Use of Chlorotoxin for Targeting of Primary Brain Tumors1

[CANCER RESEARCH 58. 4871-4879. November 1. 19981 Use of Chlorotoxin for Targeting of Primary Brain Tumors1

Liliana Soroceanu, Yancey Gillespie, M. B. Khazaeli, and Harald Sontheimer2

Dcptirlrnenls of Neurohiology. Brain Tumor Research Â¡Mhoraloriesâ€”Divisionof Ne iirosiirgery IL.S. H. S.I, Surgery [Y. GÃ¬.ami Medicine (M. lÃ¬.K.J. University of Allibitimi ut Binnin^hinn. Rirniingluiin. AÃ¬nhumii.35294

ABSTRACT this process. These include variable mutations of p53. pl6, cdk4. and RB protein, amplification or overexpression of epidermal growth Gliomas are primary brain tumors that arise from differentiated glial factor receptor or MDM2 genes (mostly found in glioblastomas that cells through a poorly understood malignant transformation. Although arise "de novo"). glioma cells retain some genetic and antigenic features common to glial cells, they show a remarkable degree of antigenic heterogeneity and Because of the lack of glioma specific markers, unequivocal diag variable mutations in their genome. Glioma cells have recently been nosis of gliomas requires tissue biopsy and relies primarily on his- shown to express a glioma-specific chloride ion channel (GCC) that is topathological criteria. Histopathologically, gliomas are highly cellu sensitive to chlorotoxin (CTX), a small peptide purified from Leiurus lar with marked pleomorphism and often contain multinucleate giant quinquestriatus scorpion venom [N. Ullrich et al.. Neuroreport, 7: 1020- cells. Vascular proliferation throughout the lesion and the presence of 1024, 1996; and N. Ullrich and H. Sontheimer, Am. J. Physiol. (Cell Physiol.), 270: C1511-C1521, 1996]. Using native and recombinant 125I- focal areas of necrosis are histological features necessary for the diagnosis of GBM.1 the most malignant and prevalent form of glioma labeled CTX, we show that toxin binding to glioma cells is specific and involves high affinity [dissociation constant (A',,i = 4.2 nM] and low affinity in adults. Immunohistochemical studies show variable expression of (Ka = 660 nM) binding sites. In radioreceptor assays, I25l-labeled CTX the astrocyte-specific antigen glial fibrillary acidic protein; protein binds to a protein with M, = 72,000, presumably GCC or a receptor that levels were found to decrease with an increasing degree of malig modulates GCC activity. In vivo targeting and biodistribution experiments nancy. Vimentin expression is a common but not a reliable marker. were obtained using I25I- and I3ll-labeled CTX injected into severe com Presently, no antigen or marker that could serve to selectively label bined immunodeficient mice bearing xenografted gliomas. CTX selectively accumulated in the brain of tumor-bearing mice with calculated brain: glioma cells is available. To date, two other reports describe proteins specifically expressed by glioma cells: brain-enriched hyaluronan muscle ratios of 36.4% of injected dose/g (ID/g), as compared to 12.4%ID/g in control animals. In the tumor-bearing severe combined binding protein (3) and the novel interleukin 13 receptor a (4). immunodeficient mice, the vast majority of the brain-associated radioac We recently identified and characterized a chloride ion channel that tivity was localized within the tumor (tumor: muscle ratio, 39.13% ID/g; seems to be abundantly expressed in glioma cells but absent in normal contralateral braimmuscle ratio, 6.68%ID/g). Moreover, 131I-labeled CTX brain tissue (5. 6). GCC activity could be recorded biophysically in distribution, visualized through in vivo imaging by gamma ray camera cultured glioma cells (5. 6) and in freshly prepared vibratome sections scans, demonstrates specific and persistent intratumoral localization of from patient biopsy tissue (7). Channels were absent in nontumor the radioactive ligand. Immunohistochemical studies using biotinylated and fluorescently brain tissue and in tumor cells of nonglial origin. Because the activity tagged CTX show highly selective staining of glioma cells in vitro, in situ, of GCC seems to be modulated by rearrangements in the cells cy- and in sections of patient biopsies. Comparison tissues including normal toskeleton (8). it has been proposed that these channels may facilitate human brain, kidney, and colon were consistently negative for CTX shape changes during glioma cell migration and invasion. Interest immunostaining. These data suggest that CTX and CTX-conjugated mol ingly, GCC expression in situ correlates with the histopathological ecules may serve as glioma-specific markers with diagnostic and thera tumor grade (7). Only 40-45% of low-grade astrocytomas (WHO peutic potential. grade I-II) express GCC, whereas >90% of all of the high-grade tumors (WHO grade III) and essentially all of the GBMs express INTRODUCTION GCC. This correlation suggests that GCC may be a candidate protein to serve as a glioma-specific marker and may be useful for diagnostic Gliomas are among the most deadly forms of cancer for which and therapeutic purposes. Biophysical studies also demonstrated that effective treatment strategies are currently lacking. Indeed, despite the CTX, a 36-amino-acid peptide isolated from scorpion venom (Leiurus overall advances in chemo and radiation therapy regimens, the median survival of glioma patients has been unaltered in the last 20 years (1, quinquestriatus), effectively inhibits currents through GCC with -80% block at 600 nM CTX (5, 6). In the present study, we have 2). This lack of success in treatment of gliomas seems to be due in part to their high resistance to radiation and chemotherapy but additionally generated CTX in synthetic and recombinant forms and have attached moieties including I25I, ml, biotin, and various fluorophores to the to their unusual ability to disperse and invade healthy brain tissue. Glioma cells display a remarkable level of heterogeneity within and CTX molecule that subsequently allowed its detection. We have used among gliomas, as manifested through clonal variation in growth these tagged CTX molecules to investigate specific binding of CTX to potential, differential genetic alterations, variable drug resistance, and glioma cells in vitro and HI situ. These studies show that iodinated expression of antigens. CTX selectively binds malignant glioma cells and can be visualized The molecular mechanisms that underlie the malignant transforma by in vivo imaging of animals bearing xenografted tumors. Moreover, tion of glial cells to become gliomas are poorly understood. As with biotinylated or fluorescently tagged CTX molecules reliably detect other cancers, a series of genetic alterations seem to be obligatory for glioma cells in patient biopsies, which suggests that this molecule could serve as a glioma-specific marker with diagnostic and thera Received 6/23/98; accepted 9/2/98. peutic potential. The costs of publication of this anide were defrayed in pan by the payment of page charges. This article must Ihereforc be hereby marked atlvenisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1The abbreviations used are: GBM. gliohlastoma multiforme: CTX. chlorotoxin: ' This work was supported by NIH Grant ROI NS 36692 and American Cancer GCC, glioma chloride channel: HRP. horseradish peroxidase: Kd. dissociation constant: Society Grant RPG-97-083. X-Gal. 5-bromo-4-chloro-3-mdolyl-b-n-galactopyranoside; TEPSA, 3-aminopropyltri- " To whom requests for reprints should be addressed, at Department of Neurobiology. ethoxysilane: %ID/g. % (percentage) of injected dose/g: UAB. University of Alabama al CIRC 545. 1719 6th Avenue South. Birmingham. AL. 35294-0021. Birmingham. 4871

MATERIALS AND METHODS control counterparts were killed at indicated time points for biodistribution studies. Individual organs were removed, placed in preweighed tubes and Cell Lines and Cell Culture. The following human glioma cell lines were processed for gamma scintillation counting. Radioactivity associated with each used: (a) D-54 MG; (bÃ¬U373MG; (r) U105MG; (dÃ¬U251MG (all from Dr. organ is expressed as %ID/g of tissue. In the case of tumor-bearing '"l-labeled D. D. Bigner, Duke University. Durham, NC); and (e) neuroglial derived CTX-injected SCID mice, tumors were carefully removed from the right side SK-MG-I (from G. Cairncross, University of Ontario, London. Ontario, Can of the brain by dissection under the microscope and weighed and counted for ada). Control cell lines were: (a) Te671, a rhabdomyosarcoma cell line from associated radioactivity. Results are presented as %ID/g of tissue and as ratios American Type Culture Collection: and (b) Balb 3T3 fibroblasts. In addition, of brain:muscle or livermuscle for the two groups of animals. Biodistribution we used primary cultured rat astrocytes as control glial cells (see below). Cells experiments were repeated three times using six to eight animals for each time were maintained in DMEM mixed 1:1 with Ham's F-12 medium, supple point in each of the tumor-bearing and control groups. mented with L-glulamine (2 mM) and 10% fetal bovine serum (Life Technol Brain I2SI Autoradiography. I25l-labeled CTX-injected tumor-bearing ogies). Cells were harvested from logarithmic phase-growth cultures by brief mice were killed at 48 h after injection of the radiolabeled material and exposure to 0.5% trypsin and 0.53 HIMEDTA (Life Technologies). The D-54 perfused by intracardiac injection with PBS followed by 4% paraformaldehyde MG cell line was transfected with the pCMVLticZll plasmid (American Type in 10% sucrose. Brains were removed and fixed an additional 24 h at 4Â°C, Culture Collection) using the standard Lipofectamine method, according to the rinsed in PBS several times, embedded in OCT free/ing medium (Miles Inc., manufacturer's instructions (Life Technologies). Stably transfected clones Naperville, IL), and sectioned at 10 Â¿Â¿mona cryotome (Zeiss HM 505E): were selected by growth in media containing neomycin (G418, 5(X) /xg/ml). sections were mounted on TEPSA-coated slides. All of the subsequent steps Verification of the incorporation of marker gene was made by immunohisto- were performed under film safelight. Slides were coated with autoradiography chemical staining for X-Gal (Promega Corp.. Madison, W[). Briefly, cells were emulsion (Eastman Kodak) and exposed for 7 days in a light-tight slide box fixed in 4% paratbrmaldehyde. rinsed in PBS. and incubated overnight in containing Drierite at 4Â°C. For developing, we used the D-19 developer X-Gal substrate solution (0.2% X-gal, 10 mM sodium phosphate (pH 7.0), 150 (Eastman Kodak); slides were dipped 3-5 min at 16Â°Cin D-19 solution, mM NaCl, 3.3 mM K4Fe(CN),,'3H2O, and 3.3 mM K,Fe(CN)6], according to the washed for 3 min in distilled water, and placed in Rapid Fix Solution (Eastman manufacturer's instructions (Promega Corp.). Kodak), prepared according to the manufacturer's instructions. Slides were Primary Astrocyte Cell Cultures. Cortical astrocytes were isolated from then washed thoroughly with water and counterstained with Mayer's hema- Sprague Dawley rat pups at postnatal day 0 or 1. Cortical tissues were toxylin. dissected free and placed in ice-cold complete saline solution, consisting of In Vitro Binding Assay. CTX was radiolabeled with ['~5I]sodium iodide at 137 mM NaCl. 5.3 mM KC1, 2.0 mM CaCI2. 1.0 mM MgCU 25.0 mM glucose, the UAB Radiolabeling Facility using the chloramine-T method. Radioiodi- and 10.0 mM HEPES acid (pH adjusted to 7.2 using NaOH). Meninges and nated CTX was separated by reversed phase high-performance liquid chroma- blood vessels were then carefully removed, and the tissues were minced and tography. and radioactivity ranged from 3.6-5.8 Ci/mmol. Cells were grown incubated for 20 min at 37Â°Cin an enzyme solution consisting of complete for 24 h in 24-well plates (2 X 10s cells/well). Subconfluent monolayers were saline solution supplemented with 0.5 mM EDTA. 1.65 mM L-cysteine, and rinsed in ice-cold binding buffer (DMEM with 2% BSA) and then incubated 30.0 units/ml papain (Worthington. Freehold. NJ). The enzyme solution was with graded amounts of l:5I-labeled CTX diluted in binding buffer added to removed, and the cells were dissociated with a Pasteur pipette in complete duplicate wells in a total volume of 400 Â¿il/wcll.For each concentration of media |Earl's MEM (Life Technologies. Grand Island. NY), 10% PCS (v/v: 125I-labeled CTX. duplicate wells received 100-fold molar excess "cold" CTX. HyClone, Logan. UT). 5(X)unitsU/ml penicillin/streptomycin, and 20.0 mM Four wells per plate received only binding buffer and were used to obtain cell glucose] supplemented with 7.5 mg/ml each. BSA and trypsin inhibitor. Cells counts. Cells were maintained at 4Â°Cin a 10% CO2:90% O2 atmosphere on an were then plated on polyornithine/laminin-coated coverslips at a density of oscillating platform for 2-3 h. After incubation, supernatants were removed 1.2 X 104/cm2 or 8-well chamber slides and maintained in complete media in and aliquots saved for counting; cells were then washed four times in binding an atmosphere of 5% CO2:95% O2 at 37Â°C.The medium was replaced after buffer without BSA and incubated with solubilization buffer (1% Triton one day. Tissues treated in this manner produced cultures that were >95% X-KX), 10% glycerol. and 20 mw HEPES (pH 7.2)1 for 30 min with agitation. positive for the glial-specific marker glial fibrillary acidic protein. After solubilization, the contents of all of the wells were transferred to Intracranial Glioma Model. Female C.B.-17 SCID mice, ages 20-28 centrifuge tubes, and insoluble debris was centrifuged in a microfuge (at weeks, were anestheti/ed by i.p. administration of ketamine (20 mg/ml) plus 13,000 X g for 20 min at 4Â°C).Aliquots from supernatants and cell pellets xylazine (0.3 mg/ml) in saline at 0.07 ml/10 g of body weight. A midline scalp were counted in a gamma scintillation counter to determine cell-associated and incision was made and an 0.5-mm burr hole was made 1.5-2 mm to the right cell-"free" radioactivity. To the wells reserved for cell counting. 450 Â¡i\of of the midline and 0.5-1.(X) mm posterior to the coronal suture. Tumor cells 1 x Trypsin-EDTA and 50 Â¡uof trypan blue were added. (IO6 cells/dose) resuspended in excipient solution were stereotactically in I25l-labeled CTX Internalizaron Assay. D-54 MG cells were harvested jected using a 250-/xl Hamilton syringe with a 30-gauge needle mounted in a with trypsin/EDTA as described above and resuspended in binding buffer in Stoelting stereotaxic apparatus. The needle was advanced vertically through polypropylene tubes at a concentration of 6 x 10''cells/tube, 10 tubes for each the burr hole to a depth of 2.5 mm and then slowly retracted 45-60 s after group. Cells were incubated (3 h at 4Â°C,95% O2:5% CO2) with various injection: the incision was then closed with a 9-mm Michel wound clip. Mice dilutions (1.5-25 nM) of I25l-labeled CTX alone, with 100-fold molar excess were returned to sterile cages and were provided with sterile lab chow and "cold" CTX (each condition in duplicate), or with cell medium alone (for cell sterile water ail libitum. Previous studies showed that the D-54 MG human counts) for 3-4 h at 4Â°Cwith agitation and 5% CO,:95% O2 atmosphere. After malignant glioma cells reliably grow in SCID mice brain and typically kill the incubation with the ligand. cells were washed three times with cold binding animals within 16-18 days (9). Thus, we elected an 8-9-day interval to allow buffer without BSA to remove nonspecific binding and then resuspended in 2 for tumor growth before reintervention through the same burr hole to deliver ml of medium, and one-half of each tube volume was transferred to new tubes. the radiolabeled CTX (15 /Â¿Cil:sl-labeled CTX in 5 fil of PBS and 75 Â¿iCi "Old" tubes were maintained at 4Â°C,new tubes were transferred at 37Â°Cand '"[-labeled CTX/mouse) to both tumor-bearing and control (litter mates) incubated for an additional 60 min, after which all of the tubes were washed SCID mice. All of the animals received 0.8% Lugol solution in their drinking once with cold medium. Aliquots from washes were counted to determine water 24 h before injection of the radiolabeled material to inhibit thyroid "free" ligand. Cells were then resuspended in ice-cold 50 mM glycine-HCl (pH uptake of the iodine-labeled ligand. 2.5) in isotonic saline and incubated for 10 min to permit dissociation of bound In Vivo Localization and Biodistribution of Radiolabeled CTX. At 24, ligand. Afterward, cells were washed three times in 0.1 ml of ice-cold medium, 48, 72, and 96 h after tumor-bearing and control SCID mice were injected with and washes from each tube were pooled (membrane-associated radioactivity is '"l-labeled CTX, they were anesthetized and imaged using a large field-of- retrieved in acid washes). Aliquots from cells and washes were counted in a view camera fitted with a pinhole collimator interfaced to a microcomputer to gamma scintillator (Comprehensive Cancer Center, UAB). Cells incubated obtain a whole body scan. Each animal was imaged up to 1200 pixels with medium alone were used to determine cell number. (determined experimentally, as a value resulting in interprÃ©tablescintigrams Radioreceptor SDS-PAGE Analysis. Radioreceptor SDS-PAGE analysis above background levels); scintigrams were analyzed using the Image Tool was performed using a modified version of the transforming growth factor-ÃŸ analysis software. Both '"I- and I25l-labeled CTX-injected animals and their receptor assay (10). Specifically, D-54 MG cells were plated overnight in 4872

Fig. I. A, Scatchard plot analysis of '~5l-labeled CTX binding to D-54 MG human glioma cell line. Data are from one representative experiment; best fit indicates two binding sites receptor kinetics (Kd = 4.3 and 660 nM). This assay was repeated four times with similar results. B, autoradiography representing SDS-PAGE analysis of lysates ob tained after cross-linking of human glioma cell membranes to I25l-labeled CTX; the blot demon strates the presence of a single band (Mt â€”¿72,000), corresponding to the CTX-binding receptor with the 72kDa. intensity considerably decreased by competitive binding of "cold" CTX (Lanes I, and 3).

O 20 40 60 80 100 120 140 160 34 BoundCTX(pMxIO2)

6-well plates at 2 x IO5 cells/ml. Cell monolayers were rinsed in binding diaminobenzidine-peroxidase substrate (Vector Laboratories), or rhodamine- buffer without BSA, then incubated for 3-4 h with 25 nM 12SI-labeled CTX conjugated streptavidin (Molecular Probes), according to manufacturer's in alone or together with 100-fold molar excess "cold" CTX on an oscillating structions. For double-staining experiments, cells were first reacted with the X-Gal substrate, as described in the "Cell Lines and Cell Cultures" section, platform at 120 rpm. Cells were washed in cold binding buffer (without BSA): 5 ju.1of 27 mM disuccinimidyl suberate solution was added to each well and then processed for immunoperoxidase staining as described above. Slides were agitated for 15 min. Cells were rinsed briefly in detachment buffer [0.25 M covered using Aquamount (Fisher) or Fluoromount (Vector Laboratories) and sucrose, 10 mM Tris, 1 mM EDTA (pH 7.4), and 0.3 mM phenylmethanesul- viewed under epifluorescence or light microscope using standard procedures. fonyl fluoride] and then scraped using a disposable teflon scraper. After Images were captured with a Leica microscope equipped with digital camera. centrifugation ( 12,000 X g for 2 min) the supernatants were removed, and cells Staining of SCID Mouse Brain Sections. For demonstration of CTX were solubilized in 100 jul of solubilization buffer [125 mM NaCl. lOmM Tris, binding to glioma cells in .vim. intracranial tumors were induced in SCID mice l mM EDTA (pH 7), 0.1% volume/volume Triton X-100] supplemented with using D-54LacZ transfected cells. At 10-15 days after tumor induction, 10 /xl/ml protease inhibitor cocktails I and II (cocktail I = 1 mg/ml leupeptin, animals were killed by intracardial perfusion of PBS. followed by 4% 1 mg/ml antipain, 5 mg/ml aprotinin, and 10 mg/ml soybean trypsin inhibitor, paraformaldehyde. Brains were removed and additionally fixed overnight in cocktail II = 1 mg/ml pepstatin, 1 mg/ml bestatin, and 30 mM phenylmeth- paraformaldehyde and 12% sucrose. After fixation, brains were rinsed several anesulfonyl fluoride in dymethyl sulfoxide]. Cells were incubated (40 min, times with PBS, sectioned coronally, and reacted overnight for X-Gal enzyme 4Â°C)in solubilization buffer with end-over-end mixing for 40 min at 4Â°C.Cells immunochemistry, as described below; brain sections were then embedded in were centrifuged at 12.000 X g for 15 min, cell debris was removed, and the OCT freezing medium and sectioned. Sections, 8-10 /xm thick, were prepared supernatant was mixed with 1 volume of electrophoresis sample buffer [100 on a cryotome, mounted on TEPSA-coated slides, and reacted for biotinylated mM Tris (pH 6.8), 20% glycerol, 2% SDS, 0.05% bromphenol blue, and 100 CTX immunoperoxidase reaction, as described below. Sections were counter- mM DTT] and boiled for 3 min. A 10-/xg sample of each reaction (determined stained with Mayer's hematoxylin. dehydrated through increasing concentra by the Bio-Rad protein assay using the Coomassie Brilliant Blue G-250 dye) tions of ethanol and coverslipped (using Aquamount) for microscopical anal were loaded on precasted 4-20% gradient Tris-acrylamide minigels (Bio-Rad) ysis. CTX binding was also observed by direct or indirect and electrophoresed at constant voltage (90 mV). After electrophoresis, gels immunofluorescence, using brain sections that had been incubated (at 4Â°C were fixed in 7% acetic acid and 25% methanol solution for 20 min, dried in overnight in the dark) with Oregon-Green5>0-tagged CTX or biotinylated CTX a vacuum-operated gel dryer (Bio-Rad, model 583). and exposed to X-ray film in a humidified chamber. For direct detection of Oregon Green-labeled CTX, for 7-10 days at -70Â°C. slides were washed two times with PBS and coverslipped. To visualize Biotinylation of CTX. EZ-Link"Â»-Sulfo-NHS-LC-Biotin (Pierce Chemi biotinylated CTX, slides were incubated (2 h, room temperature, in the dark) cals) was used to biotinylate CTX. The peptide was resuspended in 50 mM with rhodamine-linked streptavidin (Molecular Probes), washed, and cover- sodium bicarbonate buffer (pH 8.5) at a concentration of 1 mg/ml; 200 /xl of slipped. Fluorescence-tagged localization of CTX was visualized as described CTX was incubated with 15 /xl of EZ-Link-Biotin reconstituted in distilled above. water at 1 mg/ml (2-3 h at 4Â°C)with end-over-end mixing, according to the Staining of Human Tissue Sections. Freshly resected brain biopsy sam manufacturer's instructions. To remove excess biotin, the biotinylated peptide ples (obtained as waste tissues from UAB Neurosurgery, under an Institutional was dialyzed against PBS using Slide-A-LyzerÂ® (Pierce Chemicals) mem Review Board approved protocol) were embedded in OCT. sectioned (8-10 brane (M, cut-off, 3000) for 3 h in three changes of PBS at 4Â°C.Biotinylated, jam) on a cryotome (Zeiss HM505E). and mounted on TEPSA-coated slides. dialyzed CTX-biotin was stored in aliquots in the freezer and used for immu- Sections were fixed in 4% paraformaldehyde for 30 min at room temperature, nohistochemical studies at concentrations ranging between 0.5 and 1 Â¡J.M. rinsed twice in PBS, and then blocked in 0.04% H,O2 in methanol for 10 min Cytochemical and Immiinohistochemical Studies. Cells were stained to inhibit the activity of endogenous peroxidase. After PBS washes, slides were using standard immunohistochemical techniques, as previously described (9). blocked with the avidin-streptavidin blocking reagent (Vector Laboratories, Glioma cells and control cell lines were cultured overnight in eight-well Burlingame) according to the manufacturer's instructions or with normal goat chamber slide trays or on glass coverslips (2 X IO5 cells/ml): semiconfluent serum (Vector Laboratories; 20% in PBS). Slides were then incubated with cultures were fixed in 4% paraformaldehyde for 30-60 min at room temper biotinylated CTX (4 /Â¿g/ml)overnight, washed with PBS (three times, 10 min ature. After fixation, cells were washed with PBS, blocked for l h with PBS each), then incubated (2 h, room temperature) with streptavidin-HRP (Pierce; containing 2% BSA and then incubated overnight with primary antibodies 0.2 mg/ml). After PBS washes (three times, 10 min each), we used a 3,3'- (biotinylated CTX or fluorescent-tagged CTX). Control wells were incubated diaminobenzidine substrate kit (Vector Laboratories) to develop the enzymatic- with blocking buffer alone; for competition experiments, cells were preincu- reaction. Counterstaining was done using Mayer's hematoxylin; slides were bated with a 100-fold molar excess of native CTX for 2 h at room temperature subsequently dehydrated, coverslipped with Aquamount, and viewed under the before immunostaining. Cells were visualized using various secondary agents. light microscope. Control sections were either incubated with secondary anti HRP-linked streptavidin (Pierce) at 0.1 mg/ml in conjunction with 3,3'- body alone or pre-treated with CTX for 30-45 min. washed, and then incu- 4873

Control Tumor 24h 48h

240,220200160160140-120-100-80-60-40-20-

200.180-160.140-120-M0>Ã•6o:60-196h

40-J 201. 0-n m240,220-72h S\\VA\\*\V*

B .â€¢

,\ ^Â«â€¢â€¢¿^-n t "**-. â€¢¿^ t I^L:

Fig. 2. A. hiodislrihulion of 1:'I-labeled CTX al four time points (upper panels} after intracrani.il administration in tumor-hearing and contro! SCID mice. Dala represent mean values from three experiments. The results (expressed as %ID/g tissuel show persistent accumulation of '''l-labclcd CTX within the xenografted tumor tissue in the right hrain of the animals. Stomach accumulation of iodine is also visihle and is most likely a result of the dchalogenaticn ol the radioligand (see also "Results"), /fiirs. SI) ; lesi analysis was performed, and results for the righl-hrain values are shown: for all of the time points. P < 0.001 (Â»*Â».extremelysignificant). B. autoradiograms: coronal brain sections from glioma-hearing mice were obtained at 4K h alter intracranial injection of 15 Â¿Â¿Cii;Ãfil-labeledCTX. The four panels demonstrate selective accumulation of silver grains (arrows) within the human transplanted tumor cells, whereas the host tissue is virtually radioactivity free. Burs. 1.3 mm and 400 /xm (upper left and right photomicrographs, respectively) and 100 u.m (liwer photomicrographs).

4874

b b si * SÃ "i a.b. a.b.

Fig. 3. Tissue distribution of ml-labeled CTX at u. b. H.A.% I r 24 and 72 h after intracranial injection of 50 Â¿tCi I.. radiolabeled toxin. The biodistribution results Control Tumor Control Tumor (mean values from two experiments) demonstrate tumor accumulation of the radiolabeled toxin (low er two panels', bars, SD). Accumulation of ml- 200-SÂ» labelcd CTX in other organs (e.g., the stomach) is 24h^^ Control also noticeable (see comment in "Results"). The upper four panels show representative whole-body TumorI ISO-3t/lVIp.SO mouse scintigrams obtained by gamma ray camera imaging at the indicated time points. Radioactivity is seen as dark areas in the brain (h}. stomach (si). -IM and urinary bladder (a.b.). Note the difference in 100Â«QS? intensity of the brain-accumulated radioactivity be tween the control and tumor-bearing animals.

50-o^B kI ila enKJ, 1m iJ^in-Jl niifllllÃ¬llJilli3LJ-1\ â€¢¿% cnKK.200-150.100-50-, '||||Â¡|||||||| 9 ^i J?_ZlL-ttKlteKl0

*5 || 8-1"- Â¡*tj!-::72hii IJ

bated with biotinylated CTX as described in the "Cytochemical and Immuno- binding of CTX to normal rat astrocytes and Te671, a human rhab- histochemical Studies" section. domyosarcoma cell line. In agreement with our previous electrophysi- Reagents and Chemicals. Native CTX was purchased from Alomone ological characterization, these cells lacked CTX receptor sites, and Laboratories (Jerusalem, Israel). Synthetic CTX was synthesized and high- specific binding of radiolabeled toxin could not be demonstrated (data performance liquid chromatography purified (>98% purity) at the Peptide not shown). We also assessed the ability of glioma cells to internalize Synthesis Core Facility of Louisiana State University. All of the other chem bound CTX. Glioma cells in suspension were treated with I2<"l-labeled icals were purchased from Sigma, unless otherwise noted. CTX at 4Â°Cand 37Â°C,respectively. We counted radioactivity in the Statistical Analysis. Statistical analysis was performed on results from binding assays and I2il-labeled CTX biodistribution experiments, using the membrane-associated pool of the I25l-bound CTX [stripped off with a Instai software program (GraphPad, version 1.1 la, 1990). The unpaired t test glycine buffer (pH = 2.5)]; the membrane-bound fraction was 88.2% was used to obtain two-tailed />s and to establish SDs in the accumulation of of the total specific bound material at 4Â°C.in contrast to only 36.8% radioactive CTX in organs between the tumor-bearing and control animals. of the total cell-associated radioactivity after incubation for 60 min at 37Â°C(data not shown). The remainder of the cell-associated radioac RESULTS tivity was acid resistant and presumably located within the cell. These data suggest that, at physiological temperatures, significant internal In Vitro Binding of CTX. As a first step to characterize the specific CTX binding to glioma cells, we used I25l-labeled CTX and ization of CTX occurs. Radioreceptor SDS-PAGE Analysis. In an effort to begin the measured binding to cultured glioma cell lines. A representative Scatchard plot analysis of 125I-labeled CTX binding to D-54 MG characterization of the receptor molecule for CTX, radioreceptor blot analysis was performed on U-251 MG and D-54 MG cell lines (Fig. human glioma cells is shown in Fig. \A. The data are best fit with two IB). 125I-labeled CTX (25 DM)was bound to cultured glioma cells in lines, indicating the presence of two binding sites; the apparent Kd for the presence (Lanes I and 3) or absence (Lanes 2 and 4) of 100-fold this experiment is 4.3 nM for the high affinity site and 660 nM for the molar excess cold material and cross-linked using the cross-linking low affinity site. From these data, the total number of CTX binding sites were calculated to be 1,300 high affinity and 13,300 low affinity agent, disuccinimidyl suberate (Pierce). Cells were lysed in the pres sites per cell. Because these experiments were carried out at 4Â°C, ence of proteinase inhibitors, and membrane fractions were prepared internalization of the peptide and its receptor was negligible. Binding by serial centrifugation: aliquots containing equivalent amounts of assays using the D-54 MG cell line were repeated four times, and the protein were separated on SDS-PAGE and exposed to X-ray film. 125I-labeled CTX detected a single band in the membrane preparation, results were consistent with the example presented in Fig. \A: KÂ¿of 4.8, 3.9, and 5.1 nM for the high affinity site (mean Ka, 4.5 nM; which was significantly reduced in intensity by coincubation with SD = 0.57) and 590. 550, and 640 nM, respectively, for the low excess unlabeled ligand. Using the Bio-Rad KaleidoscopeÂ® affinity site (mean KÂ¿Â¡,610nM; SD = 49.66). We obtained similar prestained molecular weight marker, we estimated the size of this binding curves for six other human malignant glioma cell lines, band to be Mr â€”¿72.000.Thesedata indicate that human glioma cells namely SK-l-MG, U87-MG, U105-MG, U251-MG, U373-MG, and express a receptor molecule of A/r 72,000 for CTX. The molecular D-54 MGLacZ. Apparent CTX affinities for these cell lines ranged weight of this molecule is similar to that of members of the CLC between 3.8 and 9.8 nM for the high affinity and 490 and 960 nM for family of chloride channels recently cloned from muscle and kidney low affinity binding sites. Using the same assay, we studied the (11). 4875

Fig. 4. Fluorescence photomicrographs ol brain sections from glioma-hearing SCID mouse. The in tensely labeled cells (ii/rmv.v) represent human tumor cells stained with hiotinylated CTX and avidin-rhodamine (li and l)\ or with Oregon green-tagged CTX M and O. demonstrating specific in sint CTX label ing of human glioma cells. Bars. 100 jim (A and O. 200 /xm

In Vivo Biodistribution of I25l-labeled CTX and Brain Autora- we performed autoradiography on the tumor-bearing animals at 24, diography. Because our results demonstrated specific, high-affinity 48, and 72 h after injection of the radiolabeled ligand. Fig. 2B (four binding of CTX to glioma cells in culture, we investigated the binding lower panels) shows examples of autoradiograms obtained at 48 h characteristics of this molecule in vivo, using an animal model of after administration of the radioligand. Silver grains (indicating local human malignant glioma cells stereotactically implanted in SCID ization of 125I-labeled CTX) are found mostly within the implanted mice brains (9). Control animals were injected with saline. Nine days tumor cells and not the host brain cells, supporting the notion that after intracerebral injection both groups (tumor and control mice) radiolabeled CTX can be efficiently targeted to the tumor implants. received 15 /tCi I2sl-labeled CTX (in 6 /j.1 of PBS total volume/ In Vivo Radioimaging with 131I-labeled CTX. To obtain an in animal) intracranially administrated through the same burr hole used vivo estimate of the ability of CTX to localize to the tumor tissue we for tumor-cell or saline injection. Subsequently, the distribution of used '"l-labeled CTX intracranially delivered (50 piCi/mouse) to I25l-labeled CTX was determined by scintillation counting of tissue glioma-hearing and control SCID mice. Because of its higher y and fluid samples. As shown in Fig. 2A (four upper panels), the emittance, "'I affords the opportunity to obtain gamma-ray scinti- accumulation of radioactive CTX in the brain of glioma-hearing mice gram scans in vivo. In Fig. 3 (upper panels) representative examples was approximately 3-fold higher than in the corresponding side of the of whole-body scintigrams obtained from tumor-bearing and control brain in sham-injected animals at all of the time points studied (24-96 animals are shown (dorsal view, head of animal facing up). Accumu h). The presence of radioactivity in the left hemisphere is likely due lation of the radioactive ligand was significantly higher in the brain of to the dispersion of the injected radioligand via the ventricles; in many tumor-bearing animals compared with their age- and gender-matched instances, there was a variable (albeit smaller) amount of tumor mass control. There was some visible accumulation in the stomach (which seen in the left side of the brain, presumably because of ventricular dissemination (seen also in Fig. 5B); ratios of accumulated I25I- is observed routinely as a result of dehalogenation) as well as in the urinary bladder. Tumor-bearing animals were less active and probably labeled CTX between the right and left hemisphere ranged between reduced their food and water intake after injection of the '"[-labeled 1.5 and 3.1. In the tumor-bearing mice, retention of CTX at 24 h after radioactive ligand injection was 190<7rID/g tissue in the right side of CTX, which could account for the persistently higher radioactivity seen within their gastrointestinal track compared with that of control the brain and oSVrlD/g tissue in the left hemisphere. In contrast, in animals. Gamma camera imaging of '"l-labeled CTX was used to age- and gender-matched saline-injected mice, accumulation was 25%ID/g in the right hemisphere and 18%ID/g in the left side of the obtain qualitative information about selective retention of CTX within brain. At subsequent time points, the relative accumulation of I25I- tumor tissue. In addition, biodistribution of "'l-labeled CTX was labeled CTX in the glioma-hearing mice was consistently higher analyzed, and data for the 24- and 72-h time points are shown in Fig. within the right brain (212%ID/g, SD = 24.8 at 48 h; 76%ID/g, 3 (lower panels). In this case, we dissected the tumor separately from SD = 10.1 at 72 h; 46%ID/g, SD = 8.2 at 96 h) compared to the left the right hemisphere (designated "right brain"); we could thus estab hemisphere (64, 28, and 21<7rID/g at 48, 72, and 96 h, respectively). lish that most of the right-brain associated radioactivity was due to the In all of the cases, the differences in the right-brain accumulated accumulation of radiolabeled CTX within the tumor tissue. At 24 h radioactivity between the tumor-bearing and the control groups were after the injection of '3'l-labeled CTX, calculated tumor:muscle ratio significant (two-tailed P < 0.001). of accumulated radioactivity (mean value, 39.13; SD = 4.6) was To gain additional insight into the intratumoral distribution of CTX, significantly higher than the liver: muscle ratio (mean value, 1.12; 4876

Fig. 5. A. Â¿ÃjcZ-transfectedD-54 MG human gli oma cells demonstrate labeling with biotinylated CTX. B, coronal section of SCID mouse brain in duced with D-54 MGlMcZ glioma; mice were sacri ficed at 9 days after tumor injection and processed for LticZ enzyme histochemistry. C-F. double label ing of xenografted human glioma cells: panels C-F illustrate immune staining of LttcZ D-54 cells with biotinylated CTX (arrows), indicating that most of the human glioma cells are CTX-positive after hav ing formed tumors in SCID mice brain, whereas no CTX labeling is seen within the host tissue. Bars. 3 mm(ÃŸ). 1.5 mm (C). 100 pirn (D), 50 /im (Eand Fl. t.

â€¢¿ V

â€¢¿ ,

SD = 0.08; P = 0.072). These values had changed to 11.6 and 1.3, Once implanted, the tumors could be readily visuali/.ed by the blue respectively, 72 h after administration of "'I-labeled CTX. X-Gal label (Fig. 5B). Frozen sections of such tumors were stained Immunohistochemical Localization of CTX. Immunohistochem- with CTX-biotin and showed that most of the transplanted tumor cells ical analysis of CTX binding to glioma cells in culture and human were also positive for CTX. Representative fields of sectioned human glioma xenografts in SCID mice was pursued using two different xenografts in SCID mouse brain are shown in Fig. 5, panels C-F. approaches. Direct labeling of CTX was obtained by conjugation of Specificity of staining was tested in all of the immunohistochemical CTX to the FITC-like fluorochrome, Oregon GreenÂ® (Molecular experiments by pretreating control cells or brain sections with "cold" Probes, Inc.). This compound was tested for its ability to label glioma (untagged) CTX at 100-fold molar excess. These were consistently cells in xenografted tumors in situ (Fig. 4, A and C). Glioma cells are negative (data not shown). readily identified by fluorescent-labeled CTX within the host tissue. To show that CTX is an effective immunohistochemical marker for As a second approach, we used rhodamine-labeled streptavidin to human glioma cells, we also analyzed a variety of human biopsy and recognize biotinylated CTX. Examples that demonstrate selective autopsy tissues obtained from patients with various pathologies. Fig. labeling of glioma cells with this two-step immunohistochemical 6/\ illustrates a representative example of CTX labeling of a human approach are shown in Fig. 4 (panels B and D, again staining of D-54 GBM tissue section; control sections preincubated with native CTX MG cells xenografted intracranially into SCID mice brains). To un (100-fold molar excess) showed no positive peroxidase immuno- equivocally identify the implanted human tumor cells and differenti staining (not shown). Fig. 6C shows that the staining of brain tissue ate them from mouse host tissue, we used D-54 glioma cells that were that was unaffected by malignancy exhibits no significant CTX label transfected with the reponer gene (Lac Z) allowing identification of ing. We have analyzed lower-grade astrocytoma (WHO grade II) transfected cells by the blue X-gal reaction product. Transfected tissues for CTX immunostaining (Fig. 60). The number of CTX- D-54LacZ cells stained positive with biotinylated CTX followed by positive cells was found to increase with the malignancy grade of the HRP-linked streptavidin before injection into SCID mice (Fig. 5/4). tumors. This is in excellent agreement with our previous electrophysi- 4877

'f, Fig. 6. Imniunohistochcmical analysis of CTX ** staining of human brain tissue. Upper Â¡eftpanels demonstrate labeling with biotinylatcd CTX and streptavidin-HRP of GBM (A) and low-grade astrocyloma (C}. Significantly fewer CTX-positive cells Â«t* are found in the nonmalignant brain tissue (E). Righi panels (B, I), and F) represent H&E staining of the same tissues to illustrate characteristic cellularity for each pathology. Kur. 20 firn. A, * .

ological demonstration of enhanced expression of CTX-sensitive necessary during glioma cell migration and the invasion of surround channels with enhanced malignancy grade (7). ing normal brain tissue. It is thus likely that GCC is of functional importance in the malignant transformation of gliomas or in regulat DISCUSSION ing their highly invasive nature. On the basis of our binding studies, CTX specifically binds malig The data presented show specific and selective binding of CTX to nant glioma cells in culture, exhibiting high-affinity (KÂ¿= 4.2 nM, glioma cells in vitro and in vivo. CTX (12, 13) is a 36-amino acid 1300 sites/cell) and low-affinity (KÂ¿= 660 nw, 13,300 sites/cell) peptide that, because of four disulfide bridges, is tightly folded (14). binding sites. Given the high degree of heterogeneity typical of Its small size and compact shape are convenient for intracranial gliomas, it is remarkable that CTX binding and expression of GCC delivery. Other favorable characteristics include its capacity to be were observed across all of the malignancy grades of human gliomas recepted after radioiodination of its accessible tyrosine residues and after conjugation of its free amine group in formation of sulfo-NH- studied. The high affinity for glioma cells with little or no binding to ester bonds (for biotinylation). Because CTX inhibits GCCs in elec- normal brain cells, together with its favorable physical properties, trophysiological studies (5, 6), we assume that it binds directly to makes this molecule an attractive candidate for the targeting of chloride channels. However, it is possible that CTX binds to an gliomas. Feasibility of such an approach is supported by our in vivo independent receptor and indirectly modulates the activity of GCCs. experiments that show selective retention of radiolabeled CTX within xenografted gliomas in SCID mice. Moreover, the imaging of 'â€¢"!- For purposes of targeting glioma cells using CTX, this difference is not relevant. labeled CTX injected intracranially demonstrated specific in vivo The function of the GCC targeted by CTX is not clear. Evidence is accumulation of the radiolabeled toxin in the tumors of the glioma- accumulating that these channels may control glioma cell proliferation bearing mice, thus supporting its potential use for diagnosis or therapy (6), and may facilitate volume and cell-shape changes (8) that are of brain tumors. Additional pharmacokinetics studies after the admin- 4878

Downloaded from cancerres.aacrjournals.org on September 23, 2021. © 1998 American Association for Cancer Research. CTX TARGETING OF HUMAN GLIOMAS istration of radiolabeled CTX intracranially and i.v. will precisely 4. Murata. T., Obiri. N. I.. Debinski. W.. Purik. K. Structure of 1L-I3 receptor: analysis establish its circulatory half-life and whether multiple injections may of subunit composition in cancer and immune cells. Biochem. Biophys. Res. Com mun.. 238: 90-94. 1997. be necessary for persistent retention within the tumor. To date, the 5. Ullrich. N.. Gillespie. G. Y.. and Sontheimer. H. Human astrocytoma cells express a results of our studies using '25I-labeled CTX compare favorably with unique chloride current. Neuroreport. 7: 1020-1024, 1996. those obtained by others (15-17) using monoclonal antitumor anti 6. Ullrich. N.. and Sontheimer. H. Biophysical and pharmacological characteri/.ation of chloride currents in human astrocyloma cells. Am. J. Physiol. (Cell Physiol.). 270: bodies tagged with various radioactive molecules (18). We have not C1511-C1521, 1996. yet obtained complete toxicity data, but preliminary results indicate 7. Ullrich, N.. Bordey. A.. Gillespie. G. Y.. and Sontheimer. H. Expression of voltage- that CTX (40 mg/kg) injected into SCID mice (i.p. or i.v.) had no activated chloride currents in acute slices of human gliomas. Neurosci., K3: 1161- 1173. 1998. toxic effect for 60 days after administration of the drug. 8. Ullrich. N., and Sontheimer. H. Cell cycle dependent expression of a glioma-specific Our immunohistochemical data suggest that CTX-like molecules chloride current: proposed link to cytoskeletal changes. Am. J. Physiol. (Cell with fluorescent or biotin labels can be successfully used for immu Physiol.). 273: C1290-C1297, 1997. nohistochemical labeling of gliomas. The potential correlation be 9. Andreanski, S., Soroceanu, L.. Flotte. E. R., Chou. J.. Marken, J. M.. Gillespie. Y. G., Roizman. B., and Whitley R. J. Evaluation of genetically engineered herpes simplex tween the CTX receptor expression levels and malignancy grade viruses as oncolytic agents for human malignant brain tumors. Cancer Res.. 57: offers an attractive perspective for using CTX in the diagnosis and 1502-1509. 1997. grading of primary human malignant gliomas. We are presently eval 10. Massague. J. Identification of receptors for the type-ÃŸtransforming growth factor. uating a larger sample of human gliomas (WHO grades I-IV) as well Methods EnzymoL 146: 174-195. 1987. 11. Jentsch. T. J.. GÃ¼nther.W., Pusch. M., and Schwappach. B. Properties of voltage- as metastatic brain tumors and nontumor patient biopsy tissues for gated chloride channels of the C1C gene family. J. Physiol.. 4X2 (Suppl. P).- I9S-25S, specific binding of CTX to gather comprehensive data on the relative 1995. expression of CTX receptor and its relationship to malignancy grade. 12. DeBin. J. A.. Maggio, J. E., and Strichartz, G. R. Purification and characterization of chlorotoxin. a chloride channel ligand from Ihe venom of Ihe scorpion. Am. J. In addition to being useful glioma-specific markers, molecules that Physiol.. 264: 361-369, 1993. use CTX as a targeting moiety have the potential to be developed into 13. DeBin. J. A.. Wood. M. R., Pfenninger. K. H.. and Strichartz. G. R. A chloride cell-type-specific cytotoxic drugs. Previous approaches have used channel reconsiituled from felal rat brain growth cones. J. Membr. Biol.. 141: 7-19. 1994. radiolabeling or linkage of monoclonal antibodies to cytotoxic pro 14. Lippens, G.. Najib, J.. Wodak. S. J.. and Tartar. A. NMR sequential assignments and teins for delivery to cancerous cells (15, 19-21). Findings described solution structure of chlorotoxin, a small scorpion toxin that blocks chloride channels. in this report form the basis for our ongoing studies toward linking Biochemistry..Â«: 13-21. 1995. CTX to Saporin officinalis. a ribosomal inactivating toxin (22, 23). 15. Williams. J. A., Wessels. B. W.. Edwards. J. A.. Kopher. K. A.. Wanek, P. M.. Wharam, M. D.. Order. S. E.. and Klein. J. L. Targeting and therapy of human glioma For such approaches to be effective, internalization of the ligand is xenografts in rirÂ«utilizing radiolabeled antibodies. Cancer Res.. .W (Suppl.): 974s- essential. Results presented herein (from experiments using I25I- 979s, 1990. labeled CTX) suggest that >60% of the toxin is ultimately internal 16. Papanastassiou. V.. Pizzer. B. L.. Chandler. C. L.. Zananiri. T.. Kemshead, T.. and Hopkins. K. I.. Pharmacokinetics and dose estimates following intrathecal adminis ized at physiological temperatures, such that the aforementioned tar tration of '"l-monoclonal antibodies for the treatment of centra] system malignan geting strategy seems feasible. cies. Int. J. Radial. Oncol. Biol. Phys.. 31: 541-542. 1995. Given the high affinity of interaction between CTX and GCC, the 17. Capala. J.. Barth. R. F.. Bailey. M. Q.. Fenstermaker. R. A.. Marek. M. J.. and Rhodes. B. A.. Radiolabeling of epidermal growth factor with ""Te and in vivo specificity of binding of CTX-based moieties to human glioma cells, localization following intracerebral injection into normal and glioma-hearing rats. the small size of the CTX molecule, and its potential for chemical and Bioconjugale Chem., A: 289-295, 1997. genetic manipulation, this molecule seems to be an ideal candidate to 18. Raben. D.. Buchshaum. D. J.. Khazaeli. M. B. Rosenfeld. M. E.. Gillespie. Y. G.. engineer additional chimerical reagents to be used as diagnosis tools Grizzle. W. E.. Liu. T.. and Curici. D. T. Enhancement of radiolabeled antibody binding and tumor localization through adenoviral Iransduction of the human carci- and therapeutic modalities. These approaches may offer novel strat noembryonie antigen gene. Gene Ther.. 3: 567-580. 1996. egies in addressing this uniformly fatal disease. 19. Buchshaum. D. J., Nelson. L. Hanna. D.. and Vallera. D. A. Human leukemia cell binding and killing by anti-CD5 radioimmunotoxins. Int. J. RadiÃ¢t.Oncol. Biol. Phys., 13: 1701-1712, 1987. ACKNOWLEDGMENTS 20. Chamberlain. M. C.. Barba. D.. Kormanik. P.. Berson. A. M.. Saunders, W. M.. and Shea. M. C. Concurrent cisplatin therapy and iodine 125 brachytherapy for recurrent We thank Mark Campbell for assistance with the radiolabeling of CTX. malignant brain tumors. Arch. Neurol., 52: 162-167, 1995. Gayle Elliot for help with the use of the gamma ray camera, Katherine Mercer 21. Brown, M. T.. Coleman. R. E., Friedman. A. H., Friedman. H. S.. McLendon. for technical help in immunostaining. and Katherine Nobles and James Rob R. E., Reiman. R. E.. Felsberg, G. J.. Tien. R. D., Bigner, S. H.. Zalulsky. M. R.. inson for technical assistance. Zhao, X. G.. Wikslrand, C. J.. Pegram. C. N.. Herndon, J. E.. II, Vick, N. A., Paleologos. N.. Frederick. R. K.. Schold. S. C.. Jr.. and Bigner. D. D. Intrathecal '"1-labeled antitenascin monoclonal antibody 8IC6 treatment of patients with REFERENCES leptomeningeal neoplasms or primary brain tumor resection cavities with sub- arachnoid communication: Phase 1 trial results. Clin. Cancer Res.. 2: 963-972. 1. Kurpad. S. N.. Zhao. X. G.. Wikslrand. C. J.. Balra. S. K., McLendon. R. E.. and Bigner. D. D. Tumor antigL*ns in astrocytic gliomas. Olia, 15: 244-256, 1995. 1996. 2. Burger. P. C. Revising the World Health Organization (WHO) blue bookâ€”Histolog- 22. Tecce. R.. Nicotra. M. R.. Fraioli, R., Cuomo. M.. Trizio, D.. and Natali. P. G. ical Typing of Tumours of Ihe Central Nervous System. J. Neuro-Oncol.. 24: 3-7, Saporin 6 conjugated to monoclonal antibody selectively kills human melanoma cells. 1995. Melanoma Res.. /: 115-123. 1991. 3. Zhang. H.. Kelly. G., Zerillo, C.. Jaworski. D. M., and Hockfield. S. Expression of a 23. Laske, D. W.. Ilercil, O., Akbasak. A.. Youle. R. J.. and Oldfield. E. H. Efficacy of cleaved brain-specific extracellular matrix protein mediates glioma cell invasion in direct intratumoral therapy with targeted protein toxins for solid human gliomas in vivo. J. Neurosa.. IS: 2370-2376. 1998. nude mice. J. Neurosurg.. 80: 520-526, 1994.

4879

Liliana Soroceanu, Yancey Gillespie, M. B. Khazaeli, et al.

Cancer Res 1998;58:4871-4879.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/58/21/4871

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/58/21/4871. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.