Studies of the Significance of Tumor Antigens in Induction and Repression of Neoplastic Diseases: Presidential Address¹ Free

New cellular antigens that are virus specific have been detected in many tumors and neoplasms induced by the DNA and RNA oncogenic viruses. For convenience, these antigens may be grouped into those cell surface antigens detected by transplantation rejection technics, the so-called TR antigens, and those detected by the classical immunologic technics such as complement fixation, immunoflourescence, cytotoxicity in the presence of complement, etc. Biologic characterization of these latter antigens is based principally upon testing the antibodies produced in tumor-bearing hosts against the antigens of tumor cells. Another distinction may be made. By reference to Table 1, it may be seen that the virus-tumor systems of the DNA oncogenic viruses, that is the papovaviruses (polyoma, adenoviruses and SV40) and of the RNA viruses, principally the leukemogenic viruses and the murine lymphoma-sarcoma virus complex, appear to show distinctly different patterns in regard to the nature of their antigens and the technics used for detecting them. These differences probably relate to the unique features of replication of the DNA as opposed to the RNA viruses. Tumor cells induced by the DNA viruses usually do not release infectious virus, whereas tumors induced by RNA oncogenic viruses continually produce infectious virus except as described for certain rat and mouse neoplasms induced by MSV. The noninfectious nature of DNA virusinduced tumors may be more apparent than real, however, and will be discussed later.

Although the existence of specific transplantation resistance antigens in transplantable tumors has been known for some time, their exact role in primary tumor induction or repression was not known primarily because those TR antigens first described and characterized in carcinogen-induced neoplasms are individually distinct for each neoplasm.

As shown in the present study, supression of polyoma tumors in thymectomized C57BL mice or in susceptible C3Hf/Bi mice, by the passive transfer of sensitized, lymphoid cells as late as 30 to 35 days after virus infection, signifies that the TR antigen and the animal's immunologic reaction to it are strong determinants not only in rejection of transplantable tumors but of tumor repression in the primary host. In such cases they should be referred to as tumor rejection antigens.

That TR antigens, particularly on polyoma-induced neoplastic cells, are “weak” antigens of the histocompatibility type is suggested by the observation that most strains of mice thymectomized at 3 days of age show only very subtle signs of immunologic deficiency. Thymectomized C57BL mice, for example, are quite capable of rejecting allogeneic (H-2) skin grafts but are defective in their response to the Y-chromosome histocompatibility antigen; yet they are incapable of rejecting their own polyoma virus-induced neoplasms. Thus the animal does not have to be in a condition of gross immunologic inadequacy to provide the conditions for initiation and progressive growth of neoplasms.

The question has been considered as to whether the new cell surface antigens of the transplantation type, TR, particularly those induced by the papova viruses, are coded for by the viral or by the cellular genome. The viral specificity of such antigens, the cross-reactivity of antigens between hamster and mouse polyoma tumors and separately among the adenovirus-induced and among the SV40-induced tumors, as well as marker rescue experiments showing that the rescued viral genome will induce specific transplantation resistance all suggest that the viral genome or a fraction of it integrated in some manner in the host cell is responsible for coding of TR antigen. The DNA of cells transformed by polyoma virus contains base sequences homologous to viral DNA, suggesting also, but not proving, that transformation and acquisition of new antigenic characters are related directly to the integrated viral DNA. There is evidence also, from experiments employing the adenovirus-SV40 hybrid, that only a fraction of the viral genome, that fraction of SV40 incorporated into the adenovirus capsid, is necessary to induce the CF antigen (18). It is argued also that transcription of only a fraction of the SV40 genome is required to produce the transplantation antigen (TR) but this evidence is of a rather indirect nature.

The above arguments favoring coding for TR antigens by integrated viral genome are based upon neoplasms induced by the papova viruses as being free of infectious virus. Though most of such neoplasms assayed by sensitive in vitro systems give evidence of their virus-free status, it has been shown recently through the use of indicator cells (24) or through employment of certain polycations, such as DEAE-dextran (21), that recovery of infectious virus is possible from some “infectious virus-free” neoplasms.

Since RNA tumor viruses continue to replicate in most tumor cells, it is difficult to determine whether TR antigens as well as those detected by serologic technics are viral antigens or newly induced cellular antigens. The neoplasms and tumors induced by the leukemogenic viruses and by MTV, RSV (Schmidt-Ruppin), and MSV fall in this class. One neoplasm, an hemangiosarcoma, XM-1, although induced by the MSV (MLV) pseudotype virus and selected through serial transplantation from the original granulomatous tumor, has been shown to be free of infectious virus even in in vitro assays on mouse embryo cells employing DEAE-dextran (H. Duc-Nguyen, unpublished observations). Cells of this neoplasm were shown to contain a transplantation antigen and this TR, in all probability, represents a virus-specific, new cellular antigen. Of particular interest in the probability that TR antigen may be specified by the MSV (MLV) pseudotype, is the fact that syngeneic mice immunized with another pseudotype virus, MSV (RV), do not resist grafting of XM-1 cells. Much remains to be learned as to why the pseudotype virus MSV (MLV) has immunologic specificity whereas MSV (RV) does not. These results are similar to those reported for in vitro infections with mutants of Herpes simplex virus. New determinant antigens on the surface of infected cells were found to be specific for each of the mutant viruses studied (77).

It is not known, as yet, if the MSV-induced hemangiosarcoma XM-1 contains defective MSV that may indeed be responsible for coding of the virus-specific TR antigen described above. However the MSV-induced neoplasm of BN strain rats studied by Ting (90), which also contains an MSV-specific transplantation antigen, is known to contain MSV in a defective state, most likely represented by the MSV(O) particle (91). Likewise, sarcomas induced in mice by the Schmidt-Ruppin strain of RSV contain TR antigens, yet do not release infectious virus unless brought into contact with chicken cells. The evidence here, as in the MSV-induced tumors, favors the assumption of a relationship between TR antigens and the presence of virus-derived genetic materials in the tumor cells.

In contrast to the suppression of neoplastic cell growth in the polyoma system by specifically sensitized lymphoid cells, inhibition of growth of the MSV-induced XM-1 neoplasm appears to be affected by circulating anti-MSV antibody.

Inhibition of growth of the transplantable tumor, XM-131, an atypical granuloma, in previously X-irradiated syngeneic recipients is apparently also mediated by a humoral host response. Anti-MSV antibodies of high-titer, passively transferred after grafting of the tumor, provided inhibition of tumor cell growth.

Similar results have been recorded recently (72) showing that antiserum specific for G+ (Gross antigen) leukemic cells and with high titers of cytotoxic activity against G+ leukemic cells in vitro gives complete protection against the outgrowth of transplanted leukemic cells (G+ but not G-) even if passively transferred as late as 3 days after leukemic cell transplantation.

Thus it is clear that in some experimental situations supression of neoplastic cell growth is obtained by passive immunization with humoral antibodies apparently specific for viralinduced TR antigen. The precise role of humoral antibodies in the repression of primary neoplasms in the autochthonous host, however, is not known, and a damaging effect on solid neoplasms would not ordinarily be expected.

From what we know to date, it would appear that immunity of the cell-mediated type, that resembling delayed hypersensitivity reactions and homograft rejection, is of paramount importance and represents a very efficient mechanism in the rejection of tissues altered by neoplastic change.

Few systematic studies appear to have been made of the effect of thymic removal early in life or of other immunosuppressants on the frequency or latency of spontaneous neoplasms. Since there is preliminary evidence of immune deficits both of cell-mediated and humoral types appearing late in life following thymectomy (59, 60, 87), more needs to be learned of a possible relationship between the immune responses of old age and the development of neoplasms.

The significance of antigens, detected by serologic methods, in the conversion of a normal cell to a neoplastic cell or in immune reactions of the host against the tumor cell itself has yet to be determined. Cytotoxic activity in vitro in the presence of complement of the various antisera prepared against leukemic cells probably is an expression of TR antigens, on the basis of what is known in the mouse of the cytotoxic action of anti-H2 isoantisera. The absence of cytotoxic sensitivity of most target cells except those of lymphoid origin most probably reflects differences in antigenic receptor sites.

There has been speculation that the T-antigen (neo-antigen) detected by complement fixation or by immunofluorescence studies (IF^N) may function in the replication of viral nucleic acids or even perhaps in the repression of normal mechanisms controlling DNA synthesis and cell multiplication. These antigens appear early in the virus multiplication cycle before new infectious virus or viral coat antigens appear. In this respect they differ from other antigens. An outstanding feature of T-antigens is their specificity determined by the original inducing virus and their presence in tumors and of normal cells in culture even in widely divergent species. Thus, these should be useful in attempts to establish etiologic relationships between a given virus and a neoplasm induced by it. Several such approaches depend on seroepidemiologic studies of cancer patients and controls using isolated T-antigens of the several suspected adenoviruses or of SV40 (58, 95). However, strict controls must be used in such studies; when virus-specific antigens and antibodies from one species are used to test materials from other species, interactions between normal constituents of tissues and sera may be expected (see, for example, 31).

The observation that most carcinogen- and virus-induced neoplasms contain TR cellular antigens suggests that such new antigens are firmly associated with the conversion of a normal to a neoplastic cell rather than the consequence of such conversion. However, many such antigens are lost. This appears to be the case for some methylcholanthrene-induced sarcomas and for the polyoma-induced neoplasms of the hamster. On the other hand, TR antigens of polyoma virus-induced neoplasms of the mouse and rat are very stable; it has been impossible to select tumor cells with a deleted TR antigen even using strong selective pressures, such as passage in mice presensitized against those antigens (82). It is not known, therefore, at present, whether these specific antigens or the others discussed here are indispensible to tumor cells.

The few instances of “antigenic conversion” of tumor cells as a consequence of infection by other oncogenic viruses (70) and of “artificial heterogenization” of tumor cells with nononcogenic viruses, resulting in resistance to transplantation (32), are of interest. More needs to be learned of the nature of such changes, whether they are temporary or permanent and their relationship to the typical TR antigens described in the polyoma, adenovirus, and SV40 systems.

There have been discussed here immune reactions against tumor cells evoked by TR antigens that are either of the cellmediated type or mediated by humoral antibody. Adoptive immunization is possible with each type. Recently, Alexander and colleagues (2, 3) have described inhibitory effects in the intact tumor-bearing rat using presumably sensitized lymphocytes from xenogeneic donors (sheep and goat). The explanation of the effect is that the lymphocyte transfers information (presumably via RNA) which brings about participation of the host's defense mechanism in providing inhibition of tumor growth. Apparently the transferred lymphoid cells are then rejected by the recipient since no mention is made of graft versus host reactions. This is an unusual finding, so far unconfirmed, and is unexpected in the face of all the evidence at hand concerning the mechanisms of adoptive immunity by lymphoid cells.

If it is found that new cellular antigens of the transplantation type exist on human neoplastic cells, it is likely that repression of such growths immunologically would be by a cell-mediated mechanism. This would entail infusion of sensitized, immunocompetent lymphoid cells of allogeneic origin. The dangers inherent in this procedure, particularly in the development of severe graft versus host reactions, are obvious. Nevertheless, several groups have attempted immunotherapy in cancer patients using transfusion of lymphoid cells (9, 56) with equivocal results but in some cases with graft versus host reactions. The data do indicate that, in a few instances with acute leukemics and Hodgkin's disease patients, successful allogeneic grafting has been attained. Most recently in a patient with a pronounced cellular defect as a result of chronic mucocutaneous moniliasis, transfusion of paternal leukocytes resulted in a correction of the cellular defect and striking objective improvements in the patient (15). Progress in histocompatibility testing in man and a better understanding of the homograft reaction may indeed allow for future rational approaches in the therapy of immunologic deficiency states and cancer.

The recent findings defining more precisely mechanisms of tolerance at the cellular level (63, 68, 80) and the probability that tolerance may now be effectively developed in the adult are also signs that lead to the hope of successful transfer of allogeneic lymphoid cells in the future. Also, several lines of evidence suggest that uncommitted lymphoid cells may be converted in vitro to a specific immunologically committed status by transfer of “information” by agents such as “transfer factor” or RNA. Thus, lymphoid cells specifically sensitized to tumor antigen (TR) in vitro is a distinct possibility.