Int. J. Cancer: 15, 483-492 (1975)
DISTINCTION BETWEEN TUMOR-SPECIFIC TRANSPLANTATION ANTIGEN AND VIRION ANTIGENS IN SOLUBILIZED PRODUCTS FROM MEMBRANES OF VIRUS-INDUCED LEUKEMIC CELLS
Kenneth S . S . CnANci, Lloyd W. LAWand Ettore APPELLA Laboratory of Cell Biology, National Cancer Institute, Bethesda, Maryland 20014, USA
A membrane antigen from R B L J leukemic cells that was solubilized and partially purified isfurther characterized in this study. This soluble antigen is capable of immunizing syngeneic hosts to reject neoplastic cells and thus resembles TSTA. It also induces cytotoxic antibody in syngeneic hosts capable of specifically lysing RBL-S cells in vitro. RBL-5, however, releases infectious virus ( R L V );it was necessary therefore to rule out virus or structural virion antigens as the effective immunogen. Infectious virus was not detectable in our initial crude membrane ( C M ) material, nor in the papainsolubilized CS or the 6-150 Sephadex-chromatographed fraction. Virus-neutralizing antibody was not detected, under stringent assay conditions, in the syngeneic anti-CM sera. Antigen preparations C M , CS and the chromatographed fractions FI, F2 and F3 were assayed in a complement-fxation test against broad-reacting antisera capable of detecting virus envelope antigen and gs antigen and against syngeneic antisera. Although our antigen preparations were positive for virion antigens, CS and F2 contained an antigen that reacted only with syngeneic antiserum. These same fractions were those reactive as immunogens. On the basis of these data, it is postulated that a cellular membrane component, other than viral, functions as TSTA.
We have reported (Law and Appella, 1973) that a tumor antigen isolated and solubilized from cell membranes of Rauscher leukemogenic virus (RLV)-induced RBL-5 leukemic cells, then partially purified, was capable of immunizing syngeneic or semi-syngeneic mice so that specific rejection of leukemic cells occurred. Immunogenicity, as revealed by tumor rejection and also antigenicity as detected by in vitro assays, were found to be preserved through the solubilization and purification procedures. The model employed and the standard reference were those used in this laboratory for solubilization and purification of histocompatibility antigens (H-2) of micelimited papain digestion followed by Sephadex
G-150 column fractionation (Strober et a[., 1970; Law et al., 1971, 1972, 1974; Appella et al., 1974). This method was also used successfully in solubilizing and purifying the TSTA of membranes of sarcoma cells induced by a DNA oncogenic virus, SV40 (Drapkin et al., 1974); specific immunogenicity again was preserved by the purification procedures. An obvious possible source of TSTA in neoplasms induced by RNA oncogenic viruses of the leukemia-sarcoma complexes is infective virus or viral structural proteins since productive, non-lytic infection continues indefinitely in most of these induced neoplasms. Leukemogenic RNA viruses mature by budding from the cell surface
Received: October 8, 1974, and in revised form December 27, 1974.
48 3
CHANG ET AL.
and this surface localization may represent effective antigen. In contrast, TSTA of DNA oncogenic virus-induced neoplasms (e.g. those neoplasms induced by SV40 and polyoma viruses) are easily distinguished from virion antigens. Preliminary data from this laboratory strongly suggested that the tumor antigens isolated from RBL-5 and responsible for specific tumor rejection were of non-virion origin (Law and Appella, 1973). We now present further data that more fully characterize our solubilized antigen as a new surface component which is not a structural part of RLV. A preliminary report of some of these additional studies has been given (Chang et al., 1974a). MATERIAL AND METHODS
Animals
Female C57B1/6N (BL/6) mice were obtained from the Veterinarian Resources Branch, National Institutes of Health, Bethesda, Md., USA and were used at 2-4 months of age. Tumors
RBL-5 originally induced by RLV, and FBL-3 originally induced by Friend leukemogenic virus were carried in ascitic form in BL/6 mice. These tumors of the FMR group share a common TSTA and common antigenic determinants as detected serologically. Extraction and solubilization of antigens from tumor cells
Antigen was prepared from RBL-5 ascites cells as described previously (Law and Appella, 1973) usually starting from approximately 2 x 1 O l o cells. Instead of sonication, the nitrogen decompression method of disrupting cells was employed in some cases with equally efficient results. After limited papain digestion a crude soluble preparation (CS) was obtained followed by Sephadex G-150 column fractionation into F1 (void), F2 and F3-included fractions. Several CS preparations were obtained also from FBL-3 ascites cells. I n all the pooled F2 fractions assayed and in the CS preparation of FBL-3, the H-2 specificities 2, 5 and 28 (found on cell membranes of H-2b haplotype) were always detected by inhibition of complement-dependent cytotoxicity using monospecific alloantisera.
484
Antisera Syngeneic anti-RBL-5 crude membrane (antiCM) serum was obtained from B1/6 mice repeatedly immunized with CM prepared from RBL-5. Usually 400,ug of the C M preparation were administered SC once a week with an equal amount of Freund's adjuvant for the first immunization and without the adjuvant for the subsequent immunizations. A broad-reacting antiserum (reacting against both gs and envelope antigens of murine leukemia viruses (MuLV) Fischer rat anti-Moloney murine sarcoma virus (MSV) serum was obtained from Dr. R. E. Wilsnack (Huntingdon Research Center, Baltimore, Md., USA). An anti-RLV serum prepared from Syrian golden hamsters was obtained by immunization with the tumor cells carrying large amounts of infectious RLV (Rhim et al., 1969). This serum was also found to be broad-reacting. Guinea-pig anti-gs serum was prepared by three weekly immunizations with purified g s antigen (Oroszlan et al., 1970) mixed with Freund's adjuvant. These antisera were absorbed with BL/6 mouse liver cells to remove hetero-antibodies and were heated at 56" C for 30 min to inactivate complement. Titrations for infectious MuL V
For detection of RLV, the material was inoculated at various dilutions on NIH Swiss mouse embryo (NIH-ME) cells in culture plates; cells were pretreated with 25 ,ug/ml diethylaminoethyl (DEAE)-dextran. These plates were either ultraviolet (UV)-irradiated or subcultured, and tested for syncytial cell formation after cocultivation with XC cells. The details of these methods were described previously (Chang et al., 1974b). In vitro neutralization tests The method described previously (Chang et al., 1974b) was followed. Both RLV and MSV(RLV) at dilutions giving about 100 XC-plaque-forming units (XC-PFU) or transformed cell focusforming units (FFU) respectively, were added to the appropriate dilutions of serum to be tested. Complement fixation (CF) tests
These tests were performed in the Microtiter System using 1.8-2.0 units of guinea-pig complement and overnight fixation at 5" C (Chang et al., 19746; Hartley et a f . , 1965). Both antisera
NON-VIRION TUMOR ANTIGEN IN VIRAL LEUKEMIA
and antigens were diluted in veronal buffer, PH 7.4 and a checkerboard titration was performed. A reaction giving a reading of 3+ or higher complement fixation was taken as the endpoint. Anticomplementary activity was encountered only with antigen extracts and antisera at high concentrations. No anticomplementary effect was observed with the guinea-pig anti-gs serum at the dilutions used. The reproducibility of the test system was demonstrated by running duplicate tests on separate days and the maximum deviation was two-fold.
FBL-3 leukemic cells in B1/6 mice given a single immunization with 30pg or lOOpg of the F 2 fraction of RBL-5 antigen preparation. The 50% tumor-inducing dose (TDsO)of FBL-3 was p>0.01. A significant difference(0.05>p>0.01) in frequency of " takes " in the 100 pg group was also observed at an FBL-3 cell challenge of 1 x lo* (2+/7 compared with 13 / I4 controls). Mean days to death in parentheses.
+
TABLE I1 RESULTS
Immunogenicity of solubilized tumor membrane preparations
Since considerabls data concerning this aspect of the investigation have been published (Law and Appella, 1973), only pertinent and more recent findings will be described here. Inhibition of the growth of FBL-3 cells was observed using adoptively transferred sensitized lymphoid cells (Winn assay). The results set forth in Table I show a significant reduction in the number of takes of a 5 x lo2 cell challenge with
IMMUNOGENICITY OF FBL-3 CRUDE SOLUBLE (CS) MATERIAL (ADOPTIVE TRANSFER) FBL-3 challenge Immunization
None FBL-3 cells (1 x loe) FBL-3 (CS) (15Opg)
10' cells loa cells (No. tumor/No. mice)
+
7/7 (26.4) 6/6 (22.3) 7/7 (23.1)
7/7 (33.6)
0/7 (>90) 1/7 (52)
1 Immunizations with viable FBL-3 cells and crude soluble (CS) material were performed subcutaneously. Lymphoid cells (spleen) from immunized mice or from normal controls were mixed at a ratio of 1OO:l target-cell and challenge made intraperitoneally in C57B1/6 mice. Mean days to death in parentheses. Tumor dose 50 of FBL-3 6 lo* cells.
48 5
CHANG ET AL.
toneal challenge with 1 x103 FBL-3 cells. The immunogenicity of 150 pg was essentially that obtained with 1 x lofi viable FBL-3 cells inoculated subcutaneously. FBL-3 is not oncogenic at this cell concentration when inoculated into the subcutaneous tissues. Detection of infectious type-C viruses in leukemia RBL-5 and its membrane preparations
To rule out the presence of infectious type-C viruses in our crude membrane (CM) preparations of RBL-5, these were inoculated on NIH-ME cell cultures known to support the replication of RLV. At the same time, infectious center assays of the X-irradiated RBL-5 cells and virus titrations of freeze-thawed extracts (10%) of RBL-5 cell suspensions were carried out. The presence of RLV on NIH-ME cells was indicated by syncytial cell formation after XC cells were added to the UV-irradiated NIH-ME cells. As shown in Table 111, 0.02-0.03% of cells of RBL-5 registered as infectious centers on the indicator cells while lo" infectious virus particles could be detected in 0.4 ml of the freeze-thawed RBL-5 extract. No infectious RLV could be detected in the CM, CS, F1, F2 and F3 preparations respectively even at the concentration of 500 pg/ml. NIH-ME cells were subcultured and
TABLE 111 DETECTION OF INFECTIOUS TYPE-C VIRUSES BY XC TESTS
Test samples 0.4 ml/inoculum
X-irradiated RBL-5 cells loficells los cells 1 O4 cells
RBL-5cell extracts Freeze-thawed, supernate, I0 % Crude membrane, 500pg/ml Crude soluble, 500pg/rnl F1, 500 pg/rnl F2, 500pg/rnl F3, 500 pg/ml
XC-PFU on NIH-ME cells treated with DEAE-dextran
I88 25 3 106 0' 0 0 0
0
' The absence of infectious type-C viruses in the membrane preparations of RBL-5 was confirmed by the inability of these preparations to induce production in ME cells of progeny viruses detectable by measurement of virion-associated activity of reverse transcriptnse using oligo(dT) poly(rA) as a primer-template.
486
XC cells were added for co-cultivation. With this more sensitive technique which allows enrichment and secondary spread of viruses in the culture, no infectious RLV could be detected in CM preparations. The possibility of xenotropic viruses contaminating the RLV-producing RBL-5 cells was checked by culturing RBL-5 cells in the presence of SIRC cells (American Type Culture Collection) which are known to support the growth of some xenotropic viruses but not of RLV (Benvenistc et a[., 1974) and testing the culture supernate for the presence of virus particles possessing reverse transcriptase activity. No such viruscs could be detected. Absence of RL V-neutralizing activity in the syngeneic anti-CM serum Table IV details the results of in vitro neutralization tests with syngeneic anti-CM serum against RLV and MSV(RLV). XC-PFU and FFU assays revealed no inhibition by the serum at dilutions of 1 :4 and 1 : l O . A positive control, hamster anti-RLV serum and a negative control, serum from normal B1/6 mice, were included. These results suggest that the CM preparation used for immunization of syngeneic mice did not contain enough virion antigens to induce the production of neutralizing antibody. Also tested was the possibility of the presence in the syngeneic anti-CM serum of antibodies directed against viruses commonly found in mice either as a result of immunization with the CM containing such viral antigens or as a result of naturally occurring infections. By hemagglutination-inhibition tests and complement fixation tests (done by Dr. J. C. Parker, Diagnostic Services, Microbiological Associates, Inc., Bethesda, Md., USA) no such antibody was found against PVM, Reo 3, GDVII, K, polyoma, Sendai, ectromelia, mouse adeno, mouse hepatitis and lymphocytic choriomeningitis viruses at the dilutions used. Complement fixation tests
Several lots of our antigen preparations including CM, CS, F1, F2 and F3 preparations and of RBL-5, and the culture supernatants of RLV and MLV containing loRand 1 0 XC-PFU/ ml respectively were tested. Antisera included broad-reacting antiviral sera, i.e. Fischer rat anti-MSV(MLV) serum and
NON-VIRION TUMOR ANTIGEN IN VIRAL LEUKEMIA TABLE IV I N VITRO NEUTRALIZATION TESTS
Test serum
Dilution
RLV XC-PFUiplate
MSV(RLV) FFU/plate
Syngeneic anti-RBL-5 crude membrane
4 10 4 10 40 80 160 320 -
122 124 131 120 0 0 1 129 125
93 90 85 87 0 0 0 24 84
Normal BL/6 mice (negative control) ,,
9,
Hamster anti-RLV (positive control)
Diluent
hamster anti-RLV serum, both reacting against murine viral gs and envelope antigens. CF reactions obtained with 2 units of antisera (one dilution lower than the endpoint dilution) are presented. In addition, various lots of syngeneic anti-CM serum were also tested. Table V gives typical results obtained with some syngeneic antisera which showed relatively higher activities against both known preparations of RLV and MLV as well as various preparations of solubilized RBL-5 antigens. The antiviral sera i.e. Fischer rat and hamster sera gave positive reactions with various antigen preparations as well as with the high-titered RLV and MLV. The minimal concentration in terms of protein concentration (pglml) of antigens detectable by Fischer rat anti-MSV(MLV) serum
was 125,250,225 and 87.5 for CM, CS, F2 and F 3 respectively. (F1 fraction was excluded from the Sephadex G-150 column and not tested in this experiment). Hamster anti-RLV serum gave similar values. These results indicate the presence of viral envelope (and gs) antigens in the various antigen preparations although no infectious virus particles were found. Two lots (lots A & B) of syngeneic anti-CM serum reacted with RLV and MLV, indicating the presence of antibody specific t o viral antigens. However, the CF reactions obtained with our various RBL-5 antigen preparations, CM, CS, F2 and F3, as expressed in terms of the minimal concentration of antigen detectable by syngeneic antiserum (lot A), were 125, 62, G1.8 and 22 pg/ml respectively. Lot B antiserum gave
TABLE V
COMPLEMENT FIXATION TEST-I
Antigen preparations (lot I )
Fiscber rat anti-MSV(MLV) broad-reacting 1 :40
Hamster anti-RLV broad-reacting 1 :40
Syngeneic anti-RBL-5 crude membrane serum Lot A 1 :40
Lot B 1 :40
RBL-5 CM
+(125) +(250)
+(62.5) +(5W
F1 F2 F3
NT
NT
+(87.5) +(1/32) +w4)
+(225) +(1/16) +(1/4)
cs
RLV MLV 1
+(225)
+(125) (62) NT +K1.8) (22) +(1/16) +(1/4)
+ +
+(I251 +(62)
NT
+(< 1.8) +(1/16) +(1/4)
Minimal concentration of antigen detectable (protein concentration pglml). = not tested. Minimal CF reactive fraction of an original Concentration of a virus suspension.
* NT J
487
CHANG ET AL.
identical results except that F3 was negative. It is of interest to note that, while these syngeneic antisera reacted against RLV and MLV to the same extent as did the rat and hamster antiviral sera, their reactions against CM, CS, F2 or F 3 were generally more sensitive than those detected by antiviral sera, as shown by the lower amount of antigen detectable by the former than by the latter antisera. The minimal concentration of antigen detectable by the syngeneic antisera was considerably lower (G1.8 pg/ml) for F2 as compared with F3 (22pg/ml), CS (62pglml) or CM (125 pglml). These differential C F activities of different antisera are better illustrated by calculating the ratio of the reciprocals of the minimal concentrations of the respective antigens detectable by CF test for each antiserum (Table VII). These findings suggest the presence of non-virion antigen in the solubilized antigen preparations of RBL-5, which is especially concentrated in the F2 fraction. In another series of experiments, some lots of syngeneic anti-CM serum were found to react against some of the solubilized antigen preparations of RBL-5 but not against high-titered RLV or MLV. An example of CF reactions against these antigens obtained with one such antiserum, lot C, together with those reactions obtained with guinea-pig anti-gs serum as well as rat antiMSV(MLV) and hamster anti-RLV sera, are presented in Table VT. The finding that the syngeneic anti-CM serum reacted against CS and F2 only, without detect-
able reaction against the high-titered RLV or MLV, indicates the presence of non-virion antigen in these preparations. It is evident also that CS, F 2 and F3 preparations contained virion antigens detectable by rat anti-MSV(MLV) and hamster anti-RLV sera. Reactions with guineapig anti-gs serum indicated the presence of gs antigen in F2, F3 as well as in RLV preparations. Thus although all antigen preparations except F1 contained viral envelope and gs antigens, CS and F2 contained an antigen (or antigens) which is not virion antigen and is detectable only by syngeneic anti-CM serum. In other words, syngeneic anti-CM sera contain an antibody which reacted exclusively with non-virion antigen that is present in CS and partially purified in our F2 fraction. In order to compare the differential CF reactivities of antigens in F2, F 3 and RLV against an antiviral antiserum and a syngeneic anti-CM serum, ratios of the reciprocals of minimal concentrations of the respective antigens were computed (Table VII). These ratios indicate the relative concentrations or degrees of purification of antigens in F2, F3 and RLV with respect to both viral antigen content and non-virion antigen content. Where one of the antigen preparations showed no reactivity, no activity ratio was computable and a sign of + I - or -/ t was used. While the viral antigens in F2 remained at low levels, as compared with F3 or RLV, the relative reactivities of syngeneic antisera against F2 were considerably higher than those of F3 or
TABLE VI
COMPLEMENT FIXATION TEST-11 Antigen preparations (lot 2)
I
9
488
Fischer rat anti-MSV(MLV) broad-reacting
Hamster anti-RLV broad-reacting
Guinea-pig anti-gs
1 :40
I :40
1:10
Minimal concentration of antigen detectable (protein concentration pg/ml). Minimal CF reactive fraction of an original concentration of a virus suspension.
Syngeneic crude membrane Lot
c
I :40
NON-VIRION TUMOR ANTIGEN IN VIRAL LEUKEMIA TABLE VII DIFFERENTIAL CF ACTIVITIES OF F2, F3 AND RLV
Activity ratio
F2/F3 F2/RLV
"N",4. I I1 I
I1 F3/RLV
I I1
Sygenei anti-RBL-5 Rat anti-MSV(MLV) crude membrane serum serum Lot A Lot B Lot C
0.4 1 .o 0.14 0.5 0.36 0.5
2312
+/+/-
2 9
>9
+/0.7
-/+
-/.-
The ratio of the reciprocals of the minimal concentrations of the respective antigens (F2 and F3) detectable by CF test except for RLV in which case minimal fraction detectable was used.
RLV. By contrast, the reactivities of F3 against syngeneic antisera were not so pronounced as those of RLV. Thus the data are in accord with the concept that trace or moderate amounts of virion antigens (including envelope and gs antigens) may be present in all preparations of solubilized RBL-5 membrane antigens, but that the non-virion antigen, presumably tumor-cell membraneassociated antigen, may be partially purified and concentrated in the F2 fraction. This is suggested by the differential ratios of C F reactivity of some lots of syngeneic anti-CM sera and by reactivity with some other lots of syngeneic anti-CM sera that reacted exclusively with non-virion antigen in CS and F2 preparations. The immunological specificities of the C F activity of these RBL-5 antigens were indicated by the failure of Gross virus-induced syngeneic leukemic cells, EGG2, to absorb from the syngeneic anti-CM serum the CF antibody reacting with the F2 antigen. DISCUSSION
The protein entity from the cell membrane of RBL-5 (and FBL-3) leukemic cells that we have solubilized and partially purified has a tumor rejection capacity that is specific and therefore represents TSTA. The in vitro inhibition by soluble antigen of complement-dependent cytotoxicity as measured by 51Crrelease corresponds with in vivo activity and is also specific (Law and Appella, 1973). Since RBL-5 cells release infectious virus, it is necessary to rule out a role of infectious virus itself or of virion antigens
e.g. envelope antigen or gs antigen, in tumor rejection. Indirect evidence (Klein and Klein, 1964; Micheel and Pasternak, 1968; Steeves, 1968)and direct visualization by immuno-electron microscopy (Aoki et al., 1972) support the existence of virus-coded but non-virion cellsurface antigens of certain MuLV-induced neoplasms. However, detailed studies have not been reported in tumor antigen solubilization experiments (Martyre et al., 1973; Plata and Levy, 1974) to distinguish this type of antigen from virion antigens contaminating the preparation, nor is it known that these identified antigens act as TSTAs. There is sufficient evidence to show that MuLVs alone in all probability serve in some manner as TSTAs. Some neoplastic cells that are not detectably immunogenic nor immunosensitive t o the rejection response of MuLV-immunized mice become rapidly immunogenic and immunosensitive upon superinfection, for example with MLV (Stephenson and Aaronson, 1972). However, the evidence we present here strongly indicates the existence additionally, in an MuLV system, of a new cellular antigen specific for the virus-in this case RLV-of the TSTA type that appears to be analogous to the TSTA found in neoplasms induced by DNA-oncogenic viruses. No infectious RLV was detectable in our C M preparation since the procedure for preparation of CM from RBL-5 cells included sonication, which disrupted the virus as well as the cell membrane, and centrifugation at 16,000 rpm for 90 min, which probably eliminated most of the complete virions and structural antigens of disrupted virions. However, a certain amount of viral structural antigens was still detectable by the CF test in our CM, CS, F2 and F3 preparations. Apparently, our CM preparations did not contain either infectious RLV or enough viral envelope antigens (which are specifically able t o induce neutralizing antibody) but did contain viral structural antigens responsible for induction of RLV- or MLV-reacting CF antibody in syngeneic mice. The absence of infectious RLV in CM rules out the possibility of replication virus in vivo by introduction of an RLVcontaminated CM preparation which would have resulted in amplification of RLV production. Some pools of syngeneic anti-CM sera did not have detectable CF antibody reacting against the high-titered viruses, RLV and MLV. It is evident
489
CHANG ET AL.
that the CM used for immunization did not contain enough viral structural antigens to be effective immunogens in these cases. More exacting conditions for the preparation of CM with minimal contamination with virion antigens are presently under study. The differential reactivities observed with our F2 preparation (following Sephadex G-150 chromatography) in the C F assay support the concept that the solubilized RBL-5 membrane antigen possessing the activity of TSTA of the transplantation type is not virion structural antigen. This preparation had higher activity against syngeneic anti-CM serum than against rat or hamster anti-viral sera. The F2 preparation also reacted in the C F assay with lots of syngeneic anti-CM sera that did not react with high-titered RLV o r MLV. Since anti-CM serum was prepared in syngeneic mice it should have minimal or no anti-gs antibody; mice are usually “ tolerant ” to this antigen (Huebner et al., 1971) although apparent exceptions have been noted recently (Markham et al., 1972). It is of interest to note that the Fischer rat anti-MSV serum and the hamster anti-RLV serum at concentrations which reacted against RLV and MLV in the CF test did not manifest complement-dependent cytotoxicity against RBL-5. The syngeneic anti-CM serum which was non-reactive against both RLV and MLV did exhibit, however, cytotoxicity against RBL-5 and this activity was not blocked by pretreatment of the serum with purified, ether-disrupted RLV. Since this anti-CM serum reacted with the F2 fraction in the C F test and since the F2 fraction inhibited the cytotoxicity of anti-CM serum against RBL-5, the question of the identity of these antibodies deserves further investigation.
An approach to differentiate decisively TSTA from virion antigens is to utilize “ non-producer ” tumor cells. Recently we have studied leukemia RBL-3 which was originally induced in B1/6 mice by RLV in a similar manner as RBL-5 (McCoy et al., 1967) and maintained by in vitro culture. This line proved to be immunogenic in vivo in B1/6 mice against leukemia FBL-3 which possesses F M R tumor antigen shared by RBL-5, and vice versa. Moreover, this lymphocytic neoplasm produced no infectious RLV but was nevertheless MuLVgs-antigen positive. Therefore this can be defined as a “ non-producer ” neoplasm in the classical sense. The syngeneic anti-CM sera were cytotoxic against RBL-3 cells as well as against RBL-5 cells. This model system, therefore, may be used to study the relationship of TSTA and virion antigens. The fact that TSTA and H-2 co-chromatograph on the G-150 Sephadex column (Law and Appella, 1973; Drapkin et al., 1974) points to some similarity in molecular size and structure between the two antigens. Since it is very unlikely that we are measuring an H-2 activity in our syngeneic system, one possibility is that the TSTA is a non-H2 activity not expressed on fully differentiated normal tissue but re-expressed on neoplastic cells. The possibility that we are measuring an activity ofexternal virion membrane components is also very unlikely because of our use of a proteolytic cleavage procedure. The major iodinated components of virion membranes obtained from proteolytic cleavage have a molecular weight in the 73,000-75,000 range (Witte and Weissman, 1975). This MW is much higher than the molecular weight of TSTA or of H-2 (45,000-50,000). Upon fractionation on the G-150 Sephadex column, these two entities should separate.
DISTINCTION ENTRE L’ANTIGENE DE TRANSPLANTATION SPl?CIFIQUE DE LA TUMEUR ET LES ANTIGENES VIRIONIQUES DANS LES PRODUITS SOLUBILISES PROVENANT D E MEMBRANES DE CELLULES LEUCEMIQUES VIRO-INDUITES Dans la prisente dude, les auteurs caractirisent de faGon plus approjondie un antighe de la membrane des cellules leucimiques RBL-5 qui avait de‘ja Pte‘ solubilise‘ et partiellement purifie’. Cet antigPne soluble est capable d’immuniser des hdtes syngkniques et de les ainener ir rejeter les cellules nkoplasiques; il ressemble donc au TSTA. I1 induit aussi chez les hdtes synginiques I’anticorps rytotoxique capable de lyser spPci$quement les cellules RBL-5 in vitro. Toutejois, les RBL-5 lib2rent du virus
490
NON-VIRION TUMOR ANTIGEN IN VIRAL LEUKEMIA
infectieux ( R L V ); il ttait done ntcessaire d’tcarter la possibilitt que le virus ou les antigbnes virioniques structuraux soient responsables de cet effet immunogtne. Nous n’avons pas diceli de virus infectieux dans le mattriel initial, c’est-&dire la membrane brute ( C M ) , ni dans les prtparations solubles brutes ( C S ) solubilistes a la papaihe, ni dans les fractions chrornatographiies sur Sephadex G-150. Nous n’avons pas dttectt d’anticorps neutralisant le virus dans les se‘rums synginiques anti-CM, bien que les tests aient t t t effectuts trts soigneusement. Les priparations antigtniques CM et CS et les fractions chrornatographites FI, F2 et F3 ont t t i testtes par une ipreuve de fixation du compltment contre des anti-strums qui permettent de dttecter l’antigkne de l’enveloppe virale et I’antigtne gs, et contre des antistrums syngtniques. Bien que nos prtparations antigtniques soient positives pour les antigknes virioniques, les CS et les F2 contenaient un antigbne yui ne rtagissait qu’avec l’antistrum syngtnique. I1 s’agissait des fractions qui avaient un effet immunogdne. Sur la base de ces observations, nous Cmettons l’hypothkse qu’un composant de la membrane cellulaire, qui n’est pas de nature virale, joue le r d e de TSTA.
REFERENCES
APPELLA,E., HENRIKSEN, P., NATORI,T., TANIGAKI, N., LAW, L. W., and PRESSMAN, D., Basic structure of mouse histocompatibility antigens. Transplant.Proc., In press (1975). AOKI,T., HERBERMAN, R. B., JOHNSON,P. A., LIU,M., and STURM, M. M.. Wild-tvDe Gross leukemia virus: Classification of soluble antigens (GSA). J. Virol., 10, 1208-1219 (1972). BENVENISTE, R. E., LIEBER, M. M., and TODARO, G. J., A distinct class of inducible murine type-C viruses that replicates in the rabbit SIRC line. Proc. nat. Acad. Sci. (Wash.), 71, 602-606 (1974). CHANG,K. S. S., APPELLA, E., and LAW,L. W., Some characteristics of solubilized tumor antigen isolated from an RNA virus-transformed neoplasm. Fed. Proc., 33,791 (1974a). CHANG,K. S. S., LAW,L. W., and AOKI,T., Type-C RNA virus isolated from SJL/J mice. J. nat. Cancer Inst., 52, 771-184 (19746). DRAPKIN,M. S., APPELLA,E., and LAW, L. W., Immunogenic properties of soluble tumor-specific transplantation antigen induced by Simian virus 40. J. nat. Cancer Inst., 52, 259-263 (1974). HABEL,K., Immunological determinants of polyoma virus oncogenesis. J . exp. Med., 155, 181-193 (1962). HARTLEY, J. W., ROWE,W. P., CAPPS,W. I., and HUEBNER, R . J., Complement fixation and tissue culture assays for mouse leukemia viruses. Proc. nat. Acad. Sci. (Wash.), 53, 931-938 (1965). HOWELL,S. B., ESBER,E. C., and LAW, L. W., Cellular immunity in mice with Simian virus 40induced mKSA tumors. J. nat. Cancer Inst., 52, 1361-1363 (1974). HUEBNER, R. J., SARMA,P. S., KELLOFF,G. J., GILDEN,R. J., MEIER,H., MYERS,D. D., and PETERS,R. L., Immunological tolerance to RNA tumor-virus genome expressions : Significance of
tolerance and prenatal expressions in embryogenesis and tumorigenesis. Ann. N.Y. Acad. Sci., 181, 246-271 (1971). KLEIN,G., and KLEIN,E., Antigenic behavior of Moloney lymphomas : independence of virus release and immunosensitivity. Science, 145, 1316-1317 (1964). KLEIN,E., and SJOGREN, H. O., Humoral and cellular factors in homograft and isograft immunity against sarcoma cells. Cancer Res., 20, 452-461 (1960). LAW,L. W., and APPELLA, E., Immunologic properties of solubilized tumour antigen from an RNA virustransformed neoplasm. Nature (Lond.), 243, 83-87 (1973). LAW,L. w.,APPELLA, E., STROBER, s.,WRIGHT, P. w., and FISCHETTI, T., Induction of immunological tolerance to soluble histocompatibility-2 antigens of mice. Proc. nat. Acad. Sci. (Wash.), 69, 18581862 (1972). LAW,L. W., APPELLA, E., STROBER, S., WRIGHT, P. W., and FISCHETTI, T., Soluble transplantation antigens: Further studies of their tolerogenic properties. Transplant., 18, 487-495 (1974). LAW, L. W., APPELLA,E., WRIGHT,P. W., and STROBER, S., Immunologic enhancement of allogeneic tumor growth with soluble histocompatibility-2 antigens. Proc. nat. Acad. Sci (Wash.), 68, 3078-3082 (1971). MARKHAM, R. V., SUTHERLAND, J. C., CIMINO,E., DRAKE, W. P., and MARDINEY, M. R., Immune complexes localized in the renal glomeruli of AKR mice: the presence of MuLV gs-1 and C-type R N A tumor virus gs-3 determinants. Rev. europ. Et. elin. biol., 17, 690-694 (1972).
MARTYR& M.-C., HALLE-PANNENKO, O., and JOLL~S,P., Characterization and partial purification of normal and tumor associated transplantation
49 1
CHANG ET AL.
antigens of Rauscher leukemia cells. Ei4rop. J . Cancer, 9, 757-761 (1973). McCoy, J. L., FEFER, A., and GLYNN,J. P., Comparative studies on the induction of transplantation resistance in BALB/c and C57BL/6 mice in three murine leukemia systems. Cancer Res., 27, 17431748 (1967). G., Immunologic MICHEEL,B., and PASTERNAK, studies on Graffi-virus-infected spontaneous Gross leukemia of AKR mice. fnt. J . Cancer, 3, 603-613 (1968). S., FISHER, C. L., STANLEY, T. B., and OROSZLAN, GILDEN,R. V., Proteins of the murine C-type RNA tumour viruses: Isolation of a group-specific antigen by isoelectric focusing. J. gen. Virol., 8, 1-10, 1970. PLATA,F., and LEVY,J. P., Blocking of syngeneic effector T-cells by soluble tumour antigens. Nature (Lond.), 249, 271-274 (1974). R. J., and TING,R. C., TransRHIM,J. S., HUEBNER, formation of hamster embryo cells in vitro by
492
Rauscher leukemia virus. J. not. Cancer Inst., 42, 1053-1060 (1969). STEEVES,R. A., Cellular antigen of Friend virusinduced leukemias. Cancer Res., 28,338-342 (1968). J. R., and AARONSON, S. A., Antigenic STEPHENSON, properties of murine sarcoma virus-transformed BALB/3T3 nonproducer cells. J. exp. Meit., 135, 503-515 (1972). STROBER, S., APPELLA, E., and LAW,L. W., Serologic and immunogenic activity of mouse transplantation antigens controlled by the H-2 locus. Proc. nat. Acad. Sci. (Wash.), 67, 765-772 (1970). WINN,H. J., Immune mechanisms in homotransplantation. 11. Quantitative assay of the immunologic activity of lymphoid cells stimulated by tumor homografts. J. Immunol., 86, 228-239 (1961). WITTE, 0. N., and WEISSMAN,I. L., Membrane proteins of Moloney sarcoma-leukemia virus : analysis by restrictive proteolytic cleavage and enzymatic surface iodination. Virology, in press (1975).
DISTINCTION BETWEEN TUMOR-SPECIFIC TRANSPLANTATION ANTIGEN AND VIRION ANTIGENS IN SOLUBILIZED PRODUCTS FROM MEMBRANES OF VIRUS-INDUCED LEUKEMIC CELLS
Kenneth S . S . CnANci, Lloyd W. LAWand Ettore APPELLA Laboratory of Cell Biology, National Cancer Institute, Bethesda, Maryland 20014, USA
A membrane antigen from R B L J leukemic cells that was solubilized and partially purified isfurther characterized in this study. This soluble antigen is capable of immunizing syngeneic hosts to reject neoplastic cells and thus resembles TSTA. It also induces cytotoxic antibody in syngeneic hosts capable of specifically lysing RBL-S cells in vitro. RBL-5, however, releases infectious virus ( R L V );it was necessary therefore to rule out virus or structural virion antigens as the effective immunogen. Infectious virus was not detectable in our initial crude membrane ( C M ) material, nor in the papainsolubilized CS or the 6-150 Sephadex-chromatographed fraction. Virus-neutralizing antibody was not detected, under stringent assay conditions, in the syngeneic anti-CM sera. Antigen preparations C M , CS and the chromatographed fractions FI, F2 and F3 were assayed in a complement-fxation test against broad-reacting antisera capable of detecting virus envelope antigen and gs antigen and against syngeneic antisera. Although our antigen preparations were positive for virion antigens, CS and F2 contained an antigen that reacted only with syngeneic antiserum. These same fractions were those reactive as immunogens. On the basis of these data, it is postulated that a cellular membrane component, other than viral, functions as TSTA.
We have reported (Law and Appella, 1973) that a tumor antigen isolated and solubilized from cell membranes of Rauscher leukemogenic virus (RLV)-induced RBL-5 leukemic cells, then partially purified, was capable of immunizing syngeneic or semi-syngeneic mice so that specific rejection of leukemic cells occurred. Immunogenicity, as revealed by tumor rejection and also antigenicity as detected by in vitro assays, were found to be preserved through the solubilization and purification procedures. The model employed and the standard reference were those used in this laboratory for solubilization and purification of histocompatibility antigens (H-2) of micelimited papain digestion followed by Sephadex
G-150 column fractionation (Strober et a[., 1970; Law et al., 1971, 1972, 1974; Appella et al., 1974). This method was also used successfully in solubilizing and purifying the TSTA of membranes of sarcoma cells induced by a DNA oncogenic virus, SV40 (Drapkin et al., 1974); specific immunogenicity again was preserved by the purification procedures. An obvious possible source of TSTA in neoplasms induced by RNA oncogenic viruses of the leukemia-sarcoma complexes is infective virus or viral structural proteins since productive, non-lytic infection continues indefinitely in most of these induced neoplasms. Leukemogenic RNA viruses mature by budding from the cell surface
Received: October 8, 1974, and in revised form December 27, 1974.
48 3
CHANG ET AL.
and this surface localization may represent effective antigen. In contrast, TSTA of DNA oncogenic virus-induced neoplasms (e.g. those neoplasms induced by SV40 and polyoma viruses) are easily distinguished from virion antigens. Preliminary data from this laboratory strongly suggested that the tumor antigens isolated from RBL-5 and responsible for specific tumor rejection were of non-virion origin (Law and Appella, 1973). We now present further data that more fully characterize our solubilized antigen as a new surface component which is not a structural part of RLV. A preliminary report of some of these additional studies has been given (Chang et al., 1974a). MATERIAL AND METHODS
Animals
Female C57B1/6N (BL/6) mice were obtained from the Veterinarian Resources Branch, National Institutes of Health, Bethesda, Md., USA and were used at 2-4 months of age. Tumors
RBL-5 originally induced by RLV, and FBL-3 originally induced by Friend leukemogenic virus were carried in ascitic form in BL/6 mice. These tumors of the FMR group share a common TSTA and common antigenic determinants as detected serologically. Extraction and solubilization of antigens from tumor cells
Antigen was prepared from RBL-5 ascites cells as described previously (Law and Appella, 1973) usually starting from approximately 2 x 1 O l o cells. Instead of sonication, the nitrogen decompression method of disrupting cells was employed in some cases with equally efficient results. After limited papain digestion a crude soluble preparation (CS) was obtained followed by Sephadex G-150 column fractionation into F1 (void), F2 and F3-included fractions. Several CS preparations were obtained also from FBL-3 ascites cells. I n all the pooled F2 fractions assayed and in the CS preparation of FBL-3, the H-2 specificities 2, 5 and 28 (found on cell membranes of H-2b haplotype) were always detected by inhibition of complement-dependent cytotoxicity using monospecific alloantisera.
484
Antisera Syngeneic anti-RBL-5 crude membrane (antiCM) serum was obtained from B1/6 mice repeatedly immunized with CM prepared from RBL-5. Usually 400,ug of the C M preparation were administered SC once a week with an equal amount of Freund's adjuvant for the first immunization and without the adjuvant for the subsequent immunizations. A broad-reacting antiserum (reacting against both gs and envelope antigens of murine leukemia viruses (MuLV) Fischer rat anti-Moloney murine sarcoma virus (MSV) serum was obtained from Dr. R. E. Wilsnack (Huntingdon Research Center, Baltimore, Md., USA). An anti-RLV serum prepared from Syrian golden hamsters was obtained by immunization with the tumor cells carrying large amounts of infectious RLV (Rhim et al., 1969). This serum was also found to be broad-reacting. Guinea-pig anti-gs serum was prepared by three weekly immunizations with purified g s antigen (Oroszlan et al., 1970) mixed with Freund's adjuvant. These antisera were absorbed with BL/6 mouse liver cells to remove hetero-antibodies and were heated at 56" C for 30 min to inactivate complement. Titrations for infectious MuL V
For detection of RLV, the material was inoculated at various dilutions on NIH Swiss mouse embryo (NIH-ME) cells in culture plates; cells were pretreated with 25 ,ug/ml diethylaminoethyl (DEAE)-dextran. These plates were either ultraviolet (UV)-irradiated or subcultured, and tested for syncytial cell formation after cocultivation with XC cells. The details of these methods were described previously (Chang et al., 1974b). In vitro neutralization tests The method described previously (Chang et al., 1974b) was followed. Both RLV and MSV(RLV) at dilutions giving about 100 XC-plaque-forming units (XC-PFU) or transformed cell focusforming units (FFU) respectively, were added to the appropriate dilutions of serum to be tested. Complement fixation (CF) tests
These tests were performed in the Microtiter System using 1.8-2.0 units of guinea-pig complement and overnight fixation at 5" C (Chang et al., 19746; Hartley et a f . , 1965). Both antisera
NON-VIRION TUMOR ANTIGEN IN VIRAL LEUKEMIA
and antigens were diluted in veronal buffer, PH 7.4 and a checkerboard titration was performed. A reaction giving a reading of 3+ or higher complement fixation was taken as the endpoint. Anticomplementary activity was encountered only with antigen extracts and antisera at high concentrations. No anticomplementary effect was observed with the guinea-pig anti-gs serum at the dilutions used. The reproducibility of the test system was demonstrated by running duplicate tests on separate days and the maximum deviation was two-fold.
FBL-3 leukemic cells in B1/6 mice given a single immunization with 30pg or lOOpg of the F 2 fraction of RBL-5 antigen preparation. The 50% tumor-inducing dose (TDsO)of FBL-3 was p>0.01. A significant difference(0.05>p>0.01) in frequency of " takes " in the 100 pg group was also observed at an FBL-3 cell challenge of 1 x lo* (2+/7 compared with 13 / I4 controls). Mean days to death in parentheses.
+
TABLE I1 RESULTS
Immunogenicity of solubilized tumor membrane preparations
Since considerabls data concerning this aspect of the investigation have been published (Law and Appella, 1973), only pertinent and more recent findings will be described here. Inhibition of the growth of FBL-3 cells was observed using adoptively transferred sensitized lymphoid cells (Winn assay). The results set forth in Table I show a significant reduction in the number of takes of a 5 x lo2 cell challenge with
IMMUNOGENICITY OF FBL-3 CRUDE SOLUBLE (CS) MATERIAL (ADOPTIVE TRANSFER) FBL-3 challenge Immunization
None FBL-3 cells (1 x loe) FBL-3 (CS) (15Opg)
10' cells loa cells (No. tumor/No. mice)
+
7/7 (26.4) 6/6 (22.3) 7/7 (23.1)
7/7 (33.6)
0/7 (>90) 1/7 (52)
1 Immunizations with viable FBL-3 cells and crude soluble (CS) material were performed subcutaneously. Lymphoid cells (spleen) from immunized mice or from normal controls were mixed at a ratio of 1OO:l target-cell and challenge made intraperitoneally in C57B1/6 mice. Mean days to death in parentheses. Tumor dose 50 of FBL-3 6 lo* cells.
48 5
CHANG ET AL.
toneal challenge with 1 x103 FBL-3 cells. The immunogenicity of 150 pg was essentially that obtained with 1 x lofi viable FBL-3 cells inoculated subcutaneously. FBL-3 is not oncogenic at this cell concentration when inoculated into the subcutaneous tissues. Detection of infectious type-C viruses in leukemia RBL-5 and its membrane preparations
To rule out the presence of infectious type-C viruses in our crude membrane (CM) preparations of RBL-5, these were inoculated on NIH-ME cell cultures known to support the replication of RLV. At the same time, infectious center assays of the X-irradiated RBL-5 cells and virus titrations of freeze-thawed extracts (10%) of RBL-5 cell suspensions were carried out. The presence of RLV on NIH-ME cells was indicated by syncytial cell formation after XC cells were added to the UV-irradiated NIH-ME cells. As shown in Table 111, 0.02-0.03% of cells of RBL-5 registered as infectious centers on the indicator cells while lo" infectious virus particles could be detected in 0.4 ml of the freeze-thawed RBL-5 extract. No infectious RLV could be detected in the CM, CS, F1, F2 and F3 preparations respectively even at the concentration of 500 pg/ml. NIH-ME cells were subcultured and
TABLE 111 DETECTION OF INFECTIOUS TYPE-C VIRUSES BY XC TESTS
Test samples 0.4 ml/inoculum
X-irradiated RBL-5 cells loficells los cells 1 O4 cells
RBL-5cell extracts Freeze-thawed, supernate, I0 % Crude membrane, 500pg/ml Crude soluble, 500pg/rnl F1, 500 pg/rnl F2, 500pg/rnl F3, 500 pg/ml
XC-PFU on NIH-ME cells treated with DEAE-dextran
I88 25 3 106 0' 0 0 0
0
' The absence of infectious type-C viruses in the membrane preparations of RBL-5 was confirmed by the inability of these preparations to induce production in ME cells of progeny viruses detectable by measurement of virion-associated activity of reverse transcriptnse using oligo(dT) poly(rA) as a primer-template.
486
XC cells were added for co-cultivation. With this more sensitive technique which allows enrichment and secondary spread of viruses in the culture, no infectious RLV could be detected in CM preparations. The possibility of xenotropic viruses contaminating the RLV-producing RBL-5 cells was checked by culturing RBL-5 cells in the presence of SIRC cells (American Type Culture Collection) which are known to support the growth of some xenotropic viruses but not of RLV (Benvenistc et a[., 1974) and testing the culture supernate for the presence of virus particles possessing reverse transcriptase activity. No such viruscs could be detected. Absence of RL V-neutralizing activity in the syngeneic anti-CM serum Table IV details the results of in vitro neutralization tests with syngeneic anti-CM serum against RLV and MSV(RLV). XC-PFU and FFU assays revealed no inhibition by the serum at dilutions of 1 :4 and 1 : l O . A positive control, hamster anti-RLV serum and a negative control, serum from normal B1/6 mice, were included. These results suggest that the CM preparation used for immunization of syngeneic mice did not contain enough virion antigens to induce the production of neutralizing antibody. Also tested was the possibility of the presence in the syngeneic anti-CM serum of antibodies directed against viruses commonly found in mice either as a result of immunization with the CM containing such viral antigens or as a result of naturally occurring infections. By hemagglutination-inhibition tests and complement fixation tests (done by Dr. J. C. Parker, Diagnostic Services, Microbiological Associates, Inc., Bethesda, Md., USA) no such antibody was found against PVM, Reo 3, GDVII, K, polyoma, Sendai, ectromelia, mouse adeno, mouse hepatitis and lymphocytic choriomeningitis viruses at the dilutions used. Complement fixation tests
Several lots of our antigen preparations including CM, CS, F1, F2 and F3 preparations and of RBL-5, and the culture supernatants of RLV and MLV containing loRand 1 0 XC-PFU/ ml respectively were tested. Antisera included broad-reacting antiviral sera, i.e. Fischer rat anti-MSV(MLV) serum and
NON-VIRION TUMOR ANTIGEN IN VIRAL LEUKEMIA TABLE IV I N VITRO NEUTRALIZATION TESTS
Test serum
Dilution
RLV XC-PFUiplate
MSV(RLV) FFU/plate
Syngeneic anti-RBL-5 crude membrane
4 10 4 10 40 80 160 320 -
122 124 131 120 0 0 1 129 125
93 90 85 87 0 0 0 24 84
Normal BL/6 mice (negative control) ,,
9,
Hamster anti-RLV (positive control)
Diluent
hamster anti-RLV serum, both reacting against murine viral gs and envelope antigens. CF reactions obtained with 2 units of antisera (one dilution lower than the endpoint dilution) are presented. In addition, various lots of syngeneic anti-CM serum were also tested. Table V gives typical results obtained with some syngeneic antisera which showed relatively higher activities against both known preparations of RLV and MLV as well as various preparations of solubilized RBL-5 antigens. The antiviral sera i.e. Fischer rat and hamster sera gave positive reactions with various antigen preparations as well as with the high-titered RLV and MLV. The minimal concentration in terms of protein concentration (pglml) of antigens detectable by Fischer rat anti-MSV(MLV) serum
was 125,250,225 and 87.5 for CM, CS, F2 and F 3 respectively. (F1 fraction was excluded from the Sephadex G-150 column and not tested in this experiment). Hamster anti-RLV serum gave similar values. These results indicate the presence of viral envelope (and gs) antigens in the various antigen preparations although no infectious virus particles were found. Two lots (lots A & B) of syngeneic anti-CM serum reacted with RLV and MLV, indicating the presence of antibody specific t o viral antigens. However, the CF reactions obtained with our various RBL-5 antigen preparations, CM, CS, F2 and F3, as expressed in terms of the minimal concentration of antigen detectable by syngeneic antiserum (lot A), were 125, 62, G1.8 and 22 pg/ml respectively. Lot B antiserum gave
TABLE V
COMPLEMENT FIXATION TEST-I
Antigen preparations (lot I )
Fiscber rat anti-MSV(MLV) broad-reacting 1 :40
Hamster anti-RLV broad-reacting 1 :40
Syngeneic anti-RBL-5 crude membrane serum Lot A 1 :40
Lot B 1 :40
RBL-5 CM
+(125) +(250)
+(62.5) +(5W
F1 F2 F3
NT
NT
+(87.5) +(1/32) +w4)
+(225) +(1/16) +(1/4)
cs
RLV MLV 1
+(225)
+(125) (62) NT +K1.8) (22) +(1/16) +(1/4)
+ +
+(I251 +(62)
NT
+(< 1.8) +(1/16) +(1/4)
Minimal concentration of antigen detectable (protein concentration pglml). = not tested. Minimal CF reactive fraction of an original Concentration of a virus suspension.
* NT J
487
CHANG ET AL.
identical results except that F3 was negative. It is of interest to note that, while these syngeneic antisera reacted against RLV and MLV to the same extent as did the rat and hamster antiviral sera, their reactions against CM, CS, F2 or F 3 were generally more sensitive than those detected by antiviral sera, as shown by the lower amount of antigen detectable by the former than by the latter antisera. The minimal concentration of antigen detectable by the syngeneic antisera was considerably lower (G1.8 pg/ml) for F2 as compared with F3 (22pg/ml), CS (62pglml) or CM (125 pglml). These differential C F activities of different antisera are better illustrated by calculating the ratio of the reciprocals of the minimal concentrations of the respective antigens detectable by CF test for each antiserum (Table VII). These findings suggest the presence of non-virion antigen in the solubilized antigen preparations of RBL-5, which is especially concentrated in the F2 fraction. In another series of experiments, some lots of syngeneic anti-CM serum were found to react against some of the solubilized antigen preparations of RBL-5 but not against high-titered RLV or MLV. An example of CF reactions against these antigens obtained with one such antiserum, lot C, together with those reactions obtained with guinea-pig anti-gs serum as well as rat antiMSV(MLV) and hamster anti-RLV sera, are presented in Table VT. The finding that the syngeneic anti-CM serum reacted against CS and F2 only, without detect-
able reaction against the high-titered RLV or MLV, indicates the presence of non-virion antigen in these preparations. It is evident also that CS, F 2 and F3 preparations contained virion antigens detectable by rat anti-MSV(MLV) and hamster anti-RLV sera. Reactions with guineapig anti-gs serum indicated the presence of gs antigen in F2, F3 as well as in RLV preparations. Thus although all antigen preparations except F1 contained viral envelope and gs antigens, CS and F2 contained an antigen (or antigens) which is not virion antigen and is detectable only by syngeneic anti-CM serum. In other words, syngeneic anti-CM sera contain an antibody which reacted exclusively with non-virion antigen that is present in CS and partially purified in our F2 fraction. In order to compare the differential CF reactivities of antigens in F2, F 3 and RLV against an antiviral antiserum and a syngeneic anti-CM serum, ratios of the reciprocals of minimal concentrations of the respective antigens were computed (Table VII). These ratios indicate the relative concentrations or degrees of purification of antigens in F2, F3 and RLV with respect to both viral antigen content and non-virion antigen content. Where one of the antigen preparations showed no reactivity, no activity ratio was computable and a sign of + I - or -/ t was used. While the viral antigens in F2 remained at low levels, as compared with F3 or RLV, the relative reactivities of syngeneic antisera against F2 were considerably higher than those of F3 or
TABLE VI
COMPLEMENT FIXATION TEST-11 Antigen preparations (lot 2)
I
9
488
Fischer rat anti-MSV(MLV) broad-reacting
Hamster anti-RLV broad-reacting
Guinea-pig anti-gs
1 :40
I :40
1:10
Minimal concentration of antigen detectable (protein concentration pg/ml). Minimal CF reactive fraction of an original concentration of a virus suspension.
Syngeneic crude membrane Lot
c
I :40
NON-VIRION TUMOR ANTIGEN IN VIRAL LEUKEMIA TABLE VII DIFFERENTIAL CF ACTIVITIES OF F2, F3 AND RLV
Activity ratio
F2/F3 F2/RLV
"N",4. I I1 I
I1 F3/RLV
I I1
Sygenei anti-RBL-5 Rat anti-MSV(MLV) crude membrane serum serum Lot A Lot B Lot C
0.4 1 .o 0.14 0.5 0.36 0.5
2312
+/+/-
2 9
>9
+/0.7
-/+
-/.-
The ratio of the reciprocals of the minimal concentrations of the respective antigens (F2 and F3) detectable by CF test except for RLV in which case minimal fraction detectable was used.
RLV. By contrast, the reactivities of F3 against syngeneic antisera were not so pronounced as those of RLV. Thus the data are in accord with the concept that trace or moderate amounts of virion antigens (including envelope and gs antigens) may be present in all preparations of solubilized RBL-5 membrane antigens, but that the non-virion antigen, presumably tumor-cell membraneassociated antigen, may be partially purified and concentrated in the F2 fraction. This is suggested by the differential ratios of C F reactivity of some lots of syngeneic anti-CM sera and by reactivity with some other lots of syngeneic anti-CM sera that reacted exclusively with non-virion antigen in CS and F2 preparations. The immunological specificities of the C F activity of these RBL-5 antigens were indicated by the failure of Gross virus-induced syngeneic leukemic cells, EGG2, to absorb from the syngeneic anti-CM serum the CF antibody reacting with the F2 antigen. DISCUSSION
The protein entity from the cell membrane of RBL-5 (and FBL-3) leukemic cells that we have solubilized and partially purified has a tumor rejection capacity that is specific and therefore represents TSTA. The in vitro inhibition by soluble antigen of complement-dependent cytotoxicity as measured by 51Crrelease corresponds with in vivo activity and is also specific (Law and Appella, 1973). Since RBL-5 cells release infectious virus, it is necessary to rule out a role of infectious virus itself or of virion antigens
e.g. envelope antigen or gs antigen, in tumor rejection. Indirect evidence (Klein and Klein, 1964; Micheel and Pasternak, 1968; Steeves, 1968)and direct visualization by immuno-electron microscopy (Aoki et al., 1972) support the existence of virus-coded but non-virion cellsurface antigens of certain MuLV-induced neoplasms. However, detailed studies have not been reported in tumor antigen solubilization experiments (Martyre et al., 1973; Plata and Levy, 1974) to distinguish this type of antigen from virion antigens contaminating the preparation, nor is it known that these identified antigens act as TSTAs. There is sufficient evidence to show that MuLVs alone in all probability serve in some manner as TSTAs. Some neoplastic cells that are not detectably immunogenic nor immunosensitive t o the rejection response of MuLV-immunized mice become rapidly immunogenic and immunosensitive upon superinfection, for example with MLV (Stephenson and Aaronson, 1972). However, the evidence we present here strongly indicates the existence additionally, in an MuLV system, of a new cellular antigen specific for the virus-in this case RLV-of the TSTA type that appears to be analogous to the TSTA found in neoplasms induced by DNA-oncogenic viruses. No infectious RLV was detectable in our C M preparation since the procedure for preparation of CM from RBL-5 cells included sonication, which disrupted the virus as well as the cell membrane, and centrifugation at 16,000 rpm for 90 min, which probably eliminated most of the complete virions and structural antigens of disrupted virions. However, a certain amount of viral structural antigens was still detectable by the CF test in our CM, CS, F2 and F3 preparations. Apparently, our CM preparations did not contain either infectious RLV or enough viral envelope antigens (which are specifically able t o induce neutralizing antibody) but did contain viral structural antigens responsible for induction of RLV- or MLV-reacting CF antibody in syngeneic mice. The absence of infectious RLV in CM rules out the possibility of replication virus in vivo by introduction of an RLVcontaminated CM preparation which would have resulted in amplification of RLV production. Some pools of syngeneic anti-CM sera did not have detectable CF antibody reacting against the high-titered viruses, RLV and MLV. It is evident
489
CHANG ET AL.
that the CM used for immunization did not contain enough viral structural antigens to be effective immunogens in these cases. More exacting conditions for the preparation of CM with minimal contamination with virion antigens are presently under study. The differential reactivities observed with our F2 preparation (following Sephadex G-150 chromatography) in the C F assay support the concept that the solubilized RBL-5 membrane antigen possessing the activity of TSTA of the transplantation type is not virion structural antigen. This preparation had higher activity against syngeneic anti-CM serum than against rat or hamster anti-viral sera. The F2 preparation also reacted in the C F assay with lots of syngeneic anti-CM sera that did not react with high-titered RLV o r MLV. Since anti-CM serum was prepared in syngeneic mice it should have minimal or no anti-gs antibody; mice are usually “ tolerant ” to this antigen (Huebner et al., 1971) although apparent exceptions have been noted recently (Markham et al., 1972). It is of interest to note that the Fischer rat anti-MSV serum and the hamster anti-RLV serum at concentrations which reacted against RLV and MLV in the CF test did not manifest complement-dependent cytotoxicity against RBL-5. The syngeneic anti-CM serum which was non-reactive against both RLV and MLV did exhibit, however, cytotoxicity against RBL-5 and this activity was not blocked by pretreatment of the serum with purified, ether-disrupted RLV. Since this anti-CM serum reacted with the F2 fraction in the C F test and since the F2 fraction inhibited the cytotoxicity of anti-CM serum against RBL-5, the question of the identity of these antibodies deserves further investigation.
An approach to differentiate decisively TSTA from virion antigens is to utilize “ non-producer ” tumor cells. Recently we have studied leukemia RBL-3 which was originally induced in B1/6 mice by RLV in a similar manner as RBL-5 (McCoy et al., 1967) and maintained by in vitro culture. This line proved to be immunogenic in vivo in B1/6 mice against leukemia FBL-3 which possesses F M R tumor antigen shared by RBL-5, and vice versa. Moreover, this lymphocytic neoplasm produced no infectious RLV but was nevertheless MuLVgs-antigen positive. Therefore this can be defined as a “ non-producer ” neoplasm in the classical sense. The syngeneic anti-CM sera were cytotoxic against RBL-3 cells as well as against RBL-5 cells. This model system, therefore, may be used to study the relationship of TSTA and virion antigens. The fact that TSTA and H-2 co-chromatograph on the G-150 Sephadex column (Law and Appella, 1973; Drapkin et al., 1974) points to some similarity in molecular size and structure between the two antigens. Since it is very unlikely that we are measuring an H-2 activity in our syngeneic system, one possibility is that the TSTA is a non-H2 activity not expressed on fully differentiated normal tissue but re-expressed on neoplastic cells. The possibility that we are measuring an activity ofexternal virion membrane components is also very unlikely because of our use of a proteolytic cleavage procedure. The major iodinated components of virion membranes obtained from proteolytic cleavage have a molecular weight in the 73,000-75,000 range (Witte and Weissman, 1975). This MW is much higher than the molecular weight of TSTA or of H-2 (45,000-50,000). Upon fractionation on the G-150 Sephadex column, these two entities should separate.
DISTINCTION ENTRE L’ANTIGENE DE TRANSPLANTATION SPl?CIFIQUE DE LA TUMEUR ET LES ANTIGENES VIRIONIQUES DANS LES PRODUITS SOLUBILISES PROVENANT D E MEMBRANES DE CELLULES LEUCEMIQUES VIRO-INDUITES Dans la prisente dude, les auteurs caractirisent de faGon plus approjondie un antighe de la membrane des cellules leucimiques RBL-5 qui avait de‘ja Pte‘ solubilise‘ et partiellement purifie’. Cet antigPne soluble est capable d’immuniser des hdtes syngkniques et de les ainener ir rejeter les cellules nkoplasiques; il ressemble donc au TSTA. I1 induit aussi chez les hdtes synginiques I’anticorps rytotoxique capable de lyser spPci$quement les cellules RBL-5 in vitro. Toutejois, les RBL-5 lib2rent du virus
490
NON-VIRION TUMOR ANTIGEN IN VIRAL LEUKEMIA
infectieux ( R L V ); il ttait done ntcessaire d’tcarter la possibilitt que le virus ou les antigbnes virioniques structuraux soient responsables de cet effet immunogtne. Nous n’avons pas diceli de virus infectieux dans le mattriel initial, c’est-&dire la membrane brute ( C M ) , ni dans les prtparations solubles brutes ( C S ) solubilistes a la papaihe, ni dans les fractions chrornatographiies sur Sephadex G-150. Nous n’avons pas dttectt d’anticorps neutralisant le virus dans les se‘rums synginiques anti-CM, bien que les tests aient t t t effectuts trts soigneusement. Les priparations antigtniques CM et CS et les fractions chrornatographites FI, F2 et F3 ont t t i testtes par une ipreuve de fixation du compltment contre des anti-strums qui permettent de dttecter l’antigkne de l’enveloppe virale et I’antigtne gs, et contre des antistrums syngtniques. Bien que nos prtparations antigtniques soient positives pour les antigknes virioniques, les CS et les F2 contenaient un antigbne yui ne rtagissait qu’avec l’antistrum syngtnique. I1 s’agissait des fractions qui avaient un effet immunogdne. Sur la base de ces observations, nous Cmettons l’hypothkse qu’un composant de la membrane cellulaire, qui n’est pas de nature virale, joue le r d e de TSTA.
REFERENCES
APPELLA,E., HENRIKSEN, P., NATORI,T., TANIGAKI, N., LAW, L. W., and PRESSMAN, D., Basic structure of mouse histocompatibility antigens. Transplant.Proc., In press (1975). AOKI,T., HERBERMAN, R. B., JOHNSON,P. A., LIU,M., and STURM, M. M.. Wild-tvDe Gross leukemia virus: Classification of soluble antigens (GSA). J. Virol., 10, 1208-1219 (1972). BENVENISTE, R. E., LIEBER, M. M., and TODARO, G. J., A distinct class of inducible murine type-C viruses that replicates in the rabbit SIRC line. Proc. nat. Acad. Sci. (Wash.), 71, 602-606 (1974). CHANG,K. S. S., APPELLA, E., and LAW,L. W., Some characteristics of solubilized tumor antigen isolated from an RNA virus-transformed neoplasm. Fed. Proc., 33,791 (1974a). CHANG,K. S. S., LAW,L. W., and AOKI,T., Type-C RNA virus isolated from SJL/J mice. J. nat. Cancer Inst., 52, 771-184 (19746). DRAPKIN,M. S., APPELLA,E., and LAW, L. W., Immunogenic properties of soluble tumor-specific transplantation antigen induced by Simian virus 40. J. nat. Cancer Inst., 52, 259-263 (1974). HABEL,K., Immunological determinants of polyoma virus oncogenesis. J . exp. Med., 155, 181-193 (1962). HARTLEY, J. W., ROWE,W. P., CAPPS,W. I., and HUEBNER, R . J., Complement fixation and tissue culture assays for mouse leukemia viruses. Proc. nat. Acad. Sci. (Wash.), 53, 931-938 (1965). HOWELL,S. B., ESBER,E. C., and LAW, L. W., Cellular immunity in mice with Simian virus 40induced mKSA tumors. J. nat. Cancer Inst., 52, 1361-1363 (1974). HUEBNER, R. J., SARMA,P. S., KELLOFF,G. J., GILDEN,R. J., MEIER,H., MYERS,D. D., and PETERS,R. L., Immunological tolerance to RNA tumor-virus genome expressions : Significance of
tolerance and prenatal expressions in embryogenesis and tumorigenesis. Ann. N.Y. Acad. Sci., 181, 246-271 (1971). KLEIN,G., and KLEIN,E., Antigenic behavior of Moloney lymphomas : independence of virus release and immunosensitivity. Science, 145, 1316-1317 (1964). KLEIN,E., and SJOGREN, H. O., Humoral and cellular factors in homograft and isograft immunity against sarcoma cells. Cancer Res., 20, 452-461 (1960). LAW,L. W., and APPELLA, E., Immunologic properties of solubilized tumour antigen from an RNA virustransformed neoplasm. Nature (Lond.), 243, 83-87 (1973). LAW,L. w.,APPELLA, E., STROBER, s.,WRIGHT, P. w., and FISCHETTI, T., Induction of immunological tolerance to soluble histocompatibility-2 antigens of mice. Proc. nat. Acad. Sci. (Wash.), 69, 18581862 (1972). LAW,L. W., APPELLA, E., STROBER, S., WRIGHT, P. W., and FISCHETTI, T., Soluble transplantation antigens: Further studies of their tolerogenic properties. Transplant., 18, 487-495 (1974). LAW, L. W., APPELLA,E., WRIGHT,P. W., and STROBER, S., Immunologic enhancement of allogeneic tumor growth with soluble histocompatibility-2 antigens. Proc. nat. Acad. Sci (Wash.), 68, 3078-3082 (1971). MARKHAM, R. V., SUTHERLAND, J. C., CIMINO,E., DRAKE, W. P., and MARDINEY, M. R., Immune complexes localized in the renal glomeruli of AKR mice: the presence of MuLV gs-1 and C-type R N A tumor virus gs-3 determinants. Rev. europ. Et. elin. biol., 17, 690-694 (1972).
MARTYR& M.-C., HALLE-PANNENKO, O., and JOLL~S,P., Characterization and partial purification of normal and tumor associated transplantation
49 1
CHANG ET AL.
antigens of Rauscher leukemia cells. Ei4rop. J . Cancer, 9, 757-761 (1973). McCoy, J. L., FEFER, A., and GLYNN,J. P., Comparative studies on the induction of transplantation resistance in BALB/c and C57BL/6 mice in three murine leukemia systems. Cancer Res., 27, 17431748 (1967). G., Immunologic MICHEEL,B., and PASTERNAK, studies on Graffi-virus-infected spontaneous Gross leukemia of AKR mice. fnt. J . Cancer, 3, 603-613 (1968). S., FISHER, C. L., STANLEY, T. B., and OROSZLAN, GILDEN,R. V., Proteins of the murine C-type RNA tumour viruses: Isolation of a group-specific antigen by isoelectric focusing. J. gen. Virol., 8, 1-10, 1970. PLATA,F., and LEVY,J. P., Blocking of syngeneic effector T-cells by soluble tumour antigens. Nature (Lond.), 249, 271-274 (1974). R. J., and TING,R. C., TransRHIM,J. S., HUEBNER, formation of hamster embryo cells in vitro by
492
Rauscher leukemia virus. J. not. Cancer Inst., 42, 1053-1060 (1969). STEEVES,R. A., Cellular antigen of Friend virusinduced leukemias. Cancer Res., 28,338-342 (1968). J. R., and AARONSON, S. A., Antigenic STEPHENSON, properties of murine sarcoma virus-transformed BALB/3T3 nonproducer cells. J. exp. Meit., 135, 503-515 (1972). STROBER, S., APPELLA, E., and LAW,L. W., Serologic and immunogenic activity of mouse transplantation antigens controlled by the H-2 locus. Proc. nat. Acad. Sci. (Wash.), 67, 765-772 (1970). WINN,H. J., Immune mechanisms in homotransplantation. 11. Quantitative assay of the immunologic activity of lymphoid cells stimulated by tumor homografts. J. Immunol., 86, 228-239 (1961). WITTE, 0. N., and WEISSMAN,I. L., Membrane proteins of Moloney sarcoma-leukemia virus : analysis by restrictive proteolytic cleavage and enzymatic surface iodination. Virology, in press (1975).
