IL-4 receptors on human medulloblastoma tumours serve as a sensitive target for a circular permuted IL-4-Pseudomonas exotoxin fusion protein

Cytotoxins directed to interleukin-4 receptors have shown to mediate relatively selective cytotoxicity against a variety of human cancer cells in vitro and in vivo. In an ongoing Phase I clinical trial, a recombinant protein comprised of circularly permuted IL-4 fused to a mutated form of Pseudomonas exotoxin (the fusion protein termed IL-4(38-37)-PE38KDEL or cpIL4-PE) has shown antitumour activity against malignant glioma. Human medulloblastomas are neuroectodermal tumours that occur in children and have a poor prognosis. The goal of this study was to determine whether human medulloblastoma derived cell lines express interleukin-4 receptor and whether interleukin-4 receptor expression is accompanied by sensitivity to cpIL4-PE. Medulloblastoma cell lines express interleukin-4 receptor at the protein and mRNA levels as determined by binding, indirect immunofluorescence and RT–PCR studies. These cells expressed IL-4Rα (also known as IL-4Rβ) and IL-13Rα1 (also known as IL-13Rα′) chains, however common γc, a component of the interleukin-4 receptor system in immune cells was not detected. Consistent with the expression of IL-4R, cpIL4-PE was found to be highly and specifically cytotoxic to four of five medulloblastoma cell lines. Susceptibility of medulloblastoma cell lines to cpIL4-PE seemed to correlate closely to the functional IL-4 binding sites in general as demonstrated by 125I-IL-4 binding, but did not seem to correlate with mRNA or cell surface immunoreactive receptor protein expression. The sensitivity of medulloblastoma cells to cpIL4-PE could be eliminated by concurrent incubation with IL-4 or IL-13, but not with IL-2. None of these cell lines showed any change in proliferation upon treatment with exogenous IL-4. These studies establish the interleukin-4 receptor as a medulloblastoma-associated target for possible tumour-directed cancer therapy. Further studies are warranted to investigate interleukin-4 receptor expression in primary medulloblastoma tumours and sensitivity to cpIL-4PE in vitro and in vivo. British Journal of Cancer (2002) 86, 285–291. DOI: 10.1038/sj/bjc/6600034 www.bjcancer.com © 2002 The Cancer Research Campaign

Medulloblastomas are malignant brain tumours of neuroectodermal origin that arise in the cerebellum. They primarily occur in children within the first decade of life and carry a poor prognosis in the advanced stage of the disease (Stevens et al, 1991). New active agents are needed to improve survival and enhance the quality of life of these patients. Our previous studies have identified abundant expression of receptors for IL-4 (IL-4R) on human brain tumour cells (Husain et al, 1998;Joshi et al, 2000;Puri et al, 1994Puri et al, , 1996a. Based on our preclinical efficacy and safety studies, we have begun a Phase I clinical trial, in which a recombinant chimeric protein composed of circularly permuted IL-4 fused with Pseudomonas exotoxin (cpIL4-PE or IL4(38-37)-PE38KDEL) is administered intratumouraly in patients with high-grade glioma (Rand et al, 2000). Preliminary results suggest that cpIL4-PE given by this route is safe and has antitumour activity in some recurrent gliomas. Thus, medulloblastoma patients may benefit from targeted cancer therapeutics, which have low toxicity and high specificity.
In addition to demonstrating expression of IL-4R on cancer cells including brain tumour cells, we have extensively studied the structure and signal transduction of IL-4R on cancer cells (Kawakami et al, 2000;Leland et al, 2000;Murata et al, 1995Murata et al, , 1997aMurata et al, ,b, 1998aObiri et al, 1993Obiri et al, , 1997. We have shown that the IL-4R exists in three different types (Murata et al, 1995(Murata et al, , 1998a. Type I (classical) IL-4Rs are composed of IL-4Ra (also known as IL-4Rb) and IL-2Rg chain (g c ), while type II (alternative) receptors are composed of IL-4Ra and IL-13Ra1 subunits (Kawakami et al, 2000;Leland et al, 2000;Murata et al, 1995Murata et al, , 1997a. In some cell types all three chains are present (Type III IL-4R). From reconstitution studies, we have observed that IL-4 forms a complex with only two chains of IL-4R at a time even when all three chains were transfected in CHO-K1 cells (Murata et al, 1998b). Whether all three chains form a productive IL-4R complex is still not known. We and others have shown that type I IL-4R are expressed in immune cells e.g., T and B lymphocytes, monocytes, and haematological malignancies . Type II IL-4R is expressed in the majority of solid tumour cells examined e.g., glioblastoma, breast carcinoma, AIDS-associated Kaposi's tumour, head and neck cancer, renal cell carcinoma and others (Murata et al, 1998a). Type III IL-4R is expressed predominantly in B cells and monocytes.
Besides the dissimilarity in structure between haematological and non-haematological cells, the signalling by IL-4R in these cells is also different. In immune cells, IL-4 primarily signals through Jak1 and Jak3 tyrosine kinases while in non-immune cells IL-4 signals through Jak1 and Jak2 tyrosine kinases (Murata et al, 1995(Murata et al, , 1998a. Even though proximal signal transduction by different types of IL-4R is different, the distal signalling through signal transduction and activator of transcription (STAT) is the same. All types of IL-4R phosphorylate and activate STAT-6 protein (Murata et al, 1995(Murata et al, , 1998a.
The significance of expression of IL-4R on solid tumour cells is not known. It is also not known whether human medulloblastoma cells express IL-4R and, if they do, which chains of these receptors are present on medulloblastoma tumour cells. In the present study, we have examined the mRNA and protein expression of various receptor subunits of IL-4R by RT -PCR, radiolabelled binding, and indirect immunofluorescence assays in different medulloblastoma cell lines. In addition, we have tested the cyototoxic effect of cpIL4-PE on various medulloblastoma cell lines.

Medulloblastoma cell lines
UW-228-1, UW-228-2 and UW-228-3 cell lines were derived from a medulloblastoma tumour and generated at Washington University, Seattle, WA, USA (Keles et al, 1995). D283, and D341 cell lines were purchased from the American Type Culture Collection (ATCC), Manassas, VA, USA. All these cell lines grew as monolayers except D283, which grew mostly in suspension while D341 cell line grew as a suspension cell line. PM-RCC, renal cell carcinoma cell line was derived in our laboratory (Obiri et al, 1993) and H-9 T cells were purchased from ATCC. These cell lines were cultured in DMEM medium containing 10% foetal bovine serum (Biowhittaker, Walkersville, MD, USA), 1 mM HEPES, 1 mM non-essential amino acids, 100 units ml 71 penicillin, and 100 mg ml 71 streptomycin (Biowhittaker).

RNA extraction
Total RNA was extracted from medulloblastoma cell lines using the Trizol reagent (Gibco, Gaithersburg, MD, USA) as per manufac-turer's instructions. Briefly, 5610 6 cells were pelleted and lysed with the reagent and centrifuged after adding chloroform. The RNA from aqueous phase was separated and precipitated with cold isopropanol. The pellet was washed with 70% ethanol twice, dried and reconstituted with RNase free water. RNA was quantitated after measuring the optical densities at 260 and 280 nm in a spectrophotometer and stored at 7708C.

Reverse Transcriptase-Polymerase Chain Reaction (RT -PCR)
Levels of IL-4R subunit mRNA were determined by RT -PCR using primers and conditions previously described (Murata et al, 1997a).

Immunofluorescence assay
Twenty thousand cells from different medulloblastoma cell lines were cultured in chambered glass slides (Lab Tek-Nalge Nunc International, Naperville, IL, USA) for 48 h as described . The cells were washed with PBS and fixed in cold methanol : acetone (1 : 1) and incubated at 7208C for 2 h. The slides were then washed and rehydrated with PBS and used for indirect immunofluorescence analysis. Monoclonal antibody against IL-13Ra1 was purchased from Diaclone, Besancon, France. Mouse monoclonal antibody (M57) against IL-4Ra chain was a kind gift from Immunex Corporation, Seattle, WA, USA and polyclonal rabbit anti g c antibody was purchased from Santa Cruz, California, USA. Rehydrated cells in the chambered slide were incubated with 1% bovine serum albumin, 5% either goat or horse serum in PBS to block non-specific binding of antibody. The slides were washed and incubated either with the specified primary antibody (1 : 1500 for mAb IL-13Ra1, rabbit polyclonal and mAb for IL-4Ra) or isotype control (mouse IgG1 for IL-13Ra1 and IL-4Ra or rabbit IgG) for 2 h at room temperature. Slides were then washed three times with PBS for 5 min at room temperature and stained with secondary antibodies conjugated to either TRICT or FITC which were diluted in PBS containing 0.1% BSA as per manufacturer's recommendation. After three washes with PBS, slides were dried and layered with Vectashield anti-fluorescence fading mounting medium   (Vector Laboratories, Burlingame, CA, USA) and cover slipped. The slides were viewed in a Nikon epifluorescence microscope using appropriate filters.

125
I-IL-4 binding and displacement assay IL-4 was iodinated with IODOGEN reagent (Pierce, Rockford, IL, USA) according to manufacturer's instructions. The specific activity of radiolabelled IL-4 was *22 mCi mg 71 . The IL-4 binding assay was performed as previously described (Obiri et al, 1993;Puri et al, 1991). Briefly, tumour cells were harvested after brief incubation with versene (Biowhittaker, Walkersville, DE, USA), washed three times in Hanks balanced salt solution and resuspended in binding buffer (RPMI 1640 plus 1 mM HEPES and 0.2% human serum albumin). For the displacement assay, medulloblastoma cells (1610 6 /100 ml) were incubated at 48C with 125 I-IL-4 (100 -200 pM) with or without 200-fold molar excess of unlabelled IL-4. Following a 2 h incubation, cell bound radio-ligand was separated from unbound by centrifugation through a phthalate oil gradient and radioactivity determined with a gamma counter (Wallac). The number of receptors and binding affinities were determined as previously described (Obiri et al, 1993;Puri et al, 1991).

Protein synthesis inhibition assays
The cytotoxic activity of IL4-toxins was determined by protein synthesis inhibition assay as previously described (Puri et al, 1991). Protein synthesis inhibition directly correlates with cell death (Puri et al, 1996b). Typically, 10 4 medulloblastoma cells were cultured in leucine-free medium with or without various concentrations of cpIL4-PE for 20 -22 h  Cell proliferation assay Proliferation assays were performed as described previously . Briefly, the cells were washed, and resuspended in CM in which the FBS content was reduced to 2%. 2610 4 cells for each cell line were plated in a 96-well plate and cultured for 6 h at 378C in a 5% CO 2 incubator. IL-4 (0.1 -1000 ng ml 71 ) was added to the cultures and incubated further for 16 h. The cells were then pulsed with [ 3 H]-thymidine (1 mCi well 71 ) for an additional 8 h and frozen at 7708C. The plates were thawed, harvested and radiolabel uptake was measured by a Beta plate counter (Wallac, Gaithersburg, MD, USA).

Expression of IL-4 binding sites and IL-4 receptor chains in medulloblastoma cell lines
By radiolabelled IL-4 binding assays, all five medulloblastoma cell lines tested were found to express varying numbers of IL-4 receptors (Table 1). Subsequent RT -PCR analysis revealed that all five medulloblastoma cell lines expressed similar levels of mRNA for IL-4Ra and IL-13Ra1 chains, except D283 cell line which appear to show moderate expression for IL-13Ra1 chain as the intensity of mRNA band was lower compared to other cell lines (Table 2, Figure 1). For a positive control, PM-RCC cells expressed mRNA for IL-4Ra and IL-13Ra1 chains. PM-RCC have been previously shown to express IL-4Ra and IL-13Ra1 chains (Obiri et al, 1997). We next examined the expression of various IL-4R receptor chains by indirect immunofluorescence assays in five different medulloblastoma cells lines. As shown in Figure 2, immunofluores-cence staining for IL-4Ra and IL-13Ra1 chains was very intense in UW228-2, UW228-3 and D283 cell lines. In contrast, the immunofluorescence staining of cells with matched isotype control antibody did not reveal any staining at all. Interestingly, D283 cell line showed strong immunoreactivity with a small number of cells even though the microscopic field shown had more cells. These results indicate that IL-4R expression is heterogeneous and all cells do not express IL-4R. Similar IL-4R staining results were observed in UW228-1 and D341 cell lines (data not shown). In contrast, IL-2Rg c protein was not detected in all five medulloblastoma cell lines tested (data not shown). Medulloblastoma cell lines are highly sensitive to the cytotoxic effect of cpIL4-PE To determine whether IL-4R on medulloblastoma cell lines could be used as a therapeutic target, we tested sensitivity of these cells to cpIL4-PE. CpIL4-PE is a chimeric protein in which amino acids 38 -129 of IL-4 were fused via a peptide linker to amino acids 1 -37, which are in turn fused to amino acids 353 -364 and 381 -608 of PE, with KDEL at positions 609 -612 (Kreitman et al, , 1995bPuri et al, 1996b). This purified fusion protein was found to be highly and specifically cytotoxic to a variety of human solid tumour cell lines including glioblastoma multiforme (GBM) in vitro (Puri et al, 1996a,b). We have also shown that this cytotoxin is highly active in the regression of established human xenografts in nude mice. Thus far, cpIL4-PE has been tested against epidermoid, GBM, AIDS-KS, and breast tumour models (Husain et al, 1998(Husain et al, , 1999Kreitman et al, 1995b;Leland et al, 2000). Remarkable antitumour activity including complete disappearance of established disease has been observed. As in vitro sensitivity to cpIL4-PE seems to correlate with IL-4R expression and in vivo antitumour activity and medulloblastoma cells expressed IL-4R, we tested whether cpIL4-PE is also cytotoxic to medulloblastoma cell lines. As shown in Figure 3 and Table 1, three of five medulloblastoma cell lines expressing 4900 IL-4 binding sites/cell were highly sensitive to the cytotoxic activity of cpIL4-PE (IC 50 =1 ng ml 71 ). While one cell line (D341), which expressed *600 sites/cell was moderately sensitive (IC 50 =30 ng ml 71 ) and one cell line (D283), which expressed the lowest level of IL-4R, was not sensitive to cpIL4-PE at concentrations up to 1000 ng ml 71 . Thus, cpIL4-PE induced cytotoxicity positively correlated with IL-4 binding sites/ cell. The cytotoxic activity of cpIL4-PE was neutralized by addition of excess IL-4 and also by IL-13 in UW-228-2 and UW-228-3 cell lines. As excess of IL-2 did not neutralize the cytotoxicity of cpIL4-PE, these results indicate IL-4R are related to IL-13R on medulloblastoma cell lines and cytotoxicity mediated by cpIL-4PE is somewhat specific (Figure 4). This property of sharing of receptors for IL-4 and IL-13 has also been observed in other cancer cell lines and thus appears to be universal (Kawakami et al, 2000;Murata et al, 1997b).

IL-4 does not modulate growth of medulloblastoma cell lines
Since IL-4 has been shown to modulate growth of various tumour cell lines , we examined whether it modulated growth of human medulloblastoma cell lines. As shown in Figure 5, IL-4 did not alter [ 3 H]-thymidine incorporation in any cell line tested even with IL-4 up to a concentration of 1000 ng ml 71 . These results indicate that IL-4 does not modulate growth of medulloblastoma cell lines.

DISCUSSION
We demonstrate that medulloblastoma cell lines express IL-4R that serve as a sensitive target for IL-4R-targeted agent, cpIL4-PE. Medulloblastoma cell lines were found to over-express IL-4Ra and IL-13Ra1 chains. We have previously shown that a large number (41 of 46 samples) of glioblastoma biopsy samples or primary cell cultures are positive for IL-4Ra and IL-13Ra1 chains as determined by RT -PCR and immunofluorescence analyses Puri et al, 1996a). Our present results confirm those observations and demonstrate that medulloblastoma cell lines also express IL-4Ra and IL-13Ra1 chains in all cell lines tested. The absence of g c protein indicated that medulloblastoma cell lines express type II IL-4R as do malignant glioma cells. This configuration is similar to that observed on other tumour cells including renal cell carcinoma, AIDS-Kaposi's tumours, head and neck tumours, colon carcinoma, ovarian carcinoma and breast tumour cells (Husain et al, 1997;Kawakami et al, 2000;Kreitman et al, 1994;Leland et al, 2000;Murata et al, 1997b;Puri et al, 1996b).
It is intriguing to note that D283 cell line expressed both chains of IL-4R (IL-4Ra and IL-13Ra1) at the mRNA and protein levels by RT -PCR and indirect immunofluorescence studies, respectively. However, upon direct binding study, it only expressed about 145 IL-4 binding sites/cell. Consistent with low IL-4 binding, this cell line was not sensitive to the cytotoxic activity of cpIL4-PE. The direct binding studies, using 125 I-IL-4 did not correlate with either immunofluorescent detection of cell surface IL-4/IL-13 receptor subunits or their mRNA transcripts. Thus, these assays may not be able to predict as to which tumours are likely to respond to cytotoxin-based therapy. Therefore, other assays may be required. It is possible that D283 cell line expresses a mutated form of IL-4R, which does not allow binding of IL-4. However, this mutation does not affect its detection by RT -PCR and indirect immunofluorescence assays.
In a recent study, we have observed that one normal brain tissue and single available normal human astrocytic (NHA) primary cell culture expressed mRNA for IL-4Ra and IL-13Ra1 chains . Upon indirect immunoflorescence studies the NHA primary cell culture was shown to express these chains at the cell surface. In another study, Liu et al (2000) have also demonstrated that IL-4R are expressed in normal human astrocytes and glioma cells confirming our initial observations. Thus, it appears that NHA also express type II IL-4R. Additional samples including paediatric brain are being examined to confirm or dispute these observations.
Although expression of IL-4Ra and IL-13Ra1 chains may form a functional IL-4R complex, it is not known whether this receptor complex will internalize the cpIL4-PE after binding. It is also not known whether cpIL4-PE is cytotoxic to normal brain tissues. In our Phase I clinical trial, we did not see any microscopic damage to normal brain tissues in three patients (in which intraoperative biopsy was obtained) that received 0.2, 2 and 6 mg ml 71 doses of cpIL4-PE, respectively. In addition, in our recent study, we have found that cpIL4-PE is not cytotoxic or modestly cytotoxic to a neuronal cell line and NHA cell culture compared to 15 glioma primary cell cultures in which cpIL-4PE is highly cytotoxic (Joshi et al, unpublished results). Furthermore, it is known that resting B cells, monocytes, bone marrow cells and endothelial cells which express type I or type II IL-4 receptors are not sensitive to cpIL4-PE (Husain et al, 1997;Puri et al, 1994). These results suggest that there is a large therapeutic window in which one can target cpIL4-PE to cancer cells but cause little collateral damage to normal brain tissues or other immune and non-immune cells. Additional primary normal brain tissues particularly normal cerebellum tissue samples need to be tested for sensitivity to cpIL4-PE.
We have previously observed that none of the solid tumour cell lines including glioma cell lines secrete IL-4. Therefore, IL-4 does not appear to be an autocrine growth factor for tumour cell lines. However, exogenous IL-4 has been shown to mediate contrasting effects on various tumour cell lines. For example, IL-4 can mediate modest inhibition of proliferation of human renal cell carcinoma (Obiri et al, 1993), melanoma (Hoon et al, 1991), breast and colon carcinoma (Toi et al, 1992), and non-small cell lung carcinoma cell lines (Topp et al, 1993a, b). On the other hand, IL-4 can modestly stimulate the growth of human squamous cell carcinoma of head and neck cell lines (Myers et al, 1996). The reason for contrasting effect of IL-4 is not known. Since, medulloblastoma cell lines did not respond to IL-4, it is reasonable to conclude that IL-4 is not a paracrine growth factor for these cells. To determine the functional activity of IL-4R on tumour cells, we and others have shown that IL-4 can cause phosphorylation of JAK1, JAK2 and STAT6 proteins in human colon carcinoma and ovarian carcinoma cell lines even though IL-4 does not mediate autocrine or paracrine growth of these tumour cell lines (Murata et al, 1995(Murata et al, , 1997b. These studies suggest that IL-4 induced signal transduction does not seem to mediate growth of tumour cells. Signal transduction may modulate growth inhibition or other unknown biological effects on these cells, which are yet to be discovered. Further studies are required to determine the significance of IL-4R expression on medulloblastoma cell lines.
In conclusion, we report that, like GBM, medulloblastoma cell lines also express IL-4R, which appears to be an efficient target for cpIL4-PE. Additional primary medulloblastoma tumour samples need to be tested for their sensitivity to cpIL4-PE. In addition, the efficacy of cpIL4-PE should be tested in animal models of human medulloblastoma tumours and perhaps in the clinic for the treatment of young children who have failed standard treatment. Alternatively, this agent could be useful in paediatric patients who are most susceptible to the deleterious effects of standard treatment of radiation to developing brain.

NOTE ADDED IN PROOF
We have recently reported (Joshi et al, 2001) that 15 of 18 glioblastoma mutliforme tumours express IL-4 receptors in situ. In contrast, IL-4R protein was not detected by immunohistochemistry in 11 different normal brain tissues including sections from one paediatric brain.