a bifunctional hybrid molecule of the amino-terminal fragment of urokinase and domain ii of bikunin...

Eur. J. Biochem.253, 8172826 (1998) FEBS1998

A bifunctional hybrid molecule of the amino-terminal fragment of urokinaseand domain II of bikunin efficiently inhibits tumor cell invasion and metastasis

Hiroshi KOBAYASHI1, Dan SUGINO2, Ming Yu SHE1, Hidekazu OHI1, Yasuyuki HIRASHIMA1, Hiromitsu SHINOHARA1, Michio FUJIE3,Kiyoshi SHIBATA 3 and Toshihiko TERAO1

1 Department of Obstetrics and Gynecology, Hamamatsu University, Shizuoka, Japan2 Nissin Central Research Institute, Shiga, Japan3 Equipment Center, Hamamatsu University School of Medicine, Shizuoka, Japan

(Received11 November1997/5 January1998) 2 EJB 971599/2

Urinary trypsin inhibitor (UTI) inhibits efficiently tumor cell invasion and the formation of metastasis.The anti-metastatic effect is dependent on the COOH-terminal domain II of UTI [UTI-(782136)-peptide].To develop a molecule that binds with high affinity to the urokinase (uPA) receptor (uPAR) on tumorcell surfaces, a bifunctional hybrid molecule [uPA-(12134)-UTI-(782136)] consisting of the uPAR-bind-ing NH2-terminal fragment [UTI-(782136)-peptide] of uPA at the NH2-terminus of UTI-(782136)-pep-tide was produced inEscherichia coliby genetic engineering. The purified hybrid protein inhibited trypsinand plasmin 223-fold less effectively than UTI-(782136)-peptide and was found to bind to human tumorcells via uPAR, which was confirmed by cell binding and competition experiments. Using a modifiedBoyden chamber and an artificial basement membrane, Matrigel, it was found that the hybrid protein isvery effective at inhibiting invasion by uPAR-expressing human tumor cells. Sensitivities of tumor cellstowards the anti-invasive effect of uPA-(12134)-UTI-(782136) correlated with the density of uPAR onhuman tumor cells. Furthermore, in the spontaneous metastasis model, the hybrid protein inhibited theformation of lung and/or lymphatic metastasis by human ovarian carcinoma and choriocarcinoma cells.The hybrid protein was much more effective than uPA-(12134)-peptide, UTI-(782136)-peptide, or UTI.We conclude that this approach extends the possibility of applying recombinant protein for therapeuticuse in inhibition of human tumor cell metastasis.

Keywords:amino-terminal fragment; bikunin; hybrid protein; urinary trypsin inhibitor ; urokinase-typeplasminogen activator.

Tumor cells produce urokinase-type plasminogen activator cell surface is the preferred site for uPA-mediated protein degra-(uPA) in an enzymatically inactive proenzyme form (pro-uPA)dation [11]. Various different approaches to interfere with the[1, 2]. Secreted pro-uPA can immediately bind to the specific expression or reactivity of uPA or uPAR at the gene or proteinuPA receptors (uPAR) on tumor cell surfaces with high affinity. level were tested successfully, including the use of antisenseThe uPAR specifically recognizes pro-uPA and active high-mo-oligonucleotides, antibodies, inhibitors and recombinant or syn-lecular-mass uPA via their growth-factor-like terminal domain.thetic uPA and uPAR analogues [12215]. It has been reporteduPAR is a glycoprotein of approximately 55 kDa, its affinity forthat a soluble recombinant truncated form of the uPAR is able touPA is high (0.2 nM), and the rate of dissociation is low [3, 4].block binding of uPA to cell-surface-bound uPAR [12]. AnotherReceptor-bound uPA catalyzes the formation of plasmin on thefeasible approach to be tested is the repression of uPA or uPARcell surface to generate the proteolytic cascade that contributessynthesis by agents that block transcriptional and translationalto the breakdown of basement membranes and the extracellularfactors known to be involved in uPA/uPAR expression [12].matrix. The plasma membrane uPAR has attracted considerableWe have reported that a highly purified urinary trypsinattention because of its role in migration and tissue invasion byinhibitor (UTI) efficiently inhibits soluble and tumor-cell-sur-mononuclear phagocytes and malignant cells [528]. In some face receptor-bound plasmin [16, 17]. UTI inhibits tumor cellcell types uPAR localizes uPA to cell-cell and cell-substratuminvasion in anin vitro assay and production of experimental andcontact sites, providing the possibility of directional proteolysisspontaneous lung metastasis in anin vivomouse model [16, 182that may be involved in cell migration and invasion [9,10]. 21]. The anti-invasive effect is dependent on the anti-plasmin

It has been reported that competitive displacement of uPAactivity of UTI [16, 18221]. A UTI peptide, which inhibitsfrom uPAR resulted in decreased proteolysis, suggesting that theplasmin activity, synthesized by means of an automated peptide

Correspondence toH. Kobayashi, Department of Obstetrics andsynthesizer inhibited mouse 3LL cell invasion [21]. UTI and theGynecology, Hamamatsu University School of Medicine, Handachoeffective peptide inhibited tumor cell invasion through Matrigel.3600, Hamamatsu, Shizuoka, Japan 431-31 UTI did not inhibit tumor cell proliferation or the binding of the

Fax: 181 53 4351626. cells to Matrigel. Also, UTI did not inhibit chemotactic migra-Abbreviations.FITC, fluorescein-isothiocyanate ; uPA, urokinase-

tion of tumor cells to fibronectin. It is likely that UTI acts as atype plasminogen activator; uPAR, uPA receptor; UTI, urinary trypsinprotease inhibitor.inhibitor; PAI, plasminogen-activator inhibitor.

We attempted to synthesize conjugates between the NH2-ter-Enzymes.Plasmin (EC 3.4.21.7) ; trypsin (EC 3.4.21.4); U-plasmino-gen activator (EC 3.4.21.73). minal fragment [uPA-(12134)-peptide] of uPA and UTI. Conju-

818 Kobayashi et al. (Eur. J. Biochem. 253)

gating a physiological plasmin inhibitor to uPA-(12134)-peptide Central Research Institute. The synthetic HI-8 DNA was puri-fied by PAGE and cloned into pTV118 N vector (TAKARA) tomight target it to reduce cell-associated proteolytic activity to

the close environment of the uPAR-expressing tumor cell sur- generate the plasmid pCD17R15. The nucleotide sequence ofDNA was determined by the dideoxynucleotide methodface and subsequently may effectively inhibit tumor cell inva-

sion and metastasis [20], because the cell surface uPAR might (AutoRead Sequencing Kit; Pharmacia) using the automatedDNA sequencer (ALF DNA sequencer; Pharmacia).be a critical component of the metastatic machinery. A method

of conjugation of the UTI domain II [UTI-(782136)-peptide] to Plasmid construction. The plasmid pPPA was used as thetemplate for PCR to obtain the DNA fragment encodingthe receptor-binding of uPA-(12134)-peptide has been devel-

oped utilizing the heterobifunctional cross-linking reagent,N- uPA-(12134)-peptide using the following primers: 5′-CGTGA-GCGACTCCAAAGGCAGCAATG-3′ (Pr-1) and 5′-GGTACC-succinimidyl-3-(2-pyridyldithio) propionate [20]. The conjugate

retained its protease-inhibiting activity and bound to uPAR on ATCTGCGCAGTCATGCAC-3′ (Pr-3). The PCR product wasdigested withTaqI and KpnI to obtain a 379-bp DNA fragmentthe surface of tumor cells. We have shown that the conjugate

exhibits plasmin inhibition at the cell surface and subsequently encoding the residues102134 uPA. TheBamHI2TaqI adaptor,consisting of two complementary oligonucleotides, was synthe-inhibits tumor cell invasion through Matrigel in anin vitro inva-

sion assay [20]. sized. The synthetictac promoter carryingHindIII and BamHIcohesive ends (tac promoter GenBlock; Pharmacia), theTo extend our idea, we attempted to produce a hybrid mole-

cule consisting uPA-(12134)-peptide placed at the NH2-termi- BamHI2TaqI adaptor, the 379 bpTaqI2KpnI DNA and KpnI2HindIII vector fragment of pUC19 were ligated to construct thenus of UTI-(782136)-peptide by genetic engineering tech-

niques. Exogenously applied hybrid protein can bind to the plasmid pTAK.To obtain the UTI-(782136)-peptide DNA fragment,specific uPAR on cell surfaces with high affinity. The receptor-

bound hybrid protein localizes the protease-inhibiting activity to pCD17R15 was used as the template for PCR. The primers5′ - GGGTACCGTTGCTGCTTGCAACCTGCCGATTGTCCG -the tumor cell surface. This may be effectively a bifunctional

molecule, which, in addition to inhibiting trypsin and plasmin 3′ (Pr-4) and 5′-GTGATCAACCCGGAACACCGCAATATT-CACGG-3′ (Pr-5) were designed to introduce aKpnI site at theactivities directly, is able to block unoccupied uPAR, thereby

preventing localization of uPA activity. Our data suggest that 5′ end and aBclI site at the 3′ end of the DNA. A176-bp DNAfragment encoding the 59-amino-acid (lacking seven C-terminalinhibition of uPA, and in particular of receptor-bound uPA,

followed by inhibition of a protease cascade, might have a amino acids) of UTI-(782136)-peptide was obtained by diges-tion of the amplified product withKpnI and BclI. The DNAtherapeutic effect on neoplastic and inflammatory disease pro-

cesses. fragment was cloned into theKpnI and BamHI sites of pUC18to construct the plasmid pHIK. The plasmids pHIK and pTAKwere digested withKpnI andXmnI, respectively. The expressionplasmid pAIP, which produces the hybrid protein, was con-MATERIALS AND METHODSstructed by ligating 2356-bpKpnI2XmnI DNA of pTAK with997-bpKpnI2XmnI DNA of pHIK. The correct sequence wasCell lines and culture. Escherichia coli strain OneShot

INVaF′ (Invitrogen) was used in the uPA gene-cloning experi- verified by DNA sequencing (Fig.1).Expression of hybrid proteins in E. coli. E. coli culturesment. E. coli strain JM109 (Takara) was used for construction

of plasmids and expression of hybrid proteins.E. coli were were grown in Terrific Broth containing100 µg/ml ampicillin at37°C. The lysates were sonicated, centrifuged and lysed. Thegrown on Luria-Bertani broth or plates containing100 µg/ml

ampicillin. Terrific Broth medium was used for expression of inclusion bodies were dissolved, centrifuged and chromato-graphed on an SP MemSep1000 chromatography cartridgehybrid proteins.

Tumor cell lines used in this study were derived from human (Millipore). The hybrid protein was eluted with a gradient from0 to 1 M NaCl. Fractions containing the hybrid protein wereovarian carcinoma (HOC-I) cells [16, 17], human choriocarci-

noma (SMT-cc1) cells [20], and murine Lewis lung carcinoma pooled and concentrated by ultrafiltration (Centriplus-3 concen-trator; Amicon). The sample was loaded onto a Superdex 75 gel(3LL) cells [22, 23]. 3LL, selected for its high lung-colonization

potential was kindly provided by Chugai Pharmaceutical Co., filtration column (HiLoad 26/60 prep grade ; Pharmacia). Peakfractions were pooled and dialyzed. The sample was loaded onTokyo, Japan. The tumor was maintained by serial subcutaneous

transplantation in C57BL/6 mice. Other cell lines were estab- a Resource S column (Pharmacia). The hybrid protein waseluted with a gradient from 0 to 0.5 M NaCl and analyzed bylished in our laboratory [16221]. The tumor cells listed in

Table1 were maintained in RPMI1640 supplemented with10% 20% SDS/PAGE.The N-terminal amino acid sequence was determined byfetal calf serum (HOC-I, SMT-cc1 and 3LL) (Gibco) at 37°C in

a humidified atmosphere containing 5% CO2/95% air. Cells in automated Edman degradation using an Applied BiosystemsModel 477 A sequencer.the exponentially proliferating phase were harvested for the

experiments. Tumor cells were harvested by washing the mono- Mice. Female Balb/c nu/nu mice and specific pathogen-freefemale C57BL/6 mice were purchased from Charles River Japanlayer with NaCl/Pi (136.9 mM NaCl, 2.7 mM KCl,1.8 mM

KH2PO4, 10.1mM Na2HPO4, pH 7.4) followed by brief incuba- and used for experiments at the age of 528 weeks. Animalswere kept in a laminar-flow cabinet under pathogen-free condi-tion in trypsin/EDTA at 37°C.

Cloning and sequencing of DNA fragments encoding tions and were fed sterile food and water in the experimental-animal room. The care and use of the animals were in accor-uPA-(1−134) and UTI-(78−136). The partial cDNA for uPA

was obtained by PCR amplification of the human kidney cDNAs dance with the Institution’s guidelines.Spontaneous metastasis in nude mice (evaluation of the(Quick-Clone cDNA; Clontech). The amplified 538-bp cDNA

encoding amino acid residues12168 of uPA was cloned into antitumor effect of the hybrid protein). Since the murine 3LLcells were highly malignant, primary tumors grew rapidly, andpCRII vector (TA Cloning Kit, Invitrogen) to construct the

plasmid pPPA. lung metastases were evident in100% of the mice that receivedsubcutaneous inoculation. The C57BL/6 mice provided a con-The plasmid pCD17R15 (Nakano et al., Japanese Patent

Application Publn. No. 6-247998) contained the UTI-(782136)- ventional model in which the involvement of hybrid proteinsin the spontaneous metastasis could be investigated [24]. Thepeptide DNA and was kindly provided by S. Nakano, Nissin

819Kobayashi et al. (Eur. J. Biochem. 253)

Purification of UTI, UTI-(78 −136)-peptide and uPA-(1−134)-peptide. These compounds were purified to homo-geneity as described previously [20]. UTI-(782136)-peptide anduPA-(12134)-peptide were biotinylated according to the methodof Guesdon et al. [25] usingN-hydroxysuccinimidylbiotinami-docaproate (Sigma).

Antibodies. Rabbit polyclonal antibodies against UTI andUTI-(782136)-peptide were prepared according to the manufac-turer’s instructions. An immunoglobulin fraction was purifiedfrom a rabbit antiserum using protein-G2Sepharose chromatog-raphy (HiTrap; Pharmacia). A murine mAb, 4G12, to theNH2-terminal domain I of UTI was a kind gift from Dr HidekiYamamoto (Nissin Molecular Biology Institute Inc.). mAbsagainst uPA were kindly supplied by American Diagnostica Inc.(Greenwich CT, USA). mAbs #3921, #3471 and #394 react withthe A-chain, amino acids17234 and the B-chain of uPA,respectively [26, 27].

SDS/PAGE and western blotting. 1 µg of the purifiedhybrid protein was loaded to each well of an SDS/18% acrylam-ide gel under non-reducing and reducing conditions. Westernblotting for the detection of uPA-(12134)-UTI-(782136) usinganti-uPA Ig and anti-UTI Ig was performed as described pre-viously [20]. Gels were electroblotted onto poly(vinylidenedifluoride) membranes (Millipore), using a semi-dry electroblot-ting apparatus (Marysol) at 40 mA/gel (90 min, 23°C). Themembrane was blocked in 2% BSA in Tris/NaCl (50 mM Tris/HCl, 120 mM NaCl, pH 8.5), for1 h at 23°C with gentle agita-tion. All washings of the membrane were carried out in Tris/HCl containing 0.05% Tween 20. Primary and secondary anti-bodies were diluted1:500 in blocking buffer. Each antibodyincubation was for1 h at 23°C. The sheets were incubated withspecific antibody for uPA, and the antibody bound on the poly-(vinylidene difluoride) sheets was detected by incubation withbiotin-conjugated secondary antibody (Dako) and avidin-peroxi-Fig. 1. Plasmid construction.pHIK and pTAK were digested withKpnI

and XmnI, respectively. The expression plasmid pAIP was constructeddase (Dako), followed by addition of the enzyme substrate,by ligating the 2356-bpKpnI2XmnI DNA of pTAK with the 997-bp 4-chloro-1-naphthol (Sigma).KpnI2XmnI DNA of pHIK. The sequence was verified by DNA se- Direct binding assay. 100 µl of uPA-(12134)-UTI-(782quencing. 136) (2µg/ml) was coated onto 96-well microtiter plates

(Costar) and subjected to ELISA. Antibodies against uPA A-chain or UTI-(782136)-peptide were used as probes.100 µlspontaneous-metastasis assay measures all the steps of metasta-each antibody at various concentrations (025 nM) were addedsis from a primary tumor [19].to hybrid-protein-coated well and incubated for 2 h at 23°C.3LL cells (13106 cells/200µl NaCl/Pi) were injected in theAfter five washes,100µl of a 1 :500 dilution of biotinylatedabdominal wall of unanesthetized C57BL/6 black mice. UTIsecondary antibody (Dako) and avidin-peroxidase (Dako) were(500µg/mouse), UTI-(782136)-peptide (50µg/mouse) andadded and incubated at 23°C for 1 h. After washing,100µluPA-(12134)-UTI-(782136) (150 µg/mouse) were injected0.1 M acetate/citrate, pH 6.0,10 mg/ml tetramethylbenzidine,subcutaneously daily from day 0 to day13 after tumor inocula-0.005% H2O2 were added to each well. The color reaction wastion. Controls were injected the vehicle. 28 days after theterminated after 5 min by adjusting the solution to1.0 M sulphu-inoculation of tumor cells, mice were sacrificed and the lungsric acid. The absorbance was determined at 450 nm in a BioRadwere removed, washed in NaCl/Pi, and fixed in Bouin’s solution.Model 2550 microplate reader (BioRad).The lung tumor colonies were counted under a dissecting micro-

Competitive ELISA assay.The UTI portion of the hybridscope [19].protein was detected both immunologically and functionally (byThe human ovarian HOC-I and choriocarcinoma SMT-cc1its ability to inhibit plasmin and trypsin). The uPA-(12134)-cells were washed by centrifugation, suspended in mediumpeptide and UTI-(782136)-peptide competitive solid-phase(13107 cells/50µl), mixed with an equal volume of liquidELISA was designed for screening the transfectant supernatantsMatrigel (50µl) and injected in the abdominal wall of nudeand measuring the binding reactivity of purified uPA-(12134)-mice. 0.1 ml fluid containing107 cells was injected through aUTI-(782136) for the anti-[uPA-(17234)] Ig and anti-[UTI-27-gauge needle. Intraperitoneal administration of NaCl, UTI(782136)] Ig. 100 µl of each antibody (2µg/ml) were coated(500µg), UTI-(782136)-peptide (0.5, 5 and 50µg) or uPA-(12onto 96-well microtiter plates and subjected to ELISA. Biotiny-134)-UTI-(782136) (0.5, 5 and 50µg) was repeated daily untillated uPA-(12134)-peptide or UTI-(782136)-peptide was usedautopsy. Autopsy was performed at 50 days, or sooner if theas a probe. 50µl biotinylated peptide at1 µg/ml were addedtumor was large or the host was ill or distressed. Lung and lym-to each antibody-coated well in the presence of pro-uPA, high-phatic metastatic formation was studied by macroscopic exami-molecular-mass uPA, uPA-(12134)-peptide, UTI-(782136)-nation.peptide, or uPA-(12134)-UTI-(782136) (50µl, 021000 nM)The significance of differences in the numbers of lungand incubated at 23°C for 1 h. After five washes,100µl of ametastases detected between groups was determined using

Mann-Whitney U-test. 1:1000 dilution of avidin-peroxidase were added and incubated


at 23°C for 1 h. After washing,100 µl substrate were added to pH 7.4 (30 min, 23°C). Changes in absorbance at 405 nm weremeasured with an enzyme-immunoassay reader.each well. The absorbance was determined at 450 nm in a

Monolayers of HOC-I cells were incubated with plasminmicroplate reader. The ability of uPA-(12134)-UTI-(782136)(0.1 µM, 4 h, 4°C) and washed twice with NaCl/Pi, 0.1% BSA.to inhibit the binding of anti-UTI or anti-uPA Ig to microtiterTo investigate the inhibitory effects of UTI, UTI-(782136)-pep-plates coated with UTI-(782136)-peptide or uPA was alsotide, uPA-(12134)-peptide and hybrid protein on tumor-cell-as-assayed.sociated plasmin, each agent (021 µM) was incubated (30 min,Specific binding of the hybrid protein to tumor cells. The23°C) with cell-associated plasmin in a 96-well microtiter plateuPA-(12134)-UTI-(782136) was dialyzed against 0.1M[14], incubated (30 min, 23°C), and washed twice with NaCl/Pi,NaHCO3, pH 9.0, for at least 4 h. The indicated amount of0.1% BSA, before addition of S-2251 in Tris/HCl, 0.1% BSA,fluorescein isothiocyanate (FITC) in dimethylsulfoxide waspH 7.4. The change in absorbance at 405 nm was measured.added in small aliquots to the hybrid protein under continousWhen the cell-associated plasmin was incubated with S-2251 inrolling. The hybrid protein/FITC mixture was left for 3 h atthe absence of agents, the value at 405 nm was considered as23°C. A PD-10 column (Pharmacia) was used to purify FITC-100%.conjugated hybrid protein.

Cell-invasion assay.The cell-invasion assay was performedHOC-I cells were collected by centrifugation and suspendedas described previously [14, 16218, 21, 27]. Polycarbonatein NaCl/Pi, 0.1% BSA. Cells were treated with 50 mM glycine/filters (8-µm pore size; Costar) were coated with Matrigel, andHCl, 0.5 M NaCl, pH 3.0, to dissociate surface receptor-boundplaced in a modified Boyden chamber. Upper chamber wellsuPA, and suspended in 0.5 M Hepes, 0.1 M NaCl, pH 7.5 [13].were treated with agents at various concentrations (100 µl, 1 h,3LL cells were also used as a comparative experiment [22].23°C) in serum-free medium.100µl tumor cell suspensionAfter centrifugation (1200 rpm,10 min, 4°C), the cell pellet was(23106 cells/ml) was placed in the upper compartment of theadjusted to a density of106 cells/ml in NaCl/Pi, 0.1% BSA. Thechamber. Boyden chambers were incubated at 37°C for 12 h.cells were suspended in1.0 ml NaCl/Pi, 0.1% BSA containingFibroblast-conditioned media or fibronectin were placed in theFITC-conjugated ligands and rocked gently for1 h at 4°C. Cell-lower compartment as a source of chemoattractants. At the endassociated fluorescence (argon laser excitation at 488 nm; noof the incubation, the cells on the upper surface of the filter werewash) was immediately analyzed by means of an Epics Profileremoved with tissue paper, and the cells on the lower surface offlow cytometer (Coulter). Fluorescence data are displayed asthe filter were stained and counted under a light microscope.immunofluorescence profiles in which increasing log fluores-Each assay was carried out in triplicate.cence intensity is plotted versus cell number [26].

The cytotoxic activity of the hybrid protein was tested onInhibition of binding of FITC-pro-uPA to HOC-I cells byvarious uPAR-expressing tumor cells as described previouslyuPA-(12134)-UTI-(782136) was tested.10 nM FITC-pro-uPA[28].in the presence of competitors (021000µM) was added to the

acid-treated cells. Non-specific binding was determined in thepresence of excess parent pro-uPA (500 nM). Cell-associatedfluorescence was determined. Inhibition of binding of FITC-RESULTSuPA-(12134)-UTI-(782136) to acid-treated HOC-I cells with

Characterization of the hybrid molecule.Binding of anti-uPAthe competitors was also tested.and anti-UTI Ig to immobilized solid-phase hybrid protein wasBinding and saturation experiments.Cell-bound endoge-studied by direct ELISA. Mouse mAbs against uPA A-chainnous uPA was removed by exposure to 50 mM glycine/HCl,(mAb #3921) and uPA-(17234)-peptide (mAb #3471) and apH 3.0, 0.1 M NaCl. The cell suspension was neutralized byrabbit polyclonal antibody against UTI-(782136)-peptide re-

addition of 0.5 M Hepes, pH 7.5, 0.1 M NaCl, and centrifuged acted with immobilized uPA-(12134)-UTI-(782136), whereasas described above.

mouse mAbs to domain I of UTI (mAb 4G12) and uPA B-chainuPA-(12134)-peptide and uPA-(12134)-UTI-(782136) (mAb #394) did not react with this protein.

were labeled with Na125I using the Iodogen method according to Characterization of the hybrid protein was investigated bythe manufacturer’s instructions. The specific activity of theSDS/PAGE followed by western blot analysis using several anti-labeled proteins was approximately 20 mCi/mg. Cells (13106

bodies {mAbs #3921, #3471 and 4G12, and anti-[UTI-(782cells/ml) were incubated with [125I]uPA-(12134)-peptide {or 136)-peptide] Ig}. The hybrid protein was recognized by[125I]uPA-(12134)-UTI-(782136)} in the absence (total bind- mAbs #3921 and #3471 and polyclonal anti-[UTI-(782136)]ing) or in the presence of unlabeled ligand (non-specific bind-(Fig. 2).ing) for 2 h at 4°C [2]. The concentrations of iodinated and un- To determine if uPAR on tumor cells binds the hybrid pro-labeled ligands were 0.8 nM and100 nM, respectively. The cells tein specifically, two experiments were performed. In the firstwere washed three times with ice-cold NaCl/Pi containing 0.1% experiment, direct binding assays using flow cytometryBSA at 4°C, and the cell-bound radioactivity was determined in(Fig. 3A) and cell ELISA (data not shown) were performed toa gamma-counter. All assays were performed in triplicate.examine the region within the hybrid protein responsible forSpecific binding of the [125I]uPA-(12134)-peptide [or [125I]uPA- binding to uPAR on tumor cells. FITC-uPA-(12134)-UTI-(782(12134)-UTI-(782136)} was calculated by subtraction of non-136) was tested for its binding to human HOC-I and mouse 3LLspecific binding from total binding. The dissociation constantcells. Binding of the hybrid protein to human tumor cells was(Kd) and the number of receptors were calculated by Scatcharddose dependent and saturable, whereas it did not bind to mouseanalysis. 3LL cells (Fig. 3B). Binding of FITC-uPA-(12134)-UTI-(782

Inhibition of proteolytic activity. The inhibiting effects of 136) to the cell surface reached saturation after 60 min at 4°C.UTI, UTI-(782136)-peptide, uPA-(12134)-peptide and hybrid Specific binding approached saturation at approximately10 nM.protein on several proteases (trypsin, plasmin and human leuko- Competitive inhibition assays employing flow cytometrycyte elastase) were investigated. These agents (0220 µM) were were performed to examine the affinity of the hybrid protein forincubated (5 min, 23°C) with 6.25µM protease in a 96-well binding to the uPAR on HOC-I cells. To confirm that ATFHImicrotiter plate, followed by the addition of a chromogenic sub- bound to the cells via uPAR, competition-binding experiments

were performed in which increasing amounts of pro-uPA, high-strate (S-2251, S-2222 or S-2484) in Tris/HCl and 0.1% BSA,


Fig.4. Effects of the hybrid protein on uPA binding to HOC-I cells.(A) Effects of several competitors in FITC-conjugated uPA-(12134)-HI-8 binding to HOC-I cells. Acid-treated HOC-I cells were reacted with10 nM FITC-uPA-(12134)-HI-8 (30 min, 4°C). The dose-dependentquenching of cell-associated fluorescence by competitors [uPA-uPA-(12134)-peptide (.), pro-uPA (j), high-molecular-mass uPA (d), low-mo-lecular-mass uPA (n), UTI (r), UTI-(782136)-peptide (m), and uPA-(12134)-UTI-(782136) (s)] is shown. (B) Effects of the hybrid proteinon FITC-conjugated pro-uPA binding to HOC-I cells. Cells were reactedwith 10 nM FITC-pro-uPA (30 min, 4°C). The dose-dependent quench-ing of cell-associated fluorescence by the hybrid protein is shown. Com-petition between FITC-pro-uPA and each competitor for binding to theuPAR on HOC-I cells is shown. The fluorescence mean channels werecalculated. The cell signal in the absence of competitors was set to100%. The experiment was repeated twice with similar results. Resultsare the means of two experiments, carried out on triplicate wells. TheFig. 2. Characterization of the hybrid protein examined by westernSD values of the means were 5210% of experimental values; SD valuesblotting. Lane M, molecular-mass standards with their approximateare given for reasons of clarity.masses shown to the left. The purified proteins after SDS/PAGE under

non-reducing conditions in an18% polyacrylamide gel and western blot-ting are shown. Lane1, UTI; lane 2, deglycosylated UTI; lane 3, UTI-(782136)-peptide; lane 4, enzymatically active high-molecular-massuPA-(12134)-UTI-(782136) was also competed by pro-uPA,uPA; and lane 5, uPA-(12134)-UTI-(782136)-peptide. mAbs #3921 high-molecular-mass uPA and uPA-(12134)-peptide, but not byand #3471 react with uPA-(12134)-peptide and high-molecular-masslow-molecular-mass uPA, UTI, or UTI-(782136)-peptide. 50%uPA, mAb 4G12 reacts with UTI and deglycosylated UTI, and poly-inhibition of binding to HOC-I cells was obtained with 5 nMclonal antibody HI-8 reacts with UTI, deglycosylated UTI and UTI-

pro-uPA, 8 nM uPA-(12134)-peptide and 5 nM hybrid protein.(782136)-peptide.Thus, the cell binding of the hybrid protein was via uPA-(12134) and did not involve the UTI-(782136) portion. uPA-(12134)-peptide and uPA-(12134)-UTI-(782136) had nearly thesame affinity for the uPAR.

In a separate experiment, 50% inhibition of FITC-pro-uPAbinding to HOC-I cells was obtained with1 nM pro-uPA,1 nMuPA-(12134)-UTI-(782136) and 5 nM uPA-(12134)-peptide.UTI and UTI-(782136)-peptide did not bind to the uPAR(Fig. 4B).

Competitive inhibition of binding of FITC-uPA-(12134)-UTI-(782136) to HOC-I cells was observed to various extentsby incubation of the protein with mAb #3471 (data not shown).It is likely that a masking of the intact receptor-binding sequencewithin the growth-factor-like domain by mAb #3471 is the causeFig. 3. Direct binding of FITC-conjugated uPA-(1−134)-HI-8 to

human HOC-I cells or murine 3LL cells by flow cytometry. Tumor of the inhibition.cells [human ovarian carcinoma HOC-I cells (A) and mouse lung carci- Quantitative cell-binding experiments were performed innoma 3LL cells (B)] incubated at106 cell/ml in the absence or presencewhich increasing amounts of [125I]uPA-(12134)-UTI-(782136)of increasing concentrations of FITC-conjugated ligands [FITC-pro-uPAwere added to acid-treated HOC-I cells and the specific bindingand FITC-(12134)-HI-8)]. For measurement of non-specific binding,was determined. The binding of [125I]uPA-(12134)-peptide totumor cells at the indicated concentrations were incubated for 60 min at

human tumor cells has been analyzed by Scatchard plot analysis,4°C with FITC-(12134)-HI-8 in the presence of100-fold excess coldand the results of representative experiments are shown inhybrid protein. FITC-pro-uPA was used as a control. Values are theTable1. Saturation of the uPAR on HOC-I cells with [125I]uPA-means6SD of two experiments, carried out on triplicate wells.(12134)-UTI-(782136) revealed aKd of 1.0 nM and 72000binding sites/cell ; for [125I]uPA-(12134)-peptide, theKd in thesecells was1.1 nM and 76 000 binding sites/cell were detected,molecular-mass uPA, low-molacular-mass uPA, uPA-(12134)-

peptide, UTI, UTI-(782136)-peptide and uPA-(12134)-UTI- showing that the hybrid protein bound specifically and with highaffinity to human tumor cells.(782136) were added with FITC-uPA-(12134)-UTI-(782136)

(Fig. 4A). The cells were incubated at 4°C for 1 h with FITC-uPA-(12134)-UTI-(782136) in the presence of increasingInhibition of soluble plasmin and trypsin by the hybrid pro-

tein. The inhibition of soluble proteases by the hybrid proteinconcentrations of unlabeled competitors. The binding of uPA-(12134)-UTI-(782136) was inhibited completely by10 nM was investigated using chromogenic assays. To investigate the

specificity of inhibition of plasmin and trypsin by uPA-(12134)-unlabeled uPA-(12134)-UTI-(782136). The binding of FITC-


Fig.6. Titration of HOC-I cell-associated plasmin with the hybridprotein. The cells were incubated with the indicated concentrations ofUTI, UTI-(782136) and uPA-(12134)-UTI-(782136) for 30 min atFig. 5. Dose-dependent inhibition of plasmin by the hybrid protein.23°C. After washing the cells, the hydrolysis of the chromogenic sub-Plasmin (6.25µM) was incubated with different concentrations of UTIstrate S-2251 catalyzed by the cell-associated plasmin were measured at(r), UTI-(782136)-peptide (m), uPA-(12134)-peptide (.) or uPA-(12405 nm. The control value was set at100%. After incubation, the reac-134)-UTI-(782136)-peptide (s). The remaining protease activity wastion was stopped and the amidolytic activity was determined. Results areassayed with the chromogenic substrate S-2251. Results are the mean ofthe means of two experiments, carried out on triplicate wells. The SDtwo experiments, carried out on duplicate wells. The SD values of thevalues of the means were less than15% of experimental values; nomeans were about 5% of experimental values; no values are given.values are given. Mean6SD ID50 values were as follows: UTI,533619 nM; UTI-(782136), 618652 nM; and uPA-(12134)-UTI-(782136), 113621 nM (P,0.05 compared with ID50 value of UTI).UTI-(782136), each protease was incubated with different

concentrations of the protein. The relative enzyme activitiesobtained when plasmin was titrated with UTI, UTI-(782136)-peptide, uPA-(12134)-peptide, or uPA-(12134)-UTI-(782136) mately 20280-fold enhancement of the inhibitory effect onare shown in Fig. 5. UTI and UTI-(782136)-peptide were found tumor cell invasion compared with UTI-(782136), as assessedto be strong inhibitors of plasmin (Fig. 5) and trypsin (data notby the ID50 values.shown). Efficient inhibition was observed at molar ratios of Invasion of murine 3LL cells through Matrigel was inhibitedgreater than1 :1 of UTI/proteases (5 min, 23°C). uPA-(12134)- significantly by uPA-(12134)-UTI-(782136) with an ID50 ofpeptide showed essentially no inhibition. UTI and UTI-(782 300 nM. The half-maximum inhibition of 3LL cell invasion pro-136)-peptide were at least twofold more effective than uPA-(12 duced by UTI-(782136)-peptide occurred at 200 nM. Thus, the134)-UTI-(782136) in inhibiting plasmin and trypsin, as as-anti-invasive activity of the hybrid protein is similar to that ofsessed by the concentration giving 50% inhibition of the prote-UTI-(782136)-peptide. The observed anti-invasive activity onase-inhibitor reaction using our colorimetric assay system. human tumor-cell lines was specific because uPA-(12134)-UTI-

In a previous study, we showed that UTI inhibits tumor-cell-(782136) showed no enhanced anti-invasive activity to a murineassociated plasmin. Titration of the HOC-I cell-associated plas-3LL cell line that possesses mouse uPAR. Because of speciesmin with each agent revealed that uPA-(12134)-UTI-(782136) specificity, murine uPAR-positive cells are unable to bind humancaused approximately tenfold enhancement of the inhibition ofuPA-(12134)-peptide and the hybrid protein. To demonstratecell-bound plasmin compared with UTI or UTI-(782136)-pep- that the hybrid-protein-induced inhibition of tumor cell invasiontide (Fig. 6). in human tumor cells was mediated through the uPAR, anin

vitro invasion assay was performed in the presence of the proteinalone or with the protein and100-fold excess uPA-(12134)-Effect of the hybrid protein on tumor cell invasion and me-

tastasis. If a cell-associated uPA · plasmin is essential to the peptide. Excess uPA-(12134)-peptide reduced significantly theanti-invasive activity of uPA-(12134)-UTI-(782136) (data notinvasion process of tumor cells, tumor invasion would be sup-

pressed after all membrane receptor sites were occupied by an shown). Thus, uPAR binding was necessary for the anti-invasiveactivity of uPA-(12134)-UTI-(782136).enzymatically inactive uPA fragment, such as uPA-(12134)-

peptide, or after inhibition of membrane-associated protease We have determined the expression levels of uPA and plas-minogen-activator inhibitor (PAI)-1 before and after the treat-activities by selective protease inhibitors. Under these condi-

tions, pro-uPA synthesized and released by tumor cells cannot ment of cells with uPA-(12134)-UTI-(782136). Binding ofmAbs to uPA B-chain (mAb #394) or PAI-1 to tumor cellsbind to its own receptors, because tumor cells do not have free

uPAR. treated with uPA-(12134)-UTI-(782136) for 3 h at 37°C wasstudied by flow cytometry. The study of thein vitro effects ofWe examined the anti-invasive activity of uPA-(12134)-

UTI-(782136) on three tumor-cell lines using a modified Boy- uPA-(12134)-UTI-(782136) indicates no effect on the expres-sion of uPA or PAI-1 by the tumor cell lines tested (HOC-I andden chamber (Table1, Fig. 7). When uPA-(12134)-UTI-(782

136) was incubated with human tumor cells (HOC-I and SMT- SMT-cc1 ; data not shown).In an in vivo nude-mouse metastasis model, we quantifiedcc1) at various concentrations in the upper compartment of the

chamber, invasion by tumor cells through Matrigel was inhibited the number of lung and/or lymphatic metastases that were recog-nizable macroscopically in each treatment group (Table 2). uPA-in a dose-dependent manner. This effect was not attributable to

the physicochemical properties of uPA-(12134)-UTI-(782136), (12134)-UTI-(782136) inhibited significantly human tumormetastasis in the severalin vivo models of tumor metastasis. Insuch as high viscosity at high concentration. The ID50 values

were about1.2 nM and 10 nM in SMT-cc1 and HOC-I cells, HOC-I cells, for example, spontaneous metastasis of tumor cellsfrom the subcutaneous site to one or more ipsilateral axillary,respectively. uPA-(12134)-UTI-(782136) caused approxi-


Table 1. Effect of the hybrid protein on tumor cell invasion. The data for uPAR were analyzed by Scatchard plot (see Materials and Methods).n.d., not determined. The results are the means6SD of three experiments, carried out on duplicate chambers. The standard deviation of the ID50

values was typically less than15% of experimental values. compared to UTI value.

Tumor cells uPAR uPAR ID50 forKd binding sites-cell

UTI UTI-(782136)- uPA-(12134)- uPA-(12134)-peptide peptide UTI-(782136)

nM nM

HOC-I 1.1 78000 200 180 1000a 10a

SMT-cc1 1.0 108 000 100 220 500a 1.2a

3LL n.d. n.d. 250 200 .100a 300

a Significantly different (P,0.05) to value for UTI.

Table 2. Effect of the hybrid protein on the formation of the lungand/or lymphatic metastasis foci in tumor-bearing nude mice.13107

cells were injected subcutaneously. Results of these experiments wereevaluated by the Mann-Whitney U-test. All values with the hybrid pro-tein were significantly different from those for untreated tumor-bearingnude mice (P,0.05). Administration of hybrid protein (50µg) was re-peated daily until autopsy.

Experi- Cell line Hybrid Prevarence of Median no.ment protein lung and/or (range) of lung

lymphatic and/or lymphaticmetastasins metastasins

(%)

1 HOC-I 2 11/15 (73) 6 (3210)1 3/15 (20) 1 (02 3)

2 HOC-I 2 11/13 (85) 4 (32 5)1 3/13 (23) 0 (02 1)

3 SMT-cc1 2 13/13 (100) 9 (5212)1 2/13 (15) 2 (02 5)

4 SMT-cc1 2 10/10 (100) 10 (5220)1 4/10 (40) 2 (02 4)

nificantly the number of pulmonary and lymphatic metastaticfoci of SMT-cc1 (Table 3). The hybrid protein had a muchstronger dose-dependent inhibitory effect than UTI or UTI-(782136)-peptide. uPA-(12134)-peptide produced only a weak in-hibitory effect, even at high concentrations (500µg/mouse).

In contrast, uPA-(12134)-UTI-(782136) did not stronglyinhibit lung metastasis of murine 3LL cells. The lack of cross-

Fig. 7. Hybrid protein efficiently inhibits uPAR-expressing human reactivity of human uPA to mouse uPAR is well known. Hence,HOC-I cells. Tumor cells were added to the upper compartment of ahuman uPA-(12134)-peptide does not bind mouse uPAR. TheBoyden Chamber with a Matrigel-coated filter. After12 h incubation at inhibitory activity of uPA-(12134)-UTI-(782136) is similar to37°C, cells on the upper surface of the filter were removed, and the cells

those of UTI and UTI-(782136)-peptide, which do not need tounder the filter were counted. The number of migrated cells in thebind to uPAR. We carried outin vivo experiments with murineabsence of inhibitor was assigned a value of100% as a control. (A)tumor cell lines as a negative control. In all the groups treatedHOC-I; (B) 3LL cells. The results are the means6SD of three experi-with UTI, UTI-(782136)-peptide or the hybrid protein, therements, carried out on duplicate chambers. *,P,0.05 compared with a

control value.r, UTI; m, UTI-(782136); ., uPA-(12134); s, uPA- was no significant reduction in the size of the primary tumor.(12134)-UTI-(782136). We did not look into potential thrombotic complications in the

animals that received the inhibitors.To verify that the hybrid protein was cytotoxic on a tumor

cell line, a cell-viability assay was performed. A protein-synthe-inguinal and/or retroperitoneal lymph nodes was observed in 22of 28 of the mice injected intraperitoneally with NaCl and 6 of sis-inhibition assay and a cell-viability assay confirmed that the

hybrid protein did not kill tumor cells even at a high dose of28 of the mice injected with uPA-(12134)-UTI-(782136)(Table 2). The therapeutic efficacy of tumor cells may be heavily 20µg/ml (1 µM). The hybrid protein did not cause any visible

morphological changes of these tumor cells and did not inhibitdependent on the treatment schedule. The hybrid protein inhib-ited strongly lung and lymphatic (ipsilateral axillary and/or protein synthesis without causing cell death (data not shown).

We began administration of the peptidesin vivo on day zero toretroperitoneal) metastasis of SMT-cc1 cells, and decreased sig-


Table 3. Effect of the hybrid protein on the formation of the sequent tumor cell invasion [29]. The distribution of the uPARmetastasis foci in human tumor-bearing nude mice (dose-dependent on malignant cells and the possible role of the uPAR in tumorexperiments). Administration of anti-metastatic agents to HOC-I orprogression indicate that it may be possible to use a molecule3LL-bearing nude mice was repeated daily until autopsy or for14 days that binds with high affinity to the uPAR fused to an anti-meta-after tumor inoculation, respectively.13107 HOC-I cells and13106 3LL

static agent to selectively target tumor cells.cells were injected subcutaneously.One molecule that has been employed in the construction of

anti-metastatic agents is UTI [30235]. UTI is a 40-kDa glyco-Cell line Treatment Concen- Prevalence Mean no.protein that contains two structural domains, which acts as atration of lung (range) of

metastasis metastases Kunitz-type protease inhibitor. Residues 782136 are responsiblefor anti-metastatic activity [16221, 36, 37]. Our goal is to pro-

µg/mouse % duce a UTI-(782136)-peptide linked to uPA-(12134)-peptidethat would have substantially improved properties comparedHOC-I None 7/ 9 (78) 5 (0210)

UTI 500 1/10 (10) 1 (022) with UTI-(782136)-peptide.UTI-(782136)- We have constructed a bifunctional hybrid protein [uPA-(12peptide 50 1/10 (10) 1 (022) 134)-UTI-(782136)] composed of uPA-(12134)-peptide and

5 8/10 (80) 4 (128) UTI-(782136)-peptide by genetic engineering techniques. The0.5 7/10 (70) 7 (0211) hybrid protein was approximately 21 kDa under non-reducing

uPA-(12134)-and reducing conditions, and had full protease-inhibiting activ-UTI-(782136) 50 2/13 (15) 1 (021)ity, indicating that the UTI-(782136)-peptide portion of the5 3/13 (23) 2 (023)hybrid protein retained the protease-inhibiting property charac-0.5 11/13 (85) 5 (229)teristic of UTI-(782136)-peptide. The hybrid protein, however,

3LL None 10/10 (100) 68 (212100) was 223-fold less effective in inhibiting plasmin activity thanUTI 500 10/10 (100) 25 (3257) UTI-(782136)-peptide. The epitope expression in constant re-UTI-(782136)- gions of the hybrid protein was reduced slightly, possibly due topeptide 50 9/ 9 (100) 16 (1244) alterations in the folding of the constant region derived from theuPA-(12134)- disulfide bond formation between cysteines in UTI-(782136).UTI-(782136) 150 10/10 (100) 20 (3252)

The structure of UTI-(782136) in the hybrid molecule was com-promised to some degree, but protease-inhibiting capacity didnot seem to be strongly impaired. The hybrid proteins aredesigned to retain the binding activity for uPAR and the prote-abrogate metastasis. It is unlikely, however, that the observedase-inhibiting activity of uPI, for use in preventing cancerinhibitory effect against spontaneous metastases is a strongmetastasis.argument for their low toxicity, since the peptides and carrier

One cannot use the amount of the injected protein to com-are non-toxic.pare efficacy of the various constructs at inhibiting metastasissince their serum half-lives are likely to differ significantly. Weare trying to determine the serum concentration of each proteinDISCUSSIONto compare their relative efficacy.

The use of a drug-delivery system that attempts cell-specificWe have previously shown that both the NH2-terminal frag-ment of urokinase [uPA-(12134)-peptide] [12214, 21, 22, 26, targeting for selective toxicity is not a novel idea [38242]. Use

of a compound employing a protein-like uPA to bind to specific27] and UTI [16221, 23] inhibit tumor cell invasionin vitro.We have characterized the smallest fragment of each protein that cell-surface receptors promises to be much more specific. Once

bound to uPAR-bearing cells, the hybrid protein could inhibitretains this ability and have used these to inhibitin vitro invasion(in a number of cell lines) and experimental metastasis (princi- tumor cell invasion more efficiently than UTI-(782136)-pep-

tide. The hybrid protein binds selectively to cells bearing thepally in a 3LL murine model system) [12214, 16223, 26, 27].We have determined that the mechanism of inhibition involves uPAR, may decrease cell-associated proteolysis through compet-

itive displacement of uPA from uPAR, inhibits efficiently cell-the molecules’ abilities to block soluble and tumor-cell-associ-ated plasmin activities. We showed previously that a fusion pro- associated plasmin activity, and is suitable for large-scale pro-

duction. Not only were all of the tumors sensitive to ATFHI, buttein formed by chemical cross-linking of uPA-(12134)-peptideand UTI inhibits in vitro invasion by tumor cells [20]. In the a concentration-response effect was demonstrated in all. Any

density of uPAR on a tumor cell surface may be sufficient, andpresent study, we have extended previous observations by show-ing that a recombinant fusion protein made up of uPA-(12134)- the clinical utility of the hybrid protein might not be limited to

those tumor cells that have uPAR gene amplification.peptide and UTI-(782136)-peptide is more effective at inhibit-ing cell surface plasmin andin vitro invasion by cancer cells At least two functions of uPA bound to its receptor were

observed: a catalytic role important in generating pericellularand in vivo lung and/or lymphatic metastasis than UTI or uPA-(12134)-peptide. We demonstrate that the hybrid protein binds proteolysis; and a non-catalytic function operating in the differ-

entiation and adhesion of cells [43, 44]. Glycosylphosphatidyl-to cells via uPAR, as demonstrated by flow cytometry, that itinhibits soluble and cell-associated plasmin activity, and that the inositol-linked receptors such as uPAR can participate in signal

transduction. The potential for inhibition of tumor cell invasionhybrid protein inhibitsin vitro invasion andin vivo metastasisin a dose-dependent manner with an ID50 that correlates with and metastasis may be dependent on the continued occupation

of the uPAR. The need for continued receptor occupation sug-uPAR number/cell.Attention has focussed on anti-metastatic agents constructed gests that inhibition of tumor cell invasion and metastasis is reg-

ulated not only by the expression of uPAR and the availabilityby means of genetic engineering techniques for the therapy ofmalignant cells. One approach involves targeting uPAR, which of ligand, but also by factors that modulate uPAR turnover. One

such factor identified is PAI-1, which abrogates proteolytic ac-is present on malignant cells [12]. Malignant uPAR-bearingtumor cells have greater access to the high concentration of uPA. tivity. However, the hybrid protein did not bind PAI-1 (data not

shown). Internalization of uPA · PAI-1 complexes is mediatedOverexpression of uPAR facilitates matrix degradation and sub-


3. Appella, E., Robinson, E. A., Ullrich, S. J., Stoppelli, M. P., Corti,by uPAR, and uPA-(12134)-peptide is neither internalized norA., Cassani, G. & Blasi, F. (1987) The receptor-binding sequencedegraded, despite its ability to bind uPAR [45250]. The inter-of urokinase. A biological function for the growth-factor modulenalization of FITC-uPA-(12134)-UTI-(782136) in the presenceof proteases,J. Biol. Chem. 262, 443724440.of active PAI-1 in HOC-I cells has been tested by immunofluo-

4. Vassalli, J.-D., Baccino, D. & Belin, D. (1985) A cellular bindingrescence (data not shown; [20]). The fluorescent spots seen site for theMr 55 000 form of the human plasminogen activator,represent patches of FITC-uPA-(12134)-UTI-(782136) bound urokinase,J. Cell Biol. 100, 86292.to uPAR. Positive staining on the cell surface was observed even5. Sitrin, R. G., Todd, R. F. III, Mizukami, I. F., Gross, T. J., Shollen-after incubation of the cells with active PAI-1 for 5 h at 37°C, berger, S. B. & Gyetko, M. R. (1994) Cytokine-specific regulation

of urokinase receptor (CD87) expression by U937 mononuclearindicating that FITC-uPA-(12134)-UTI-(782136) is neitherphagocytes,Blood 84, 126821275.internalized nor degraded (data not shown). This may be the

6. Vassalli, J.-D., Wohlwend, A. & Belin, D. (1992) Urokinase-cata-reason why the results obtained in thein vitro andin vivo assayslyzed plasminogen activation at the monocyte/macrrophage cellare much more dramatic than expected for inhibition of cell-surface: a localized and regulated proteolytic system,Curr. Top.associated plasmin activity by the hybrid protein.Microbiol. Immunol. 181, 65286.In target cells, enhancement of cell proliferation is mediated7. Kirchheimer, J. C. & Remold, H. G. (1989) Functional characteris-

by uPA or uPA-(12134)-peptide. Binding of uPA to some tumor tics of receptor-bound urokinase on human monocytes: catalyticcells triggers cell proliferation. uPA-(12134)-peptide is respon- efficiency and susceptibility to inactivation by plasminogen acti-sible for growth-factor-like activity in some cells. Fucosylation vator inhibitors,Blood 74, 139621402.

8. Kirchheimer, J. C. & Remold, H. G. (1989) Endogenous receptor-within the growth-factor-like domain may serve as a molecularbound urokinase mediates tissue invasion of human monocytes,trigger in eliciting this response [51, 52]. The addition of theJ. Immunol. 143, 263422639.hybrid protein also stimulates proliferation of tumor cells, as

9. Ploug, M., Rønne, E., Behrend, N., Jensen, A. L., Blasi, F. & Dano,confirmed by thymidine-uptake and cell-growth assays (Schmitt,K. (1991) Cellular receptor for urokinase plasminogen activator:

M., personal communication). However, the hybrid protein did carboxyl-terminal processing and membrane anchoring by glyco-not significantly stimulate proliferation of tumor cells in anin sylphosphatidylinositol,J. Biol. Chem. 266, 192621933.vivo animal model. The viability of tumor cells was unaffected10. Behrendt, N., Ronne, E. & Dano, K. (1995) The structure and func-by the hybrid protein. Our data and previous observations tion of the urokinase receptor, a membrane protein govering plas-

minogen activation on the cell surface,Biol. Chem. Hoppe-Seylerindicate that the hybrid protein has little cytotoxicity toward hu-376, 2692279.man and mouse cancer cells, fibroblasts and endothelial cells

11. Crowley, C. W., Cohen, R. L., Lucas, B. K., Liu, G., Shuman, M.(Kobayashi, H., unpublished data).A. & Levinson, A. D. (1993) Prevention of metastasis by inhibi-We have to consider the possibility that tumor cells develoption of the urokinase receptor,Proc. Natl Acad. Sci. USA 90,

back-up systems that might render the invasion and proliferation 502125025.process of tumor cells in cancer independent of the uPA/uPAR12. Schmitt, M., Wilhelm, O., Janicke, F., Magdolen, V., Reuning, U.,system. Nevertheless, in some human tumor cell lines, inhibition Ohi, H., Moniwa, N., Kobayashi, H., Weidle, U. & Graeff, H.of tumor metastasis was demonstrated by animal experiments. (1995) Urokinase-type plasminogen activator (uPA) and its recep-

tor (CD87) : a new target in tumor invasion and metastasis,J.Thus, uPA-(12134)-UTI-(782136) or peptide analogues withObstet. Gynecol. 21, 1512165.higher affinities might provide a therapeutically promising basis

13. Kobayashi, H., Ohi, H., Shinohara, H., Sugimura, M., Fujii, T.,for the prevention of tumor cell invasion and metastasis. ThisTerao, T., Schmitt, M., Goretzki, L., Chucholowski, N., Jänicke,inhibiting effect is postulated to occur via binding to uPAR. TheF. & Graeff, H. (1993) Saturation of tumour cell surface receptorscurrent study indicates that the anti-metastatic activity of the for urokinase-type plasminogen activator by amino-terminal frag-

hybrid protein should be explored further. In humans, there will ment and subsequent effect on reconstituted basement membranesbe competition for binding of the hybrid protein when injected invasion,Br. J. Cancer 67, 5372544.to humans, although the human uPA-(12134)-peptide molecule 14. Kobayashi, H., Gotoh, J., Fujie, M., Shinohara, H., Moniwa, N. &

Terao, T. (1994) Inhibition of metastasis of Lewis lung carcinomawill not bind to mouse host cells when injected into animals.by a synthetic peptide within growth factor-like domain of uro-Assessment of systemic absorption is important in the preclinicalkinase in the experimental and spontaneous metastasis model,Int.testing of ATFHI because the uPAR is abundantly expressed onJ. Cancer 57, 7272733.monocytes, macrophages, or kidney, and the kidney may be the

15. Wilhelm, O., Werdle, U., Hohl, S., Rettenberger, P., Schmitt, M. &dose-limiting organ of toxicity for the intravenous administra- Graeff, H. (1994) Recombinant soluble urokinase receptor as ation of uPA-containing recombinant proteins. It would be inter- scavenger for urokinase-type plasminogen activator (uPA). Inhibi-esting to construct a mouse uPA/UTI hybrid protein and test the tion of proliferation and invasion of human ovarian cancer cells,ability of such a molecule to inhibit metastasis of mouse carci- FEBS Lett. 337, 1312134.

16. Kobayashi, H., Fujie, M., Shinohara, H., Ohi, H., Sugimura, M. &noma cells in mice. In this situation, studies of how the hybridTerao, T. (1994) Effects of urinary trypsin inhibitor on the inva-protein is targetted in the animals would be interesting. Thesion of reconstituted basement membranes by ovarian cancerpharmacokinetic evaluation of hybrid proteins should be studiedcells, Int. J. Cancer 57, 3782384.in animals and humans. Our current efficacy study suggests

17. Kobayashi, H., Shinohara, H., Takeuchi, K., Itoh, M., Fujie, M.,strongly that further studies should be pursued to test hybridSaitoh, M. & Terao, T. (1994) Inhibition of the soluble and the

proteins for their potential anti-metastatic properties in human tumor cell receptor-bound plasmin by urinary trypsin inhibitorclinical trials. and subsequent effects on tumor cell invasion and metastasis,

Cancer Res. 54, 8442849.18. Kobayashi, H., Shinohara, H., Ohi, H., Sugimura, M., Terao, T. &

Fujie, M. (1994) Urinary trypsin inhibitor (UTI) and fragmentsREFERENCESderived from UTI by limited proteolysis efficiently inhibit tumor

1. Dano, K., Andreasen, P. A., Grondahl-Hansen, J., Kristensen, P. I., cell invasion,Clin. Exp. Metastasis 12, 1172128.Nielsen, L. S. & Skriver, L. (1985) Plasminogen activators, tissue19. Kobayashi, H., Shinohara, H., Gotoh, J., Fujie, M., Fujishiro, S. &degradation, and cancer,Adv. Cancer Res. 44, 1392266. Terao, T. (1995) Anti-metastatic therapy by urinary trypsin inhibi-

2. Stoppelli, M. P., Tacchetti, C., Cubellis, M. V., Corti, A., Hearing, tor in combination with an anti-cancer agent,Br. J. Cancer 72,V. J., Cassani, G., Appella, E. & Blasi, F. (1996) Autocrine satura- 113121137.tion of pro-urokinase receptors on human A431 cells, Cell 45, 20. Kobayashi, H., Gotoh, J., Hirashima, Y., Fujie, M., Sugino, D. &

Terao, T. (1995) Inhibitory effect of a conjugate between human6752684.


urokinase and urinary trypsin inhibitor on tumor cell invasionin 38. Ballance, D. J., Marshall, J. M., Cottingham, I. R., Steven, J., Berry,vitro, J. Biol. Chem. 270, 836128366. S., Cederholm-Williams, S. A., Goodey, A. R. & Courtney, M.

21. Kobayashi, H., Gotoh, J., Kanayama, N., Hirashima, Y., Terao, T. & (1992) A hybrid protein of urokinase growth-factor domain andSugino, D. (1995) Inhibition of tumor cell invasion through plasminogen activator inhibitor type 2 inhibits urokinase activityMatrigel by a peptide derived from the domain II region in urinary and binds to the urokinase receptor,Eur. J. Biochem. 207, 1772trypsin inhibitor,Cancer Res. 55, 184721852. 183.

22. Kobayashi, H., Gotoh, J., Shinohara, H., Moniwa, N. & Terao, T.39. Cavallaro, U., delVecchio, A., Lappi, D. A. & Soria, M. R. (1993)(1994) Inhibition of the metastasis of Lewis lung carcinoma by A conjugate between human urokinase and saporin, a type-1 ribo-antibody against urokinase-type plasminogen activator in the some-inactivating protein, is selectively cytotoxic to urokinaseexperimental and spontaneous metastasis model,Thromb. receptor-expressing cells,J. Biol. Chem. 268, 23186223190.Haemost. 71, 4742480.

40. Xiang, J., Koropatnick, J., Qi, Y., Luo, X., Moyana, T., Li, K. &23. Kobayashi, H., Gotoh, J., Hirashima, Y. & Terao, T. (1996) Expres-Chen, Y. (1993) Production of a bifunctional hybrid moleculesion of urokinase and urinary trypsin inhibitor in metastatic andB72.3/metallothionein-1 by protein engineering,Immunology 78,non-metastatic murine Lewis lung carcinoma cells,Int. J. Oncol.5742581.8, 7572763.

41. Sarosdy, M. F., Hutzler, D. H., Yee, D. & von Hoff, D. D. (1993)24. Hollas, W., Blasi, F. & Boyd, D. (1991) Role of the urokinase recep-In vitro sensitivity of human bladder cancers and cell lines totor in facilitating extracellular matrix invasion by cultured colonTP-40, a hybrid protein with selective targeting and cytotoxicity,cancer,Cancer Res. 51, 369023695.J. Urol. 150, 195021955.25. Guesdon, J. L., Ternynck, T. & Avrameas, S. (1979) The use of

42. Sforzini, S., Bolognesi, A., Meazza, M., Marciano, S., Casalini, P.,avidin-biotin interaction in immunoenzymatic techniques,J.Histochem. Cytochem. 27, 113121139. Dürkop, H., Tazzari, P. L., Stein, H., Stirpe, F. & Ferrini, S. (1995)

26. Kobayashi, H., Schmitt, M., Goretzki, L., Chucholowski, N., Differential sensitivity of CD301 neoplastic cells to geloninCalvete, J., Kramer, M., Günzler, W. A., Jänicke, F. & Graeff, H. delivered by anti-CD30/anti-gelonin bispecific antibodies,Br. J.(1991) Cathepsin B efficiently activates the soluble and the tumor Haematol. 90, 5722577.cell receptor-bound form of the proenzyme urokinase-type43. Niedbala, M. J. & Stein-Picarella, M. (1993) Role of protein kinaseplasminogen activator (pro-uPA),J. Biol. Chem. 266, 51472 C in tumor necrosis factor induction of endothelial cell urokinase-5152. type plasminogen activator,Blood 81, 260822607.

27. Kobayashi, H., Moniwa, N., Sugimura, M., Shinohara, H., Ohi, H. &44. Waltz, D. A. & Chapman, H. A. (1994) Reversible cellular adhesionTerao, T. (1993) Effects of membrane-associated cathepsin B on to vitronectin linked to urokinase receptor occupancy,J. Biol.the activation of receptor-bound prourokinase and subsequent

Chem. 269, 14 746214750.invasion of reconstituted basement membranes,Biochim. Biophys.45. Olson, D., Pöllänen, J., Høyer-Hansen, G., Rønne, E., Sakaguchi,Acta 1178, 55262.

K., Wun, T.-C., Appella, E., Dano, K. & Blasi, F. (1992) Internal-28. Puri, R. K., Leland, P., Kreitman, R. J. & Pastan, I. (1994) Humanization of the urokinase-plasminogen activator inhibitor type-1neurological cancer cells express interleukin-4 (IL-4) receptorscomplex is mediated by the urokinase receptor,J. Biol. Chem.which are targets for the toxic effects of IL4-pseudomonas exo-267, 912929133.toxin chimeric protein,Int. J. Cancer 58, 5742581.

46. Nykjær, A., Petersen, C. M., Møller, B., Jensen, P. H., Moestrup, S.29. Kariko, K., Kuo, A., Boyd, D., Okada, S. S., Cines, D. B. &K., Holtet, T. L., Etzerodt, M., Thøgersen, H. C., Munch, M.,Barnathan, E. S. (1993) Overexpression of urokinase receptor in-

creases matrix invasion without altering cell migration in a human Andreasen, P. A. & Gliemann, J. (1992) PurifiedA2-macroglobu-osteosarcoma cell line,Cancer Res. 53, 310923117. lin receptor-LDL receptor-related protein binds urokinase/plas-

30. Hochstrasser, K., Schonberger, O. L., Rossmanith, I. & Wachter, E. minogen activator inhibitor type-1 complex,J. Biol. Chem. 267,(1981) Kunitz-type protease inhibitors derived by limited proteol- 14543214546.ysis of inter-A-trypsin inhibitor, V, Hoppe-Seyler Z. Physiol. 47. Li, H., Kuo, A., Kochan, J., Strickland, D., Kariko, K., Barnathan, E.Chem. 362, 135721362. S. & Cines, D. B. (1994) Endocytosis of urokinase-plasminogen

31. Wachter, E. & Hochstrasser, K. (1981) Kunitz-type protease inhibi- activator inhibitor type1 complexes bound to a chimeric trans-tors derived by limited proteolysis of the inter-alpha-trypsin inhib- membrane urokinase receptor,J. Biol. Chem. 269, 815328158.itor, IV, Hoppe-Seyler’s Z. Physiol. Chem. 362, 135121355. 48. Conese, M., Olson, D. & Blasi, F. (1994) Protease nexin-1-urokinase

32. Albrecht, G. J., Hochstraser, K. & Salier, J.-P. (1983) Kunitz-typecomplexes are internalized and degraded through a mechanismprotease inhibitors derived by limited proteolysis of the inter-al-that requires both urokinase receptor andA2-macroblobulin recep-pha-trypsin inhibitor, IX,Hoppe-Seyler’s Z. Physiol. Chem. 364,tor, J. Biol. Chem. 269, 17886217892.170321708.

49. Nykjær, A., Kjøller, L., Cohen, R. L., Lawrence, D. A., Garni-33. Reisinger, P., Hochstrasser, K., Albrecht, G. J., Lempart, K. & Salier,Wagner, B. A., Todd, R. F. III, van Zonneveld, A.-J., Gliemann,J.-P. (1985) Human inter-alpha-trypsin inhibitor : localization ofJ. & Andreasen, P. A. (1994) Regions involved in binding of uro-the Kunitz-type domains in the N-terminal part of the moleculekinase-type-1 inhibitor complex and pro-urokinase to the endo-and their release by a trypsin-like proteinases,Biol. Chem. Hoppe-

Seyler 366, 4792483. cytic A2-macroblobulin receptor/low density lipoprotein receptor-34. Balduyck, M., Laroui, S., Mizon, C. & Mizon, J. A. (1989) A pro- related protein,J. Biol. Chem. 269, 25668225676.

teoglycan related to the urinary teypsin inhibitor (UTI) links the50. Cones, M., Nykjær, A., Petersen, C. M., Cremona, O., Pardi, R.,two heavy chains of inter-alpha-trypsin inhibitor,Biol. Chem. Andreasen, P. A., Gliemann, J., Christensen, E. I. & Blasi, F.Hoppe-Seyler 370, 3292336. (1995) A2-macroblobulin receptor/Ldl receptor-related protein

35. Hochstrasser, K., Albrecht, G. J., Schonberger, O. L. & Wachter, E. (Lrp)-dependent internalization of the urokinase receptor,J. Cell(1989) Kunitz-type protease inhibitors derived by limited proteol- Biol. 131, 160921622.ysis of the inter-alpha-trypsin inhibitor, VIII,Hoppe-Seyler’s Z. 51. Rabbani, S. A., Mazar, A. P., Bernier, S. M., Haq, M., Bolivar, I.,Physiol. Chem. 370, 168921696.

Henkin, J. & Geltzman, D. (1992) Structural requirments for the36. Kobayashi, H., Gotoh, J., Fujie, M. & Terao, T. (1994) Characteriza-growth factor activity of the amino-terminal domain of urokinase,tion of the cellular binding site for the urinary trypsin inhibitor,J. Biol. Chem. 267, 14151214156.J. Biol. Chem. 269, 20642220647.

52. Sreerama, S., Kumar, A., Johnson, A., Purblitz, S. & Idell, S. (1995)37. Kobayashi, H., Gotoh, J., Hirashima, Y. & Terao, T. (1996) Inter-A-Urokinase receptor in human malignant mesothelioma cells: roletrypsin inhibitor bound to tumor cells is cleaved into the heavyin tumor cell mitogenesis and proteolysis,Am. J. Physiol. 268,chains and the light chain on the cell surface,J. Biol. Chem. 271,

11362211367. L9722L982.

a bifunctional hybrid molecule of the amino-terminal fragment of urokinase and domain ii of bikunin...

Documents