Eukaryotic Expression Cloning with an Antimetastatic

[CANCER RESEARCH 59, 3812–3820, August 1, 1999] Eukaryotic Expression Cloning with an Antimetastatic Monoclonal Antibody Identifies a Tetraspanin (PETA-3/CD151) as an Effector of Human Tumor Cell Migration and Metastasis1

Jacqueline E. Testa,2,3 Peter C. Brooks,4 Jian-Min Lin,5 and James P. Quigley3 Department of Pathology, State University of New York at Stony Brook, Stony Brook, New York 11794

ABSTRACT interactions with different matrix proteins. Our efforts to identify novel metastasis-associated antigens have, therefore, focused on the tumor cell A monoclonal antibody (mAb), 50-6, generated by subtractive immuniza- surface. As our model system, we have used a highly metastatic human tion, was found to specifically inhibit in vivo metastasis of a human epider- epidermoid carcinoma cell line, HEp-3, which disseminates to host lung moid carcinoma cell line, HEp-3. The cDNA of the cognate antigen of mAb 50-6 was isolated by a modified eukaryotic expression cloning protocol from tissue. The distinct characteristics and behavior patterns of this tumor cell a HEp-3 library. Sequence analysis identified the antigen as PETA-3/CD151, line have been described by Ossowski in a series of papers published a recently described member of the tetraspanin family of proteins. The cloned between 1980 and 1998 (7–9). These cells are very aggressive and readily antigen was also recognized by a previously described antimetastatic anti- give rise to metastasizing tumors in both the chicken embryo (10–12) and body, mAb 1A5. Inhibition of HEp-3 metastasis by the mAbs could not be in the nude mouse model (13, 14). As such, these cells should possess attributed to any effect of the antibodies on tumor cell growth in vitro or in distinct surface antigens that are functionally involved in mediating tumor vivo. Rather, the antibodies appeared to inhibit an early step in the formation cell dissemination. of metastatic foci. In a chemotaxis assay, HEp-3 migration was blocked by Brooks et al. (11) generated several mAbs6 against HEp-3 cell surface both antibodies. HeLa cells transfected with and overexpressing PETA-3/ CD151 were more migratory than control transfectants expressing little proteins using an approach termed subtractive immunization. Their pro- CD151. The increase in HeLa migration was inhibitable by both mAb 50-6 tocol allowed them to produce mAbs with no preconceived notion as to and mAb 1A5. PETA-3 appears not to be involved in cell attachment because the identity or function of the targeted antigen. Two of the antibodies, adhesion did not correlate with levels of PETA-3 expression and was unaf- DM12-4 and 1A5, inhibited spontaneous HEp-3 metastasis in the chicken fected by mAb 50-6 or mAb 1A5. The ability of PETA-3 to mediate cell embryo metastasis assay by 86 and 90%, respectively. Neither antibody migration suggests a mechanism by which this protein may influence metas- affected primary tumor growth on the chorioallantoic membrane or tumor tasis. These data identify PETA-3/CD151 as the first member of the tet- cell growth in vitro, indicating that the mAbs specifically blocked met- raspanin family to be linked as a positive effector of metastasis. astatic behavior. The identification of the antigens recognized by the mAbs was not reported or was unknown. INTRODUCTION In the present study, another monoclonal antibody generated by Metastasis is a complex, multistep cascade of cellular events including subtractive immunization, mAb 50-6, was used to clone and charac- migration of tumor cells through the surrounding stroma, entry into the terize a cell surface, metastasis-associated antigen expressed on circulatory system, and finally arrest, extravasation, and growth at a HEp-3 cells. This antibody inhibits both spontaneous and experimen- distant secondary site (reviewed in Refs. 1–6). Given the complexity of tal HEp-3 metastasis. Eukaryotic expression cloning of the antigen the metastatic process, it is not surprising that a number of proteins have identifies it as PETA-3/CD151, a member of the tetraspanin family of been associated with tumor cell dissemination including transcription proteins. We show that PETA-3/CD151 appears to be required at an factors, signaling proteins, adhesion molecules, proteases, motility fac- early step in the formation of metastatic foci. Furthermore, this protein tors, and others (1–6). Although nuclear, cytoplasmic, and secreted mediates tumor cell migration but does not appear to affect cell proteins have been associated with metastatic potential, tumor cell dis- adhesion to various purified matrix proteins. The work described semination is executed via the physical interactions of the cancer cell herein identifies PETA-3/CD151 as the first member of the tet- surface with various host tissue elements. Not only is the cell membrane raspanin family to be linked as a positive effector of metastasis. the interface at which cell-cell and cell-substrate contacts are made, it is the portal through which external signals must pass. Activation of cell surface receptors, by mutation or by ligand binding, initiates intracellular MATERIALS AND METHODS signaling cascades that influence expression of genes that promote the Cell Lines and Hybridomas. Human breast adenocarcinoma cells (MDA- malignant phenotype (2, 3, 6). Proteases expressed on or bound to the cell MB-231), human cervical carcinoma (HeLa), human fibrosarcoma (HT1080), membrane mediate degradation of tissue barriers. Integrins mediate tu- and monkey kidney cells (COS-7) were obtained from the American Type mor cell motility and transmit environmental cues by virtue of their Culture Collection (Rockville, MD). Metastatic human epidermoid carcinoma cells (HEp-3) were obtained from solid tumors serially passaged on the CAMs Received 1/14/99; accepted 6/2/99. of chicken embryos (10, 11). All cells were maintained as monolayer cultures The costs of publication of this article were defrayed in part by the payment of page in DMEM (Life Technologies, Inc., Gaithersburg, MD) supplemented with charges. This article must therefore be hereby marked advertisement in accordance with 10% FBS (HyClone, Logan, UT), sodium pyruvate, penicillin/streptomycin, 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by Grants RO1 CA60800 (to J. E. T.) and RO1 CA65660 and nonessential amino acids (Life Technologies, Inc.; growth medium). (to J. P. Q) from the National Cancer Institute at the NIH. Cultures were grown in a humidified atmosphere of 5% CO2 at 37°C. 2 To whom requests for reprints should be addressed, at Department of Vascular Hybridomas producing mAb 50-6 and mAb 1A5 were generated by sub- Biology, VB-1, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037. Phone: (619) 784-7188; Fax: (619) 784-7323; E-mail: [email protected]. tractive immunization (11). Cultures of each hybridoma were maintained in 3 Present address: Department of Vascular Biology, VB-1, The Scripps Research one part DMEM, one part Hybridoma SFM (Life Technologies, Inc.), supple- Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037. mented with 2.5% alpha calf serum (HyClone), sodium pyruvate, penicillin/ 4 Present address: Department of Biochemistry and Molecular Biology, Norris Cancer Center, Topping Tower, Room 5409, 1441 Eastlake Avenue, University of Southern California, Los Angeles, CA 90033. 6 The abbreviations used are: mAb, monoclonal antibody; CAM, chorioallantoic 5 Present address: Matrix Pharmaceuticals, Inc., 34700 Campus Drive, Fremont, CA membrane; FBS, fetal bovine serum; FN, fibronectin; LN, laminin; TM4SF, transmem- 94555. brane 4 superfamily; VN, vitronectin. 3812

membrane was made transparent with a drop of paraffin oil, and 200 ␮g of purified mAb or normal mouse IgG (Sigma Chemical Co., St. Louis, MO) in 0.1 ml PBS were inoculated into the blood vessel with a 30-gauge needle. The window was sealed, and after an additional 6 days of incubation, the eggs were opened; the primary tumors were excised, trimmed of CAM tissue, and weighed as a measure of tumorigenicity. The lungs of the embryos were removed, finely minced, and passaged onto the CAMs of a second set of 10-day-old embryos. These embryos were incubated for an additional 7 days to allow any HEp-3 cells in the lungs to multiply. The “lung tumors” arising from the transferred lungs were then excised and finely minced, and the presence of HEp-3 was determined biochemically by quantitating human urokinase-type plasminogen activator activity present in de- tergent extracts of the lung tumors (8–11). Antibody inhibition of HEp-3 experimental metastasis was determined by coinoculating 0.1 ml of PBS containing tumor cells (2.0 ϫ 104) and 200 ␮gof purified mAb or normal mouse IgG (Sigma) directly into a prominent blood vessel (prepared as described above). For the time course study of inhibition of HEp-3 experimental metastasis, the antibodies were inoculated at different times before or after inoculation of the tumor cells, as indicated. The inoculated embryos were incubated for an additional 6 days, after which the lungs were excised, finely minced, and transferred to prepared CAMs of a second set of embryos. The assay was then completed as described above. Eukaryotic Expression Cloning. A custom-made, unidirectional cDNA library was constructed in the eukaryotic expression vector pcDNA I (Invitrogen, San Diego, CA) using poly(A)ϩ RNA isolated from metastatic HEp-3 cells. Eukaryotic expression cloning in COS monkey kidney cells was conducted as described previously (15) with some modifications. The first two rounds of transfection and immunoselection were performed as described except that COS cells were transfected using the calcium phosphate method (16). Plasmids recovered at the end of the second round were used to transfect COS cells growing on tissue culture plates. Twenty-four h later, the transfected cells were detached from the plates with nonenzymatic cell dissociation solution (Sigma) and plated onto polycarbonate membranes (90 mm diameter, 0.4 ␮ pore size; Millipore, Bedford, MA). After an additional 24 h, the cells (attached to the membranes) were washed three times with PBS, fixed with 0.25% glutaraldehyde in PBS for 5 min at room temperature, washed, quenched with 1.0 M glycine (pH 8.0) for two h at room temperature, washed again, and then incubated with 10% normal goat serum in PBS (blocking solution) for1hatroom temperature. The membranes were then incubated with mAb 50-6 (1 ␮g/ml in blocking solution) overnight at 4°C with gentle agitation. As a control, one membrane was incubated with an isotype- matched control antibody (IgG1; Sigma; 1 ␮g/ml in blocking solution). The primary antibody was removed by washing the membranes 3 ϫ 10 min in PBS, and the membranes were then incubated with biotin-conjugated goat anti-mouse

Fig. 1. mAb 50-6 recognizes a Mr 29,000 cell surface antigen. Live, unfixed HEp-3 IgG (Southern Biotechnology Associates, Birmingham, AL; diluted 1:500 in cells were immunostained with normal mouse IgG (a, NM) or mAb 50-6 (b) and analyzed blocking solution) for 1.5 h at room temperature with gentle agitation. The by flow cytometry. Horizontal axis, fluorescence intensity; vertical axis, numbers of cells membranes were washed and incubated with horseradish peroxidase-conjugated analyzed. HEp-3 lysates were also analyzed by Western blotting (c). Proteins (10 ␮g) were resolved by SDS-PAGE, transferred to nitrocellulose, then incubated with mAb 50-6 streptavidin (Southern Biotechnology Associates; diluted 1:500 in blocking solu- or normal mouse IgG. The signal was visualized with a peroxidase-conjugated goat tion) for 45 min at room temperature with gentle agitation. After three washes in anti-mouse IgG by chemiluminescence. Left, molecular weight standards (in thousands). PBS, the membranes were developed with chloronaphthol to identify immunopositive COS transfectants. No color reaction was seen on cells incubated with control IgG1. With the aid of a dissecting microscope and a flame-drawn glass streptomycin, and nonessential amino acids (Life Technologies, Inc.). Cultures microcapillary pipette, ϳ30 strongly immunopositive cells were detached from the were grown in spinner flasks in a humidified atmosphere of 5% CO2 at 37°C. membranes and transferred to a microcentrifuge tube. Episomal plasmid DNA was mAb Purification. Conditioned media from the hybridoma cultures were recovered from these cells as described previously (17) and used to transform centrifuged at 5000 ϫ g for 20 min and then pumped over a column of bacteria (MC1061/p3; Invitrogen, San Diego, CA) by electroporation. The result- GammaBind Plus Sepharose (Amersham Pharmacia Biotech, Piscataway, NJ). ing colonies were pooled, grown in liquid culture, and plasmid DNA was isolated The columns were washed with 10 column volumes of PBS, and the mAbs with a Qiagen Plasmid kit (Qiagen, Chatsworth, CA). Plasmid DNA was frac- were eluted with 0.1 M glycine (pH 3.0). Purified mAbs were dialyzed against tionated by resolving 1 ␮g on an agarose gel. The lane was cut into six segments, PBS, filter sterilized, then aliquoted and stored at Ϫ20°C. and DNA was isolated from each with a Gene Clean kit (Bio 101, La Jolla, CA) Effect of mAb 50-6 on HEp-3 Growth in Vitro. HEp-3 cells were plated and used to electroporate MC1061/p3 (Invitrogen). Plasmids were isolated from into six-well culture plates (2.0 ϫ 105 cells/well) in the presence of 50 ␮g/ml cultures of bacterial cells transformed with DNA from each of the six gel slices and of mAb 50-6 or normal mouse IgG. At 24, 48, and 72 h, the cells in two wells the fraction containing cDNA clones which directed the synthesis of a cell surface from each culture condition were trypsinized and counted. antigen recognized by mAb 50-6 was identified by transfecting and immuno- Inhibition of HEp-3 Metastasis in the Chicken Embryo Assay. Antibody staining COS cells (growing on poly-L-lysine-coated coverslips) as described inhibition of HEp-3 spontaneous metastasis in the chicken embryo was conducted above. DNA from one positive gel fraction was used to transform bacteria, and as described previously (11). Briefly, tumor cells were inoculated through a mini-prep DNA from 10 individual colonies was used to transfect COS cells for window in the eggshell onto the surface of CAMs of 10-day-old chicken embryos immunocytochemical analysis. Immunopositive cells were detected in 1 of the 10 (SPAFAS, Preston, CT). The window was sealed, and the embryos were returned transfected cultures. The positive clone contained an insert of ϳ1.5 kb, as deter- to the incubator. Twenty-four h later, a second window was carefully cut in the mined by restriction endonuclease digestion. The insert was sequenced with the T7 eggshell directly over a prominent blood vessel. The underlying, nonliving shell Sequenase Quick-Denature plasmid sequencing kit (version 2.0; Amersham Phar- 3813

Table 1 Inhibition of HEp-3 metastasis by mAb 50-6 Inhibition of spontaneous metastasis was determined by Brooks et al. (11), inhibition of experimental metastasis was determined by i.v. inoculating 11-day-old chicken embryos with 0.1 ml of PBS containing mAb 50-6 (200 ␮g) and HEp-3 cells (2 ϫ 104). The inoculated embryos were incubated for an additional 7 days, at which time the embryonic lungs were removed, finely minced, and transferred to the CAMs of a second set of embryos, which were incubated for an additional 7 days. The extent of lung metastasis in both assays was measured by determining the amount of human urokinase-type plasminogen activator activity, which has been shown to be a quantitative reflection of the numbers of human cells in the lung (8–10). Data are the combined results of two separate spontaneous assays and two separate experimental metastasis assays. Primary tumor weighta Lung metastasis (huPA in lung) Antibody No. of embryos (mean Ϯ SEM) (mean Ϯ SEM) % inhibition Spontaneous metastasis Normal mouse IgG 11 274.1 (Ϯ 51.2) 1015.8 (Ϯ 249.4)b 0 mAb 50-6 13 253.9 (Ϯ 46.7) 261.2 (Ϯ 117.2)b 74.30% Experimental metastasis Normal mouse IgG 11 2169.3 (Ϯ 601.8)c 0 mAb 50-6 10 937.8 (Ϯ 299.9)c 56.80% a P ϭ 0.689. b P ϭ 0.022. c P ϭ 0.044. macia Biotech, Arlington Heights, IL). The nucleotide sequence was compared in a 96-well culture plate and cultured for 36 h. The cells were then washed three with the National Center for Biotechnology Information database. times with PBS, fixed with 0.25% glutaraldehyde in PBS for 5 min at room Transfection of HeLa Cells with the Cloned cDNA. HeLa cells were temperature, washed again, then quenched with 1.0 M glycine (pH 8.0) for2hat cotransfected with the cloned PETA-3 cDNA and pSV2neo using the calcium room temperature. The plates were washed three times with PBS and used phosphate method (16). Controls were cotransfected with vector (pcDNA I) immediately or stored in 0.1% sodium azide in PBS at 4°C. Stored plates were containing no insert and pSV2neo. Forty-eight h later, Geneticin (G418; Life washed three times with PBS prior to use to remove the sodium azide. Technologies, Inc.) was added at a concentration of 400 ␮g/ml of medium, and the For the assay, wells were incubated with 0.2 ml of blocking solution overnight cultures were incubated for an additional 12 days. The resulting G418-resistant at 4°C. Purified mAb or isotype-matched control antibody (0.1 ml, 1 ␮g/ml in colonies were detached from the culture plates with nonenzymatic cell dissociation blocking solution) was added to the appropriate wells and incubated for2hat solution (Sigma), pooled, washed three times with serum-free DMEM, then room temperature. The plates were washed three times with PBS, and then blocked with 10% normal goat serum in PBS (blocking solution), on ice, for 30 horseradish peroxidase-conjugated goat anti-mouse IgG (Southern Biotechnology min. The cells were then incubated with 1 ␮g/ml mAb 50-6 or an isotype-matched Associates) was added to the wells (0.1 ml, 1 ␮g/ml in blocking solution) and control (IgG1; Sigma) for1honice, washed three times with blocking solution, incubated for2hatroom temperature. The plates were washed, and 0.1 ml of the then incubated with a phycoerythrin-conjugated goat anti-mouse IgG (Southern substrate, o-phenylene diamine (0.34 mg/ml, 0.1 M sodium citrate, pH 4.5, 0.012% Biotechnology Associates; diluted 1:500 in blocking solution) for 30 min on ice in H2O2) was added. After a 10-min incubation at 37°C, the plates were read at 405 the dark. The labeled cells were washed three times in PBS, and overexpressing nm using a Titer Tek Multiscan plate reader. The nonspecific signal from the PETA-3 transfectants were isolated by a fluorescence activated cell sorter. Control isotype-matched control was subtracted from the experimental wells. Cell surface transfectants were also selected with a fluorescence-activated cell sorter. The cells ␣ ␤ levels of the 3 1 integrin on HEp-3, HeLa, MDA-MB-231, and HT1080 cells collected were and then subcloned by limiting dilution. Approximately 50 sub- were measured in the same manner, using mAb 1992 (Chemicon, Temecula, CA), clones isolated from each of the two populations of cells were expanded in culture, which specifically recognizes the heterodimer. then rescreened by whole-cell ELISA to verify the levels of PETA-3 expression. Migration (Chemotaxis) Assay. HEp-3 cells were detached from culture In vitro growth rates of the HeLa transfectants were determined by plating 2 ϫ 104 plates with Versene (Life Technologies, Inc.), washed twice with serum-free cells into each of two wells in a 24-well plate. At 48-, 72-, and 96-h time points, DMEM (Life Technologies, Inc.), and resuspended in AIM-V medium (Life the cells were trypsinized and counted. Results are reported as numbers of Technologies, Inc.). Cells (1.4 ϫ 104) were added to the upper reservoir of cells/well. BioCoat control chambers (uncoated; 8 ␮m pore size; Becton Dickinson Whole-Cell ELISA. The levels of cell-surface PETA-3 expression were Labware, Bedford, MA) in AIM-V medium, and 50 ␮g/ml of mAb 50-6, mAb measured by whole-cell ELISA. Subconfluent cultures HEp-3 or HeLa cells were 1A5, or normal mouse IgG (Sigma). The lower reservoirs contained DMEM detached from the culture plates, washed three times in serum-free DMEM, then supplemented with 10% FBS (HyClone) and 50 ␮g/ml of the appropriate resuspended in growth medium. Cells (2.0 ϫ 104/0.1 ml) were added to each well antibody. After 6 or 12 h of incubation, the microporous inserts were fixed with 10% neutral buffered formalin and stained with hematoxylin. Cells on the upper surfaces of the membranes were removed with cotton swabs, and the membranes were excised and mounted on microscope slides in Permount. Cells on the underside of one quadrant of each filter were counted. Experi- ments were conducted in triplicate. Chemotaxis assays with the HeLa transfectants were similarly conducted, except that 2.0 ϫ 104 cells were added to the upper reservoirs, and the experiments were terminated after 18 h. Western Blot Analysis. Cells were lysed in Triton X-100 lysis buffer [0.5% Triton X-100, 0.1 M Tris (pH 8.0), 5 mM EDTA, 10 ␮M E64, 20 units/ml aprotinin, and 20 ␮g/ml soybean trypsin inhibitor (all from Sigma)] on ice for 10 min with vortexing at 5-min intervals. Insoluble material was removed by centrifugation at 12,000 ϫ g for 5 min at 4°C. The protein concentration of the cleared lysates was measured with the bicinchoninic acid system (Pierce Chemical Co., Rockford, IL). Proteins were resolved on 10% SDS-PAGE gels and then transferred to nitrocellulose. The blots were blocked with a solution of 5% nonfat milk, 5% FBS, and 0.1% Tween 20 in PBS (Western blocking solution) for1hatroom temperature. The blots were then incubated with either mAb 50-6, mAb 1A5, or normal mouse IgG (1 ␮g/ml in Western blocking solution) overnight at 4°C. The blots were washed three times for 5 Fig. 2. Effect of mAb 50-6 on HEp-3 growth in vitro. HEp-3 cells were plated into min with 0.1% Tween 20 in PBS and then incubated with horseradish perox- six-well culture plates (2.0 ϫ 105 cells/well) in the presence of 50 ␮g/ml of mAb 50-6 or normal mouse IgG (NM). At 24, 48, and 72 h, the cells in two wells from each culture idase-conjugated goat anti-mouse IgG (diluted 1:2500 in Western blocking condition were trypsinized and counted. Data are reported as numbers of cells/well. solution) for2hatroom temperature. The blots were then washed three times 3814

Fig. 3. Immunocytochemical localization of PETA-3 on HEP-3 cells and COS transfectants. Cells were plated onto poly-lysine-coated glass coverslips, fixed 24 h later, and then immunostained with mAb 50-6 or normal mouse IgG. Signals were detected with horseradish peroxidase-conjugated goat anti-mouse IgG and chloronaphthol. All images were photographed with trans- mitted light. a, HEp-3 cells incubated with mAb 50-6. b, HEp-3 cells incubated with normal mouse IgG. c, COS cells transiently transfected with PETA-3 cDNA and incubated with mAb 50-6. d, COS cells transiently transfected with PETA-3 cDNA and incubated with normal mouse IgG. e, COS transiently transfected with vector (pcDNA I) alone and incubated with mAb 50-6. Bars, 20 ␮m.

for 10 min with 0.1% Tween 20 in PBS, and the signals were visualized with tration of 2 ϫ 105 cells/ml in PBS/0.1% BSA containing 50 ␮g/ml of mAb the ECL system (Amersham Pharmacia Biotech, Arlington Heights, IL) ac- 50-6, mAb 1A5, or normal mouse IgG (Sigma). After a 5-min incubation at cording to the manufacturer’s directions. room temperature, 0.1 ml of the cell suspension was added to the appropriate Immunoprecipitation. HEp-3 cells were in lysed in Brij lysis buffer [1% Brij well and incubated for 15 and 30 min at 37°C. At the end of the incubation ␮ 98, 25 mM HEPES (pH 7.5), 150 mM NaCl, 5 mM MgCl2,10 M E64, 20 units/ml period, unattached cells were removed by gently washing the plates three times aprotinin, and 20 ␮g/ml soybean trypsin inhibitor (all from Sigma)] for 1 h with with PBS. Adherent cells were fixed with 0.25% glutaraldehyde for1hat constant rocking at 4°C (18). Insoluble material was removed by centrifugation at room temperature, followed by three washes in PBS. Cell adhesion was ϫ ␮ 12,000 g for 5 min at 4°C. The lysate was precleared with GammaBind Plus quantitated adding 50 l of crystal violet (0.1% in H2O) to each well. After 10 Sepharose beads. Aliquots of the extract representing 107 cells were incubated min at room temperature, the plates were washed in PBS, and the dye with either 25 ␮g of mAb 1A5, 10 ␮g of mAb 1992 (Chemicon, Temecula, CA), incorporated by the attached cells was released by adding 0.1 ml of 10% acetic ␣ ␤ ␮ which specifically recognizes the 3 1 integrin heterodimer, or 25 g of normal acid, then quantitated by spectrophotometry in a Titer Tek Multiscan plate mouse IgG (Sigma). Fifty ␮l of packed GammaBind Plus Sepharose beads were reader at 595 nm. added to each sample, and the mixtures were incubated at 4°C overnight with constant rocking. The beads were then washed with the Brij lysis buffer, and the RESULTS immune complexes were eluted with Laemmli sample buffer at 95°C for 2 min.

The eluted proteins were resolved on a 10% SDS-PAGE gel, transferred to mAb 50-6 Recognizes a Mr 29,000 Cell Surface Antigen. MAb nitrocellulose. Because biotinylation of both mAb 1A5 and mAb 50-6 destroys 50-6 was generated by subtractive immunization (11) using intact their immunoreactivity, PETA-3 was detected by incubating the blots with unla- HEp-3 cells as the immunogen. Localization of the antigen to the cell beled mAb 1A5, followed by horseradish peroxidase-conjugated goat anti-mouse surface of HEp-3 cells was determined by whole-cell ELISA (not IgG as described above. shown) and flow cytometry (Fig. 1). When compared with cells Cell Attachment to Purified Matrix Proteins. For the cell attachment incubated with normal mouse IgG (Fig. 1a), the fluorescence intensity assays, 96-well plates precoated with purified FN were purchased from Becton Dickinson Labware (Bedford, MA). Purified VN (Becton Dickinson Labware, signal from live, nonpermeablized HEp-3 cells immunostained with Bedford, MA) was used to coat 96-well plates as described previously (19). mAb 50-6 (Fig. 1b) is significantly higher. The molecular weight of Cells were detached from culture plates with nonenzymatic cell dissociation this cell surface protein was determined by Western blot analysis. As solution (Sigma) and washed three times in PBS containing 0.1% heat dena- seen in Fig. 1c, mAb 50-6 recognizes a single broad protein band

tured (60°C for 30 min) BSA. The cells were then resuspended to a concen- having an apparent molecular weight of Mr 29,000. 3815

length HEp-3 cDNA library. COS cells were subjected to several rounds of transfection and immunoselection with mAb 50-6. In the final round of the cloning protocol, immunopositive cells showing a strong cell surface signal were selected for isolation of cDNA clones because these cells would more likely harbor full-length inserts (see “Materials and Meth- ods”). Indeed, a clone that directed the synthesis of a cell surface antigen recognized by mAb 50-6 was isolated and sequenced. The open reading frame of this clone encodes a core protein having a predicted molecular mass of 27.8 kDa. There are four distinct hydrophobic domains, and a single N-glycosylation site in a large hydrophilic region that separates the third and fourth hydrophobic domains. Comparison of the sequence of this clone with the National Center for Biotechnology Information database identified the metastasis-associated antigen as PETA-3/CD151, a recently described Member of the tetraspanin family of proteins, which is also known as the TM4SF. The coding sequence was identical to that reported by Fitter et al. (21). The cell surface distribution of PETA-3 on HEp-3 cells and COS cells transfected with the cloned cDNA was Fig. 4. PETA-3/CD151 is recognized by mAb 50-6 and mAb 1A5. COS-7 cells were transiently transfected with the cloned PETA-3/CD151 cDNA or vector (pcDNA I) alone. analyzed by immunocytochemistry. On HEp-3 cells incubated with mAb Lysates of HEp-3 cells and the transfected COS cells were resolved by SDS-PAGE and 50-6, there was a strong staining pattern over the entire surface of the cell analyzed by Western blotting. a, Western blot of HEp-3 cell lysate (10 ␮g; Lane 1) and membrane, including filopods extending from the cell body (Fig. 3a). The lysates from control (COS/pcDNA I; 20 ␮g; Lane 2) and PETA-3-transfectants (COS/ PETA-3; 20 ␮g; Lane 3) probed with mAb 50-6. b, Western blot of HEp-3 cell lysate (0.5 same cell surface staining pattern was seen on COS cells transiently ␮g; Lane 1) and lysates from control (COS/pcDNA I; Lane 2; 2.5 ␮g) and PETA-3 transfected with PETA-3 cDNA (Fig. 3c). No signal was observed on transfectants (COS/PETA-3; Lane 3; 2.5 ␮g) probed with mAb 1A5. In both panels, the signals were visualized with a peroxidase-conjugated goat anti-mouse IgG by chemilu- HEp-3 or PETA-3-transfected COS cells incubated with normal mouse minescence. Arrow, PETA-3, Mr 29,000. A Mr 25,000 signal is also seen in COS cell IgG (Fig. 3, b and d) or on control COS transfectants incubated with mAb lysates probed with both mAb 50-6 and mAb 1A5. Right, molecular weight markers. 50-6 (Fig. 3e). Western blot analysis of the cell lysates from PETA-3- transfected COS cells also demonstrated that the encoded protein comigrates with the M 29,000 antigen expressed on HEp-3 cells (Fig. 4A). mAb 50-6 Inhibits Spontaneous and Experimental Metastasis r Brooks et al. (11) had described an antimetastatic mAb, mAb 1A5, of HEp-3 Cells. To determine whether the M 29,000 cell surface r which also recognizes a M 29,000 HEp-3 cell surface antigen. To protein was involved in HEp-3 dissemination, the effect of mAb 50-6 on r spontaneous metastasis was measured using an in vivo model, the chicken determine whether this mAb also recognizes PETA, mAb 1A5 was embryo metastasis assay (11, 20). In the spontaneous metastasis assays, used to probe Western blots of COS cells transiently transfected with purified antibody (mAb 50-6 or normal mouse IgG) was inoculated i.v. the PETA-3/CD151 cDNA (Fig. 4B). No signal was observed in 24 h after metastatic HEp-3 cells were implanted onto prepared CAMs, lysates from control transfectants (Lane 2); however, mAb 1A5 rec- and the extent of tumor cell dissemination to the lungs was measured (see ognized PETA-3/CD151 expressed by the COS cells (Lane 3). A “Materials and Methods”). As shown in Table 1, spontaneous metastasis minor Mr 25,000 band, immunostained with mAbs 50-6 and 1A5, was was reduced by 74% in embryos that received mAb 50-6. The effect of also evident in lysates of transfected COS cells (Fig. 4) and in HEp-3 mAb 50-6 on experimental metastasis also was tested by coinoculating lysates (seen upon prolonged exposure of the Western blot; data not the tumor cells with the antibody directly into the circulatory system. In shown). This is likely the nonglycosylated form of the protein, be- these experiments, dissemination to the embryonic lungs was inhibited by cause cultivation of HEp-3 cells with tunicamycin or treatment of 57% when compared with controls. The antimetastatic properties of mAb HEp-3 cell lysates with N-glycanase generates a Mr 25,000 band that 50-6 cannot be attributed to a cytostatic or cytotoxic effect, because this is immunoreactive with mAb 1A5 (data not shown). This is smaller antibody had no effect on proliferation of HEp-3 cells in vitro (Fig. 2), nor than the predicted mass of 27.8 kDa, and may be due to the hydro- did it result in a decrease in the size of the primary tumor on the CAM phobic nature of the protein and/or the presence of disulfide bonds (Table 1). These results indicate that mAb 50-6 specifically blocks a resulting from 15 cysteine residues. These results demonstrate that step(s) in the metastatic cascade. both mAb 50-6 and mAb 1A5 recognize PETA-3 and suggest that the

Eukaryotic Expression Cloning Identifies the Mr 29,000 Metas- epitope(s) is resident in the protein core. tasis-associated Antigen as PETA-3/CD151. The antibody inhibition PETA-3/CD151 Mediates an Early Event in the Formation of studies demonstrated a functional role for the Mr 29,000 protein in Metastatic Foci. To examine the possible mechanisms by which mediating metastasis. To identify the protein, mAb 50-6 was used in a PETA-3/CD151 effects metastasis, a time course study of inhibition eukaryotic expression cloning strategy to isolate the cDNA from a full- of experimental metastasis was carried out. As shown in Table 2,

Table 2 Time course of inhibition of HEp-3 experimental metastasis Purified mAb 1A5 (180–200 ␮g) was inoculated i.v. into 11-day-old chicken embryos either 2 h before or 2, 6, 10, or 20 h after inoculation of 2 ϫ 104 HEp-3 cells. As a control, normal mouse IgG was inoculated 2 h before the HEp-3 cells. The inoculated embryos were incubated for an additional 6 days, at which time the embryonic lungs were removed, finely minced, and transferred to the CAMs of a second set of embryos, which were incubated for an additional 7 days. The extent of lung metastasis was measured by determining the amount of human urokinase-type plasminogen activator activity, which has been shown to be a quantitative reflection of the numbers of human cells in the lung (8–10). Lung metastasis (huPA in lung) Antibody Time of antibody inoculation No. of embryos mean Ϯ SEM % inhibition Normal mouse IgG Ϫ2 h 47 1522 Ϯ 54 0 mAb 1A5 Ϫ2 h 36 295 Ϯ 60 81 mAb 1A5 ϩ6 h 21 401 Ϯ 36 74 mAb 1A5 ϩ10 h 8 1471 Ϯ 36 3 mAb 1A5 ϩ20 h 20 1408 Ϯ 67 7 3816

Fig. 5. Inhibition of HEp-3 migration (chemotaxis) by mAb 50-6 and mAb 1A5. HEp-3 cells were detached from culture plate with Versene (Life Technologies, Inc.), washed twice with serum-free DMEM (Life Technologies, Inc.), and resuspended in AIM-V medium (Life Technologies, Inc.). Cells (1.4 ϫ 104) were added to the upper reservoir of the migration chambers in AIM-V medium and 50 ␮g/ml of mAb 50-6, mAb 1A5, or normal mouse IgG. The lower reservoir contained DMEM supplemented with 10% FBS and 50 ␮g/ml of the appropriate antibody. After6hofincubation, the microporous inserts were fixed with 10% neutral buffered formalin and stained with hematoxylin, and cells on the underside of one quadrant of each filter were counted. Experiments were conducted in triplicate; bars, SEM.

HEp-3 dissemination to the embryonic lungs was markedly inhibited when mAb 1A5 was administered as early as 2 h before or as late as 6 h after inoculation of the tumor cells. In contrast, there was little effect on HEp-3 colonization of the lungs when mAb 1A5 was administered 10 or 20 h after the tumor cells were inoculated. The inability of mAb 1A5 to inhibit lung colonization at the latter time points (when tumor cells would likely have extravasated already) indicates that the antibody does not block metastasis by inhibiting cell growth at the secondary sites. This is consistent with our observations that the anti-PETA-3 mAbs have no effect on HEp-3 growth in vivo or in vitro (Table 1; Fig. 2). These experiments suggest that PETA- 3/CD151 mediates HEp-3 dissemination by affecting an earlier event, such as cell adhesion to the vessel wall, extravasation, and/or tumor cell migration to selective sites of secondary growth. PETA-3/CD151 Is Involved in Cell Migration in Vitro. The tetraspanins have been described as molecular facilitators that influence cell adhesion and migration, presumably through their interactions with integrins (reviewed in Refs. 22–24), and it is possible that PETA-3/CD151 affects migration of HEp-3 cells. Therefore, the antimetastatic mAbs were tested for their ability to inhibit HEp-3 migration in a chemotaxis assay. Fig. 5 demonstrates that both mAb 50-6 and mAb 1A5 inhibited HEp-3 migration by approximately 45 and 44%, respectively, when compared with controls (P ϭ 0.034). To further demonstrate the role of PETA-3/CD151 in tumor cell migration, HeLa cells were transfected with the cloned cDNA or with vector alone. Two clones from each group were selected for testing in Fig. 6. Effects of PETA-3/CD151 expression on HeLa cell migration. HeLa cells were cotransfected with the cloned PETA-3/CD151 cDNA and pSV2neo. Controls were co- a chemotaxis assay on the basis of their PETA-3 expression levels transfected with pcDNA I and pSV2neo. Two PETA-3 overexpressing clones (PB6HI and (Fig. 6a), their similar rates of growth in vitro (Fig. 6b) and similarity PC3HI) and two control clones (NB11LO and NB17LO) were selected. a, the amount of cell surface PETA-3 expressed by the HeLa transfectants and HEp-3 cells was determined by in morphological appearance (not shown). The levels of cell surface whole-cell ELISA on fixed, nonpermeabilized cells and is expressed as the absorbance PETA-3 on the two overexpressing clones, PB6HI and PC3HI, were (OD) of the chromogen at 405 nm (b). In vitro growth rates of the HeLa transfectants were about 2-fold higher than the HEp-3 cells, whereas the control trans- determined by plating 2 ϫ 104 cells into each of two wells in a 24-well plate. At the times indicated, the cells were trypsinized and counted. Results are reported as number of cells fectants, NB11LO and NB17LO, expressed approximately one-quarter per well. c, inhibition of HeLa migration (chemotaxis) by mAb 50-6 and mAb 1A5. HeLa to one-half the amount detected on HEp-3 cells (Fig. 6a). The growth cells were detached from culture plate with Versene (Life Technologies, Inc.), washed rates of the four clones were identical (Fig. 6b). Migration by the twice with serum-free DMEM (Life Technologies, Inc.), and resuspended in AIM-V medium (Life Technologies, Inc.). Cells (2.0 ϫ 104) were added to the upper reservoir of overexpressing clones PB6HI and PC3HI was significantly higher than the migration chambers in AIM-V medium and 50 ␮g/ml of mAb 50-6, mAb 1A5, or normal mouse (NM) IgG. The lower reservoirs contained DMEM supplemented with 10% the underexpressing control transfectants NB11LO and NB17LO (Fig. ϭ ␮ FBS (HyClone) and 50 ␮g/ml of the appropriate antibody. After 18 h of incubation, the 6c; P 0.001). When mAb 50-6 or mAb 1A5 (50 g/ml) was added microporous inserts were fixed and stained, and the cells on the underside of one quadrant to the chambers, migration of clone PB6HI was significantly reduced of each filter were counted. Experiments were conducted in triplicate; bars, SEM. 3817

background bands observed in the immunoprecipitated samples are due to the reactivity of the secondary antibody to mouse IgG proteins and IgG fragments as seen in the control lanes. When the PETA-3-transfected HeLa clones were analyzed by ␣ ␤ whole-cell ELISA, there was no detectable 3 1 signal. Weak signals were seen with the control HeLa transfectant (Fig. 7). This suggests that PETA-3 may interact with a different integrin in the transfected HeLa cells to effect the observed increase in cell migration (Fig. 6c). PETA-3/CD151 Appears Not to Be Associated with Cell Adhe- sion. Because cell adhesion is a prerequisite to cell migration, it is possible that the inhibitory effects of mAbs 50-6 and 1A5 in the chemotaxis assays are related to antibody inhibition of cell attachment. Therefore, to quantitate the role of PETA-3 in cell attachment, HEp-3 cells and the HeLa transfectants were tested for adhesion to ␣ ␤ Fig. 7. Whole-cell ELISA of expression of the 3 1 integrin heterodimer by equal various purified matrix proteins in the presence of mAb 50-6, mAb numbers (2 ϫ 104) of MDA-MB-231 human breast carcinoma cells, HT1080 human fibrosarcoma cells, HEp-3 cells, and the PETA-3 overexpressing HeLa clones (PB6HI and 1A5, or normal mouse IgG. As shown in Fig. 9, at the 30-min time PC3HI) and control HeLa clones (NB11LO and NB17LO). mAb 1992 was used to point, HEp-3 attachment to both FN (Fig. 9a) and VN (Fig. 9b) was specifically identify the heterodimer. Data represent the amount of chromogen released in the assay and are expressed as the absorbance (OD) at 405 nm. unaffected by the presence of either of the antimetastatic mAbs. Likewise, attachment of the overexpressing and weakly expressing

␣ ␤ Fig. 8. PETA-3 expressed by HEp-3 cells associates with the 3 1 integrin heterodimer. ␣ ␤ HEp-3 cell lysates were incubated with mAb 1A5 (Lane 2), mAb 1992 (specific for the 3 1 heterodimer; Lane 3), or normal mouse IgG (NM; Lane 4). Immune complexes were collected on GammaBind Plus Sepharose, resolved by SDS-PAGE, and transferred to nitrocellulose. The blot was incubated with unlabeled mAb 1A5 as a primary antibody, followed by horseradish peroxidase-conjugated goat anti-mouse IgG. Signals were visualized by chemiluminescence. As a control, a NM “immunoprecipitate” was probed with NM IgG and peroxidase-conjugated goat anti-mouse IgG (Lane 5). A sample of the original HEp-3 lysate .(PETA-3 signal. Left, molecular weight markers (in thousands ,ء .is shown in Lane 1

(P ϭ 0.034) by 51.6 and 52.8%, respectively, whereas migration of ϭ clone PC3HI was significantly reduced (P 0.007) by 36.7 and 50.6%, respectively, lowering the motility of these cells to the level of control NB11LO cells. Neither mAb 50-6 nor mAb 1A5 inhibited ϭ migration of the weakly expressing clones NB11LO (P 0.729) or ϭ NB17LO (P 0.337), indicating that migration by these cells is mediated by a PETA-3-independent mechanism. ␣ ␤ PETA-3/CD151 has been shown to associate with the 3 1 integrin in several cell types (25, 26). Using a monoclonal antibody, mAb ␣ ␤ 1992, which specifically recognizes the 3 1 heterodimer, we have ␣ ␤ detected, by whole-cell ELISA, the 3 1 integrin on HEp-3 cells (Fig. 7) at levels equivalent to known positive controls, MDA-MB-231, a human breast carcinoma cell line, and HT1080, a human fibrosarcoma ␣ ␤ cell line (18). A physical interaction between 3 1 and PETA-3 in HEp-3 cells was demonstrated by the ability of mAb 1992 to coim- munoprecipitate these two molecules from HEp-3 cell lysates. West- ern blot analysis of immunoprecipitates obtained with mAb 1992 shows a Mr 29,000 protein that immunostains with mAb 1A5 (Fig. 8, Fig. 9. Effects of mAb 50-6, mAb 1A5, and PETA-3 expression on cell adhesion to various Lane 3). This protein comigrates with the authentic antigen present in matrix proteins. HEp-3 cells, PETA-3-overexpressing HeLa clones (PB6HI and PC3HI) and control HeLa clones (NB11LO and NB17LO) were added to the wells of 96-well plates coated the HEp-3 lysates (Fig. 8, Lane 1) and in immunoprecipitates obtained with either fibronectin (a) or vitronectin (b) in the presence of 50 ␮g/ml mAb 50-6, mAb 1A5, with mAb 1A5 (Fig. 8, Lane 2). No PETA-3 signal was observed in or normal mouse IgG and incubated for 30 min. Nonadherent cells were gently washed away, and the amount of cell adhesion was determined by crystal violet staining. Data are expressed control normal mouse IgG “immunoprecipitates” probed with mAb as the absorbance (OD) at 595 nm of the incorporated crystal violet dye and are representative 1A5 (Fig. 8, Lane 4) or normal mouse IgG (Fig. 8, Lane 5). The of two separate experiments conducted in triplicate; bars, SEM. 3818

HeLa transfectants to FN or VN was not blocked by mAb 50-6 or of preferred secondary tumor growth (37–39). Our studies on the time mAb 1A5. There was also no correlation between levels of PETA-3 course of inhibition of HEp-3 experimental metastasis (Table 2) expression and cell attachment. Similar results were obtained when suggest that PETA-3 is involved in an early step in the formation of the adhesion assays were terminated after 15 min (not shown) and metastatic foci, such as arrest, extravasation, and/or migration into the when cell attachment was tested on plates coated with laminin, type I connective tissue stroma of the secondary organ. Although PETA-3 is collagen, and type IV collagen (not shown). These data suggest that known to be expressed on endothelial cells (25, 28), neither mAb 50-6 PETA-3 is not functionally involved in cell attachment. nor mAb 1A5 react with the endothelium of the chicken embryo, the host in the present metastatic model.7 Thus, the antimetastatic effect DISCUSSION of mAb 50-6 and mAb 1A5 is the result of antibody binding to the HEp-3 cells themselves and not to host endothelial cells. The metastatic process has been correlated with expression of a Tumor invasion of tissue elements is one hallmark of the malignant wide variety of cellular proteins, including adhesion molecules, phenotype and is dependent on the ability of tumor cells to transiently growth factors, motility factors, proteases, transcription factors, and adhere to various matrix proteins and to migrate into the surrounding signaling molecules. However, a correlative association does not stroma. TM4SF proteins are known to associate with other tetraspanins, necessarily imply that a protein is functionally relevant to tumor cell integrins, and potential signaling molecules and are believed to facilitate metastasis. In the present study, mAb 50-6 was used to immunolog- the formation and stabilization of these macromolecular complexes and ically target a cell surface antigen that is mechanistically involved in thus influence a number of cellular functions including migration and tumor cell dissemination. This mAb was generated by subtractive adhesion (Refs. 25, 33, and 40–44; reviewed in Refs. 22–24). In the immunization against a highly metastatic human epidermoid carci- present report, several experiments demonstrate a positive role for noma cell line, HEp-3, without bias as to the nature or identity of the PETA-3 in mediating cell migration: (a) we were able to inhibit HEp-3 antigen. The ability of mAb 50-6 to block metastasis specifically chemotaxis with mAb 50-6 and mAb 1A5 (Fig. 5); (b) we showed that (Table 1) strongly suggests that its cognate antigen is functionally HeLa cells transfected with and overexpressing PETA-3 were more involved in the metastatic process. Using mAb 50-6 in a modified migratory than control transfectants (Fig. 6c); and (c) the increase in eukaryotic expression cloning protocol, the full-length cDNA encod- motility by PETA-3-transfected HeLa clones was inhibitable by both ing the Mr 29,000 metastasis-associated HEp-3 antigen was isolated, mAb 50-6 and mAb 1A5 (Fig. 6c). The results of our antibody inhibition sequenced, and identified as PETA-3/CD 151 (21), also known as studies are consistent with recent observations that random migration of SFA-1 (27). We also demonstrated that PETA-3 is recognized by a endothelial cells (25) and polymorphonuclear chemotaxis (26) are sensi- previously described metastasis-inhibiting antibody, mAb 1A5 (Fig. tive to inhibition by anti-PETA-3 mAbs. The mechanism by which 4), also generated by subtractive immunization (11), but the cognate PETA-3 influences migration of these cells is apparently related to the antigen of which was unknown. The unbiased selection of two inde- association of this tetraspanin with the ␣ ␤ integrin (25, 26) and poten- pendently generated anti-PETA-3 mAbs that block metastasis indi- 3 1 tial signaling molecules (26). In the present study, we have demonstrated, cates that the subtractive immunization approach may be a powerful by coimmunoprecipitation, a physical association between PETA-3 and tool to identify functionally important antigens. ␣ ␤ in HEp-3 cells (Fig. 8). However, there is little or no detectable PETA-3 is a TM4SF protein. Also known as tetraspanins, members 3 1 ␣ ␤ expressed on our HeLa transfectants (Fig. 7). TM4SF proteins can of this protein family are characterized by having four hydrophobic, 3 1 interact with several different integrin molecules, primarily those in the transmembrane domains, two short cytoplasmic tails, and one small ␤ class (22–24), and it is possible that PETA-3 expressed by the HeLa and one large extracellular loop (reviewed in Refs. 22–24). PETA-3 is 1 cells associates with another integrin to effect the observed increase in expressed by a variety of cell types including the basal cells of the motility. epidermis, epithelial cells, skeletal, smooth and cardiac muscle, The effects of PETA-3 on migration of HEp-3 and HeLa cells appear Schwann cells, platelets, and endothelial cells (28). Ya´n˜ez-Mo´ et al. (25) have shown that PETA-3 expressed by cultured endothelial cells not to be related to changes in cell adhesion. We found no correlation is localized at cell-cell junctions. However, our immunocytochemical between levels of cell surface PETA-3 expression and adhesion to wells ␤ analysis of fixed, nonpermeabilized HEp-3 cells show that PETA-3 is coated with the 1 substrates FN (Fig. 9a), LN, collagen type I, or ␤ distributed over the entire cell surface (Fig. 3a). Furthermore, this collagen type IV (not shown) or the 5 substrate VN (Fig. 9b). In expression pattern was not restricted to HEp-3 cells because immu- addition, there was no significant difference in adhesion when cells were nostaining over the entire cell surface was also observed on COS-7 plated onto these same substrates in the presence of mAb 50-6 or mAb transfectants expressing PETA-3 (Fig. 3b). Whether there is a func- 1A5 (Fig. 9 and data not shown). Ya´n˜ez-Mo´ et al. (25) found that tional relevance associated with the different patterns of PETA-3 endothelial cell adhesion to FN, LN, and collagen type I increased distribution remains to be determined. slightly but significantly in the presence of their anti-PETA-3 mAbs. The Several TM4SF proteins have been implicated as regulators of cell reason for the differences between the results of this latter study and our proliferation. Cell proliferation can be either stimulated (29–31) or observations remains to be determined. slowed (32) in the presence of anti-TM4SF antibodies. Transfection with Several members of the tetraspanin family of proteins have been several tetraspanin family members also retards growth (32–35). Our associated with the metastatic phenotype, but these associations have anti-PETA-3 mAbs 50-6 and 1A5 were shown not to affect in vitro been, for the most part, negative. KAI-1/CD82 expression suppressed growth of HEp-3 cells (Fig. 2; Ref. 11) nor in vivo growth in the primary experimental metastasis of rat prostate tumor cells (45), decreased mo- tumor (Table 1; Ref. 11) or at the secondary site (Table 2). In addition, the tility and invasion of colon carcinoma cells (44), and decreased invasion in vitro growth rates of HeLa transfectants overexpressing PETA-3 are no and metastasis of mouse melanoma cells (46). Likewise, experimental different than control clones (Fig. 6b). These results indicate that PETA-3 metastasis of mouse melanoma was reduced in cells expressing motility- does not affect tumor cell proliferation but functions specifically in one or related protein (MRP)-1/CD9 (33). In addition, (over)expression of CD9 more steps in the metastatic process itself. slowed growth and blocked migration of CHO cells, and human lung Metastatic success by tumor cells has been shown to be dependent adenocarcinoma and myeloma cells (33). CD63 expression also resulted on initial arrest in the secondary organ (36), as well as events that occur after extravasation, such as migration through the stroma to sites 7 Quigley, unpublished results. 3819

Downloaded from cancerres.aacrjournals.org on September 26, 2021. © 1999 American Association for Cancer Research. PETA-3/CD151 IS AN EFFECTOR OF MIGRATION AND METASTASIS in decreased in vivo growth of human melanoma cells (34) and NIH3T3 21. Fitter, S., Tetaz, T. J., Berndt, M. C., and Ashman, L. K. Molecular cloning of cDNA cells (35) and blocked experimental metastasis of human melanoma cells encoding a novel platelet-endothelial cell tetra-span antigen, PETA-3. Blood, 86: 1348–1355, 1995. (34). Claas et al. (32) recently cloned the rat homologue of CO-029, a 22. Hemler, M. E., Mannion, B. A., and Berditchevski, F. Association of TM4SF proteins tetraspanin that appears to affect metastasis by altering the homing pattern with integrins: relevance to cancer. Biochim. Biophys. Acta, 1287: 67–71, 1996. of tumor cells. Transfection of BSp73AS cells, a weakly metastatic rat 23. Maecker, H. T., Todd, S. C., and Levy, S. The tetraspanin superfamily: molecular facilitators. FASEB J., 11: 428–442, 1997. pancreatic adenocarcinoma cell line, with the homologue shifted the 24. Wright, M. D., and Tomlinson, M. G. The ins and outs of the transmembrane 4 metastatic burden from the lymph nodes to the lungs and resulted in an superfamily. Immunol. Today, 15: 588–594, 1994. 25. Ya´n˜ez-Mo´, M., Alfranca, A., Cabanas, C., Marazuela, M., Tejedor, R., Ursa, M. A., increased survival rate of animals inoculated with the transfectants. A Ashman, L. K., de Landazuri, M. O., and Sanchez-Madrid, F. Regulation of endo- monoclonal antibody to the CO-029 homologue partially reduced a thelial cell motility by complexes of tetraspan molecules CD81/TAPA-1 and CD151/ consumptive coagulopathy associated with expression of this protein; PETA-3 with ␣3␤1 integrin localized at endothelial lateral junctions. J. Cell Biol., 141: 791–804, 1998. however, the effect of the antibody on metastatic dissemination was not 26. Yauch, R. L., Berditchevski, F., Harler, M. B., Reichner, J., and Hemler, M. E. Highly reported (32). In contrast to the reports cited above, we have, in the stoichiometric, stable, and specific association of integrin ␣3␤1 with CD151 provides present study, exploited the techniques of subtractive immunization and a major link to phosphatidylinositol 4-kinase, and may regulate cell migration. Mol. Biol. Cell, 9: 2751–2765, 1998. eukaryotic expression cloning to detect, clone, and identify PETA-3/ 27. Hasegawa, H., Utsunomiya, Y., Kishimoto, K., Yanagisawa, K., and Fujita, S. SFA-1, CD151 as a metastasis-associated antigen that appears to contribute a novel cellular gene induced by human T-cell leukemia virus type 1, is a member of positively to the metastatic phenotype. PETA-3 does not affect tumor cell the transmembrane 4 superfamily [published erratum appears in J. Virol., 71: 1737, 1997]. J. Virol., 70: 3258–3263, 1996. proliferation but rather appears to be specifically involved in an early step 28. Sincock, P. M., Mayrhofer, G., and Ashman, L. K. Localization of the transmembrane 4 in the formation of secondary metastatic lesions. The ability of PETA-3 superfamily (TM4SF) member PETA-3 (CD151) in normal human tissues: comparison to mediate tumor cell migration provides a possible mechanism for the with CD9, CD63, and ␣5␤1 integrin. J. Histochem. Cytochem., 45: 515–525, 1997. 29. Bell, G. M., Seaman, W. E., Niemi, E. C., and Imboden, J. B. The OX-44 molecule role of this protein in effecting metastatic dissemination. Our studies couples to signaling pathways and is associated with CD2 on rat T lymphocytes and identify PETA-3 as the first member of the tetraspanin family of proteins a natural killer cell line. J. Exp. Med., 175: 527–536, 1992. to be linked as a positive effector of metastasis. 30. Nojima, Y., Hirose, T., Tachibana, K., Tanaka, T., Shi, L., Doshen, J., Freeman, G. J., Schlossman, S. F., and Morimoto, C. The 4F9 antigen is a member of the tetra spans transmembrane protein family and functions as an accessory molecule in T cell REFERENCES activation and adhesion. Cell. Immunol., 152: 249–260, 1993. 31. Lebel-Binay, S., Lagaudriere, C., Fradelizi, D., and Conjeaud, H. CD82, member of 1. Ahmad, A., and Hart, I. R. Mechanisms of metastasis. Crit. Rev. Oncol./Hematol., 26: the tetra-span-transmembrane protein family, is a costimulatory protein for T cell 163–173, 1997. activation. J. Immunol., 155: 101–110, 1995. 2. Woodhouse, E. C., Chuaqui, R. F., and Liotta, L. A. General mechanisms of 32. Claas, C., Seiter, S., Claas, A., Savelyeva, M., and Schwab, M. Association between metastasis. Cancer (Phila.), 80: 1529–1537, 1997. the rat homologue of CO-029, a metastasis-associated tetraspanin molecule, and 3. Boyd, D. Invasion and metastasis. Cancer Metastasis Rev., 15: 77–89, 1996. consumptive coagulopathy. J. Cell Biol., 141: 267–280, 1998. 4. Zetter, B. Angiogenesis and tumor metastasis. Annu. Rev. Med., 49: 407–424, 1998. 33. Ikeyama, S., Koyama, M., Yamaoko, M., Sasada, R., and Miyake, M. Suppression of 5. DeClerck, Y. A., Imren, S., Montgomery, A. M., Mueller, B. M., Reisfeld, R. A., and cell motility and metastasis by transfection with human motility-related protein Laug, W. E. Proteases and protease inhibitors in tumor progression. Adv. Exp. Med. (MRP-1/CD9) DNA. J. Exp. Med., 177: 1231–1237, 1993. Biol., 425: 89–97, 1997. 34. Radford, K. J., Mallesch, J., and Hersey, P. Suppression of human melanoma cell 6. Denhardt, D. T. Oncogene-initiated aberrant signaling engenders the metastatic growth and metastasis by the melanoma-associated antigen CD63 (ME491). Int. J. phenotype: synergistic transcription factor interactions are targets for cancer therapy. Cancer, 62: 631–635, 1995. Crit. Rev. Oncogenesis, 7: 261–291, 1996. 35. Hotta, H., Hara, I., Miyamoto, H., and Homma, M. Overexpression of the human 7. Kim, J., Yu, W., Kovalski, K., and Ossowski, L. Requirement for specific proteases melanoma-associated antigen ME491 partially suppresses in vivo malignant pheno- in cancer cell intravasation as revealed by a novel semiquantitative PCR-based assay. types of H-ras-transformed NIH3T3 cells in athymic nude mice. Melanoma Res., 1: Cell, 94: 353–362, 1998. 125–132, 1991. 8. Ossowski, L., and Reich, E. Experimental model for quantitative study of metastasis. 36. Cheng, H. C., Abdel-Ghany, M., Elble, R. C., and Pauli, B. U. Lung endothelial dipeptidyl Cancer Res., 40: 2300–2309, 1980. peptidase IV promotes adhesion and metastasis of rat breast cancer cells via tumor cell 9. Ossowski, L., and Reich, E. Loss of malignancy during serial passage of human surface-associated fibronectin. J. Biol. Chem., 273: 24207–24215, 1998. carcinoma in culture and discordance between malignancy and transformation pa- 37. Koop, S., MacDonald, I. C., Luzzi, K., Schmidt, E. E., Morris, V. L., Grattan, M., rameters. Cancer Res., 40: 2310–2315, 1980. Khokha, R., Chambers, A. F., and Groom, A. C. Fate of melanoma cells entering the 10. Testa, J. E. Loss of the metastatic phenotype by a human epidermoid carcinoma cell microcirculation: over 80% survive and extravasate. Cancer Res., 55: 2520–2523, 1995. line, HEp-3, is accompanied by increased expression of tissue inhibitor of metallo- 38. Hangan, D., Uniyal, S., Morris, V. L., MacDonald, I. C., von Ballestrem, C., Chau, proteinase 2. Cancer Res., 52: 5597–5603, 1992. T., Schmidt, E. E., Chambers, A. F., Groom, A. C., and Chan, B. M. Integrin VLA-2 11. Brooks, P. C., Lin, J. M., French, D. L., and Quigley, J. P. Subtractive immunization (␣2␤1) function in postextravasation movement of human rhabdomyosarcoma RD yields monoclonal antibodies that specifically inhibit metastasis. J. Cell Biol., 122: cells in the liver. Cancer Res., 56: 3142–3149, 1996. 1351–1359, 1993. 39. Luzzi, K. J., MacDonald, I. C., Schmidt, E. E., Kerkvliet, N., Morris, V. L., 12. Gordon, J. R., and Quigley, J. P. Early spontaneous metastasis in the human epider- Chambers, A. F., and Groom, A. C. Multistep nature of metastatic inefficiency: moid carcinoma HEp3-chick embryo model: contribution of incidental colonization. dormancy of solitary cells after successful extravasation and limited survival of early Int. J. Cancer, 38: 437–444, 1986. micrometastases. Am. J. Pathol., 153: 865–873, 1998. 13. Ossowski, L., Russo, H., Gartner, M., and Wilson, E. L. Growth of a human 40. Domanico, S. Z., Pelletier, A. J., Havran, W. L., and Quaranta, V. Integrin ␣6A ␤1 carcinoma (HEp3) in nude mice: rapid and efficient metastasis. J. Cell. Physiol., 133: induces CD81-dependent cell motility without engaging the extracellular matrix 288–296, 1987. migration substrate. Mol. Biol. Cell, 8: 2253–2265, 1997. 14. Ossowski, L., Russo-Payne, H., and Wilson, E. L. Inhibition of urokinase-type 41. Radford, K. J., Thorne, R. F., and Hersey, P. Regulation of tumor cell motility and plasminogen activator by antibodies: the effect on dissemination of a human tumor in migration by CD63 in a human melanoma cell line. J. Immunol., 158: 3353–3358, 1997. the nude mouse. Cancer Res., 51: 274–281, 1991. 42. Miyake, M., Koyama, M., Seno, M., and Ikeyama, S. Identification of the motility- 15. Seed, B., and Aruffo, A. Molecular cloning of the CD2 antigen, the T-cell erythrocyte receptor, by a rapid immunoselection procedure. Proc. Natl. Acad. Sci. USA, 84: related protein (MRP-1), recognized by monoclonal antibody M31–15, which inhibits 3365–3369, 1998. cell motility. J. Exp. Med., 174: 1347–1354, 1991. 16. Wigler, M., Pellicer, A., Silverstein, S., Axel, R., Urlab, G., and Chasin, L. DNA- 43. Shaw, A. R., Domanska, A., Mak, A., Gilchrist, A., Dobler, K., Visser, L., Poppema, mediated transfer of the adenine phosphoribosyltransferase locus into mammalian S., Fliegel, L., Letarte, M., and Willett, B. J. Ectopic expression of human and feline ␤ cells. Proc. Natl. Acad. Sci. USA, 76: 1373–1376, 1979. CD9 in a human B cell line confers 1 integrin-dependent motility on fibronectin and 17. Davis, S., Aldrich, T. H., Jones, P. F., Acheson, A., Compton, D. L., Jain, V., Ryan, laminin substrates and enhanced tyrosine phosphorylation. J. Biol. Chem., 270: T. E., Bruno, J., Radziejewski, C., Maisonpierre, P. C., and Yancopoulos, G. D. 24092–24099, 1995. Isolation of angiopoietin-1, a ligand for the TIE2 receptor, by secretion-trap expres- 44. Takaoka, A., Hinoda, Y., Satoh, S., Adachi, Y., Itoh, F., Adachi, M., and Imai, K. sion cloning [see comments]. Cell, 87: 1161–1169, 1996. Suppression of invasive properties of colon cancer cells by a metastasis suppressor 18. Berditchevski, F., Bazzoni, G., and Hemler, M. E. Specific association of CD63 with KAI1 gene. Oncogene, 16: 1443–1453, 1998. the VLA-3 and VLA-6 integrins. J. Biol. Chem., 270: 17784–17790, 1995. 45. Dong, J. T., Lamb, P. W., Rinker-Schaeffer, C. W., Vukanovic, J., Ichikawa, T., Isaacs, 19. Deng, G., Curriden, S. A., Wang, S., Rosenberg, S., and Loskutoff, D. J. Is plasmin- J., and Barrett, J. C. KAI1, a metastasis suppressor gene for prostate cancer on human ogen activator inhibitor-1 the molecular switch that governs urokinase receptor- chromosome 11p11.2 [see comments]. Science (Washington DC), 268: 884–886, 1995. mediated cell adhesion and release? J. Cell Biol., 134: 1563–1571, 1996. 46. Takaoka, A., Hinoda, Y., Sato, S., Itoh, F., Adachi, M., Hareyama, M., and Imai, K. 20. Ossowski, L., and Reich, E. Antibodies to plasminogen activator inhibit human tumor Reduced invasive and metastatic potentials of KAI1-transfected melanoma cells. Jpn. metastasis. Cell, 35: 611–619, 1983. J. Cancer Res., 89: 397–404, 1998.

3820

Downloaded from cancerres.aacrjournals.org on September 26, 2021. © 1999 American Association for Cancer Research. Eukaryotic Expression Cloning with an Antimetastatic Monoclonal Antibody Identifies a Tetraspanin (PETA-3/CD151) as an Effector of Human TumorCell Migration and Metastasis

Jacqueline E. Testa, Peter C. Brooks, Jian-Min Lin, et al.

Cancer Res 1999;59:3812-3820.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/59/15/3812

Cited articles This article cites 45 articles, 24 of which you can access for free at: http://cancerres.aacrjournals.org/content/59/15/3812.full#ref-list-1

Citing articles This article has been cited by 41 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/59/15/3812.full#related-urls

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

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

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/59/15/3812. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.