Znt. J . Cancer: 22, 55-62 (1978)

TUMOR-SPECIFIC A N D TUMOR-ASSOCIATED MEMBRANE ANTIGENS OF ROUS SARCOMA VIRUS TRANSFORMED HAMSTER FIBROBLASTS Paolo M. COMOOLIO, Marilena BERTINIand Maria PRAT Department of Human Anatomy, School of Medicine, University of Turin, Turin; and Chair of Histology and Embryology, School of Medicine, University of Trieste, Trieste, Italy

Hamster fibroblasts transformed by an env- strain of Rous sarcoma virus (RSV) express a t their surface tumor-associated antigens of unknown origin and a tumor-specific antigen (VCSA) which i s not expressed by hamster fibroblasts transformed by unrelated D N A or R N A oncogenic viruses. This antigen was detectable by rabbit antibodies and a complement-dependent 5tCr-release cytotoxkity assay and i s common to RSV-transformed cells of different animal species. By comparing the antiVCSA serum with antisera directed against purified gp85, gs-proteins, reverse tranrcriptase or detergentlysed virus particles, it was shown that VCSA i s not a known virion structural protein. Moreover, VCSA expression does not correlate with viral replication since it is not detectable in chick embryo fibroblasts productively infected with the transformationdefective virus RAV-1 which shares virus structural genes with RSV. Finally, in hamster cells transformed by an RSV mutant, temperature-sensitive for the ability to transform the host cell, VCSA expression at the cell surface correlates with the expression of the transforming gene.

Following viral transformation, a malignant cell can express on its surface a variety of antigens not found on the corresponding normal cell. These have a different genetic origin, structure and biological properties (for recent reviews see Comoglio and Forni, 1976; Herberman, 1977). The whole problem as it now stands can be summarized as follows. Neoplastic transformation is almost invariably associated with the appearance of antigens coded for by cellular genes derepressed as a result of the interference of the oncogenic agent with cellular regulatory genes (Baldwin et al., 1974; Coggin and Anderson, 1974; Uriel, 1976). It is also generally agreed that new antigenic specificities expressed at the surface of transformed cells may be due to the insertion of virus structural proteins into the plasma membrane (Kurth, 1975; Nowinski and Klein, 1975). These antigens are virus-specific, are i.e. expressed by different cells transformed by the same virus, and their appearance is obvious in permissive systems where budding virus particles are assembled at the cell surface. On the other hand, a virus-specific cell surface antigen which is believed not to be a structural virus component has been described in the avian system. Chicken cells transformed by Rous sarcoma virus (RSV) express a surface antigen, which was defined as tumor-specific (TSSA) since it is absent in cells

productively infected by transformation-defective strains (Kurth and Bauer, 1975; Kurth, 1976). In chicken cells TSSA is expressed in association with the major envelope glycoprotein (gp85), but support for the idea that it is not a virion protein-although under the control of a viral gene-came from the finding that (1) it is expressed by RSV-transformed mammalian cells that do not produce virus particles, and (2) it lacks (unlike gp85) virus sub-group specificity. It has recently been found, however, that even though mammalian cells are not permissive for virus replication, gp85 and other viral proteins can be synthetized and expressed at their cell surface (Aupoix and Vigier, 1976). Moreover, antigenic determinants common to gp85 of different RSV subgroups have been identified (Rohrschneider et al., 1975). Thus, the possibility that TSSA might be identical to group-specific antigenic determinants of gp85 somehow expressed only at the surface of transformed cells has t o be considered. This seems not to be the case, since TSSA is also expressed by a quail cell line transformed by an RSV strain lacking the genetic information for gp85 (Bauer et al., 1977). This paper reports the antigenic composition of hamster fibroblasts transformed in vitro by RSV mutants either defective for the synthesis of gp85 (env-) or temperature-sensitive for the ability to transform the host cell (ts). These cells were found to express at their surface a virus-specific antigen, which is likely to be the mammalian counterpart of the avian TSSA since it is not one of the known virus structural proteins, and its expression is under the control of the viral gene responsible for neoplastic transformation. MATERIAL AND METHODS

Cells and viruses

The cell lines used in this work are listed in Table I. The following four sublines were derived from BHK21jc13 (C13) Syrian hamster fibroblasts: (I) B46 cells were a clone isolated in this laboratory from B4 fibroblasts transformed by the Bryan envstrain of RSV. These cells are negative for gp85 expression, as tested by radioimmunoprecipitation Received: February 20, 1978 and in revised form May 10, 1978.

56

COMOGLIO ET AL.

TABLE I CONTROL AND Cell line

C13 B46 RS 213 14-B PY PM-1 PM-3 xcit 16Q(-) CEF CEF '-RAV-1

CEF '-PR-A

VIRUS-TRANSFORMED CELL

Animal species

Hamster Hamster Hamster Hamster Hamster Hamster Hamster Rat Quail Chicken Chicken Chicken

' Primary cultures of C/E chicken embryo fibroblasts. -

LINES

Transformed by

References

-

Stoker and Macpherson (1964) Pitts (1971) Montagnier and Vigier ( I 967) Biquard and Vigier (1972) Stoker and Macpherson (1964) Monti-Bragadinand Ulrich (1972) Monti-Bragadin et al. (1970) Svoboda (1964) Murphy (1977)

RSV(BH) RSV(SR-D) RSV-ts(Fu-19) Polyoma virus Murine sarcoma virus SV-40 RSV(PR-C) RSV(BH)

-

RSV(PR-A) Infected by the non-transforming Rous associated

with specific antisera (M.P. and P.M.C., unpublished observations); (2) RS 2/3 was a clone transformed by the Schmidt-Ruppin D (SR-D) strain of RSV and selected for its high expression of proteins of the gs complex; (3) 14-B cells were transformed by the ts mutant FU-19 (Biquard and Vigier, 1967); (4) Py cells were transformed by Polyoma virus. PM-I and PM-3 were lines of Syrian hamster fibroblasts transformed in vivo by Murine sarcoma virus (MSV) or simian virus 40 (SV-40) respectively. XC/t cells were a subline derived from a rat sarcoma induced by RSV of the Prague-C (PR-C) strain. 16 Q(-) was a continuous line of quail fibroblasts infected in vitro by the env- Bryan strain of RSV; these cells were transformed and permissive for virus replication. For references of the various cell lines see Table 1. Secondary cultures of chicken embryo fibroblasts (CEF) gs-, chf-, transformed by RSV subgroup A (PR-A) or infected by the Rous associated virus-1 (RAV-I) were also used.

-

virus-I.

blastosis virus (AMV). (2) Anti-gs: a polyvalent antiserum directed against all the gs proteins of AMV cores. (3) Anti-B77: a polyvalent antiserum raised against purified B77-Prague-C RSV disrupted by detergent. A goat antiserum (anti-pol) raised against reverse transcriptase purified from AMV (obtained from US National Institutes of Health) was also used. Analytical procedures

The cytotoxic activity of the antiserum was titrated by the 61Cr-release assay as described by Wright and Law (1971). Freeze-dried guinea-pig 01 frozen rabbit sera were used as sources of complement with similar results; controls included cells incubated with antiserum alone or complement alone. Antibody binding to antigens solubilized by detergent from cells was performed as described in detail previously (Prat and Comoglio, 1976) with the following modifications: membrane proteins Antisera The anti-B46 serum was produced in rabbits were solubilized by 2 % sodium deoxycholate primed by intramuscular injection of 2.5 x lo7 (DOC) in 10 mM TrisHCI buffer PH 8.2, and the live B46 cells, carefully washed to minimize the test was done in Tris buffer saline (TBS) in the amount of calf serum of the culture medium bound, presence of 0.2 % deoxycholate. Microtiter wells and boosted by two identical doses of cells at coated with the appropriate amount of antigen 20-day intervals. The antiserum was adsorbed for were made to react for 3 h at 4" C with 200 pl 30 min in ice with equal volume of packed sheep of the test antiserum diluted 1:lOO with TBS, then erythrocytes, to remove non-specific cytophilic washed and incubated as above with lo5 CPM of immunoglobulins and antibodies against Forssman 1251-labelledgoat anti-rabbit IgG antibodies (specific antigen (Aupoix et al., 1973). It was then chromato- activity 1.7 ,uCi/pg). graphed through a column of calf serum insolubilized RESULTS by covalent binding to Sepharose 4B to remove contaminant anti-calf serum antibodies (Tarone and Comoglio, 1977). Finally, I-ml aliquots were Tumor-associated antigens common to hamster cells exhaustively adsorbed with C13 fibroblasts by transformed by unrelated oncogenic viruses incubation three times with 2 x 1 0 * live cells at The rabbit antiserum, raised against RSV-trans0" C for 60 min. formed hamster fibroblasts, after exhaustive adThe following Iabbit antisera, kindly supplied sorption with the non-transformed parental hamster by Drs. D. Bolognesi and M. Halpern, were also line C13 (anti-B46: see '' Methods "), had no cytoused : (1) anti-gp85 : a monospecific anti-serum toxic activity against normal hamster cells, but raised against gp85 purified from avian myelo- still killed the transformed B46 cells (Fig. 1). When

TUMOR ANTIGENS OF RSV-.TRANSFORMED FIBROBLASTS

tested against hamster cells (Py) transformed by the Polyoma virus-an unrelated DNA viius-a positive reaction was also observed (Fig. l), showing that this antiserum contained antibodies directed against cross-reactive tumor-associated antigen(s). Similarly, it reacted with two other hamster cell lines transformed by different oncogenic DNA or RNA viruses: PM-1 and PM-3 cells (data not shown). Experiments were then psrformed by testing this antiserum against 61Cr-labelled B46

R. a - B 46 ( C 13 ads.)

U

401

57

cells after further adsorption with Py-, PM-Ior PM-3-transformed fibroblasts. As shown in Figure 2a, Py cells adsorbed about 35% of the cytotoxic activity; this percentage did not increase as the number of the adsorbing cells increased over a certain value. On the other hand, as expected, complete adsorption was observed if the antiserum, adsorbed with adequate amounts of Py cells, was tested against Py target cells (Fig. 2b). Complete adsorption was obtained with B46 cells. PM-1- and PM-3-transformed fibroblasts behaved similarly in adsorbing specifically about 35-45 % of the cytotoxic activity of the antiserum (Table 11). This set of experiments clearly showed that antiB46 serum contained antibodies against two distinct cell-surface antigens: (1) one common to all transformed hamster lines tested, and (2) one expressed only by cells transformed by RSV. The VCSA: a virus-inducedcell surface antigen

$ 30CJ aJ -

The anti-B46 serum was then further adsorbed with Py cells. After this treatment, it reacted specifiaJ cally against the B46 line used for immunization L (Fig. 3a). The cytotoxic activity of this doubly L 20adsorbed antiserum was not removed by further V c adsorption with Py (Fig. 3b), PM-I- or PM-3transformed fibroblasts (Table 11). 2 On the other hand, the specific cytotoxic activity 10was completely removed by adsorption with B46 cells, as expected (Fig. 3b), and with the other C 13 hamster lines transformed by RSV of a different subgroup, RS 2/3 and 14-B (Table 11). 10 20 40 80 Moreover, when this doubly adsorbed antiserum s e r u m dilution-' was tested against fibroblasts of different animal species transformed by RSV, namely rat XC/t FIGURE 1 - Complement-dependent cytotoxic activity. of rabbit anti-B46 serum, adsorbed as described in cells transformed by the PR-C strain, chicken '' Methods " and tested against 61Cr-labelled B46 (o), fibroblasts transformed by the PR-A strain and the 16 Q(-) line of quail fibroblasts transformed Py (A) or C13 ( 0 ) target cells. v)

r-

R - a n t i 846 (C13 ads.) target Py target B 4 6

a

b

2 adsorbing cells x

FIGURE 2 - Adsorption of the cytotoxic activity of anti-B46 serum against B46 or Py cells. The antiserum, at a dilution giving 50% of maximum lysis, was incubated for 30 min on ice with increasing ) amounts of B46 (0), P ~ ( A or C13 ( 0 ) cells and tested against Wr-labelled B46 (a) or Py target cells (b).

4

6

0

58

COMOGLIO ET AL. R.anti-VCSA

R. anti - V C S A

b 2,100.

U

c ._ > .c

u

40

O

.-

v

X

$ -

c

:3 c

2

50.

u

a, L

L

I

-u

FIGURE3 - Cytotoxic activity of rabbit anti-B46 serum further adsorbed with Py cells (antiVCSA). (a) anti-VCSA serum tested against B46 (0), Py ( A ) or C13 ( 0 ) 61Cr-labelled target cells. (b) antiVCSA serum tested against 51Crlabelled B46 target cells after adsorption with increasing amounts of B46 (0), Py (A) or C13 ( 0 ) cells.

75.

.-c

2c

.

m

251

a

f

1c

Py, C 1 3

a 10

d

21.

h

20 40 80 serum dilution-'

2 4 8 adsorbing cells x ~ O - ~

by the env- Bryan strain, clearly positive results were obtained (Table 111). These experiments showed that the antigen, recognized by the C13- and Py-adsorbed antiserum, was common or cross-reactive to all RSV-transformed lines, indicating that its expression was under control of the viral genome. For this reason we have called this doubly adsorbed antiserum anti-VCSA and the antigen identified VCSA (virus cell surface antigen). Non-identity between VCSA and RS V structural proteins

VCSA is not a virus structural protein. This was demonstrated by comparing the cytotoxic activity on different target cells of anti-VCSA serum with that of: (1) an antiserum raised against detergentdisrupted purified RSV virion particles (anti-B77); (2) a polyvalent antiserum raised against purified internal core proteins of the gs-complex (anti-gs); (3) a monospecific antiserum against purified gp85 (anti-gp85) ; (4) a monospecific antiserum raised against purified virus reverse transcriptase (antipol) (see " Methods "). When the RSV-transformed hamster cells were used as targets, no correlation was found between the cytotoxic activity of antiVCSA serum and that of the four antisera specifically raised against virion structural proteins (Table IV). The possibility that the low cytotoxic activity of these antisera was due to antibodies of an immunoglobulin class ineffective in the complement-dependent lysis was ruled out by testing the anti-B77 serum against chicken embryo fibroblasts productively infected and transformed by RSV. Here a consistent amount of virus structural proteins was expressed on the cell surface and the antiserum showed a good cytotoxicity. The anti-pol serum was negative against the two cell lines tested. Similarly, the anti-gs serum was not significantly cytotoxic. This observation is consistent with the prominent intracellular localization of core p proteins of the gs-complex.

TABLE I1 EXPRESSION OF CROSS-REACTING ANTIGENS AT THE SURFACE OF HAMSTER CELLS TRANSFORMED BY UNRELATED ONCOGENIC VIRUSES Specific adsorption Cell line

Transformed by

Anti-846 serum

C13 RSV B46 RS 213 RSV 14-B RSV Polyoma virus PY PM- 1 Mouse sarcoma virus PM-3 SV40

0 100

97&2 98&5 35&5 45 & 3 40&2

I

of

Anti-VCSA serum

0 100

95*5 89k3 0 0 0

' % of reduction of the C'-dependent cytotoxic activity against 846 cells after exhaustive adsorption. The antisera were tested at a dilution giving 50% of the maximum lysis. - Rabbit anti-846 serum preadsorbed with C13 cells and sheep red blood cells. a Rabbit anti-B46 serum preadsorbed as above and with Py cells.

TABLE I11 COMPLEMENT-DEPENDENT CYTOTOXICITY OF RABBIT ANTI-VCSA SERUM 1 Transformed by

C13 CEF B46 RS 213 14-B XCit 16 Q(-) CEF-PR-A PY

PM-I

PM-3

'

Hamster Chicken RSV(BH) Hamster RSV(SR-D) Hamster RSV(SR-D) Hamster RSV(PR-C) Rat RSV(BH) Quail Chicken RSV(PR-A) Hamster Polyoma virus Hamster Mouse sarcoma virus Hamster SV40

% $'Cr released a

0 0

4015 42k6 4012 30k3 6OL-5 3318 0 0 0

Rabbit antiserum raised against 846 cells and exhaustively adsorbed with C13 and Py cells. - a Maximum lysis ( I j l O serum dilution) f SD.

59

TUMOR ANTIGENS OF RSV-TRANSFORMED FIBROBLASTS TABLE IV COMPLEMENT-DEPENDENT CYTOTOXICITY OF RABBIT ANTISERA AGAINST RSV VIRION PROTEINS AND VCSA

% "Cr released ' by C' and Target cells

Anti-gs

Anti-gpS5

0 0

0 0 5t4 NT NT NT

B46 RS 213 14-B

3k1

CEF-PR-A C13 CEF

1f0.5 0 0

Anti-B77

0

1 +0.5 812 5212 0 0

Anti-pol

3h0.5 1 h0.5 NT NT NT NT

Anti-VCSA

40415 42f6 4012 33 418 0 0

' Maximum lysis (1/10 serum dilution)*so.

The anti-gp85 serum was negative, as expected, against the B46 line, transformed by the env- Bryan strain, which is defective for gp85 synthesis; a very weak reaction was observed only with the 14-B line. Similar results were obtained with the anti-B77 serum. The conclusion that the specific cytotoxic activity displayed by the anti-VCSA serum towards RSVtransformed cells not due to antibodies directed against RSV structural proteins was stressed by further experiments where the binding activity of this antiserum and the one of antisera against virus structural proteins were compared in radioimmunoassay. In fact, a significant binding between the anti-VCSA serum and antigens solubilized from B46 cells was observed (Fig. 4a). On the contrary, no binding was measured when the antiserum was tested against chicken cells, infected with RAV-1, that were actively producing virus particles. In the mirror experiment (Fig. 4b and cf, when the anti-gs and anti-B77 sera were tested, a high level

R. a - V C S A

of binding was measured with antigens solubilized from RAV-1 infected cells, whereas only a weak reaction with antigens solubilized from B46 cells was observed. Since gs antigens are known to be expressed in the cytoplasm of RSV-transformed hamster cells (Aupoix and Vigier, 1976) this result was to be expected. Correlation between VCSA expression and neoplastic transformalion To investigate whether the expression of VCSA correlates with neoplastic transformation, we performed cytotoxicity and adsorption experiments on a hamster cell line transformed by the RSV ts mutant FU-19. This is a mutant of RSV-SR-D obtained by mutagenization with 5-fluorouracil which shows, in permissive cells, a loss of ability to transform the host cell at high temperature (41" C: restrictive temperature), whereas it retains its capacity to replicate, showing that the only gene function sensitive to temperature is that of

R. a - B 77

R.a-gs

a

FIGURE4 - Binding of antiVCSA and anti-virion protein sera

to plastic wells coated with decreasing concentrations of antigens solubilized from B46 (0)or CEFM v - 1 cells (A). (a) anti-VCSA. (b) anti-gs. ( c ) anti-B77. The indirect binding assay was performed with 1261-labelIedimmunoglobulins purified from a goat anti-rabbit IgG serum. Wells coated with bovine serum albumin (m) were used to calculate the background.

eiiz!9 00

10

1

100

10

1

100

10

concentration of coating antigens ( p g / m l )

1

COMOGLIO ET AL. R.anti-VCSA

R. anti -VCSA

a

40

10

20 40 serum dilution-'

FIGURE 5 - Expression of VCSA at the cell surface of hamster cells transformed by the FU-19 t s mutant. (a) Cytotoxic activity of the anti-VCSA serum on W r labelled 14-B cells grown at the permissive temperature of 37" C (0) or at the restrictive temperature of 41"C ( 0 ) . (b) Adsorption of the cytotoxic activity of anti-VCSA serum against 61Cr-labelled B46 target cells by increasing numbers of 14-B cells grown at 37" C (0)or at 41"C (0). 6

80

the sarc gene, the gene responsible for induction and maintenance of neoplastic transformation (Biquard and Vigier, 1972). The 14-B cell line used in our experiments was a clonal line obtained after transformation in vitro of BHK~13,21 cells with this mutant. When grown at 41" C these cells switch from a transformed to a normal phenotype (Aupoix et al., 1974). The anti-VCSA serum lysed 14-B cells grown at 37" C; however, when the temperature was raised to 41" C for a minimum of 12 h, the degree of lysis was dramatically reduced (Fig. 5a). These results were confirmed by adsorption experiments performed by testing anti-VCSA serum against B46 cells after adsorption with 14-B cells, grown either at 37" C or at 41" C. Those cells grown at 37" C completely adsorbed the cytotoxic activity of the antiserum, while those grown at 41" C were largely ineffective (Fig. 56). In controls performed with a rabbit anti-hamster serum, the cytotoxic antibodies were adsolbell by 14-B cells grown at either temperature; moreover, the hamster cell line RS 2/3 transformed by the wild-type RSV-SR-D were lysed and able to adsorb completely the cytotoxic activity of the anti-VCSA serum when grown at either 37" C or 41" C (data not shown). Further experiments were performed by testing the anti-VCSA serum on chicken embryo fibroblasts either transformed by RSV or infected by the transformation-defective strain RAV-1. Since chicken cells are permissive for virus replication, under both conditions virus structural proteins

TABLE V COMPLEMENT-DEPENDENT CYTOTOXlCITY OF RABBIT ANTISERA AGAINST RSV VIRION PROTEINS A N D VCSA

Target cells

CEF CEF CEF ' Maximum

Infected by

-

RAV-1 RSV(PR-A)

7

8

log,o adsorbing cells

% 6'Cr released

' by C

AntLB77

Anti-VCSA

9 35&8 52k3

312 311 35 &6

lysis ( 1 j l O serum dilution) f SD.

and

were synthetized and mature virions assembled and released from the cell surface. As shown in Table V, the anti-VCSA serum was significantly cytotoxic for the transformed cells only. The anti-B77 serum, o n the contrary, did show cytotoxic activity for either RAV-1-infected or RSV-transformed cells. The precise nature of the target molecule for these lytic antibodies is unknown, since the gp85 of RAV-1 and RSV-B77 differs in subgroup specificity and the anti-gs serum we used did not lyse intact cells. The expression of gp85 group-specific antigenic determinants at the surface of virusinfected or transformed cells will be further investigated. Control chicken embryo fibroblasts (gs-, chf-) were not lysed by either antiserum. DISCUSSION

The experiments reported show that xenogeneic antisera obtained by immunizing rabbits with RSV-transformed hamster fibroblasts after appropriate adsorption recognize, in the W r release cytotoxicity assay, at least two tumor-membrane antigens: (1) an antigen common to hamster cells transformed by unrelated DNA or RNA oncogenic viruses, and (2) a different one, called VCSA (virus cell surface antigen) which is virus-specific and is also expressed by ceils of different animal species transformed by RSV. Since the first antigen was found o n hamster cells transformed by unrelated viruses, we can reasonably assume that it is coded for by a cellular gene, whose nature is unknown. Many tumor cells transformed by oncogenic viruses express at their surface a Forssman-like antigen (Fogel and Sachs, 1962; O'Neil, 1968; Burger, 1971; Aupoix et al., 1973). However, adsorption of our antiserum with Forssman-positive sheep red blood cells (see "Methods") did not remove the cytotoxic antibodies directed against hamster fibroblasts transformed by RSV, Polyoma virus, MSV o r SV-40. Re-expression of embryonal antigens on malignantly transformed cells has been demonstrated in a variety of other spontaneous and experimental animal tumors (for review see: Baldwin etal., 1974; Coggin and Anderson, 1974; Uriel, 1976). This seems

TUMOR ANTIGENS OF RSV-TRANSFORMED FIBROBLASTS

to be due to derepression of cellular genes, normally repressed in differentiated cells, as a consequence of malignant transformation, or of the entry of the cell into the mitotic cycle (Bertini et al., 1974; Burk and Drewinko, 1976). Embryonal antigens in both chemical- and virus-induced tumors almost invariably cross-react between different tumors within the same species (Baldwin el al., 1972; Thompson and Alexander, 1973 ; Comoglio and Forni, 1973; Comoglio ef al., 1975). The wide cross-reaction observed in our system suggests that the antigens shared by RNA and DNA virustransformed fibroblasts belong to this class, although we have not checked their in vivo expression. The presence of embryonal antigens in avian sarcomas and in RSV-induced mouse tumors has, indeed, already been reported (Kurth and Bauer, 1973). This cross-reacting tumor antigen is clearly distinct from VCSA. This conclusion is supported by the specificity of the cytotoxic activity of the anti-VCSA serum (doubly adsorbed with control and polyoma-transformed hamster cells), which is restricted to cells transformed by RSV. The VCSA is exposed at the surface of hamstel cells transformed by RSV only; moreover it is also found on RSV-transformed fibroblasts from different animal species, namely chicken, quail and rat. This shows that its expression is under the control of the viral genome. However, the VCSA is not a virus structural protein. This possibility is ruled out by the following observations: (1) the anti-VCSA serum has been raised against a cell transformed by an env- strain of RSV, which lacks genetic information for virus envelope glycoproteins (Scheele and Hanafusa, 1971; De Giuli et al., 1975); (2) the mammalian cell lines lysed by the rabbit anti-VCSA serum were not lysed by rabbit antisera specifically directed against virus structural proteins of the whole virion, of the internal core, or of the envelope; (3) the binding activity of the anti-VCSA serum, tested in radioimmunoassay against antigens solubilized from transformed or untransformed virus-producing cells, was not correlated to that of the antisera against virus structural proteins; (4) in 14-B cells, transformed by the RSV t s mutant FU-19, the expression of the VCSA at the cell surface is temperature-sensitive, unlike virus structural proteins, which are synthetized at both permissive and restrictive temperatures (Biquard and Vigier, 1972); (5) finally, the VCSA is detectable only on RSV-

61

transformed chicken embryo fibroblasts, but not on chicken embryo fibroblasts infected by the transformation-defective Rous associated virus-1, where virus particles are assembled and released from the cell surface. From the above findings we can conclude that the VCSA is a membrane antigen, the expression of which is under the control of the vixal genome, and that it is not a virion structural protein. The only RSV gene not coding for virion structural proteins is the sarc gene, the gene responsible for induction and maintenance of neoplastic transformation (for review see Weiss, 1976). The fact that the VCSA was expressed on ts mutant-transformed cells only at permissive temperature, and that it was not detectable on cells infected by the transformation defective strain, strengthens the conclusion that its expression is under the control of the sarc gene. In this respect, VCSA seems to be the mammalian counterpart of the avian TSSA (Kurth and Bauer, 1975; Kurth, 1976). The VCSA itself may be the product of the sarc gene; alternatively it may be coded for by a cellular gene under control of the sarc gene product. In the second hypothesis the cellular gene involved should have been conserved during evolution in order to explain the immunological cross-reaction of the VCSA observed in different animal species. Recently, a tumor-specific non-virion antigen of 61),000 mol.wt. has been associated with the sarc gene expression in transformed chicken and hamster cells (Brugge and Erikson, 1977). Its localization within the cell is, however, still unknown. The elucidation of the relationship between this antigen, the avian TSSA and the VCSA, which are expressed at the cell surface, will contribute to an understanding of the role played by the plasma membrane in malignant transformation. ACKNOWLEDGEMENTS

This work was supported by funds of the Italian National Research Council (C.N.R.). The authors are indebted t o Dr. F. Tatb, Dr. S. Barlati and Dr. P. Vigier for generously supplying the cell lines used in these experiments, and to Dr. M. Halpern for providing the antisera against virion proteins. The excellent technical assistance of Miss M. R. Amedeo and Mrs. M. Mura-Lagna is gratefully acknowledged.

ANTIGENES MEMBRANAIRES TUMORISPECIFIQUESET ASSOCIES A LA TUMEUR SUR LES FIBROBLASTES DE HAMSTER TRANSFORMES PAR LE VIRUS DU SARCOME DE ROUS Les fibroblastes de hamster transform& par une souche P I I Y - du virus du sarcome de Rous (RSV) expriment B leur surface des antigenes associts A la tumeur d’origine inconnue et un antigene tumorispecifique (VCSA) que n’expriment pas les fibroblastes de hamster transform& par des virus oncogenes B A D N ou A ARN non apparent&. Cet antigene a pu &re decele au moyen d’anticorps de lapin et dans un test de cytotoxicite dependant du complement bast sur le relargage du Cr”; il Btait commun aux cellules d’espbces animales diffkrentes qui avaient 6te transformees par le RSV. En comparant le sCrum antiVCSA et des antidrums diriges contre la gp85, les protkines gs, la transcriptase inverse ou les particules virales lysees par un detergent purifiees, on a montre que le VCSA n’est pas une proteine structurale virionique connue. De plus, son expression n’est pas en correlation avec la replication virale puisqu’on ne le deckle pas dans les fibroblastes d’embryon de poulet infect& productivement avec le RAV-I, qui est dkpourvu d’activite de transformation et qui a des genes structuraux communs avec le RSV. Enfin, dans les cellules de hamster transform& par un mutant du RSV, dont la capacite de transformer la cellule h6te est thermosensible, I’expression du VCSA A la surface cellulaire est en corr6lation avec I’expression du gene transformant.

62

COMOGLIO ET AL. REFERENCES

AUPOIX, M., SIMKOVIC, D., and GAZZOLO,L., Tumor specific cell surface antigen and Forssman antigen associated with rat and hamster cells transformed with different strains of avian sarcoma virus. Neoplasma, 20, 481-490 (1973). AUPOIX,M., BIQUARD, J. M. and CACHARD, A., Cell surface antigen induced by avian tumor viruses in hamster cells transformed by a temperature-sensitive mutant of Rous sarcoma virus. Int. J . Cancer, 14, 61 1-616 (1974). AUPOIX,M., and VIGIER,P., Expression of viral proteins in mammalian cells transformed by avian sarcoma viruses. Int. J. Cuncer. 18, 787-797 (1976). BALDWIN,R. W., EMBLETON, M. J., PRICE,M. R., and VOSE,B. M., Embryonic antigen expression on experimental rat tumors. Tmnsplant. Rev., 20, 77-99 (1974). BALDWIN,R. W., GLAVES, D., and VOSE, B. M., Embryonic antigen expression in chemically induced hepatomas and sarcomas. Int. J . Cancer, 10,233-239 (1972). BAUER, H., IGNIATOVIC, J., RUBSAMEN, H., and HAYAMI, M., Transformation associated cell surface antigens in virus and chemically transformed avian cells. Med. Microhiol. Immunol., 164, 197-205 (1977). BERTINI,M., FORNI,G., and COMOGLIO, P. M., A tumor associated membrane antigen transiently expressed by normal cells during mitosis. Clin. exp. Immunol., 18, 101-108

KURTH. R., Avian cell transformation and expression o f avian sarcoma virus-specific tumor antigens. Nature (Lond.), 264, 261-263 (1976).

KURTH,R., and BAUER,H., Aviaii oncornavirus-induced tumor antigens of embryonic and unknown origin. Virology, 56, 496-504 (1973). KURTH,R.,and BAUER,H., Avian RNA tumor viruses. A model for studying tumor associated cell surface alterations. Biochem. Biophys. Acta, 417, 1-24 (1975). MONTAGNIER, L., and VIGIER,P., Transformation de cellules d e hamster par deux souches, non defective et defective d u virus de Rous. C . R. Acad. Sci. (Paris), 264, sCrie D , 193-196 (1967).

MONTI-BRAGADIN, C., CONVENTI, L., and MELONI,G. A,, Sulla produzione di SV40 negli eterocarionti d a fusione di cellule sensibili con cellule trasformate sottoposte ad irradiazione UV. Boll. Soc. Ital. biol. Sper., 95,565-569 (1970). MONTI-BRAGADIN, C., and ULRICH,K., Rescue of the genome of the defective murine sarcoma virus from a non-producer hamster tumor cell line, PM-I, with murine and feline leukemia viruses as helpers. Int. J. Cancer, 9, 383-392 (1972). NOWINSKI, R . C., and KLEIN,P., Anomalous reactions of mouse alloantisera with cultured tunior cells. 11. Cytotoricity is caused by antibodies t o leukemia virus. J . Immunol., 115,

(1 974).

1261-1268 (1 975).

BIQUARD, J. M., and VIGIER,P., Characteristics of a conditional mutant of Rous Sarcoma Virus defective in ability to transform cells a t high temperature. Virology, 47,444-455

O'NEIL,C. H., An association between viral transformation and Forssman antigen detected by immune aderence in cultured BHK-21 cells. J. Cell Sei., 3, 405-409 (1968). PUTS, J. D., Growth properties o f normal and transformed cells. In: G. Wolstenholme and J. Knight (eds.) Growth, control in cell cultures, CIBA Foundation Symposium, Churchill, Edinburgh (1971). PRAT,M., and COMOGLIO, P. M., A solid-state conipetitive binding radioimmunoassay for measurement of antigens solubilized from membranes. J . Immunol. Methods, 9, 267-

(1967).

BRUGGE, J., and ERIKSON, R., Identification of a transformation-specific antigen induced by an avian sarconla virus. Nature (Lond.), 269, 346-348 (1977). BURGER,M. M., Forssman antigen exposed on surface membrane after viral transformation. Nature (Lond.), 231, 125-127 (1971).

BURK,R., and DREWINKO,B., Cell cycle dependency of tumor antigens. Cancer Res., 36, 3535-3538 (1976). COGGIN,J. H.,, and ANDERSON, W. C., Cancer differentiation and embryonic antigens: some central problems. Advanc. Cancer Res., 19, 106-159 (1974). COMOGLIO, P. M., BERTINI,M., and FORNI,G., Evidence for a membrane carrier molecule common to embryonal and tumor-specific antigenic determinants expressed by a ntouse transplantable tumor. Immunology, 29, 353-364 (1975). COMOGLIO, P. M., and FORNI,G., Plasma cell and tumorassociated membrane antigens of mouse plasmacytoma MOPC-315and MOPC-460. Int. J . Cancer,l2,613-625(1973). COMOGLIO, P: M., and FORNI,G . , Molecular basis of the immunogenicity of cell surface tumor antigens. Exp. Cell. Biol., 44, 150-169 (1976). DE GI U L I,C., KAWAI,S., DALES,S., and HANAFUSA, H., Absence of surface projections on some non-infectious forms of RSV. Virology, 66, 253-260 (1975). FOGEL,M., and SACHS,L. Studies o n the antigenic coniposition of hamster tumors induced by polyoma virus and normal hamster tissues in vivo and in vitro. J . nat. Cancer Inst., 29, 239-249 (1962). HERBERMAN, R. B., Immunogenicity of tumor antigens. Biochem. Biophys. Acta, 473, 73-92 (1977). KURTH,R.,Tumor virus protein at the cell surface. Nature (Lond.), 256, 613-614 (1975).

272 (1976).

ROHRSCHNEIDER, L. R., BAUER,H., and BOLOGNESI, D. P., Group-specific antigenic determinants on the large envelope glycoprotein of avian oncornaviruses. Virology, 67, 234-241 ( 1 975).

SCHEELE, C., and HANAFUSA, I-I., Proteins of helper-dependent RSV. Virology, 45, 401-410 (1971). STOKER,M., and MACPHERSON, I., Syrian hamster fibroblast cell line BHK/2l and its derivative Nature (Lond.), 203, 1355-1357 (1964).

SVOBODA, J., Malignant interactions of Rous virus with mammalian cells in vivo and in vitro. Nut. Cancer Inst. Monogr., 17, 277-292 (1964). TARONE, G. and COMOGLIO,P. M., Plasma membrane proteins exposed on the outer surface of control and Rous sarcoma virus-transformed hamster fibroblasts. Exp. Cell Res., 110, 143-152 (1977). THOMPSON,D., and ALEXANDER, P., A cross-reacting embryonic antigen i n the membrane of rat sarcoma cells which is immunogenic in the syngenic host. Brit. J. Cancer, 27, 35-44 (1973).

URIEL, J., Cancer, retrodifferentiation and the myth of Faust. Cancer Res., 36, 4269-4275 (1976). WEISS, R., Molecular analysis of the oncogene. Nature (Lond.), 260, 93 (1976). WRIGHT,P. W. and LAW, L. W., Quantitative in vitro measurement of simian virus 40 tumor-specific antigens. Proc. Nut. Acad. Sci. ( W a s h ) , 68, 973-977 (1971).