Reactivities of Mouse Monoclonal Antibody K2.7 to Renal Cancers in Complement Dependent Cytotoxicity and Antibody Dependent Cell-Mediated Cytotoxicity

0022-5347/95/154 1-0288$03.00/0 THE JOURNAL OF UROLoCY Copyright 0 1995 by AMERICANU ~ O m i r A LASSOCIATION,

Vol. 154,288-292. July 1996 Printed rn U . S d

INc.

REACTMTIES OF MOUSE MONOCLONAL ANTIBODY K2.7 TO RENAL CANCERS IN COMPLEMENT DEPENDENT CYTOTOXICITY AND ANTIBODY DEPENDENT CELL-MEDIATED CYTOTOXICITY T0SHIAK.I KINOUCHI, NEIL H. BANDER

AND

TOSHIHIKO KOTAKE

From the Department of Urology, The Center for Adult Diseases, Osaka, Japan, Laboratory of Urologic Oncology, The New York HospitalCornell Medical Center, New York, New York and Ludwig Institute for Cancer Research, New York Unit at Memorial Sloan-Ketteriw Cancer Center, New York, New York

ABSTRACT

Immunohistochemical analysis by indirect immunoperoxidase staining demonstrated that monoclonal antibody (mAb) K2.7, derived from a mouse immunized with a renal cell carcinoma (RCC) cell line OS-RC-2, reacted with 89 of 95 renal cancer tissues (94%). Only 1 gastric and uterine cancer tissue showed positive staining among 87 cancer specimens from 9 different organs. Among normal human tissues, the renal tubule, testis, epithelium of the uterine endometrial gland and Fallopian tube, grey matter of cerebrum and cerebellum, and foreskin showed positive staining. Serological analysis by protein A mixed hemadsorption (PA) assay demonstrated that mAb K2.7 reacted with 25 of 31 RCC cell lines (81%),but with only 2 of 50 other cell lines from different organs. The specific antitumor activities of mAb K2.7 against RCC cell lines were investigated in vitro by complement dependent cytotoxicity (CDC) and antibody dependent cell-mediated cytotoxicity (ADCC)assays. In the CDC assay, all of the 9 RCC cell lines reactive serologically with mAb K2.7 were killed by mAb K2.7 with normal human serum, and the killing activity of mAb K2.7 correlated well with the reactivity of mAb K2.7 in the PA assay. mAb K2.7 showed the same killing activity against each of 3 RCC cell lines with sera from 9 patients with low and high stage renal cancers, as well a s with normal human serum. In the ADCC assay, mAb K2.7 with peripheral blood leukocytes (PBLs)from 4 healthy donors showed cytotoxic activity against RCC cell lines. Peripheral blood leukocytes from the same 9 renal cancer patients also showed significant killing activity against the 3 RCC cell lines. These findings suggest the potential utility of mAb K2.7 for specific immunotherapy of renal cancer. KEY WORDS:carcinoma, renal cell; antibodies, monoclonal; antibody-dependentcell cytotoxicity

The prognosis of advanced renal cell carcinoma (RCC) is very poor because no adequately effective systemic treatment modalities are available. However, in some renal cancer patients, metastatic lesions have spontaneously disappeared1 or appeared at long intervals aRer nephrectomy.2 And serum antibodies reactive with autologous renal cancer cells have been found in some patients3 These findings suggest t h a t patients with renal cancers might have antitumor immunity to renal cancer. In malignant melanoma, IgG3 mAb R24 reactive with GD3 ganglioside showed strong antitumor activity in vitro and in viv0.4 Houghton e t al. reported major tumor regression in 3 of 12 patients with metastatic malignant melanoma in a Phase I trial with mAb R24.6 Furthermore, treatment with IgG3 mAb 3F8 reactive with GD2 ganglioside produced clinical responses in 7 of 17 patients with neuroblastoma and malignant melanoma.6 These results suggest that mouse IgG3 mAb may be useful for cancer immunotherapy. Unfortunately, only 2 reported mAbs reactive with renal cancer tissues were of the IgG3 isotype, namely mAb K2.7 and mAb BA-2.7 mAb K2.7s was established by immunizing mice with the RCC cell line OS-RC-2.9 In the present study, we demonstrate that mAb K2.7 is reactive against RCC cell lines in complement dependent cytotoxicity (CDC) and antibody dependent cell-mediated cytotoxicity (ADCC) assay using se-

rum and peripheral leukocytes (PBLs) from renal cancer patients as a source of complement and effector cells. MATERIALS AND METHODS

Accepted for publication February 3. 1995. uests for reprints: Department of Urology, The Center for Aid%iseases. Osaka, 1-3-3 Nakamichi, Higashinari-ku, Osaka 537,Japan. This work was supported in art by a Grant-in-Aid for Scientific Research from the Ministry of Aealth and Welfare of Japan (2-21,.

Cell lines. Almost all cell lines of human cancers were obtained from the laboratory of Human Cancer Immunology at Memorial Sloan-Kettering Cancer Center, New York, New York. A498, ACHN and Caki-1 were obtained from American Tissue Type Collection, Rockville, Maryland. OUR-10 was a gift from Dr. M. Matsuda, Osaka University Medical School, Osaka, Japan. OS-RC-2 to -9 and EBV-B, a n EBV transformed B cell line from the donor of OS-RC-2, were maintained in the Department of Urology, The Center for Adult Diseases, Osaka, Japan. Serological analysis. Protein A mixed hemadsorption (PA) assay was performed as described previously.10 Briefly, the serological assay was performed on cells previously plated in Terasaki plates. Antibodies were incubated with target cell8 a t room temperature for 40 minutes. The target cells were then washed, and indicator cells consisting of protein A conjugated human RBCs were added for 40 minutes. Preparation of d b K2.7. mAb K2.7 was prepared as described previously.H Briefly, spleen cells from a BALBk mouse immunized with OS-RC-2 were fused with NS-1, and the supernatants were screened serologically with OS-RC-2 by the PA assay. The immunoglobulin isotype was determined by immunodiffusion. mAb was purified using a MAPS-I1 kit (Bio-Rad, Richmond, California). lmmunohistocliemical staining. Normal and malignant tissues of various organs were obtained from surgical and au-

288

HUMAN CDC AND ADCC TO RCC

289

topey specimens. The tissues were promptly frozen in liquid TABLE2. Reactivities of mAb K2.7 with human tumom by i n d W fitlogen and stored at -8OC. Cryostat sections were fixed immunopemxidme staining with cold acetone or 2% paraformaldehyde in PBS for 10 Tumors Reactintie minutes. Assays were performed by the indirect immm~Kidney 89/95 pemxidase methods with a Vectastain ABC kit for mouse I ~ G Bladder w19 (vector Lab., Burlingame, California)or pemxidasecoqjugated 016 0/5 rabbit anti-mouse Ig ( D M 0 Corp., h t a Barbara,C&o&) 0115 as a second antibody. Ascites from mice bearing NS-1 cells V15 or normal mouse sera were routinely included as negative 0111 controls. 016 Complement dependent cytotoxicity (CDC)assay. Target Of5 v5 cells were cultured in Terasaki plates overnight a t 3°C. Ten Number of specimens with positive staining/numberof specimens tested. microliters (10 p1.) of diluted mAb K2.7 was added and inmbated for 40 minutes at 37c. Without washing, 10 d.of k2 diluted human sera as a source of complement was added and incubated for 4 hours at 37C. After fixation with meth- not shown). Of 87 tumor specimens from 9 Werent organs, anol for 10 minutes, viable cells were counted microscopically only 1 gastric and uterine cancer tissue showed positive by staining with Giemsa solution. mAb K2.7 or diluted hu- staining. man sera alone showed less than 10% killing activity. Analysis of the reactivity of mAb fc2.7 with cultured cell Antibody dependent cell mediated cytotoxicity (ADCC) as- lines by PA assay. Specificity analysis was investigated serosay. Human PBLs separated by the lymphocyte separating logically by PA assay on a range of human cell lines. As medium (Organon Teknika, Durham, North Carolina), di- shown in table 3,25 of 31 RCC cell lines (81%)reacted with luted mAb K2.7 and "Cr-labeled target cells were incubated mAb K2.7.Of50 cell lines f h m other organs, only 1cell line in 96-well round bottom microplates (Corning 25850, Corn- each of colon and ovarian cancer reacted with mAb K2.7. The ing, New York) a t 37C for 4 hours. Radioactivities of the Epstein-Barr virus transformed B cell line from the patient supernatants collected by a supernatant collection system donating the 0s-RC-2 cell line was not reactive. (Skatron VA Leir, Norway) were counted and percent qb- Complement dependent cytotoxicity assay. Figure 1 shows toxicity was calculated as the reactivities of mAb K2.7 with RCC cell linea in the PA assay and CDC activities of mAb K2.7 with normal human Experimental group release - Spontaneous release serum as a source of complementagainst RCC cell lines. Nine x 100 of 11RCC cell lines reacted serologidy with mAb K2.7. In Maximum release - Spontaneous release the CDC assay, the 9 RCC cell lines which reacted with mAb mAb K2.7 alone did not show any significant killing activity K2.7 in the PA assay were killed by mAb K2.7 with normal human serum.Killing activities were almost parallel to the to any target cells. reactivities with mAb K2.7 in the PA assay. When the sera from 9 patients with renal cancers, including low and high RESULTS &ages, were used as sources of complement in the CDC assay Analysis of the specificity of mAb E2.7 by indirect immunoperoxidase staining of cryopreserved tissues. The specificity of mAb K2.7 was investigated in cryopreserved tissues by the TABLE3. Reactivities of mAb K2.7 with human cell lines by indirect immunoperoxidase method. The reactivity of mAb PA assay K2.7 was first tested on normal human tissues from various Cell lines Reactivitiesa sources. As shown in table 1, among normal human tissues, the renal tubule, testis, epithelium of uterine endometrial Renal (3Ub SK-RC-1, 6,7.9, 18, 28, 29, 39, 42, 44, gland and Fallopian tube, grey matter of cerebrum and cer0 .0 .. 45,47,48,49,52,56,59,61,62 ebellum, and foreskin were stained by mAb K2.7. Tissues 00000000 OSRC-2, 3,4, 5,6, 7 , 8 , 9 from other normal organs were not stained. A positive reacOUR-10, A498,ACHN, Caki-1 0000 tion with mAb K2.7 was observed on the cell membrane and Bladder (5) 639V, RT4,263J, 647V,T24 00000 in the cytoplasm, but not in the nucleus of the cell. As shown Colon (5) in table 2, among human tumors, 89 of 95 renal cancer 0000. HT29, SW403,SW1116, M7609, CaCO2 tissues (94%),including low and high grade tumors, showed Gastric (4) 0000 almost diffusely positive staining. More than 90% of metaNUGC-2, 3.4, MNK-1 Lung (7) static lesions of renal cancers also stained positively (data

..........

TABLE1. Reactivities of mAb K2.7 with normal human tissues by indirect immunopemxidase staining Tissues Kidney Glomerulus Tubule hostate

Ureter

%activitiesa

0 0

0

Tissues Esophagus Stomach Colon PalllXWS Liver

0 Bladder 0 Thymu Testis 0 Thyroid Uterus Parotid gland Cervix 0 Breast Endometrium 0 Lung Myometrium 0 Spleen Fallopian tube 0 Adrenal gland Placenta 0 Skin Cerebrum 0 Foreskin Cerebellum a Reactivities are as follows: 0, positive; 0, negative.

Reactivities^

0 0 0 0 0 0 0 0 0 0

0 0 0 0

HM-1, KNS-62, SK-LUI, QG-90, F'C-I LUCI-3, 10 Melanoma (7) SK-MEG23,28.31,33,37,173, MeWo

00000 00

0000000

Astroeytoma (5)

00000 SK-MG1,3,4, U25lMG. U173MG Pallmas (3) 000 Panc-l,2, A s p 1 ovary (3) 00. OV-3,6, 2774 Uterus & Cervix (2) 00 uT2, ME180 Breast (2) 00 MCF7, BT-20 prostate (1) U Pc-3 Hematologic (5) 00000 Daudi, MOLT-4, HPB-ALL, 5111, EBV-B' Fibroblast (1) WI-38 0 Reactinties are as followe: 0, positive; 0, negative. Number of cell linea tested. 'EBV h.ansformed B cell line h m the donor of OS-RC-2.

,.

292

290

HUMAN CDC AND ADCC TO RCC HUMAN CDC AND ADCC T O RCC

FIG. 1. A, reactivities of mAb K2.7 with RCC cell lines by PA assay. Nine RCC cell lines reacted with mAb K2.7, but 2 cell lines did not. B, complement dependent cytotoxicity of mAb K2.7 with human normal serum. Cell lines reactive with mAb K2.7 were all killed. Killing activity of mAb K2.7 against cell lines correlated with reactivity of mAb K2.7 in serological assay. against 3 RCC cell lines, SK-RC-7, -56 and -18 Figure 2, all of the 9 sera, independent of the clinical stage of renal cancer, showed the same killing activities as normal human serum to each of the 3 cell lines. Antibody dependent cell mediated cytotoxicity assay. The optimal concentration of mAb K2.7 in ADCC assay was first determined with SK-RC-7 as a target and PBLs from a healthy donor as effector cells. Maximum killing was observed with 5 microgram/ml. of mAb K2.7, and this optimal dose was used in further experiments (data not shown). Peripheral blood leukocytes from 4 healthy donors showed significant cytotoxicity against RCC cell lines in the presence of mAb K2.7, but these PBLs showed different killing activities against the same cell line, as shown in Table 4. When the PBLs from one donor were used in ADCC assays against various RCC cell lines, the ADCC activities depended on the serological reactivities with

mAb K2.7 in the PA assay as shown in Figure 1. The ADCC activities of PBLs from the 9 patients with renal cancers were investigated against 3 RCC cell lines, SKRC-7, -56 and -18 as targets. As shown in Table 5, PBLs from all patients showed significant cytotoxicity against the 3 RCC cell lines, independent of the clinical stage of renal cancer. Renal toxicity of mAb K2.7. The clinical application of mAb K2.7 reactive with renal tubules may result in renal toxicity in renal cancer patients. mAb K2.7 reacted with renal tubules of rat immunohistochemically. We examined whether rat complement could work with mAb K2.7 to kill RCC cell lines. Three RCC cell lines, SK-RC-7, -56, -18 were killed equally by mAb K2.7 with rat serum or human serum (data not shown). As an animal model for evaluating the renal toxicity of mAb K2.7, 1 mg./kg. of mAb K2.7 was adminis-

FIG. 2. Complement dependent cytotoxicity of mAb K2.7 with sera from 9 patients with renal cancers to 3 RCC cell lines. Human normal serum was used as positive control. All 9 patients had same killing activity against each target as healthy donor.

HUMAN CDC AND ADCC TO RCC dependent cell-mediated cytotatiety activities of PBLs from 4 k d t h y donors against RCC cell lines

TABLE 4. Antibody

% Cytobxicity of F'Eb from 4 healthy donors

2 3 4 + + + -

1

Ta%ets +a

-b

SIC-RC- 7 79 19 63 9 41 12 42 10 -18 81 ND 2 6 6 4 3 9 8 -56 63 8 62 8 76 26 47 3 -29 36 2 NDC ND ND -49 12 1 ND ND ND -1 22 12 ND ND ND a Percent eytobxicity at an effeetor:target ratio of 1W1. c'ancentration of & m.7 was 5 ~ J m l . bpercent eytobxicity of PBL alone without mAb K2.7 (natural k&r activity).

Not done. 6 K2.7 alone did not show any killing.

TABLE5 . Antibody &pendent cell mediated cytotoxieity activities of PBLS from 9 patients with r e d cancers against 3 RCC cell lines % cytotoxici~

Patients

Robson stages

SK-RC-7 +a

I

6

I

49 49 50

I I I I IIIa Ivb 9 a

-b

IVb

6

87

62 62 52 23

7

17 17 2 15 20 6 1 6

-18

+

-56

-

67 17 53 16 30 7 6 4 6 8 8 5 4 71 35 79 37 49 2 43 20

+

-

44 39 31 53 82 66 50 38 33

4 5 4 4 15 32 11 1 17

See Table 4,Footnote a. See Table 4,Footnote b.

tered intravenously into rats, and levels of serum BUN and creatinine were measured twice a week for 1month. No renal toxicity was observed, and rat kidneys demonstrated no damage on histopathologic examination (data not shown). DISCUSSION

We previously demonstrated that mAb K2.7 reacted with 9 of 15 RCC cell lines by PA assay, all of the 34 normal kidney tissues and 22 of 25 renal cancer tissues by indirect immunoperoxidase staining.8 In the present study, the reactivity of mAb K2.7 was further examined on a large panel of cell lines, and normal and neoplastic tissues. An important consideration in mAb immunotherapy is whether sera and PBLs from cancer patients can kill cancer cells by the CDC and ADCC mechanisms. Ortaldo et al.ll detected ADCC activities in 2 of 4 malignant melanoma patients, but did not describe patient characteristics. Hellstrom et al.12 could not detect ADCC activities in advanced melanoma patients. These results seem to be inconsistent with ours; however, factors such as the immunologicalstatus of patients with each cancer should be carefully considered. This point requires further study in a large number of RCC patients. The study of CDC activities using sera from cancer patients has not been reported elsewhere. In this report, sera from all RCC patients, including low and high stages, showed the same CDC activities as that of the healthy donor. Based on these findings, mAb K2.7 may be useful for immunotherapy of renal cancer. The antitumor activity of mAb may depend on the isotype of the Ig and the number of binding sites per cell. Herlyn and Koprowskil3 initially observed that IgGZa mAbs were most effective in ADCC. However, when they prepared isotype variants of mAb BR55-2 (IgG3) reactive with the Y oligosaccharide and compared ADCC, CDC activities and in vivo antitumor effects among these variants, they concluded that

291

among the Ig isotypes, IgG3 and IgG2a were the moat efficient isotypes for cancer immunotherapy.14 In general, it is rare to obtain IgG3 mouse mAbs. because the number of IgG3-secreting lymphocytes in a normal m o w is much lower than other Ig isotypes. Woodhouse and Morgan16 tried to obtain IgG3 mouse mAbs by immunizing mice with a tumor extract bound to lectinaqjugated agarose beads and succeeded in obtaining 4 IgG3 mAbs. Dippold et al.16 have established IgG3 mAb R24 which showed reactivity restricted to melanoma and astrocytoma cell lines. Puke1 et al." characterized the antigen recognized by mAb R24 as the disialoganglioside GD3. Welt e t al.4 investigated ADCC and CDC activity of mAb R24 against 22 melanoma cell lines which showed striking variations in mAb R24 binding. The ADCC and CDC functions mediated by mAb R24 required a critical number of GD3 antigenic sites on the cell surface. mAb R24 did not show antitumor effectson cell lines having fewer than this critical number of mAb R24 binding sites, and the antitumor activity of mAb R24 depended on the number of antibody molecules bound to the cell surface. In the present study, we did not determine the number of mAb K2.7 binding sites on the surface of the RCC cell lines. But serological analysis by the PA assay showed the relative number of antibody binding sites. In our study, cell lines that reacted strongly with mAb K2.7 in the PA assay were more effectively killed in the ADCC and CDC assays. This point is important for clinical studies because the quantity of antigens reactive with mAb K2.7 in renal cancer tissues may vary among patients or in primary and metastatic lesions. Nevertheless, more than 90%of RCC patients, whether primary or metastatic, showed reactivity with mAb K2.7.18 The ADCC activity against the same target cell appeared to vary among effector cells from Merent individuals. Ortaldo et d.l1examined ADCC activities of mAb MB3.6 (IgG3) reactive with GD3 by various effector cells from peripheral blood. Large granular lymphocytes (LGLs) showed strong ADCC activity, but monocytes and T cells did not. These authors also examined ADCC activities of mAb 9.2.27 (IgG2a) reactive with the core protein of chondmitin sulfate proteoglycan; only LGL showed ADCC activity similar to that of mAb MB3.6. These results suggest that ADCC activities may depend on the ratio of LGLs in PBLs from individuals. A concern about the clinical application of mAb K2.7 is its reactivity with normal renal tubules, which may result in renal toxicity. In a rat model for evaluating the renal toxicity of mAb K2.7,no renal toxicity was observed. mAb R24 reacts with GD3 ganglioside and shows positive staining to human renal tubules, but it did not cause renal toxicity in a clinical Phase I trial. In the clinical trial of immunotherapy with mouse mAb, the production of human anti-mouse Ig antibody (HAMA) is an almost inevitable limitation of this therapeutic approach. The reconstruction of m o w mAb to chimeric or humanized mAb may be the future problem. The ability of mAb K2.7 to bind more than 90% of renal cancers and induce cytotoxicity mediated by complement and cellular mechanisms makes this mAb a good candidate for immunotherapy trials. Acknowledgment. We thank Dr. Eiichi Nakayama for critical reading of the manuscript. REFERENCES

1. deKernion, J. B.,Ramming, K. P. and Smith, R. B.: The natural historv of metastatic renal cell carcinoma: a computer analysis. J.-Urol., 1u): 148, 1978. 2. MeNichols. D. W.. & m a , J. W. and DeWeerd, J. H.: Renal cell carcinoma: lo&& s&val and late recurrence. J. Urol., 126 17,1981. 3. Ueda, R., Shiku, H., Pfreundschuh, M., Takahashi, T.,Li, L. T. C., Whitmore, W. F., Jr., Oettgen, H. F. and Old, L. J.: Ceil sukace antigens of human rend cancer defined by autologous typing. J. Exp. Med., 1M): 564, 1979.

292

HUMAN CDC AND ADCC T O RCC

human antibody-dependent cellular cytotolricity with murine 4. Welt, S., Carswell. E. A,, Vogel, C. A.. Oettgen, H. F. and Old, monoclonal antibodies. J. Immunol., 138 3566, 1987. L. J.: Immune and nonimmune effector functions of IgG3 mouse monoclonal antibody R24 detecting t h e disialoganglio- 12. Hellstrom, I., Brankovan. V. and Hellstrom, K E.: Strong antitumor activities of IgG3 antibodies to a human melanoma. side GD3 on the surface of melanoma cells. Clin. Immunol. Immunopathol., 45: 214. 1987. associated ganglioside. Roc. Natl. Acad. Sci. U.S.A., 82: 1499, 5. Houghton. A. N., Mintzer, D., Cordon-Cardo, C.. Welt, S.. 1985. Fliegel, 9..Vadhan, S.. Carswell, E.. Melamed, M. R., Oettgen. 13. Herlyn, D. and Koprowski, H.: IgG2a monoclonal antibodies H. F. and Old, L. J.: Mouse monoclonal 1 6 3 antibody detectinhibit human tumor growth through interaction with effector ing GD3 ganglioside: a phase I trial in patients with malignant cells. Proc. Natl. Acad. Sci, U.S.A., 79: 4761, 1982. melanoma. Roc. Natl. Acad. Sn. U.S.A.. 82: 1242, 1985. 14. Scholz, D., Lubeck, M., Loibner, H., McDonald-Smith, J., 6. Cheung, N. K. V., Lazarus, H., Miraldi, F. D., Abramowsky. Kimoto, Y., Koprowski, H. and Steplewski, 2.: Biological atC. R.. Kallick. S., Saarinen. U. M.,Spitzer, T.. StrandJord, tivities in the human system of isotype variants of oligosacS. E., Coccia, P. F. and Berger, N. A,: Ganglioside GD2 specific charide-Y-specific murine monoclonal antibodies. Cancer Immonoclonal antibody 3F8: a phase I study in patients with munol. Immunother., 33: 153, 1991. neuroblastoma and malignant melanoma. J. Clin. Oncol., 5: 15. Woodhouse, C. S. and Morgan, A. C., Jr.: Murine monoclonal 1430. 1987. IgG3 antibodies to human colorectal tumor-associated anti7. Platt, J. L., LeBien. T. W. and Michael, A. F.: Stages of renal gens: production and characterization of antibodies active in ontogenesis identified by monoclonal antibodies reactive with both antibody-dependent cellular cytotoxicity and complelymphohematopoietic differentiation antigens. J. Exp. Med., ment-mediated cytolysis. Cancer Res., 4 9 2776, 1989. 157 155, 1983. Li, L. T. C., Ikeda, H., Oettgen, 16. Dippold, W. G., Lloyd, K. 0.. 8. Kinouchi, T.. Nakayama, E.. Ueda. R., Ishiguro. S.. Uenaka. A,. H. F. and Old, L. J.: Cell surface antigen of human malignant Oda, H. and Kotake. T.: Characterization of a kidney antigen melanoma: definition of six antigenic systems with mouse defined by a m o u e monoclonal antibody K2.7. J. Urol., 137: monoclonal antibodies. Roc. Natl. Acad. Sci. U.S.A., 77: 6114, 151. 1987. 1980. 9. Kmouchi, T.. Kotake. T.. Mori, Y.and Abe, T.: Human renal cell C. S., Lloyd, K. O., Travassos, L. R., Dippold, W. G., carcinoma: establishment and characterization of a new cell 17. -el, Oettgen, H. F. and Old, L. J.:GD3-a prominent ganglioside of line IOS-RC-2). In Vitm Cell. Dev. Biol.. 21: 195. 1985. 10. Pfreundschuh. M., Shiku, H., Takahashi, T., Ueda. R.. human melanoma: detection and characterization by mouse monoclonal antibody. J. Exp. Med., 155: 1133, 1982. Ransohoff. J.. Oettgen. H.F. and Old, L. J.: Serological analysis of cell surface antigens of malignant human brain tumors. 18. Kinouchi, T., Saiki. S., N a w , T., Uenaka, A., Kotake, T., Shiku, H. and Nakayama, E.: Correlation of c-myc expression with Roc. Natl. Acad. Sci. U.S.A., 71: 5122, 1978. pleomorphism in human renal cell carcinoma. Cancer Res.,49: 11. Ortaldo. J. R.. Woodhouse. C.. Morgan, A. C., Herberman, R. B., 3627, 1989. Cheresh. D. A. and Reisfeld, R.: Analysis of effector cells in