Znt. J. Cancer: 18, 243-249 (1976)

SEROLOGICALLY DETECTABLE SPECIFIC AND CROSS-REACTIVE ANTIGENS ON THE MEMBRANE OF A POLYOMA VIRUS-INDUCED MURINE TUMOR Isaac P. WITZl, Nora LEE and George KLEIN Department of Microbiology, George S. Wise Center for Life Sciences, Tel Aviv University, Raniat Aviv, Tel Aviv, Israel: and Department of Tumor Biology, Karolinska Institutet, Stockholm, Sweden

SUMMARY

With the aid of an assay measuring complementdependent cytotoxicity mediated by syngeneic antibodies, we performed a serological analysis of surface antigens of a polyoma-virus-induced murine tumor (SEYF-a). In vivo propagated SE YF-a ascites tumor cells expressed a specific membrane antigen in addition to various other cross-reacting antigens. Among these we could identify at least four separate specificities. Two of these were present on MuLVinduced lymphoma cells, the first on Moloney-virusinduced YAC cells and the second on Gross-virusinduced GHA cells. The third cross-reacting antigen was detected on EL-4 cells. At least one additional specificity was present on two methylcholanthreneinduced murine sarcomas. Normal syngeneic lymphoid cells were insensitive to cytotoxicity mediated by the anti-tumor antisera. Quantitative and perhaps also qualitative diferences between the antigenic expression of in vivo propagated and cultured SE YF-a cells were indicated. These studies show that hyperimmune sera produced in syngeneic mice against transplanted tumors may contain a considerable number of antibody specificities, only some of which are speciJic for the tumor. Furthermore the results also suggest that polyoma-virus-inducedtumors may possess individually distinct antigenic specificities, over and above the known cross-reacting TSTA or TSSA type antigen. A bewildering number of antigenic specificities have been identified on the surface of experimental and human tumor cells. Methods used to pinpoint them include a variety of serological reactions, cellmediated responses (with different effector cell populations) and combinations of the two. It is now obvious that for further progress some kind of an " antigenic dictionary " will have to be developed for various well-defined tumor cell types. It may be recalled that, for example in the field of human transplantation, progress was only achieved after workers reached agreement on the identification and nomenclature of the major histocompatibility antigen system in man. Complex as it is, the HL-A antigen system is the product of loci on a single chromosome,

whereas a much vaster antigenic world appears to be concealed behind the present broad designation of " tumor-associated antigens ". It seems prudent to start from serological analysis. This is not to underrate the significance of cellmediated reactions, but to facilitate the identification of antigenic " departure points " reproducible in different laboratories by the easy exchange of serum reagents. The highly informative mapping of lymphocyte and other differentiation-associated antigens by Boyse, Old and their co-workers (Old and Boyse, 1973) may illustrate the benefit of such an approach. In the choice of suitable tumor systems, it appears reasonable to start from virally or chemically induced mouse tumors where a reasonable amount of background information exists concerning in vivo recognition of TSTA-i.e. rejection-inducing antigens in the syngeneic host. Methylcholanthrene- and polyoma-virus-induced sarcomas in mice are probably the prototypes which have been best studied so far. In the choice of serum reagents, the syngeneic mouse is preferable to allogeneic or xenogeneic hosts. While, at a later stage, after appropriate antigen purification, monospecific sera produced in xenogeneic hosts may be highly useful, this does not appear to be a proper point of departure because of the multitude of specificities, many of them not tumor-related at all, recognized by the foreign host. Using this approach Ting and co-workers (Ting and Herberman, 1974; Ting et al., 1974) analysed syngeneic antisera directed against a Friend-virusinduced leukemia and an SV40 tumor, and demonstrated that these antisera contained antibodies against multiple antigenic specificities. The present study is an attempt to characterize the antigenic specificities on the membrane of SEYF-a cells, a polyoma-virus-induced murine tumor (Sjogren, 1964) by antibodies present in syngeneic or semi-syngeneic antisera. It became evident that the Received: March 15, 1976.

244

WIT2 ET AL. TABLE I MURINE TUMOR CELLS USED IN THE PRESENT STUDY

Tumor

Strain in which originally induced and genotype

Oncogenic agent

Histological characteristic

Mode of propagation of cells used in study

A.BY (H-Z6)

Polyoma virus Moloney virus Gross virus Benzo(a)pyrene Methylcholanthrene Methylcholanthrene Methylcholanthrene

Sarcoma Lymphoma Lymphoma Lymphoma Sarcoma Sarcoma Sarcoma

Ascites and culture Sjogren (1964) Ascites Klein and Klein (1964) Ascites Klein ef a/. (1 962) Ascites Gorer (1950) Bataillon et a/.(1975) Culture Bataillon et al. (1975) Culture Bataillon et al. (1 975) Culture Wiener et al. (1973) Culture

SEYF-a YAC GHA EL-4 MC57-M MC57-T MC57-G A9HT

A (H-2")

C3H (H-Zd) C57B1 (H-2b) C57B1 (H-26) C57BI (H-Z6) C57Bl (H-2*) C3H (H-Zd)

Reference

A malignant derivative of the A9 line (the A9 cell is an 8 azaguanine-resistant derivative of the L-cell line).

syngeneic host recognizes a considerable number of antigens on the target cell surface and we have attempted to define and characterize some of the reactions by fractionated absorption.

MATERIAL AND METHODS

Tumor cells SEYF-a is the ascitic form of a murine sarcoma induced in the A.BY strain by polyoma virus (Sjogren, 1964). SEYF-a cells are also maintained in culture. The cultured line is serially propagated in RPMI 1640 medium with 10% fetal calf serum. The line was maintained for approximately 12 months in vitro prior to use and grew as a typical malignant SEYF-a ascites tumor upon reinoculation into A.BY mice, thereby proving its malignancy. Ascites or cultured SEYF-a cells were used to produce antisera and as target cells in cytotoxicity assays. Details on other tumor cells used in this study are given in Table I.

Sera against ascites or cultured SEYF-a cells were produced in A.BY or F, hybrids of A.BY mice by multiple inoculations of cells irradiated with 6,000 rad. Mice were injected once every 14-21 days with 5 x 106-5x lo7cells and bled 7-10 days later through the retro-orbital sinus. Periodical bleeding of mice started after six, seven or eight injections. The serum of mice belonging to several immunization groups (each consisting of 10-30 mice) was pooled into batches of about 10 ml. The different serum batches were derived from the same hyperinimunized animal groups and represent pools from different bleedings. In this study we used three pools of antiserum directed against ascites SEYF-a cells (Nos. 26, 28 and 34) and two pools of antiserum directed against cultured cells (Nos. 27 and 33). The schedule of hyperimmunization in terms of number of immunizing inoculations and number of cells inoculated was similar in the groups immunized with cultured and with ascites cells. The antisera were stored frozen at -20" C until use.

g

P O O L NQ

P O O L NQ 2 8 1I

u 0

i

R e c i p r o c a l of

Hyperimmune antisera

2

4

a n t i s e r u m dilution

34

8

1 6 3 2

FIGURE 1 The activity of two pools (Nos. 28 and 34) of anti-ascites SEYF-a antiserum towards various murine Ascites tumor cells. w-w: A-A : cultured SEYF-a; SEYF-a; o-o: YAC; 8-8: GHA; A-A: EL-4; n-n: MC57-M; V-V: MC.57-T; 0-0 : MC57-G ; V-V: A9HT.

245

POLYOMA TUMOR ANTIGENS

centration (Ran et al. 1976). The suspension was incubated for 45 min at 37" C. At the end of the incubation period, cytotoxicity was determined.

GHA

Absorptions

1 2 4

a

Exhaustive absorption aimed at eliminating antibody activity towards a certain cell type was carried out as follows: 1 vol of undiluted antiserum was incubated with 1/3 or '/2 volume of absorbing cell pellet for 1-2 h at 4"C with occasional stirring. The absorbing cells were discarded after centrifugation. This procedure was repeated until complete elimination of cytotoxic activity against the absorbing cells was achieved. 163264

1 2 4

0 163264 RESULTS

R eci pr ocal of antiserum d ilu tio n

FIGURE2 Reactivity of anti-ascites SEYF-a antiserum pool No. 34 towards YAC and GHA cells. 0-0: Unabsorbed; 0-0 : negative towards YAC cells; x -x : negative towards YAC and GHA cells.

Cytotoxicity assays

Assays to measure complement-dependent lysis (CdL) of cells mediated by the hyperimmune antiSEYF-a antisera were performed in two steps using trypan-blue exclusion. Twenty p1 of antiserum dilution were mixed with 5p1 of cell suspension containing lo7 cells/ml. The mixture was incubated for 30min at room temperature and the cells were then spun down and washed once. Thirty p1 of fetal rabbit serum as a source of complement at a dilution of 1:6 for all cells and 1:20 for ascites SEYF-a cells were then added to the cell pellet. The reason for the difference in complement concentration was that ascites SEYF-a cells are coated in vivo with potentially cytotoxic antibodies and hence lyse following addition of exogeneous complement at high con-

MC 57- M

€ L-4

4

I n view of the reactivity towards MuLV-induced lymphoma cells YAC and GHA, we carried out exhaustive absorptions of the antisera with such lymphoma cells. Absorptions were repeated until no reactivity against the absorbing cells could be detected (Fig. 2). It should be noted that antisera with no detectable activity against the Moloney-virusinduced YAC were still reactive with Gross-virusinduced GHA cells (Fig. 2). Figure 3 compares the reactivity towards EL-4, SEYF-a and MC57-M of the antisera before and after exhaustive absorption with Moloney-virusinduced YAC cells. The absorption had only minor effects on reactivity towards the three target cells or none at all.

SEYF-a

7. 2

Figure 1 demonstrates the CdL curves obtained by two antiserum pools towards various in vivo propagated and cultured tumor cells. As can be seen, the antisera directed against ascites SEYF-a cells exhibit a wide cross-reactivity with many of these tumors. The antisera did not kill lymph-node cells from normal A.BY mice.

FIGURE3 Reactivity of anti-ascites SEYF-a antiserum pool No. 34 towards EL-4, MC57-M and ascites SEYF-a cells. 0-0: Unabsorbed; 04: negative towards YAC cells.

8 16 32 6k 128

2

4

8 16 32 64 128

R e c i p r o c a l of a n t l s e r u r n

dilution

2 4

B

1 6 32 64128

246

WITZ ET AL.

EL- 4

l0I

FIGURE 4 Reactivity of anti-ascites SEYF-a antiserum pool No. 34 towards EL-4; MC57-M and ascites SEYF-a 0-0: Unabsorbed; cells. x -x : negative towards YAC and GHA cells.

Reclprocsl 0 1

antiserum dilution

As shown in Figure 4, absorbed sera which reacted against neither YAC nor GHA cells still killed EL-4, SEYF-a and MC57-M cells, all of the H-2b genotype. The antisera which no longer reacted against MuLV-induced tumor cells were exhaustively absorbed with EL-4 cells until no activity against the absorbing cells could be detected (Fig. 5). The absorbed antisera still reacted with SEYF-a and MC57-M cells (Fig. 6).

SEYF-a

0'

0.5

The absorbed antisera with no demonstrable activity for YAC, GHA and E L 4 cells were further exhaustively absorbed with a mixture of MC57-M 2

4

8 16 32 64 128

2

4

a ib 32

64 128

Reciprocal Of antiserum dilution

1a

EL- 4

U

0.5

1 2 4 8 16 32 64 128 Reciprocal of antiserum dilution FIGURE 5 Reactivity of anti-ascites SEYF-a antiserum pool No. 34 towards EL-4 cells. o--0: Unabsorbed; A-A: negative towards YAC, GHA and EL-4 cells.

FIGURE 6 Reactivity of anti-ascites SEYF-a antiserum pool No. 34 towards MC57-M and ascites SEYF-a cells. 0-0 : Unabsorbed; A-A : negative towards YAC, GHA and EL-4 cells.

and MC57-T cells, until no activity against these cells could be detected (Fig. 7). These absorbed sera retained their activity for ascites SEYF-a cells (Fig. 8) but did not react with about 10 other cultured and ascites murine tumor-cell lines (to be published). These results indicate the presence of an apparently specific serologically detectable membrane antigen of ascites SEYF-a cells. Syngeneic antisera were also raised against cultured SEYF-a cells. Figure 9 compares the activity of these antisera to the activity of the antisera raised against ascites cells, towards the following cells: ascites SEYF-a; cultured SEYF-a, YAC, GHA, E L 4 and MC57-M. The following results were obtained: both types of antisera had a similar cytotoxic activity towards cultured SEYF-a and MC57-M cells. It may be noted that the antisera against cultured SEYF-a cells had a very low titer against ascites SEYF-a cells indicating

POLYOMA TUMOR ANTIGENS MC57.M

1.0,

V

,

MC57T

I

1

2

4

8

16 32 6 4 128 1 2 4 8 16 32 6 4 128 Reciprocal o f antiserum dilution

FIGURE7 Reactivity of anti-ascites SEYF-a antiserum pool No. 34 towards MC57-M and MC57-T cells. 0-0: Unabsorbed; A-A : negative towards YAC, GHA, EL.-4, MC57-M and MG57-T cells.

perhaps that certain surface structures are more highly immunogenic on the in vivo propagated cells than on the cultured ones. The antibody titer against the MuLV-induced tumors YAC and GHA was higher in the antisera against the cultured SEYF-a cells than in the antisera against the ascites tumors. On the other hand, the titer against EL-4 was higher in the antisera against ascites SEYF-a cells. These results indicate quantitative and perhaps qualitative differences in the antigenic or immunogenic properties of cultured and in vivo propagated SEYF-a cells.

247

EL-4 cells and the other on cultured chemically induced sarcomas. We have no information as to the nature of these antigens. There were only relatively small amounts of detectable antibodies to fetal antigens in the anti SEYF-a antisera or none at all. Absorption ofthe antisera with embryonic tissue had in general no effect on the cytotoxicity titer against various target cells. We are aware, however, of the fact that the spectrum of embryonic tissue antigens changes very rapidly. Thus the possibility exists that such antigens were not expressed on the embryonic tissue used for absorption because of a restriction to certain stages of development and differentiation. In addition to cross-reacting surface antigens we detected on SEYF-a cells an antigen which is apparently specific for these cells. Preliminary unpublished results have indicated that it is not expressed on two other polyoma-virus-induced tumors. This finding probably eliminates the possibility that this surface antigen is virus-specific and indicates that it is an individual tumor-associated and perhaps tumor-specific antigen. Results showing the existence of individual tumorspecific antigens in virally induced tumors have already been published (Weiss, 1969; Morton et al., 1969). These results and the ones reported in the present paper suggest that virus-induced neoplasms may carry, in addition to their known cross-reactive antigens, individually distinct antigens comparable to the chemically induced tumor systems but normally overshadowed by the cross-reactive antigen demon-

DISCUSSION

The present paper demonstrated that in vivo propagated poly oma-virus-induced SEYF-a celI s express several cross-reacting serological specificities. At least one of these was detected on YAC cells, a murine leukemia induced by the Moloney virus. Another non-cross-reactive specificity was detected on GHA cells-a Gross virus-induced lymphoma. These antigens may represent either virally determined surface antigens and/or MuLV antigens. The latter have indeed been demonstated on cultured chemically induced murine sarcomas (Grant et al., 1974). It was also recently reported by Nowinski and Klein (1 975) that various mouse alloantisera showed unexpected reactions against cultured tumor cells lacking the corresponding alloantigens. This reactivity was due to antibodies directed against envelope proteins of MuLV present on the surface of these tumor cells. Expression of MuLV antigens of SEYF-a may be the result of exogeneous contamination of transformed cells with MuLV. However, no data are available to support this possibility. In addition to MuLV antigens we detected at least two other cross-reacting antigens, one present on

POOL NQ28

P O O L NO_ 3 4

2 4 8 16 32 2 4 8 16 32 Reciprocal of a n t i s e r u m dilution

FIGURE 8

Reactivity of two absorbed anti-ascites SEYF-a antiserum pools (Nos. 28 and 34) towards ascites SEYF-a cells. The absorbed antisera were negative towards YAC, GHA, EL-4, MC57-M and MC57-T cells. Two independent assays are shown.

248

WITZ ET AL.

S E Y F-a CU L T URE

SEYF-a ASCITES

FICUKE 9

2

4

8

1632

2

4

8

1632

2

4

8 1 6 3 2

1

Reactivity of anti-ascites and cultured SEYF-a antisera towards various murine tumor cells (specified on the Figure). U-0: Antisera against cultured SEYF-a cells; O-•: antisera against ascites SEYF-a cells. A-pool No. 27 (anti-culture) and pool No. 28 (anti-ascites). B-pool No. 33 (anticulture) and pool No. 34 (antiascites).

0.5 a

SEYF-a CULTURE

MC 57-hh

ASCITES

-.++PI=--W

2

4

8

1632

2

4

8 1632

2

4

8

1632

Reciprocal of antiserum dilution

strated in rejection tests. Individual antigenic specificities may become apparent when fractionated absorptions are used as was done in the present study, or when the animal is tolerant to the common antigen. More experiments are required, however, to establish this point firmly. We found some striking differences between in vivo and in vitro propagated SEYF-a cells. Apparently the latter had a higher capacity to induce antibody formation against MuLV-induced cells. This may be due to the fact that the Moloney cell surface antigen is found at a higher concentration in cultured cells than in in vivo propagated cells as originally described by Cikes et al. (1973) and confirmed by Dorval et al. (personal communication).

It was also interesting to note that antisera against cultured SEYF-a cells had almost no cytotoxicity toward ascites cells, although they were sensitive to CdL mediated by the antisera against the ascites cells. This means that if the cultured cells express the specific SEYF-a antigen, as they apparently do, it is not expressed in immunogenic form, or that cultured cells express lower amounts of the SEYF-a antigen than ascites cells. In this case the SEYF-a antigen behaves in a similar way to the genetically determined H-2 antigens. It is not unlikely that the reason for the decreased expression of the SEYF-a antigen on in vitro propagated cells is due to the higher expression of the MuLV-associated antigens acting as space occupants.

POLYOMA TUMOR ANTIGENS ACKNOWLEDGEMENTS

The study was supported by NIH Grant No. SR01-CA-14054-02 to George Klein and by Contract NOI-CB-43858 of the Division of Cancer Biology and Diagnosis, National Cancer Institute, NIH, to Isaac P. Witz.

249

Part of the study was undertaken during the tenure of an American Cancer Society-Eleanor Roosevelt-International Cancer Fellowship awarded by the International Union Against Cancer to I. P. Witz while he was on sabbatical leave at the Department of Tumor Biology, KaroIinska Institutet, Stockholm, Sweden.

ANALYSE SEROLOGlQUE DES ANTIGENES SPECIFIQUES ET A REACTIONS CROISEES DE LA MEMBRANE DES CELLULES D’UNE TUMEUR MURINE INDUITE PAR LE VIRUS POLYOME Nous avons effectut une analyse serologique des antigknes de la surface des cellules d’une tumeur murine induitepar le viruspolyome ( S E YF-a) en mesurant la cytotoxicitt des anticorps syngtniques like au compliment. Les cellules de la tumeur ascitique SE YF-a propagkes in vivo expriment un antigkne membranaire sptccifique ainsi que divers autres antigenes ci rtactions croiskes. Parmi ces derniers, nous avons p u identifier au moins quatre sptcificitks distinctes. Deux de ces antigknes ttaient prtsents sur des cellules de lymphome induit par le MuLV, le premier sur des cellules YAC induites par le virus de Moloney et le second sur des cellules GHA induites par le virus de Gross. Le troisieme a ttk dttectt sur les cellules EL-4; il y en avait au moins un quatrieme sur des sarcomes murins induits par le mkthylcholanthrlne. Les cellules lymphoides syngkniques normales sont insensibles ci la cytotoxicitt mkdike par les antistrunis antitumoraux. Des difftrences d’expression antigknique, non seulement quantitatives mais peut-&re aussi qualitatives, sont apparues entre les cellules S E YF-a propagtes in vivo et les m h e s cellules en culture. Ces etudes montrent que les skrums hyperimmuns produits chez des souris syngtniques contre les tumeurs transplantkes peuvent contenir un nombre considkrable de spicificitks dont quelques-unes seulement sont sptcifiques pour la tumeur. De surcroit, les rtsultats conduisent a penser que les tumeurs induites par le virus polyome peuvent posskder des spteificitts antigkniques in~ividuel~ement uniques en plus des antigknes du type TSTA ou TSSA, dont on connait les rkactions croiskes.

REFERENCES

BATAILLON, G., PROSS,H., and KLEIN,G., Comparative in virro sensitivity of two methylcholanthrene-induced murine sarcoma lines to humoral and cellular immune cytotoxicity. Int. J . Cancer, 16, 255-265 (1975). CIKES,M., FRIBERG, S., JR., and KLEIN,G., Progressive loss of H-2 antigens with concomitant increase of cell-surface antigen(s) determined by Moloney leukemia virus in cultured murine lymphomas. 1.nut. Cancer Inst., 50, 347-362 (1973). CORER, P. A., Studies in antibody response of mice to tumor inoculation. Brit. J. Cancer, 4, 372-379 (1950). GRANT,J. P., BIGNER,D. D., FISCHINGER, P. J., and BOLOGNESI, D. P., Expression of murine leukemia virus structural antigens on the surface of chemically induced murine sarcomas. Proc. nut. Acad. Sci. ( W a s h . ) , 71, 50375041 (1974). KLEIN, E., and KLEIN, G., Antigenic properties of lymphomas induced by the Moloney agent. J. nut. Cancer Insr., 32, 547-568 (1964). KLEIN,G., SJOGREN, H. O . , and KLELN, E., Demonstration of host resistance against isotransplantation of lymphomas induced by the Gross agent. Cuncer Res., 22,955-961 (1962). MORTON,D. L., MILLER,G . F., and WOOD, D. A., Demonstration of tumor-specific immunity against antigens unrelated to the mammary tumor virus in spontaneous mammary adenocarcinomas. J . not. Cuncer Inst., 42,289-301 (1 969). NOWINSKI, R. C., and KLEIN,P. A., Anomalous reactions of mouse alloantisera with cultured tumor cells. 11. Cytotoxicity

is caused by antibodies to leukemia viruses. J. Imrnunol., 115, 1261-1268 (1975). OLD, L. J., and BOYSE,E. A., Current enigmas in cancer research. Harvey Lect., 67,273-315 (1973). RAN, M., KLEIN, G., and WITZ, I. P., Tumor-bound immunoglobulins. Evidence for the in vivo coating of tumor cells by potentially cytotoxic anti-tumor antibodies. I n t . J. Cancer, 17, 90-97 (1976). SJOGREN, H. D., Studies on specific transplantation resistance t o polyoma virus induced tumors. I. Transplantation resistance induced by polyoma virus infection. J . nut. Cancer Inst., 32, 361-364 (1964). TING,C. C., and HERBERMAN, R. B., Serological analysis of immune response to Friend virus-induced leukemia. Cancer Res., 34, 1676-1683 (1974). TING, C. C., ORTALDO, J. R., and HERBERMAN, R. B., Serological analysis of the antigenic specificities of simian virus 40-transformed cells and their relationship to tumorassociated transplantation antigen. 1. naf. Cancer Insr., 52, 815-821 (1974). WEISS,D., Immunological parameters of the host-parasite relationship in neoplasia. Ann. N. Y. Acad. Sci., 164,431-448 (1969). WIENER,F., KLEIN,G., and HARRIS,H., The analysis of malignancy by cell fusion. IV. Hybrids between turnour cells and a malignant L-cell derivative. J. Cell Sci., 12, 253-261 (1973).