An Approach to Tumor Virus Identification

BRIEF PROPOSAL AN APPROACH TO TUMOR VIRUS IDENTIFICATION PHYLLIS B. BLAIR, Ph.D.* Viruses are recognized as causative agents in the development of cancer in several animals, including the chicken, mouse, rat, rabbit, hamster, and monkey. In man, no tumorigenic virus has yet been identified, but we can assume that viruses are responsible for at least some ofthe neoplasias ofman also. Traditional methods ofidentifying agents which cause disease may be oflimited application in the search for tumorigenic viruses in man; not only is experimentation impossible or difficult, but also the possibility ofan extended latent period, such as frequently characterizes virus-induced tumors in other species, may make identification ofa causative agent difficult in a long lived species such as man. Several experimental techniques have already been utilized in the search for tumorigenic viruses in man. Successful attempts have been made to induce neoplasia in other species by using known human viruses. Various immunologic techniques have been used to identify specific tumor antigens. Some ofthese tumor antigens may be viral in origin [1-3], but they havenotyet been related to anyhumanvirus. Electron microscopists have attempted to identify viral particles in tumor biopsies, in tumor cells cultured in vitro, andinmaterial centrifuged from bodyfluids. However, virus-likeparticles frequently can be distinguished from cellular particles and debris only with difficulty [4, 5]. Successful identification of a tumorigenic virus in the human will require the application ofsome ofthese as well as other techniques in a comprehensive approach. We may develop such an approach by applying the general principles derived from the study oftumorigenic viruses in other species. Viruses already identified as oncogenic in other animals can be grouped into two generalclasses [6]. Themembersofonegroup sharesomecharacteristicswidbtthemyxoviruses and can be classified with them as compound helical RNA-containing viruses. They are large viruses which appear to replicate in the cytoplasm ofinfected cells and to mature by a budding process at the cell surface. This group ofviruses includes members ofthe avian leukosis complex, the Rous sarcoma virus, the mouse leukemia viruses, and the mouse mammary tumor virus. These viruses produce leukoses, leukemias, sarcomas, and mammary carcinomas as natural infections in at least two species (the chicken and the mouse), and some ofthem have induced tumors in other species (hamster, rat, and mon- * Department ofBacteriology and Immunology, and the Cancer Research Genetic Laboratory, University ofCalifornia, Berkeley. I73 key) under experimental conditions. In general, virus multiplication occurs in the tumors, and antigenicity associated with viral particles is found in the tumors. The second class oftumorigenic viruses includes small DNA-containing viruses such as polyoma, SV40, and the adenoviruses. In structure these are simple cubic viruses, and they replicate in the nucleus ofinfected cells. Tumors induced by these viruses possess distinctive tumor-specific antigens as a result ofvirus infection and/or transformation; however, this tumor-associated antigenicity is not necessarily related to antigens ofthe virus itself. The viruses usually do not propagate in the tumors they induce, and therefore identification in tumors ofa new tumor-inducing virus ofthis type may be difficult. Since viruses in both ofthese classes are involved in neoplasia in several species, one can predict that similar viruses in these classes will be identified as tumorigenic agents in man and also in domestic animals. This prediction can be tested by applying available techniques to bolate such viruses. Since the large RNA viruses are probably the easiest to identify, a search for tumorigenic viruses ofthis type in man or domestic animals would be a logical beginning and will be considered herein. These hypothetical RNA tumor viruses will probably share with the known RNA tumor viruses the following characteristics: diameter ofapproximately 100 mpi; buoyant density in the range of 1.16-1.22 gm/ml; single-stranded, high-molecular-weight RNA (larger than that reported for normal cell components or for other viruses) [7, 8]; reproduction in the cytoplasm of infected cells and maturation at the cell surface; and virus replication in neoplastic cells. We will dius be able to apply avariety ofestablishedisolation , purification, and identification procedures in our search. Let us outline a basic procedure using availabletechniques. Possibleviralparticles present in tumor tissues or in fluids (such as serum or milk) obtained from tumorous animals or from cancer patients can be...