxins using two different clones (Fig. 2 ). The scFv immunotoxins were expressed in BL21(DE3) E. coli cells after induction with 1 mmol/L IPTG. The majority of the scFv toxin (∼90%) was found in the inclusion bodies. Both proteins were purified by ion-exchange Q-Sepharose and mono Q columns after renaturation from inclusion bodies. The purity of both immunotoxins was confirmed by SDS-PAGE. Both proteins showed ∼67-kDa band in SDS-PAGE (Fig. 3 ). The yield of anti-IL-13Rα2(scFv)-PE38 for clones 4 and 7 was 1.1 and 4 mg from 1 L bacterial culture, respectively.

Cytotoxicity against IL-13Rα2-Expressing Tumor Cells

As the immunotoxin-induced cell death is caused by toxic activity of PE, which inhibits protein synthesis, the ability of anti-IL-13Rα2(scFv)-PE38 to inhibit protein synthesis was used as a measure of its cytotoxic activity. It has been shown that IL-13Rα2 chain plays a critical role in ligand binding and internalization, and high specific binding to IL-13Rα2 would exert improved cytotoxic activity of immunotoxins to IL-13Rα2-expressing cancer cells. Consistent with this interpretation, anti-IL-13Rα2(scFv)-PE38 derived from clone 4 was more cytotoxic to PM-RCC and U251 cell lines with IC₅₀ of 1.5 and 1 ng/mL, respectively, compared with clone 7 (P < 0.05; Fig. 4A and B ). The IC₅₀ of anti-IL-13Rα2(scFv)-PE38 clone 7 was 20 ng/mL in PM-RCC cell line and 10 ng/mL in U251 cell line. Both immunotoxins did not show any detectable cytotoxicity (>1,000 ng/mL) in antigen-negative human astrocyte normal cell line (data not shown). In addition, cytotoxicity mediated by both immunotoxins was neutralized by recombinant ECDα2 protein, whereas IL-13 was not able to attenuate their cytotoxicities.

We also did cell viability studies to support protein synthesis inhibition results. Our data show that anti-IL-13Rα2(scFv)-PE38 decreases cell viability of target U251 glioma cells in a dose-dependent manner (data not shown). The IC₅₀ values [a concentration of IL-13-PE38 and anti-IL-13Rα2(scFv)-PE38 clones 4 and 7 causing 50% decrease in cell viability] of 0.8, 8, and 80 ng/mL were observed, respectively. These data are consistent with protein synthesis inhibition assay.

To determine whether the binding site of IL-13Rα2 on cell surface is shared by scFv antibodies and IL-13, we did binding assays using radiolabeled IL-13. Both immunotoxins were not able to displace [¹²⁵I]IL-13 binding to PM-RCC cells, whereas unlabeled IL-13 or ECDα2 attenuated radiolabeled IL-13 binding (Fig. 4C). These results suggest that both immunotoxins would not be affected by endogenous IL-13, which may be expressed by normal and cancer cells (17).

Antitumor Activity of Anti-IL-13Rα2(scFv)-PE38 in Nude Mice Bearing Human Glioblastoma Xenografted Tumor

We have shown previously that administration of another agent IL-13-PE38 by various routes induced growth arrest and/or complete regression of human glioblastoma multiforme tumor xenografts (3). Although IL-13-PE38 by i.v. route was capable of causing regression of established s.c. tumors, the half-life of IL-13-PE38 in tumor-bearing nude mice is very short (t_1/2α = 9 min and t_1/2β = 75 min). To achieve a better therapeutic activity, a twice daily i.p. administration was explored. This route of administration showed better antitumor effect of IL-13-PE38. In addition, as glioblastoma is a localized intracranial disease and blood-brain barrier does not allow access of therapeutic concentration of protein drug to tumor, a direct i.t. administration of drug was undertaken. This route of administration provided optimal antitumor response. Therefore, for current studies, we chose both i.p. and i.t. routes of administration of immunotoxins. To determine the antitumor activity of anti-IL-13Rα2(scFv)-PE38, several different doses of both immunotoxins were administered to nude mice bearing U251 tumors. Anti-IL-13Rα2(scFv)-PE38 isolated from both clones was injected either i.p. (50 or 100 μg/kg twice daily for 5 days) or i.t. [25 or 125 μg/kg once daily for 3 alternate days (days 5, 7, and 9)] when s.c. tumors were established. As shown in Fig. 5A , U251 tumors treated with excipient control grew linearly and mean tumor size reached 261 ± 38 mm² by day 36. On the other hand, mice treated by i.p. route with IL-13 cytotoxin or anti-IL-13Rα2(scFv)-PE38 showed suppressed tumor growth during the treatment schedule (days 5-9 after implantation). Although animals treated with IL-13-PE38 (25 or 50 μg/kg) or clone 4 derived anti-IL-13Rα2(scFv)-PE38 showed significant tumor regression by day 36 (P < 0.05 versus control), antitumor activity mediated by IL-13-PE38 was significantly higher compared with clone 4 derived immunotoxin (P < 0.05). Although there was no significant difference in tumor size between clones 4 and 7 derived anti-IL-13Rα2(scFv)-PE38 (50 μg/kg) treated mice, antitumor activity of clone 4 derived anti-IL-13Rα2(scFv)-PE38 was significantly higher compared with clone 7 derived anti-IL-13Rα2(scFv)-PE38 at 100 μg/kg dose (P < 0.05).

We next examined the efficacy of anti-IL-13Rα2(scFv)-PE38 and IL-13-PE38 immunotoxins when injected by i.t. route. As shown in Fig. 5B, tumors in excipient only injected control mice grew aggressively (214 ± 38 mm² by day 29). However, animals treated with clone 4 derived anti-IL-13Rα2(scFv)-PE38 showed significant tumor regression at both dosages (P < 0.05, 25 or 125 μg/kg dose versus control). Clone 7 derived anti-IL-13Rα2(scFv)-PE38 also caused significant tumor regression but only at the higher dose (P < 0.05, 125 μg/kg dose versus control). Thus, similar to i.p. route, antitumor activity of clone 4 derived anti-IL-13Rα2(scFv)-PE38 was significantly higher than clone 7 derived anti-IL-13Rα2(scFv)-PE38 at both dosages (P < 0.05) tested. Despite impressive antitumor effects, no visible toxicity was observed in both immunotoxins treated groups of mice in both i.p. and i.t. routes.

For comparison, i.t. route of administration of IL-13-PE38 also showed better antitumor activity. After administration of 125 μg/kg dosage of IL-13-PE38, four of five mice showed complete regression of tumors. Mean tumor size from all animals by day 29 was 6 ± 13 mm² (P < 0.05 versus control).

Nonspecific Toxicity in Mice

Both anti-IL-13Rα2 directed immunotoxins were also evaluated for their nonspecific toxicity in nude mice because mice are not expected to express high levels of IL-13Rα2 in any organ. Five mice in each group received i.p. injection for 5 days or 3 alternate days with various doses of immunotoxins. The mice were observed for 3 weeks. The mortality data are shown in Table 1 . Almost all of the deaths occurred within first 5 days of immunotoxin administration. Anti-IL-13Rα2(scFv)-PE38 immunotoxins showed remarkable safety profile. No mortality was observed and both immunotoxins were well tolerated up to maximum dose (200 μg/kg/d) given either by alternate day schedule or daily schedule. In contrast, the maximum tolerated dose of IL-13 cytotoxin for 3 alternate-day administration was 150 μg/kg, whereas maximum tolerated dose for five daily i.p. injections was 100 μg/kg/d (data not shown).

Toxicology of Anti-IL-13Rα2(scFv)-PE38

To further assess immunotoxin-related organ toxicities in normal tissues, blood serum chemistry was done after i.p. administration of three different doses of anti-IL-13Rα2(scFv)-PE38. Athymic nude female mice were used for these studies. Blood samples were collected on day 6 or 7 (1 day after the completion of injection schedule). Animals in clone 4 derived anti-IL-13Rα2(scFv)-PE38 showed elevation of lactate dehydrogenase at 200 μg/kg dosage only (1,669 compared with 680 units/L upper limit of normal). No other abnormality was observed. No abnormal changes were observed in mice treated with clone 7 derived anti-IL-13Rα2(scFv)-PE38. Similarly, serum chemistry values remained within the normal range on i.p. administration of up to 100 μg/kg/d IL-13-PE38 (data not shown). However, after five daily i.p. injections of 150 μg/kg/d dose IL-13-PE38, aspartate aminotransferase and alanine aminotransferase levels were slightly elevated.

Discussion

In the current study, we report, for the first time, production of two highly purified fusion proteins, in which scFv portion of IL-13Rα2 mAb was linked to mutated PE. This was made possible by isolation of two clones of single-chain antibody by phage display technology. These immunotoxins were found to be specifically cytotoxic to cell lines expressing IL-13Rα2 chain in vitro and in vivo. Both immunotoxins mediated regression of established s.c. human glioma tumors in nude mice without any significant organ toxicity. Although both immunotoxins mediated statistically significant regression of established s.c. tumors, the antitumor activity of clone 4 anti-IL-13Rα2(scFv)-PE38 was significantly higher than clone 7 anti-IL-13Rα2(scFv)-PE38 at two doses studied. Similarly, both immunotoxins at high doses (125 μg/kg) also mediated higher antitumor activity when injected directly into s.c. tumors.

Several immunotoxins have been produced and some evaluated and found successful in clinical trials (1–3). These immunotoxins have been targeted to a wide variety of hematologic malignancies including leukemias and lymphomas (1, 2). Significant antitumor activity has been observed with immunotoxins targeting the IL-2 receptor and CD33 antigen. Denileukin diftitox targeting IL-2 receptor and Mylotarg targeting CD33 has been approved by the U.S. Food and Drug Administration for the treatment of cutaneous T-cell lymphoma and acute myelogenous leukemia, respectively (19, 20). These successes with immunotoxins were realized only for hematologic malignancies but not for solid tumors. Among several reasons for lack of success in solid tumors include limited availability of targeted agents to tumors, and selective expression of cell-surface antigens or receptors, although highly expressed in tumor cells but also expressed at lower levels in normal tissues. In that regard, it is worth noting that IL-13Rα2 chain expression was not detected in normal brain tissues (6, 10), whereas it is induced on a variety of cancer cell types and some types of fibroblasts derived from pulmonary asthma and fibrosis. Because of the differential expression of IL-13Rα2, IL-13 cytotoxin (IL-13-PE38) has shown survival benefit in phase I/II clinical trials in patients with glioblastoma multiforme at doses that do not produce significant toxicity (3, 12, 13). As IL-13-PE38 is expected to bind IL-13Rα1 chain, which is ubiquitously expressed, it is possible that nonspecific organ toxicities will be observed at higher doses. To overcome these limitations, we produced anti-IL-13Rα2(scFv)-PE38, which showed highest specificity to IL-13Rα2. Because anti-IL-13Rα2(scFv)-PE38 does not bind IL-13Rα1, these immunotoxins showed minimum toxicity as observed by animal survival and serum chemistry changes. Thus, it is safe to conclude that because most normal cells express very weak or nondetectable level of IL-13Rα2, anti-IL-13Rα2(scFv)-PE38 may be administered systemically and safely exerting its selectivity (3, 6–9). In addition, because anti-IL-13Rα2(scFv)-PE38-mediated cytotoxicity was not neutralized by IL-13, it is predicted that the antitumor activity will not be blocked by endogenous IL-13, which may be expressed by tumors, peripheral normal cells, or circulating IL-13.

One additional advantage of using anti-IL-13Rα2(scFv)-PE38 for cancer therapy is that scFv mAb is selected from a human synthetic library. Because of this characteristic, there will be limited or no immune response against the antibody portion of fusion protein. In previous studies, mouse mAb have been used to make immunotoxins (1, 2). As human anti-mouse antibodies make complexes with circulating therapeutic mAb, it has been extremely difficult to achieve efficacious level of circulating therapeutic mAb or immunotoxin (21, 22). To overcome these limitations, modifications in antibody portion of immunotoxin have been made. Chimeric antibodies or “humanized” antibodies have been produced to attenuate the foreign nature of these therapeutic proteins and also prevent the hypersensitive reactions to xenogeneic proteins in the host (23, 24). However, PE38 used in this study is a bacterial foreign protein that is certain to be immunogenic in immunocompetent hosts, although the scFv is human origin. Future studies will focus on generating immunotoxins replacing PE by nonimunogenic molecules such as cyclophosphamide or other plant-derived toxins that cause cell death or damage (1, 25).

It is worth noting that both anti-IL-13Rα2(scFv)-PE38 did not show detectable cytotoxicity to normal astrocyte cells that do not express IL-13Rα2. In addition, both anti-IL-13Rα2(scFv)-PE38 did not induce thymidine uptake or enhance protein synthesis in TF-1 cell line, whereas IL-13-PE38 induced both protein synthesis and thymidine uptake because of functional nature of IL-13 (14). To further establish specificity of anti-IL-13Rα2(scFv)-PE38 immunotoxins, we did cytotoxicity assays in the presence of ECDα2 and IL-13. The cytotoxicity of these immunotoxins was neutralized by ECDα2 but not by IL-13. In addition, the binding of radiolabeled was neutralized by IL-13 and ECDα2 but not by anti-IL-13Rα2(scFv)-PE38 immunotoxins. Furthermore, clones 4 and 7 immunotoxins were selected by ECDα2 protein panning. Based on these results, we conclude that these immunotoxins are highly specific for IL-13Rα2. Despite the specificity, anti-IL-13Rα2(scFv)-PE38 did not show higher cytotoxicity to IL-13Rα2-expressing glioma and renal cell carcinoma cell lines compared with IL-13-PE38. The IC₅₀ of clones 4 and 7 derived anti-IL-13Rα2(scFv)-PE38 was ∼10- and 100-fold higher compared with IL-13-PE38 in cytotoxicity assays, respectively. Consistent with in vitro results, anti-IL-13Rα2(scFv)-PE38 did not mediate higher antitumor activity compared with IL-13-PE38 in vivo nude mice model of human brain tumors when the immunotoxins were administrated by i.p. or i.t. route to s.c. tumor-bearing mice. This may be due to lower affinity of antibody portion of fusion protein to target antigen. It is known that antibody fragments isolated by phage display technology often show lower binding activity compared with conventionally isolated mAb (26). Therefore, in vitro affinity maturation such as mutagenesis of scFv antibody will need to be done to increase the affinity and avidity to target cells.

The mechanism of differential activity of clones 4 and 7 derived antibody toxin is not completely clear. As both anti-IL-13Rα2(scFv)-PE38 immunotoxins did not compete for binding of radiolabeled IL-13, we could not assess whether these immunotoxins bind IL-13Rα2 chain with similar or dissimilar affinities. However, because clone 4 derived anti-IL-13Rα2(scFv)-PE38 immunotoxin showed higher cytotoxicity to two different cancer cell lines, it is hypothesized that this protein has higher binding affinity to IL-13Rα2. Additional direct binding studies are needed to fully explore these and other possibilities.

Our current study focused on xenografted s.c. brain tumor model in which tumor cells expressed high levels of IL-13Rα2 chain. We are interested to examine efficacy of IL-13Rα2-directed single-chain fusion immunotoxins in brain tumors, which are lower expressers of IL-13Rα2 chain, and in intracranial tumor models to simulate human disease. In these models, it is envisioned that anti-IL-13Rα2(scFv)-PE38 will be administered by convection-enhanced delivery using an Alzet pump to bypass the blood-brain barrier. Recent studies have shown development of brain tumor models that replicate the pathologic characteristics of human tumors (27). Our future studies will examine efficacy of immunotoxins in these and other brain tumor models.

To summarize, we have produced and evaluated two anti-IL-13Rα2(scFv)-PE38 immunotoxins, which show a specific cytotoxic activity against glioma and renal cell carcinoma cell lines and no detectable cytotoxicity to normal astrocytic cell line. Both immunotoxins showed specific antitumor activity in mouse xenograft models when given by i.p. or i.t. routes. In addition, both anti-IL-13Rα2(scFv)-PE38 were well tolerated without any systemic abnormalities. Because of these characteristics, anti-IL-13Rα2(scFv)-PE38 may be a useful agent for the therapy of IL-13R-expressing tumors including malignant glioma. Future studies should be done to evaluate safety and efficacy of anti-IL-13Rα2(scFv)-PE38 in different types of tumors that express IL-13Rα2 chain.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.