Download - Targeting of the Receptor Protein Tyrosine Phosphatase B ... · MAB-ZAP, IgG-SAP, DAT-SAP, 7A9B4-SAP, and 7E4B11-SAP were purchased (or created in the case of 7A9B4-SAP and 7E4B11-SAP)

Targeting of the Receptor Protein Tyrosine Phosphatase B

with a Monoclonal Antibody Delays Tumor Growth

in a Glioblastoma Model

Erik D. Foehr, Gustavo Lorente, Jane Kuo, Rosie Ram, Karoly Nikolich, and Roman Urfer

AGY Therapeutics, Inc., South San Francisco, California

Abstract

The receptor protein tyrosine phosphatase B (RPTPB) is afunctional biomarker for several solid tumor types. RPTPBexpression is largely restricted to the central nervous systemand overexpressed primarily in astrocytic tumors. RPTPBis known to facilitate tumor cell adhesion and migrationthrough interactions with extracellular matrix componentsand the growth factor pleiotrophin. Here, we show that RPTPBis expressed in a variety of solid tumor types with lowexpression in normal tissue. To assess RPTPB as a potentialtarget for treatment of glioblastoma and other cancers,antibodies directed to RPTPB have been developed andprofiled in vitro and in vivo. The recombinant extracellulardomain of human short RPTPB was used to immunize miceand generate monoclonal antibodies that selectively recognizeRPTPB and bind to the antigen with low nanomolar affinities.Moreover, these antibodies recognized the target on livingtumor cells as measured by flow cytometry. These antibodieskilled glioma cells in vitro when coupled to the cytotoxinsaporin either directly or via a secondary antibody. Finally,in vivo studies showed that an anti-RPTPB immunotoxin(7E4B11-SAP) could significantly delay human U87 gliomatumors in a mouse xenograft model. Unconjugated 7E4B11provides a modest but statistically significant tumor growthdelay when delivered systemically in mice bearing U87 gliomatumors. (Cancer Res 2006; 66(4): 2271-8)

Introduction

There are currently >500 new oncology drugs in clinical trials.This reflects the huge medical need and intense competition forinnovative drugs to novel targets. Cell surface receptors are primetargets of therapy due to the fact that they often play vital rolesin tumor biology and are amenable to both small molecule andmonoclonal antibody (mAb) drugs. Receptors that are overex-pressed or display abnormal signaling often distinguish a tumorcell from normal cells. However, the pharmacologic targeting ofreceptors is often complicated by the reality that some normaltissues express the cancer target and cause unwanted organ uptakeor dangerous side effects. Therefore, the search for a uniquelytumor-specific target goes on. Despite this hurdle, there are severalinstances of successful strategies to target receptor tyrosine kinases(RTK), these include Herceptin (Genentech, South San Francisco,

CA) and Cetuximab (Imclone Systems, New York, NY), antibodiesto the epidermal growth factor receptor (EGFR) family of RTKs (1).Few examples exist of the successful targeting of receptor proteintyrosine phosphatases (RPTP) by either small molecule orantibody-based therapeutics.One such promising RPTP target is RPTPh, a protein that is

overexpressed in astroctyomas (2–4). In astroctyomas, increasedexpression levels of RPTPh correlate with a poor clinical prognosis(3). RPTPh has a shown role in the migration and adhesion oftumor cells, and one of its primary ligands, pleiotrophin, is knownto enhance tumor growth and angiogenesis (2, 5–7). Until recently,very little was known regarding the intracellular signaling ofRPTPh. Reports indicate that pleiotrophin can bind and inactivateRPTPh and thereby lead to increased tyrosine phosphorylation ofh-catenin and Fyn (8, 9). The extracellular domain of RPTPh hasthree characterized domains: a carbonic anhydrase, fibronectin III,and glycine-serine rich region (10–13). These regions differentiallyinteract with various cell adhesion molecules, such as NgCAM,Contactin, and tenascin C, to facilitate cell attachment, repulsion,and signaling (14, 15). Several splice variants of RPTPh have beendescribed; these include the soluble matrix protein Phosphacanand long and short receptor forms (10, 11, 16). The short receptorform lacks the 806-amino-acid, glycine-rich domain in theextracellular portion of the receptor. This truncation produces areceptor with a unique junction and a modified ligand-bindingprofile (4, 17). Both long and short RPTPh are expressed in gliomatumor tissue (4).Antibody targeting of RPTPh is a promising approach for the

treatment of cancer. RPTPh tissue expression is largely centralnervous system (CNS) and developmentally restricted (12, 16, 18). Ithas a shown role in tumor biology and is amenable to therapeutictargeting. The present study describes the generation andcharacterization of murine antibodies to RPTPh. The expressionprofile of RPTPh in both normal and tumor tissue (other thanbrain tumors) is described. We also detail the efficacy of anti-RPTPh antibodies in both cell-based and in vivo tumor studies.

Materials and Methods

Materials. Human U87 (American Type Culture Collection, Manassas,

VA) cells were cultured in DMEM supplemented with 10% fetal bovineserum (Invitrogen, Carlsbad, CA) and 100 units/mL penicillin/streptomycin.

Mouse mAbs against the COOH terminus of RPTPh were obtained from BD

Transduction Labs (Franklin Lakes, NJ). Direct immunotoxin conjugates

MAB-ZAP, IgG-SAP, DAT-SAP, 7A9B4-SAP, and 7E4B11-SAP were purchased(or created in the case of 7A9B4-SAP and 7E4B11-SAP) from Advanced

Targeting Systems (San Diego, CA).

Mouse mAbs. Custom mouse mAbs were generated to the extracellulardomain of recombinant human RPTPh short form. The extracellular

domain for short RPTPh (residues 26-774), including the unique splice

junction, was expressed with a COOH terminus 6� His tag using

Note: Supplementary data for this article are available at Cancer Research Online(http://cancerres.aacrjournals.org/).

Requests for reprints: Erik D. Foehr, AGY Therapeutics, Drug Discovery, 270 EastGrand Avenue, South San Francisco, CA 94080. Phone: 650-228-1173; Fax: 650-615-4544; E-mail: [email protected].

I2006 American Association for Cancer Research.doi:10.1158/0008-5472.CAN-05-1221

www.aacrjournals.org 2271 Cancer Res 2006; 66: (4). February 15, 2006

Research Article

Research. on March 14, 2020. © 2006 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

baculovirus. The protein was purified from the media with twochromatography steps: immobilized metal affinity (Ni2+-NTA FF, Qiagen,

Valencia, CA) followed by anion exchange (Q Sepharose FF, GE Healthcare,

Pittsburgh, PA; ref. 4). Mouse hybridomas were generated by Anaspec (San

Jose, CA) using BALB/c mice with the protein immunogen in Freund’sadjuvant. Spleen cells from immunized mice were fused with a mouse

myeloma cell line. Supernatants were screened by ELISA using the RPTPhprotein immunogen and isotyped, and RPTPh-specific IgG producing

hybridomas were expanded, subcloned, and cultured. Promising hybrid-oma lines were grown, and the antibodies were purified from culture

media using Protein A/G Montage Purification columns (Millipore,

Billerica, MA).

Immunohistochemistry. Tissue MicroArray slides were used to study

the expression RPTPh. Slides were placed on a heat block at 45jC for 4 to

6 hours and dewaxed using EZ-DeWax solution (Biogenex, San Ramon,

CA). Slides were then placed in a bath containing Target Retrieval Solution

(Innogenex) and simmered for 15 minutes in a microwave. Slides were

stained with either the COOH terminus RPTPh antibody (Transduction

Labs) or the custom made RPTPh mouse mAbs (Anaspec). The slides

were processed using anti-mouse immunohistochemistry kits with 3,3V-diaminobenzidine colorimetric end-point detection and hematoxylin

counterstain (Innogenex).

ELISA. To determine the specificity of antibody containing hybridoma

supernatants and purified antibody preparations, 96-well Maxisorp plates

(Nalgene, Rochester, NY) were coated in 0.5 mmol/L sodium bicarbonatesolution containing 1 Ag/mL antigen for 1 hour at room temperature. The

plates were then washed with PBS and blocked with bovine serum albumin

for 1 hour. Antibodies were added to triplicate wells for 1 hour and then

washed thrice with PBS containing 0.1% Tween 20 to remove unbound ornonspecific proteins. The anti-mouse IgG-horseradish peroxidase detection

antibody was then added for 1 hour before the final washes in PBS

containing 0.1% Tween 20. The immune complex was incubated briefly with

TMB substrate (Sigma-Aldrich, St. Louis, MO) and stopped with 0.1 N HCl,and absorbance was detected at 490 nm on a plate reader. Triplicate

experiments were done, and error bars represent SDs.

Affinity measurements/surface plasmon resonance. Biacore uses asensor chip technology for monitoring interaction between two or more

molecules in real time, without the use of labels. These studies were

undertaken by Biacore contract services. The antibodies were captured on

the sensor via anti-mouse IgG preimmobilized on the chip surface. Therunning buffer was HBSEP [50 mmol/L HEPES (pH 7.4), 150 mmol/L NaCl,

3 mmol/L EDTA, and 0.005% Surfactant P-20], and the analysis temperature

was 25jC. The recombinant human short RPTPh was injected, using an

automated method, up to 10 minutes at flow rates ranging from 10 to50 AL/min. Binding data was fit to a 1:1 binding model using Biacore

software (BIAevaluation v 4.1) to obtain the kinetic and affinity constants.

Antibody-mediated internalization assay. U87 cells were plated at3,000 per well (30,000/mL) onto black-walled 96-well plates (BD, Franklin

Lakes, NJ) and incubated overnight at 37jC in 5% CO2. The cells were

treated with 200 ng/well primary antibody prepared in Opti-MEM (Life

Technologies). The antibody treatments include the RPTPh ectodomainantibodies and control IgG. The cells were processed and stained with anti-

RPTPh endodomain mAb (Transduction Labs) and detected with anti-

mouse Alexa 488 secondary antibody. The cells were imaged using an

ArrayScan platform using Zeiss optics and Cellomics imaging algorithms.The punctuate nature of the staining indicates receptor-antibody internal-

ization. Images representative of triplicate experiments are presented.

Colony formation in soft agar. To test the effect of 7E4B11 on colonyformation, we used a cell transformation detection assay kit from

Chemicon (Temecula, CA). Briefly, 24-well plates were prepared with 0.8%

base agar in cell culturemedia (DMEM, 10% fetal bovine serum, 1% penicillin/

streptomycin). U87 cells were harvested and mixed at 4,000 per well in 0.4%top agar solution in cell culture media. Medium containing 20 Ag/mL IgG1

(Cymbus Biotechnology, Eastleigh, NH), 20 Ag/mL EGFR-528 (Santa Cruz

Biotechnology, Santa Cruz, CA; ref. 19), or 20 Ag/mL 7E4B11 was added in

triplicate to the appropriate wells. The cells were incubated for 21 days at37jC in 5% CO2 with medium treatment changes every 3 to 4 days. At that

time, colonies were stained and visualized using light microscopy. Triplicateexperiments were done, and representative images were shown.

Immunotoxin-directed cell cytotoxicity. To evaluate if our RPTPhmouse antibodies can act as an immunotoxin, we tested indirect and direct

immunotoxin-mediated cell cytotoxicity assays. For indirect immunotoxinassay, U87 cells were plated at 3,000 per well (30,000/mL) onto white-walled,

96-well plates (BD Falcon) and incubated overnight at 37jC in 5% CO2. The

cells were treated with 200 ng/well primary antibody prepared in Opti-MEM

(Life Technologies). The antibody treatments include the RPTPh antibodiesas well as EGFR-528 (Santa Cruz Biotechnology; ref. 19), CD71 (Transduc-

tion Labs), and IgG1 (Cymbus Biotechnology). Half of the test wells ( four of

eight) were subsequently treated with saporin-conjugated secondary

antibody (MAB-ZAP) at 200 ng/well. The cells were then incubated for 3days 37jC in 5% CO2. At that time, the number of viable cells was assessed

using the luminescence-based detection reagent Cell Titer Glo (Promega,

Madison, WI) and read on a luminometer. For direct immunotoxin-mediated cell cytotoxicity assay, U87 cells were prepared the same as above

but treated directly with the saporin toxin–conjugated antibodies IgG-SAP,

DAT-SAP, 7A9B4-SAP, and 7E4B11-SAP (Advanced Targeting Systems).

Triplicate experiments were done, and error bars represent SDs.Tumor xenografts. Female athymic nude mice were 11 to 12 weeks old

on day 1 of the study. The U87 glioblastoma line used for this study was

maintained in athymic nude mice. A tumor fragment (1 mm3) was

implanted s.c. into the right flank of each test mouse. Tumors weremonitored twice weekly and then daily as their volumes approached 120 to

160 mm3, at which time the treatment began. IgG-saporin and 7E4B11-

saporin doses were prepared on each day of dosing by dilution with saline.Mice were sorted into eight groups with 10 mice per group. The antibody-

treated groups received 15 and 30 Ag/mouse, 2�/wk � 3 i.t. Each animal

was euthanized when its neoplasm reached the predetermined end point

size (1,5000 mm3). The log-rank test was employed to analyze the significantdifferences of TTE. These studies were done by Peidmont Research Center

in compliance with the Guide for Care and Use of Laboratory Animals and

accredited by Association for Assessmentand Accreditation of Laboratory

Animal Care International.

Results

RPTPB antigen expression and immunization. To provide thebest source material for therapeutic antibody development, wesubcloned the entire extracellular domain of short RPTPh intobaculovirus expression vector. This construct has a polyhistidinetag introduced into the COOH terminus of the extracellulardomain. The protein expressed from the baculovirus constructswas purified using conventional Nickel affinity column and otheravailable methods to obtain the highest purity proteins (Fig. 1A).The short RPTPh antigen was tested for the ability to bindpleiotrophin and modulate cell adhesion (data not shown; ref. 4).Short RPTPh protein has a domain structure very similar toPhosphacan, an extracellular matrix proteoglycan with anti-adhesive properties. In fact, short RPTPh and Phosphacan havecomparable anti-adhesive abilities when tested in cell-based assays(4). These studies indicate that the short RPTPh antigen used forimmunization has retained structural and functional character-istics that may enhance the chances of generating functionalantibodies.Specificity and binding of RPTPB antibodies. Antibodies to

RPTPh were screened by ELISA to determine specificity. In theELISA experiment, a panel of purified antibodies was tested forbinding to recombinant short RPTPh extracellular domain proteinantigen (short RPTPh-ECD, black columns) or a unrelated butsimilarly expressed and purified control antigen, (NR-ECD, graycolumns). The control protein used for this experiment is calledGPR56 (20) and is a similarly glycosylated protein produced

Cancer Research

Cancer Res 2006; 66: (4). February 15, 2006 2272 www.aacrjournals.org



recombinately in a fashion nearly identical to RPTPh antigen. Inthis example, four of the five RPTPh antibodies tested gave a robustand specific signal for short RPTPh-ECD and did not recognizeGPR56 (Fig. 1B). In addition, we tested the ability of the antibodiesto cross react with rat chondroitin sulfate proteoglycan mix thatincludes Phosphacan. Phosphacan is essentially the extracellulardomain of RPTPh. Our lead antibodies did not significantly crossreact with rat Phosphacan, indicating species selectivity (datanot shown). The hybridomas corresponding to three of the fourpositive antibodies were selected and used for further study.

To measure binding of RPTPh antibodies to native epitopes, weused flow cytometry. Live human U87 glioma cells were incubatedwith the three lead candidate antibodies and detected using anti-mouse IgG-Alexa 488. Live cells were gated from dead cells (stainedwith propidium iodide) and measured for RPTPh-specific signalintensity (Fig. 1C). An IgG1 isotype control (top) was comparedwith the three lead candidate RPTPh antibodies 1B9G4, 7A9B5,and 7E4B11. All three lead candidate RPTPh antibodies bind nativeepitope with a similar profile as indicated by the shift influorescence intensity peak.

Figure 1. Antigen production and antibodyselectivity. A, the protein encoding theextracellular domain of short RPTPhwas purified from the media with twochromatography steps. Immobilized metalaffinity (Ni2+-NTA FF) followed by anionexchange (Q Sepharose FF). Arrow,recombinant human short RPTPh in itspurified form. This material was used toimmunize mice and for screening purposes.B, a subset of RPTPh antibodies generatedfrom the immunization of mice with shortRPTPh were tested by ELISA for selectivityto either recombinant short RPTPh(black columns ) or a nonspecific controlprotein (gray columns ). An IgG1-negativecontrol isotype was included as a reference.Bars, SD. C, flow cytometry analysis ofRPTPh antibody binding to live U87 gliomacells. Human U87 glioma cells were stainedwith control IgG1, 1B9G4, 7A9B5, or7E4B11 as indicated and detected with afluorescent secondary antibody. Live cellswere gated from dead cells and thefluorescent intensity of the cell populationmeasured using flow cytometry. D,immunohistochemistry analysis of normaland tumor tissue for RPTPh expression.Normal colon (i) had modest expressionin some duct cells, whereas colonadenocarcinoma (ii ) was positive forRPTPh. Normal breast (iii ) had lowexpression in some duct cells, whereasinvasive ductal carcinoma of the breaststained intensely for RPTPh. Normal lung(v ) did not stain for RPTPh, but lungadenocarcinoma displayed strong RPTPhimmunoreactivity. Normal skin (vii ) did notstain with RPTPh, but melanoma (viii) wasimmunoreactive. Representative of thepattern of staining multiple tissue sectionsusing COOH terminus–specific RPTPhantibody (Transduction Labs).

Targeting RPTPb with mAbs




To determine the affinity of each lead candidate antibody for theRPTPh antigen, we undertook surface plasmon resonance meas-urements using BIAcore technology. Each of the antigens andantibodies were initially tested to determine the binding specificity,antibody capture levels, and chip surface regeneration. Once theproper binding conditions were achieved, a concentration series ofeach antigen was injected over their respective antibody surfacesusing an automated method. Regeneration solution was used toremove the antigen from the amine coupled antibody surface orthe antibody/antigen from the captured surface. Binding data werefit to a 1:1 binding model using BIAcore software (BIAevaluationv4.1) to obtain the kinetic and affinity constants (Table 1). All threelead candidate antibodies had similar low nmol/L affinities(KD = f8 nmol/L).RPTPB expression in normal and tumor tissue. There are a

few published reports of RPTPh overexpression in humancarcinomas, including astrocytoma and melanoma, but little elseis known about RPTPh expression in peripheral tissues and tumors(21, 22). We and others have studied the expression of RPTPh inhigh-grade astrocytomas (2–4). In our current work, we examinedthe expression of RPTPh in various tumor types (Table 2; Fig. 1D).Using an antibody to the COOH terminus of RPTPh, we surveyedseveral tumor types that revealed RPTPh overexpression (Table 2).Consistent with published reports, RPTPh is overexpressed inmelanomas (21). Lung, colon, breast, and prostate carcinomasalso stood out as having a significant number of tumor sampleswith RPTPh overexpression versus control tissue. Other tumortypes, such as transitional carcinoma of the bladder and squa-mous cell carcinomas of the oral cavity and pharynx, were note-worthy. In addition, lymphoma also indicated up-regulation ofRPTPh. A representative subset of the abovementioned cancersis presented in Fig. 1D , with matched normal controls. RPTPhis expressed most consistently in colon, breast, lung, and skincancer. In some of these tissues, subsets of cells in the corre-sponding normal tissue (i.e., duct/mucosa cells) stained positive forRPTPh.RPTPB immunotoxin-mediated cell cytotoxicity. The critical

criteria for selecting a lead antibody are the ability to bindnative epitope, become internalized, and kill tumor cells. To showantibody-mediated receptor internalization, U87 cells weretreated with anti-RPTPh ectodomain antibodies, and theinternalization of the receptor-antibody complex was measured(Fig. 2A). U87 cells treated with anti-RPTPh antibody but notcontrol IgG treated show a punctate staining pattern indicative ofinternalization.To show immunotoxin-mediated cell killing, glioma cells are

treated with RPTPh antibodies as well as positive controlantibodies (anti-EGFR-528 and anti-CD71), with known ability tobind, internalize, and kill target cells. The negative controlsincluded an isotype control antibody and media alone. Followingprimary antibody treatment, the cells are incubated with a saporin

toxin–conjugated anti-mouse IgG antibody (MAB-ZAP). If theprimary antibody recognizes its target and gets internalized, thenthe toxin-antibody complex is delivered and kills the cells. Figure2B shows that the lead RPTPh antibodies (1B9G4, 7A9B5, and7E4B11), as well as the positive control antibodies, effectively killtumor cells when coupled to the anti-mouse IgG immunotoxin. Theisotype control antibody was unable to bind and internalize andthus could not kill these cells. Because 7A9B5 and 7E4B11 seemedto work slightly better than 1B9G4 and their correspondinghybridoma lines are more productive, they were selected for furtherevaluation.Purified RPTPh-7E4B11 and RPTPh-7A9B5 antibodies were

directly conjugated to saporin and evaluated in cell culture for theability to kill glioma cells. In this experiment, glioma cellsare treated with these RPTPh immunotoxins as well as controlimmunotoxins. The negative controls included an IgG controlantibody (Neg. Ctrl) or media alone (Vehicle). The positivecontrol DAT-SAP antibody (Pos. Ctrl) targets the dopaminetransporter expressed on astrocytoma cells. If the immunotoxinrecognizes its target and gets internalized, then the toxin payloadis delivered and kills the cells. Figure 2C shows that theimmunotoxins 7A9B5-SAP and 7E4B11-SAP, as well as the positivecontrol antibody, effectively kill tumor cells when directly coupledto the toxin. The isotype control antibody was unable to bind andinternalize and thus could not kill these cells.RPTPh has been shown to play a role in tumor cell invasion,

migration, and adhesion (2). However, to date, it is unclear ifRPTPh activity promotes tumor growth. Using short interferingRNA to knock down RPTPh, we did not see an effect on U87cell proliferation (data not shown). In addition, studies usingpleiotrophin, one of RPTPh’s ligands, failed to show a clearenhancement of cell proliferation (data not shown). In contrast,pleiotrophin does enhance colony formation in soft agar assays (5),supporting the notion that RPTPh may play a role in tumor growthin complex cellular systems. To better understand the role ofRPTPh in tumor cell growth, we tested whether the lead RPTPhantibody, 7E4B11, could function as an unconjugated antibody to

Table 1. Affinity of RPTPh antibodies

Anti-RPTPh Ab ka (M�1 s�1) kd (s�1) KD (nM)

1B9G4 8.2 � 104 6.7 � 10�4 8.27A9B5 7.6 � 104 6.4 � 10�4 8.5

7E4B11 6.9 � 104 6.5 � 10�4 9.4

Table 2. RPTPh expression in tumor tissue

Cancer tissue Histology Positive tumors

Breast Adenocarcinoma 6 of 15 (40%)

Ovary Cystadenocarcinoma 6 of 10 (60%)

Endometrium Adenocarcinoma 2 of 7 (28%)Stomach Adenocarcinoma 4 of 5 (80%)

Colon Adenocarcinoma 10 of 11 (91%)

Pancreas Adenocarcinoma 0 of 9 (0%)Liver Hepatocarcinoma 3 of 6 (50%)

Renal/pelvis Transitional carcinoma 4 of 8 (50%)

Kidney Renal carcinoma 8 of 15 (53%)

Bladder Transitional carcinoma 14 of 20 (70%)Prostate Adenocarcinoma 6 of 14 (43%)

Skin Melanoma 3 of 5 (60%)

Esophagus Adenocarcinoma 5 of 6 (83%)

Lip/tongue/mouth Squamous 23 of 28 (82%)Paratoid Mixed tumor 4 of 7 (57%)

Larynx Squamous 5 of 5 (100%)

Pharynx Squamous 6 of 6 (100%)

Lymph node Lymphoma 4 of 7 (57%)Lung Squamous/adenocarcinoma 2 of 10 (20%)

Cancer Research




Figure 2. RPTPh antibody-mediated internalizationand tumor cell killing. A, antibody-mediatedinternalization was shown by treating U87 glioma cellswith control IgG (i) or RPTPh ectodomain antibodies(ii ). The cells were then processed and stained forRPTPh using an antibody specific to the endodomainof RPTPh and detected with a fluorescence-conjugated secondary antibody. The punctate stainingindicates internalized receptor antibody complex.B, for indirect immunotoxin experiments, U87 gliomacells are treated with RPTPh antibodies as well aspositive control antibodies, anti-EGFR-528 andanti-CD71. The negative controls included mediaalone and an isotype control IgG1. Following primaryantibody treatment, the cells are subsequentlyincubated with media (black columns ) or a saporintoxin–conjugated anti-mouse antibody (gray columns ).Bars, SD. C, for direct immunotoxin experiments,purified 7E4B11 and 7A9B5 antibodies were directlyconjugated to saporin and evaluated in cell culture forthe ability to kill glioma cells. In this experiment, gliomacells are treated with 7E4B11 and 7A9B5immunotoxins along with control immunotoxins,nonspecific IgG-SAP (Neg. Ctrl ), DAT-SAP(Pos. Ctrl ), and vehicle. Bars, SD. D, U87 cellssuspended in soft agar were treated with mediacontaining IgG1 (20 Ag/mL), 7E4B11 (20 Ag/mL), orEGFR-528 (20 Ag/mL). The cells were incubated for21 days with medium treatment changes every 3 to4 days. At that time, colonies were stained and imagedusing light microscopy as shown here.





inhibit tumor cell growth. The 7E4B11 antibody (and other RPTPhantibodies) failed to inhibit U87 cell proliferation when simplygrown in standard culture conditions over a number of days (datanot shown). We then set out to test if 7E4B11 could inhibit orprevent colony formation in soft agar. U87 cells were seeded intosoft agar, and then the cells were grown in the presence of test andcontrol antibodies (Fig. 2D). Control IgG1 antibody did not inhibitcolony formation in soft agar. In contrast, the positive controlEGFR antibody and 7E4B11 had fewer and smaller colonies.Therefore, 7E4B11 is able to modulate RPTPh function and therebyinhibit U87 tumor formation in soft agar.In vivo efficacy of 7E4B11 immunotoxin. As part of the in vivo

proof-of-concept studies for RPTPh targeted immunotoxin, weundertook U87 tumor growth delay studies using the 7E4B11antibody. These studies are designed to determine the ability of7E4B11 to target and delay tumor growth and are summarized inTable 3. The immunotoxin consists of 7E4B11 directly linked tosaporin (7E4B11-SAP).U87 tumors were engrafted in Matrigel s.c., and treatments

began when the tumors reached 140 mm3. The activities of 7E4B11-SAP were compared with that of vehicle (PBS) and nonspecificIgG1-SAP delivered i.t. Mice were euthanized when tumors reachedthe end-point volume (1,500 mm3). Nontargeted IgG1-SAP treat-ments at 15 Ag/dose produced only 2% tumor growth delay (TGD).However, at 30 Ag/dose, IgG1-SAP produced 24% TGD. At 15 and 30Ag/dose, 7E4B11-SAP produced 25% and 73% TGD, respectively,and were highly significant (P < 0.001; Fig. 3A). All tumors reachedthe end-point volume by day 44. The median time to end point(TTE) of PBS-treated mice was 18.6 days. Whereas the 7E4B11-SAP(30 Ag/dose) treated mice had a TTE of 32.1 days (Fig. 3B). Theseresults show that 7E4B11-SAP has significant antitumor activity.We then set out to determine the antitumor activity of

unconjugated 7E4B11 in the human U87 glioblastoma xenograftmodel. Nude mice received i.p. injections of 7E4B11, control IgG,or PBS twice weekly for 2 weeks. All treatments began on day 1 ingroups of 10 mice bearing well-established (130 mm3) U87glioblastomas; i.p. 7E4B11 (20 Ag/dose) treatment produced 27%tumor growth delay (P < 0.05) relative to the vehicle, whereas theIgG treatment group did not show statistically significantantitumor activity. In addition, 7E4B11 treatment yielded two60-day survivors with partial regression responses. The 7E4B11unconjugated antibody was well tolerated, and no toxic deathswere recorded (data not shown).The cross-reactivity of 7E4B11 to peripheral tissue, especially

those tissues involved in absorption, secretion, and toxicity, isimportant to establish. Especially, if an antibody toxin conjugateis to be generated. To this end, we studied, colon, kidney, liver,

small intestine, and stomach tissues for cross-reaction to 7E4B11(Fig. 3C). No cross-reactivity was seen in these tissues. Theastrocytoma-positive control displays extensive staining, whereasthe isotype-negative control IgG did not stain.

Discussion

Astrocytomas are the most common and often deadly form ofbrain tumor. RPTPh expression in astrocytomas has been describedand correlates with increased malignancy (3). One approach for thetreatment of astrocytoma is the delivery of therapeutic antibodiesi.c. This avoids the difficulties in penetrating the blood-brain barrierand may be an effective way to target and destroy tumor cells thatremain after surgical resection. RPTPh has many hallmark featuresof a functional tumor biomarker. RPTPh seems to be very CNSspecific. This suggests the opportunity to target peripheral tumorsfor destruction while avoiding toxicity to normal tissue. In the caseof astrocytoma/glioblastoma, RPTPh is overexpressed and could bean excellent biomarker and therapeutic target. Others and we haveshown the overexpression of both long and short RPTPh isoformsin astrocytoma using functional genomics and immunohistochem-istry (2–4). The role of short and long RPTPh in cell adhesion andmigration suggests an important function in tumor cell biology (4).Therefore, the development of RPTPh antibodies may prove usefulfor the treatment of astrocytoma and other cancers. RPTPh seemsto be nonessential to normal tissues, as shown by the lack ofphenotype of the knockout mouse (23). These mice are normalexcept for sensitivity to EAE (murine multiple sclerosis model) andhave impaired remyelination following insult (24). In contrast,tumor cells seem to require RPTPh for adhesion, migration,and growth (2, 4, 25). These observations would suggest thatRPTPh can be safely targeted by therapeutic antibodies.Because both long and short RPTPh are expressed in astrocyto-

mas and facilitate tumor cell adhesion and migration, we developedantibodies to the ectodomain of RPTPh and characterized them forbinding and efficacy in tumor cell killing. The entire ectodomain ofshort RPTPh was used to immunize mice and generate antibodies.Therefore, a better characterization of the antibody-binding sitesand understanding of ligand-receptor interactions will be under-taken. We show that 7E4B11 and several other RPTPh antibodieshave low nanomolar affinity. This seems to be adequate for in vivoefficacy. However, higher-affinity interactions could facilitateincreased uptake of immunotoxin or enhance the presumed effectson signal transduction (altered adhesion, invasion, and growth).The RPTPh antibodies described here are selective, have

reasonable affinities, bind native surface RPTPh, and get internal-ized (either actively or passively). RPTPh is probably continuously

Table 3. Treatment response summary for 7E4B11-SAP study

Agent Route Schedule Median TTE % TGD Max % BW loss No. TR

Vehicle i.t. 2�/wk � 3 18.6 0 0 0

Gemcitabine 160 mg/kg i.p. Q3d � 4 33.2 78 6; day 9 1

IgG-SAP 30 Ag/mouse i.t. 2�/wk � 3 23 24 8.2; day 16 07E4B11-SAP 30 Ag/mouse i.t. 2�/wk � 3 32.1 73 13.7; day 9 1

Abbreviations: BW, body weight; TR, treatment-related death.

Cancer Research




recycled (much like other receptors) and thereby could passivelydeliver the immunotoxin, alternatively, antibodies to RPTPh mayactively facilitate internalization. Active internalization could resultfrom the formation of specific receptor conformations resulting

from antibody-receptor interactions. Interestingly, RPTPhnormally has constitutive phosphatase activity, and when it bindspleiotrophin, RPTPh becomes inactive (8). Therefore, a receptorconformation that is in the off state (lacks phosphatase activity)

Figure 3. A, effect of 7E4B11-SAP on tumorgrowth delay in established U87 xenografts. Micereceived i.t. injections twice a week for 2 weeksonce tumors had reached a mean tumor volume of130 mm3. The immunotoxins 7E4B11-SAP (30 Ag/dose) and IgG-SAP (30 Ag/dose) were testedalongside the vehicle control (PBS) for tumorgrowth delay of established tumors. Points, meantumor volume versus time; bars, SD. B, points,mean time to end point (growth of tumor to1,500 mm3). C, examination of normal humantissue for 7E4B11 reactivity: (i ) astrocytoma-positive control, (ii ) colon mucosa, (iii ) colonmuscularis mucosa, (iv ) kidney cortex renaltubular, (v ) kidney cortex glomerulus, (vi ) kidneymedulla renal tubular, (vii ) liver hepatocytes, (viii )liver portal area, (ix ) small intestine myentericplexus and musculars propria, (x ) small intestinevilli, (xi) stomach muscularis propria, (xii ) stomachmuscosa, (xiii ) astrocytoma IgG1-negative control.Magnification, �40. Representative of replicatetissue samples from different subjects, in astudy conducted by LifeSpan. Additional normalhuman tissue is also presented in SupplementaryFig. S1.





may be the one rapidly internalized. This is in contrast to themodels of antibody-mediated internalization suggested for recep-tor tyrosine kinases. Some EGFR antibodies bind and activate thisreceptor and thereby trigger internalization (1).RPTPh seems to be important for maximum tumor growth in

situations of contact independence or in complex cellular environ-ments. RPTPh binds to matrix proteins, such as NCAMs andtenascin (15–17). These interaction may be necessary for tumorgrowth or survival (resistance to anoikis) in in vivo settings. Thecomplex nature of RPTPh biology is consistent with the studies ofpleiotrophin and the protumor role it plays. Wherein, pleiotrophinpromotes colony formation and in vivo tumor growth but lacksdirect cell proliferative activity in cell culture models (5, 7). In ourhand, the RPTPh antibody-7E4B11 seems to slow tumor growth insoft agar and tumor growth in vivo .The ability of RPTPh antibodies to target and kill U87 tumors

provides proof of concept for the development of better RPTPhimmunotoxins. Currently, several small molecular peptide andprotein toxins are being evaluated for efficacy as antibodyconjugates (26, 27). Ideally, the antibody-toxin conjugate shouldbe selective to tumor cells and have a linker that is cleaved insidethe cell and a toxin that is potent. These characteristics ensure thatonly tumors get killed and normal tissues are spared. Saporin is aribosomal inhibiting protein toxin and, although useful for research

studies, has not proved itself for clinical application, due to its sizeand immunogenecity (28). Immunotoxins must not be immuno-genic (as is often the case for protein toxins) and should retain thedesired characteristics of the antibody (stability, affinity, selectivity,and function). Systemic delivery of an immunotoxin is risky if theantibody is not sufficiently targeted or if the toxin has undesiredside effects. A ‘‘naked,’’ unconjugated antibody is ideal, but manyof the same hurdles still need to be overcome (29). The nakedantibody must be highly tumor selective and able to modulate itstarget such that tumor cells die and normal cells are not affected.In many cases, this can be a ‘‘tall order’’ and unwanted side effectsdue to dysregulation of signaling pathways can be just asdevastating as toxin-mediated side effects. This is because manyantibody targets play important roles in normal cellular functions.In many ways, RPTPh is an ideal target (tumor-selectiveexpression, functional relevance) and is nonessential for normaltissues. The further study of RPTPh and its antibodies mayfacilitate development of therapeutics for the treatment of cancer.

Acknowledgments

Received 4/7/2005; revised 11/23/2005; accepted 12/9/2005.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.

References1. Grumwald V, Hidalgo M. Developing inhibitors of theepidermal growth factor receptor for cancer treatment.J Natl Cancer Inst 2003;95:851–67.

2. Mueller S, Kunkel P, Lamszus K, et al. A role forreceptor tyrosine phosphatase zeta in glioma cellmigration. Oncogene 2003;22:6661–8.

3. Ulbricht U, Brockmann MA, Aiger A, et al. Expressionand function of the receptor protein tyrosine phospha-tase zeta and its ligand pleiotrophin in human astrocy-tomas. J Neuropathol Exp Neurol 2003;62:1265–75.

4. Lorente G, Nelson A, Mueller S, et al. Functionalcomparison of RPTPh long and short receptor spliceforms: implications for glioblastoma treatment. Neuro-oncol 2005;2:154–63.

5. Chauhan AK, Li YS, Deuel TF. Pleiotrophin transformsNIH3T3 cells and induces tumors in nude mice. ProcNatl Acad Sci U S A 1993;90:679–82.

6. Maeda N, Nishiwaki T, Shintani T, Hamanaka H, NodaM. 6B4 proteoglycan/phosphacan, an extracellularvariant of receptor-like protein-tyrosine phosphatasezeta/RPTPbeta, binds pleiotrophin/heparin-bindinggrowth-associated molecule (HB-GAM). J Biol Chem1996;271:21446–52.

7. Grzelinski M, Bader N, Czubayko F, Aigner A.Ribozyme-targeting reveals the rate-limiting role ofpleiotrophin in glioblastoma. Int J Cancer 2005;117:942–51.

8. Meng K, Rodriguez-Pena A, Dimitrov T, et al.Pleiotrophin signals increased tyrosine phosphorylationof beta-catenin through inactivation of the intrinsiccatalytic activity of the receptor-type protein tyrosinephosphatase beta/zeta. Proc Natl Acad Sci U S A 2000;97:2603–8.

9. Pariser H, Ezquerra L, Herradon G, Perez-Pinera P,Deuel TF. Fyn is a downstream target of the pleio-trophin/receptor protein tyrosine phosphatase beta/zeta-signaling pathway: regulation of tyrosine phos-phorylation of Fyn by pleiotrophin. Biochem BiophysRes Commun 2005;332:664–9.

10. Krueger NX, Saito HA. Human transmembraneprotein-tyrosine-phosphatase, PTP zeta, is expressed inbrain and has an N-terminal receptor domain homol-ogous to carbonic anhydrases. Proc Natl Acad Sci U S A1992;89:7417–21.

11. Levy JB, Canoll PD, Schelessinger J. The cloning of areceptor-type protein tyrosine phosphatase expressedin the central nervous system. J Biol Chem 1993;268:10573–81.

12. Canoll PD, Petanceska S, Schlessinger J, MusacchioJM. Three forms of RPTPh are differentially expressedduring gliogenesis in the developing rat brain andduring glial cell differentiation in culture. J Neurosci Res1996;44:199–215.

13. Nishiwaki T, Maeda N, Noda M. Characterization anddevelopmental regulation of proteoglycan-type proteintyrosine phosphatase zeta/RPTPbeta isoforms. J Bio-chem 1998;123:458–67.

14. Grumet M, Milev P, Sakurai T, et al. Interactions withtenascin and differential effects on cell adhesion ofneurocan and phosphacan, two major chondroitinsulfate proteoglycans of nervous tissue. J Biol Chem1994;269:12142–6.

15. Milev P, Friedlander DR, Sakurai T, et al. Interactionsof the chondroitin sulfate proteoglycan phosphacan, theextracellular domain of a receptor-type protein tyrosinephosphatase, with neurons, glia, and neural celladhesion molecules. J Cell Biol 1994;127:1703–15.

16. Barnea G, Grumet M, Milev P, et al. Receptor tyrosinephosphatase h is expressed in the form of proteoglycanand binds to the extracellular matrix protein tenascin.J Biol Chem 1994;269:14349–52.

17. Adamsky K, Schilling J, Garwood J, Faissner A, PelesE. Glial tumor cell adhesion is mediated by binding of theFNIII domain of receptor protein tyrosine phosphatasebeta (RPTPh) to tenascin C. Oncogene 2001;20:609–18.

18. Hayashi N, Oohira A, Miyata S. Synaptic localizationof receptor-type protein tyrosine phosphatase zeta/betain the cerebral and hippocampal neurons of adult rats.Brain Res 2005;1050:163–9.

19. Masui H, Kawamoto T, Sato JD, Wolf B, Sato G,Mendelsohn J. Growth inhibition of human tumor cellsin athymic mice by anti-epidermal growth factorreceptor monoclonal antibodies. Cancer Res 1984;44:1002–7.

20. Shashidhar S, Lorente G, Nagavarapu U, et al. GPR56is a GPCR that is overexpressed in gliomas andfunctions in tumor cell adhesion. Oncogene 2005;24:1673–82.

21. Normann SA, Golfino JG, Scheck AC. Expression of areceptor protein tyrosine phosphatase in human glialtumors. J Neurooncol 1998;36:209–17.

22. Goldmann T, Otto F, Vollmer E. A receptor-typeprotein tyrosine phosphatase PTPzeta is expressed inhuman cutaneous melanomas. Folia Histochem Cyto-biol 2000;38:19–20.

23. Harroch S, Palmer M, Rosenbluth J, et al. No obviousabnormality in mice deficient in receptor proteintyrosine phosphatase h. Mol Cell Biol 2000;20:7706–15.

24. Harroch S, Furtado GC, Brueck W, et al. A critical rolefor the protein tyrosine phosphatase receptor type Z infunctional recovery from demyelinating lesions. NatGenet 2002;32:411–4.

25. Sakurai T, Friedlander DR, Grumet M. Expression ofpolypeptide variants of receptor-type protein tyrosinephosphatase beta: the secreted form, phosphacan,increases dramatically during embryonic developmentand modulates glial cell behavior in vitro . J Neurosci Res1996;43:694–706.

26. Frankel AE, Kreitman RJ, Sausville EA. Targetedtoxins. Clin Cancer Res 2000;6:326–34.

27. Dyba M, Tarasova NI, Michejda CJ. Small moleculetoxins targeting tumor receptors. Curr Pharm Des 2004;10:2311–24.

28. Gottstein C, Winkler U, Bohlen H, Diehl V, Engert A.Immunotoxins: is there a clinical value? Ann Oncol 1994;5:97–103.

29. Brekke OH, Sandlie I. Therapeutic antibodies forhuman diseases at the dawn of the twenty first century.Nat Rev 2003;2:52–62.

Cancer Research




2006;66:2271-2278. Cancer Res Erik D. Foehr, Gustavo Lorente, Jane Kuo, et al. Glioblastoma Modelwith a Monoclonal Antibody Delays Tumor Growth in a

βTargeting of the Receptor Protein Tyrosine Phosphatase

Updated version

http://cancerres.aacrjournals.org/content/66/4/2271

Access the most recent version of this article at:

Cited articles

http://cancerres.aacrjournals.org/content/66/4/2271.full#ref-list-1

This article cites 26 articles, 11 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/66/4/2271.full#related-urls

This article has been cited by 6 HighWire-hosted articles. Access the articles at:

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

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. (CCC)Click on "Request Permissions" which will take you to the Copyright Clearance Center's

.http://cancerres.aacrjournals.org/content/66/4/2271To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/content/66/4/2271.full#ref-list-1

http://cancerres.aacrjournals.org/content/66/4/2271.full#related-urls

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



Download - Targeting of the Receptor Protein Tyrosine Phosphatase B ... · MAB-ZAP, IgG-SAP, DAT-SAP, 7A9B4-SAP, and 7E4B11-SAP were purchased (or created in the case of 7A9B4-SAP and 7E4B11-SAP)

Top Related