Common Human Melanoma-associated Antigen(s) Detected by Monoclonal Antibodies

Transcription

1 [CANCER RESEARCH 40, , July 980] /80/0040-OOOOS02.00 Common Human Melanoma-associated Antigen(s) Detected by Monoclonal Antibodies Stefan Carrel, Roberto S. Accolla, Anna L. Carmagnola, and Jean-Pierre Mach Unit of Human Cancer Immunology, Lausanne Branch. Ludwig Institute for Cancer Research, 066 Epalinges-sur-Lausanne, Switzerland ABSTRACT Hybridoma cells have been derived from a fusion between mouse myeloma cells (P3-NSI/ Ag4) and spleen cells from a mouse immunized with membrane-enriched fractions from the human melanoma cell line Me-43. Of the 26 hybrids obtained, seven secreted antibodies which reacted with the melanoma cell line used for immunization as assessed by an indirect antibody-binding radioimmunoassay. The specificity of the an tibodies produced by the seven positive hybrids was further investigated on 6 melanoma cell lines, 5 other tumors, and 4 lymphoblastoid cell lines. The antibodies from four positive hybrids showed a broad reactivity, whereas those from three hybrids reacted exclusively with melanoma cells. The antibod ies from two of these three hybrids, a-mel/5 and a-mel/4, seem to be directed against common melanoma antigen(s) since they reacted with all (with one exception) of the 6 melanoma cell lines tested, whereas antibodies from the third hybrid, a-mel/7, reacted only with five of the 6 melanoma lines. Reciprocal binding inhibition tests using [3H]leucine-labeled antibodies showed that a-mel/5 and a-mel/4 antibod ies were directed against different antigenic determinants. INTRODUCTION Xenoantisera have often been used in the search for common tumor-associated antigen(s) on melanoma cells (2, 3, 6, 3-5, 8, 20, 2). Since xenoantisera usually contain large amounts of anti-species antibodies, extensive absorption of anti-human melanoma sera with human normal tissue is re quired before any tumor specificity can be demonstrated. The method of KÃ hlerand Milstein ( ) allowing the production of monoclonal antibodies against single antigenic determinants potentially represents a great improvement for the selection of antibodies reacting with tumor-associated antigens. Recently, several investigators have tried to produce monoclonal anti bodies against various human tumor antigens or markers, including melanoma (2, 9, 25), neuroblastoma (8), colon carcinoma (7), and carcinoembryonic antigen (). Koprowski ef al. (2) reported the production of several hybrids secreting antibodies binding to melanoma cells. The antigens recognized by these antibodies appeared to be expressed only on some of the melanoma cell lines tested. Yeh ef al. (25) described 3 hybrids which produced monoclonal antibodies restricted to an autologous melanoma antigen. This paper analyzes the specificity of the monoclonal anti bodies secreted by hybrids obtained from fusion between mouse myeloma cell line P3-NSI/Ag4 (0) and spleen cells from a mouse immunized with membrane-enriched fraction of To whom requests for reprints should be addressed. Received November 7, 979; accepted March 28, 980. the melanoma cell line Me-43. It will be shown that antibodies from 3 of 26 hybrids obtained appear to have a restricted reactivity towards melanoma cells as assessed by indirect antibody binding RIA2 (9, 22). Moreover, the results indicate that antibodies from 2 of 3 hybrids are directed against 2 different common antigenic determinants expressed on all (with one exception) of the 6 different melanoma cell lines tested, while the third hybrid product is directed against a determinant expressed only on some lines. MATERIALS AND METHODS Cell Lines. All human cells used in this study including 6 melanoma, 5 non-melanoma tumor, and 4 lymphoblastoid lines were grown in Roswell Park Memorial Institute Medium 640 supplemented with 0% fetal calf serum. The BALB/c myeloma cell line P3-NSI/ Ag4 used for fusion was maintained in MEM containing 0% horse serum. This cell line is resistant to 8-azaguanine (20 /ig/ml) and does not grow in HAT selective medium (5, ). Preparation of Melanoma Cell Membranes. Membrane-en riched fractions from the melanoma cell line, Me-43, were prepared by N2 cavitation (7). Briefly, washed cells at a minimal concentration of 5 x 07 cells/ml in serum-free Ros well Park Memorial Institute Medium 640 were exposed to a pressure of 0 atmospheres N2 for 20 min at 4Â under gentle stirring. After return to normal atmosphere value, the disrupted cells were collected and centrifuged at 400 x g for 5 min. The supernatant was centrifuged at 20,000 x g for 30 min. The pellet obtained after the 20,000 x g spin was used to immunize the recipient animal. The protein concentration in this fraction was determined by the microbiuret method. Immunization. BALB/c mice were immunized i.p. by 2 injec tions of membrane-enriched fraction from the melanoma cell line Me-43 (0. mg of membrane protein in 0.2 ml of 0.9% NaCI solution) in complete Freund's adjuvant. After 20 days, the mice were boosted i.v. with the same amount of membrane proteins in 0.9% NaCI solution. Three days later, the mice were killed, and their spleen cells were used for the fusion. Cell Fusion. Cell fusion was performed by incubating 07 P3-NSI/Ag4 mouse myeloma cells with 08 mouse spleen cells in 0.3 ml of 40% polyethylene glycol (M.W. 000) for 3 min at 37Â (6). The cells were centrifuged for 5 min at 200 x g; then 5 ml serum-free MEM were added dropwise to dilute the polyethylene glycol. After fusion, the cells were washed and resuspended in 00 ml of HAT medium and distributed in 96 wells with diameters of.6 cm. Antibody Detection RIA. The production of specific antibod- 2 The abbreviations used are: RIA, radioimmunoassay; MEM, Eagle's minimal essential medium; HAT. hypoxanthine-aminopterin-thymidine; ing ratio. a-, anti-; BR, bind JULY

2 S. Carrel et al. Â esin HAT-selected hybrids against the cells used for immuni zation was tested by an antibody-binding assay essentially as described by Williams (22). Briefly, 5 x 0s target cells in 00 /ti of medium were incubated for 2 hr at room temperature with 00 /il of culture fluid from the different hybrids in U-bottomed microtest plates. The plates were centrifuged at 200 x g for 3 min, and the supernatant was removed. After 3 washings with 00 /il of medium, 00 \i\ of 25l-labeled rabbit anti-mouse F(ab'>2 (5 ng of protein corresponding to about 25,000 cpm) were added and incubated for hr. The cells were then washed 3 times with medium and transferred to tubes for y counting. Positive hybrids detected by this method were then cloned by a limiting dilution system ( ) in a 96-well plate, and a repre sentative clone of each specificity was chosen for further studies. Quantitative ferent numbers Absorption Test. Suspensions containing dif of cells (2.5 to 40 x 05/00 /til) were added to equal volumes of culture fluid and incubated for hr at room temperature. Cells were centrifuged, and the remaining binding activity of the culture fluid was tested in a RIA as described above. Reciprocal Binding Inhibition Test. Antibodies from the various clones were internally labeled by the addition of 0 /ici of [3H]leucine to cultures of 06 hybrid cells in ml of leucinefree MEM. After 6 hr of incubation at 37Â,culture fluids were harvested, and 25 \i\ of each 3H-labeled antibody were incu bated for 2 hr with 5 x 05 melanoma cells. The cells were centrifuged and counted in a /i-liquid scintillation counter. For competition analysis, 25 Â i\of unlabeled antibodies of the various hybridoma were allowed to react for 0 min with the target cells before the addition of 25 jil of 3H-labeled antibodies from each hybrid. After 3 hr, the cells were washed and counted. RESULTS Detection of Antibody-secreting Hybrids. From the fusion between spleen cells from a mouse immunized with melanoma membranes and the mouse myeloma cell line P3-NSI/Ag4, we obtained a total of 26 growing hybrids. In an initial screen ing, 7 of them were found to produce antibodies binding to the immunizing melanoma cell line Me-43, as determined by RIA. After cloning, the positive hybrid products were further ana lyzed for their binding capacity against a large panel of mela noma and non-melanoma cell lines. The 7 selected positive clonal products described here were designated as a-mel/5, a-mel/7, a-mel/, a-mel/2, a-mel/4, a-mel/34, and a- Mel/37. Table summarizes the RIA results obtained with these reagents on 6 different melanoma cell lines. The results are expressed as a BR which represents the total number of cell-bound counts divided by the number of cell-bound counts using as control culture fluid of the mouse myeloma P3x63 AG8 producing an Igd (K) immunoglobulin of unknown spec ificity. The background counts varied for each cell line tested, ranging between 65 and 205 cpm. Table 2 summarizes the RIA results obtained with 5 non-melanoma tumor cell lines, includ ing 5 colon carcinomas, 4 breast carcinomas, 2 cervical car cinomas, one endometrial carcinoma, one choriocarcinoma, one glioblastoma, and one hypernephroma. Table 3 summa rizes the results obtained with 4 lymphoblastoid cell lines including 7 B-cell lines, 4 T-cell lines, one non-t-non-b line, and 2 myeloid lines. It can be seen that a-mel/5 antibodies gave a BR of 3 or more on all melanoma cell lines tested except for Me-25 and no more than on all non-melanoma cell lines. The highest BR was obtained for the 3 melanoma lines Me-43, IGR3, and MelEI-78, with values ranging from 20 to 22. Values ranging from 8 to 3 were also obtained with 5 other melanoma cell lines, indicating that the antigen detected by a-mel/5 was relatively abundant on the surface of several melanoma cell lines. a-mel/7 antibodies showed a restricted specificity for a limited number of melanoma cell lines with BR ranging from 3 to 7 for 6 melanoma lines. No binding was observed with all control cell lines tested. a-mel/4 antibodies appeared highly specific for melanoma, since they gave a BR of 3 to 24 with all the 6 melanoma cell lines tested with no significant binding (BR- ) for the 29 control non-melanoma cell lines tested. It is noteworthy that markedly elevated BR's of 8 to 24 were obtained with the majority (0 of 6) of the melanoma cell lines. Chart summarizes the results obtained with these 3 hybrid- Tablei Binding of hybridoma anti-melanoma antibodies to melanoma cell lines hybridomaa-me/ a-mel/ a-mel/ a-mel/ produc of Melanoma (cpm) linesme-43mel-67mp-6mel-57igr3ca-375mel-2amme-8csk-mel-25sk-mel-mp-8mel-2apme-25-mel-ei-78me-85me-33bra cell Total number of cell-bound counts divided by the number of cell-bound counts using P3x63 culture fluid. BKG, background counts bound by P3x63 culture fluid. '' Cell lines expressing la-like antigens, demonstrated by immunoprecipitation analysis and microcytoxicity with a rabbit anti-la serum (4) CANCER RESEARCH VOL. 40

3 Monoclonal Anti-Human Melanoma Antibodies Table 2 Binding of hybridoma anti-melanoma antibodies to various non-melanoma tumor cell Hiñes BP" of hybridoma product lumor linesend-cme-80cama- cell a-mel/4 a-mel/7 «-Mel/ n a-mel/2 (cpm) < 3 58Co- 02 5' HT-29Br-38CCL-989G40hHeLa"Col-Ca'Hyp-5'Co- 57 <32<a-Mel/ BKG < i ) 28 < 0 25Co- < 2BT-208MCF-78a-Mel/5 3 <2 3< < 3 3a-Mel/3452 '' Total number of cell-bound counts divided by the number of cell-bound counts using P3x63 culture fluid. " BKG, background counts bound by P3x63 culture fluid. '' Endometrial carcinoma cell line expressing la-like antigens (4). Cervical carcinoma. 8 Breast carcinoma. Colon carcinoma. 9 Choriocarcinoma. '' Glioblastoma. ' Hypernephroma. Table 3 Binding of hybridoma anti-melanoma antibodies to various lymphoblastoid cell lines. BP" of hybridoma product BKG6 (cpm) Lympoid linesbjabcnalm-640ccem8ramoscnamalvackruschclipscmolt-48hsb-28likck-562'hl-60'jurkat8a-mel/5 cell -Mel/a-Mel/7 c, 2 4 2< 3< < < 4a-Mel/ a-Mel/4<a-Mel/ a-Mel/ l Ulal numuur UPt,tMi~uuunu uuuiuo uwiuou uy me numue " BKG, background counts bound by P3x63 culture fluid. ' B-cell lines. " Non-B-non-T-cell lines. 8 T-cell lines. Myeloid cell lines. orna products on the 45 human cell lines tested. It can be seen that antibodies from a-mel/5 and a-mel/4 have a similar reactivity for all melanoma cell lines tested, suggesting that they recognize common melanoma antigen(s), whereas anti bodies from a-mel/7 have a reactivity restricted to 4 to 5 melanoma cell lines. Quantitative Absorption Experiments. In order to further demonstrate the specificity of the 2 hybrid products, a-mel/5 and a-mel/4, for melanoma cells, 2 series of quantitative absorption experiments were performed. Increasing numbers of cells from 3 melanoma lines (Me-43, IGR3, MP6) and from 2 control tumor lines (HT-29, Hyp-5) were added to 200 /il of culture fluid from hybrid a-mel/5 (at :5 dilution) and the a- Mel/4 (at :0 dilution). The remaining binding activity in the supernatant for IGR3 melanoma cells was then tested by RIA. Chart 2 demonstrates that as few as 5 x 05 melanoma cells from Me-43, IGR3, and MP-6 cell lines absorbed 50% of the binding activity of a-mel/5, while 6 x 06 colon carcinoma cells (HT-29) or hypernephroma cells (Hyp-5) gave no signifi cant inhibition. Similarly, Chart 3 shows that 5 x 05 melanoma cells from lines MP-6 or IGR3 and x 06 cells from melanoma line Me-43 inhibited the binding of a-mel/4 up to 50%, while 6 x 06 colon carcinoma cells (HT-29) or hypernephroma cells (Hyp-5) gave no significant inhibition. Reciprocal Binding Inhibition Experiments. It was of inter est to determine whether the 2 monoclonal antibodies a-mel/ 5 and a-mel/4, which appeared to be directed against com mon antigens expressed on nearly all melanoma cell lines tested, were directed against identical or different antigenic determinants. We therefore performed a series of binding in hibition experiments in which ['HJIeuicne-labeled antibodies from these 2 clones were tested for their binding capacity to IGR3 melanoma cells in the presence of an excess of unlabeled antibodies from various clones. Chart 4a shows that the binding JULY

4 S. Carrel et al. CPM OO LYMPHOID CELLS Chart Reactivity spectrum of 3 melanoma-specific monoclonal antibodies. Results are given as cpm bound per 5 x 0s cells. For each cell type, 3 bars are shown, representing (left to right) hybridomas a-mel/5. a-mel/7, and a-mel/4, respectively. of labeled a-mel/5 antibodies was inhibited by unlabeled an tibodies from the same clone, but not by a-mel/4 antibodies. As a control, unlabled a-mel/, a-mel/37, and a-mel/7 an tibodies used as 20 times concentrated culture fluid were also found to be ineffective in inhibiting the binding of labeled a- Mel/5. Chart 4b shows the reciprocal experiment, where the binding of labeled a-mel/4 antibodies was inhibited only by unlabeled a-mel/4 but not by a-mel/5 antibodies. These results clearly indicate that the antibodies from the 2 clones a-mel/5 and a-mel/4 are reacting with different antigenic determinants present on the same melanoma cells. DISCUSSION A single fusion between spleen cells from a mouse immu nized with crude membrane fractions of human melanoma cells and the myeloma cell line P3-NSI/Ag4 produced 7 hybrids secreting antibodies reacting with the immunizing cells. Exten sive specificity analysis showed that the antibodies from 3 of the 7 hybrids reacted with melanoma cells and not with the 29 control cell lines tested. The antibodies from 2 hybrids, a-mel/ 5 and a-mel/4, displayed a broad reactivity with all melanoma cells tested (with one exception for a-mel/5), suggesting that they recognized antigenic structures which are common to melanoma cells. In contrast, a-mel/7 antibodies reacted with only a proportion of melanoma cell lines, suggesting the exist ence of another more restricted melanoma antigen. The ab sence of reactivity of all 3 melanoma-specific antibodies for different B-cell lines ruled out the possibility that they reacted with the la-like antigens known to be present on some mela noma cell lines (23, 24). The melanoma specificity of the 2 antibodies directed against a putative common melanoma-associated antigen(s) was con firmed by a series of quantitative studies showed that pretreatment absorption experiments. The with 5 x 0s melanoma cells removed 50% of the binding activity of both monoclonal anti CANCER RESEARCH VOL. 40

5 Monoclonal Anti-Human Melanoma Antibodies 2800IJC IGR3 4(a);igg(\\(b)9t* Mel/5 [3 H] oã -Mel/ 320-[SH]* 2000 R x 3000-p IO NUMBER OF CELLS ( X0"5)USED FOR ABSORPTION Chart 2. Quantitative absorption of a-mel/5 antibodies. Binding assay using IGR3 melanoma cells as targets. Each point represents cpm bound by a-mel/5 after absorption of 50 /il (:5) with the number of cells indicated. ', cpm of 25Ilabeled rabbit anti-mouse antibodies bound by the unabsorbed culture fluid; â Melanoma Me-43; A, IGR3; â,mp6 cells; A. Colon Carcinoma HT-29; G. Hypernephroma Hyp-5 cells. S Q. U C IGR3 t Chart 4. Competition of binding to IGR3 melanoma cells between a-mel/5 and a-mel/4 antibodies. Antibodies from both clones were internally labeled with [3H]leucine. a, labeled a-mel/5: ). without unlabeled culture fluid; 2. after addition of unlabeled a-mel/5; 3. after addition of a-mel/4; 4. after addition of a-mel/ ; 5, after addition of a-mel/37; 6. after addition of a-mel/7. b. labeled a-mel/4: ;, without unlabeled culture fluid; 2. after addition of unlabeled o- Mel/5; 3. after addition of unlabeled a-mel/4; 4. after addition of unlabeled a- Mel/; 5, after addition of unlabeled a-mel/37; 6. after addition of unlabeled a-mel/7. detected by our monoclonal reagents were concomitantly ex pressed on all melanoma lines tested. Thus, we cannot exclude the possibility that the 2 antibodies are detecting different determinants localized on the same melanoma-associated an tigen. Further biochemical studies are therefore in progress to use these antibodies to characterize and purify these common melanoma antigens. ACKNOWLEDGMENTS The authors wish to thank Dr. J.-C. Cerottini for helpful discussion and reviewing the manuscript and express their gratitude for the excellent technical assistance of S. Salvi and E. Duruz. *t *t REFERENCES NUMBER OF CELLS ( X0"5)USED FOR ABSORPTION Chart 3. Quantiative absorption of a-mel/4 antibodies. Binding assay using IGR3 melanoma cells as targets. Each point represents the cpm bound by a-mel/ 4 after absorption of 50 /il (:0) with the number of cells indicated. â. cpm of 25l-labeled rabbit anti-mouse antibodies bound by the unabsorbed culture fluid; â Melanoma Me-43; A. lgr3; â,mp6 cells; A, Colon Carcinoma Ht-29; D, Hypernephroma Hyp-5 cells. bodies, whereas up to 80 times more non-melanoma tumor cells were unable to reduce the binding capacity of a-mel/5 and a-mel/4 to melanoma cells. In addition, it was demon strated by reciprocal binding inhibition experiments that the a-mel/5 and a-mel/4 antibodies were directed against differ ent antigenic determinants present on the membrane of mela noma cells. This lack of competition was reciprocal; hence, it seems unlikely that a different binding affinity of the 2 antibod ies would be responsible for this observation. Finally, it should be noted that the 2 antigenic structures. Accolla. R. S.. Carrel, S., and Mach. J.-P Monoclonal antibodies specific for carcinoembryonic antigen (CEA) produced by two hybrid cell lines. Proc. Nati. Acad. Sei. U. S. A.. 77: , Byrstryn. J. C. Release of cell-surface tumor-associated antigens by viable melanoma cells from humans. J. Nati. Cancer Inst , Carrel. S.. Dent. P. B., and Liao, S. K. Demonstration of the specificity of a monkey antiserum against human melanoma: evidence that the cytotoxic antibodies from the specific antiserum belong to the IgM class. Cancer 4. Immunol. Immunother., in press Carrel. S., Gross. N.. Heumann, D.. and Mach, J.-P. Expression of "la-like" antigens on cells from a human endometrial carcinoma cell line, END-. Transplantation (Baltimore), , Cowan, N. J., Secher, D. S, and Milstein. C. Intracellular immunoglobulin chain synthesis in non-secreting variants of a mouse myeloma: detection of inactive light chain messenger RNA. J. Mol. Biol , Fritze, D.. Kern. D. H., Drogemuller, C. R., and Pilch. Y H. Production of antisera with specificity for malignant melanoma and human fetal skin. Cancer Res., , Herlyn. M., Steplewski, Z., Herlyn. D., and Koprowski, H. Colorectal carci noma-specific antigen: detection by means of monoclonal antibodies. Proc. Nati. Acad. Sei. U. S. A., , Kennet!, R. H., and Gilbert, F. Hybrid myelomas producing antibodies against a human neuroblastoma antigen present on fetal brain. Science (Wash. D. C.), Klinman, N. R. The mechanism of antigenic stimulation of primary and secondary clonal precursor cells. J. Exp. Med.. / , 972. JULY

6 S. Carrel et al. 0. Kà hler, G., Howe, S. C., and Milstein, C. Fusion between immunoglobulin secreting and non-secreting myeloma cell lines. Eur. J. Immunol., , Kà hler, G., and Milstein, C. Continuous cultures of (used cells secreting antibody of predefined specificity. Nature (Lond.), 256: , Koprowski, H. Steplewski. Z. Herlyn, D., and Herlyn. M. Study of antibodies against human melanoma produced by somatic cell hybrids. Proc. Nati. Acad. Sei. U. S. A., , Liao, S. K., Kwong, P. C., Thompson, J. C., and Dent, P. B. Spectrum of melanoma antigens on cultured human malignant melanoma cells as de tected by monkey antibodies. Cancer Res., , McCabe, R. P., Ferrone, S., Pellegrino, M. A., Kern, D. H.. Holmes, E. C., and Reisfeld, R. A. Purification and immunological evaluation of human melanoma-associated antigens. J. Nati. Cancer Inst., , Metzgar, R. S., Bergoc, P. M., Moreno, M. Y., and Seigier, H. F. Melanomaspecific antibodies produced in monkey by immunization with human mela noma cell lines. J. Nati. Cancer Inst., , Pontecorvo, G. Production of mammalian somatic cell hybrids by means of polyethylene-glycol treatment. Somatic Cell Genet., : , Schmidt-Ulrich. R., Ferber, E., Knuefermann, H., Fischer, H., and Hoelzl- Wallach, D. F. Analysis of the proteins in thymocyte plasma membrane and smooth endoplasmic reticulum by sodium dodecylsulfate-gel electrophoresis. Biochim. Biophys. Acta, , Sorg, C., Brüggen,J., Seibert, E., and Macher, E. Membrane-associated antigens of human malignant melanoma IV: changes in expression of anti gens on cultured melanoma cells. Cancer Immunol. Immunother., , Steplewski, Z., Herlyn, M., Herlyn, D.. Clark. W., and Koprowski. H. Reac tivity of monoclonal anti-melanoma antibodies with melanoma cells freshly isolated from primary and metastatic melanoma. Eur. J. Immunol., , Stuhlmiller. G. M., and Seigier, H. F. Characterization of a chimpanzee antihuman melanoma antiserum. Cancer Res , Viza, D., and Phillips, J. Identification of an antigen associated with malignant melanoma. Int. J. Cancer, 6: 32-37, Williams. A. F. Differentiation antigens of the lymphocyte cell surface. Mol. Immunol.. 6: 93-6, Wilson, B. S., Indiveri, F., Pellegrino, M. A., and Ferrone, S. DR (la-like) antigens on human melanoma cells. Serological detection and immunochemical characterization. J. Exp. Med., 49: , Winchester, R. J., Wang. C., Gibofsky. A., Kunkel, H. G., Lloyd. K. O., and Old, L. J. Expression of la-like antigens on cultured human malignant melanoma cells lines. Proc. Nati. Acad. Sei. U. S. A., , Yeh. M.-Y., Hellstrà m, I., Brown, J. P., Warner, G. A., Hansen, J. A., and Hellstrà m, K. E. Cell surface antigens of human melanoma identified by monoclonal antibody. Proc. Nati. Acad. Sei. U. S. A , CANCER RESEARCH VOL. 40