After a series of studies, I came to the conclusion in a roundabout way that it might be possible to show the effect of the agent in mice which were within the age of greatest susceptibility. The livers of mice, suckling on mothers with and without the agent, and the milk or milk fractions, when separated on substituted cellulose columns and then injected into suckling mice without the agent, produced a differential effect in the liver phospholipides. The interesting thing is that this biochemical difference disappears when the mice cease breast feeding, which indicates that the agent may insult the tissue and become integrated or incorporated into the metabolism or structure of the cell. This alteration, through hormones or in some other way, may then produce tumors only after the breast tissue has developed.

Dr. Muhlbock: I am glad that two speakers have supported my views. When I accepted the invitation to attend this meeting a few months ago, I talked with Dr. Beard, and he said, "I will give you hell." So I did not know what would happen to me.

I think the situation is quite clear. I am not a virologist; indeed, I do not know much about virology. I have applied half my lifetime to the mammary cancer of mice and to hormones, and Dr. Beard has stated he does not know anything about mammary cancer, but I think he knows about virus particles.

I could say, "What is the sense of my coming here and talking?" I think it is very sensible indeed, because we have to get together and learn from each other so that we will be prevented from looking at only one angle of the cancer problem. Then we learn that there is not only one single problem but many to be solved. As a non-virologist in a meeting consisting mainly of virologists, I would like to stress this point: There are other factors which are instrumental in the development of cancer. If we keep this in mind, I think sometime in the future we will come together and say, "Now, we agree altogether."

Experimental Alteration of Host Response to an Occult Tumor Virus: Induction of Thymic Lymphosarcomas in X-Irradiated Mice

In the past several years it has become well established that the lymphosarcomas and lymphatic leukemias (collectively termed lymphomas) which arise "spontaneously" in mouse strains AK and C58 are due to a virus. The agent is transmissible by cell-free filtrates (CFF) and centrifugates (CFC) which, when inoculated at birth, elicit the same kind of neoplasms in mice of normally nonsusceptible strains (1) and greatly accelerate their development in mice of the strain of origin (2). The presence of the thymus is essential for tumor development in filtrate-injected mice (8-6). Serial passage of filtrable material from tumors with short latent periods significantly increases the activity of the agent, as judged by greater tumor incidence, progressively shorter latent periods, and capacity to induce tumors when the agent is inoculated into older mice (6", 7). Electron micrographs of ultra thin sections of such filtrate-induced lymphomas reveal characteristic particles in the cytoplasm or in the extracellular spaces adjacent to the plasma membrane of thymic tumor cells (8).

Meanwhile it had long been known that lymphomas may be induced in very high incidence in normally nonsusceptible strains of mice by exposure to certain exogenous carcinogens (9). The morphologic identity of the virus-induced and carcinogen-induced lymphomas suggested the possibility that the action of the carcinogens in the inducible strains might also be mediated by an occult virus. Such an induction sequence might be imagined if the carcinogen were, for example, able to suppress "carrier" host resistance, or to release the virus from a bound intracellular site, or both. This possibility has now been investigated for one external agent, X irradiation. While the evidence still leaves much to be desired, the data from work with different strains in different laboratories are at least strongly suggestive. I shall limit my presentation to a summary of this evidence and to some tentative interpretations which are put forth to stimulate further discussion and experimentation.

Most of the pertinent work has been done with strains C57BL, BALB/c, C3H, and their reciprocal F1 hybrid crosses. "Spontaneous" lymphoma incidence in untreated mice of these strains is essentially zero up to 18 months of age, rising to perhaps 0.5 to 4 percent in older mice. The lymphomogenic efficacy of X irradiation is most striking in strain C57BL mice, in which, under optimal experimental conditions, a tumor incidence of between 90 and 100 percent is regularly attainable (10). The first deaths due to induced thymic lymphomas in this strain occur at about 90 to 100 days after X-ray exposure, and incidence reaches a plateau by 8 to 12 months. Strains BALB/C and C3H are somewhat less susceptible to comparable doses of X radiation (9,11).

CFF and CFC, prepared from radiation-induced C57BL thymic lymphomas by the method of Gross (1), have been inoculated into isologous and F1 hybrid mice at birth. Controls have included untreated mice of the same genotypes and mice inoculated at birth with heated (56°, 30 min.) lymphoma CFF, CFF from normal thymus, and CFF from other, nonlymphomatous neoplasms of isologous origin. Filter checks and transplantation characteristics of CFF-induced lymphomas have been used to exclude tumor- cell contamination of the filtrates and centrifugates. The observed lymphoid tumor yield from fresh lymphoma CFF, though very significantly higher than that in any of the control groups (12), still falls far short of the incidence attainable in irradiated mice of these strains, and the onset of these lymphomas is late.

Serial passage of CFF from filtrate-induced lymphomas through newborn mice results in a very appreciable augmentation of tumor incidence and reduction of latent period (12). Similar results have been independently obtained with CFF from radiation-induced lymphomas arising in strain C3H mice (13, 14). Heated lymphoma CFF and CFF from normal thymus and other types of neoplasms were inactive, whereas CFF prepared from a long-established transplantable lymphoma, originally radiation-induced in the thymus of a strain C57BL mouse, was as active as CFF from freshly induced lymphomas (12). In one experiment, CFF prepared from a virus-induced lymphoma of strain AK mice was inactive in strain C57BL mice inoculated at birth (12), which suggested that the agent in strain C57BL lymphoma CFF, though perhaps closely related, was not identical with the lymphoma virus of strain AK. In another experiment,3 CFF from radiation-induced C57BL lymphomas failed to elicit tumors after inoculation into newborn C3H mice. Attempts to date to demonstrate viruslike particles in radiation-induced thymic lymphoma cells by electron microscopy have been generally unsuccessful (15), though such particles have been found in the extracellular spaces of lymphomas induced in C3H mice by CFF prepared from X-ray-induced C3H lymphomas.