In examining the morphological conversion of the cell, we can now consider another case: that in which a single type of virus is used to infect cells of different types. Conversion still takes place in the various cell types, but the product of it, depending on the cell type, is different. For instance, a virus fine isolated by Temin transforms the standard fibroblasts to round cells, the heart fibroblasts to pale triangular cells, and the lung epithelial cells to bricklike cells.

These observations on morphological conversion show, therefore, that the morphological type of a cell infected by the Rous virus can vary considerably, depending on the genotype of the virus and on the original morphological type of the cell before infection. This finding may have some relevance on the problem of the participation of viruses in the origin of cancer in general. In fact, a possible objection to the hypothesis that the cancers are produced by viruses may be found in the considerable variety of morphological cancer types, since one might consider unlikely the existence of a correspondingly large number of different cancer viruses. The results of the studies on the morphological conversion here reported show that this is not true, and that a considerable variety of morphological types can arise when a single virus, with its mutants, infects various types of cells found in an organism.

The described conversion has a practical importance in that on it is based a very efficient method of assay for the virus. This assay method was developed by Temin and Rubin (4). It consists in counting the number of foci of transformed cells produced by a virus sample on a monolayer culture of chicken-embryo fibroblasts under suitable conditions. This assay method is as sensitive or more sensitive than methods based on infecting animals; there is, nevertheless, reason to believe that it detects about only 1 to 10 percent of the potentially infective virus particles.

We shall now turn to other consequences of the infection of a cell by the Rous virus. There appear to be changes of physiological type, other than morphological. Several such physiological changes have been reported. One is an increased growth potentiality of the cells, which becomes manifest only under conditions of crowding of the cultures: The converted cells appear to be able to grow further when the normal cells stop growing (Temin, personal communication). Another reported change is an increase in acid production by cultures of converted cells compared to usual cells. Whether this increase is due to an increased aerobic glycolysis of the cells or to greater crowding of the cultures is still uncertain. Also an increase in the ribonucleic acid (RNA) content of the converted cells has been claimed, but not conclusively proved.

Finally, infection by the Rous virus converts the normal cells into virus-carrying and virus-producing cells (5). Virus production by the infected cells begins about 14 hours after infection; the rate of virus release increases with time and reaches a maximum 3 or 4 days after infection. Virus release continues through a number of cell generations without preventing the continued growth and division of the cells. This is shown by the fact that cell populations deriving from a single infected cell under conditions preventing extracellular reinfection produce the same amount of virus as an equivalent number of primarily infected cells, and is even more directly demonstrated by the occurrence of division of infected cells kept individually under continuous observation, after virus release has started. It is not known, however, whether all the descendant cells can continue to multiply and release virus indefinitely.

The property of the Rous virus to infect a cell in such a way, that it will release virus at a constant rate and continue to divide normally for many generations, distinguishes it from the majority of the animal viruses; the Rous virus because of this property is considered a moderate virus. It is likely that other viruses, including possibly all tumor-producing viruses, belong to this class; conclusive evidence is still, however, lacking.

From the in vitro study of the Rous virus, another relevant property becomes apparent: The virus has a very restricted host range. This property of the virus can be detected also in the animal. The virus originally isolated by Rous was restricted in its action to the connective tissue of a special line of chickens. During subsequent cultivation, variations arose that extended the host range to the connective tissue of other lines of chicken, and later of ducks, and to brain tissue of chickens. In vitro, mutants of extended host range were isolated by Temin (personal communication); these mutants are able to grow in genetically resistant chicken cells, and some of them, more rare, are able to grow in duck cells. Thus the in vitro experimentation reveals the same general situation as occurs in animals and, in addition, shows that the changes in host range observed in animals are due to the selection of special virus mutants.

I wish now to make some considerations of general and of a somewhat speculative nature on the essential aspects of the process of infection of a normal cell by the Rous virus and of its conversion into a cell with new properties, which we can define as a Rous cell.

A first point that I wish to discuss is the conversion phenomenon. That a virus can convert a cell by infecting it and can impress upon it new properties has been known for some time. Very clear cases of conversion occur in bacteria infected by temperate bacteriophages; a classical one is that of nontoxigenic strains of Corynebacterium diphtheriae, which upon infection with a bacteriophage of suitable genetic constitution are converted to toxin producers. In this and other known instances the mere presence of the phage genome inside the host cells is the only requirement for conversion. One may, however, wonder how legitimate it is to compare results obtained in bacteriophage work with findings in the Rous field. Bacteriophages are different from the Rous virus in that their genome is constituted by deoxyribonucleic acid (DNA); furthermore it is established that the genome of temperate phages enters in an intimate relationship with the genome of the host cell; in the Rous virus, on the contrary, there is an RNA virus, for which any intimate interaction with the chromosomal genie apparatus of the cells is not only unproved, but probably unlikely. However, it has been shown that the bacteriophage genome can cause conversion of the host cell even if it does not establish a contact with the host genome (6); thus the Rous virus genome can act from some position in the cell not necessarily associated with the chromosomes. Even with this limitation, the genome of the Rous virus can be regraded as integrated with the cell.

One may wonder about the mechanism by which the conversion is carried out. To this effect it is essential to consider that the virus probably brings into the cell a piece of nucleic acid of the length of 5,000 to 10,000 nucleotides. Such a piece of nucleic acid is equivalent in length to a single gene in bacteriophage. It is therefore likely that the virus brings into the cell the information corresponding to one or very few genes; as a consequence, not more than a few new virus-controlled proteins are synthesized in the infected cells. A virus-controlled protein may become incorporated in the surface of the cell, as experimentally found for cells infected with influenza virus (7), and then may produce conversion by changing the structure of the cell surface and consequently its shape and function. This idea finds support in phenomena occurring in lysogenic and colicinogenic bacteria, which carry either a phage or phagelike genome; phenomena of this type are the conversion of the surface antigen in Salmonella, by a bacteriophage (6), and the association of the colicin protein with the surface antigen, also in Salmonella (8).

The second point of general interest is the nature of the Rous cells, and particularly whether the Rous cells, obtained by converting normal cells by the Rous virus, are truly cancerous. We lack direct conclusive evidence on this point, and it should be appreciated that direct evidence is difficult to obtain. Transplantation experiments would be needed to test the ability of Rous cells produced in vitro to grow as cancer cells in the animal ; however, the continued release of virus from the Rous cells would complicate these experiments, since, in the animals receiving the transplant, a cancerous growth could be induced by the virus carried with or produced by the transplanted cells.

Some more indirect experimental evidence on the nature of the Rous cells is however available: The development of Rous cell foci in vitro occurs in a way that closely resembles, in timing and probability of infection, the development of virus-induced small tumors on the chorioallantoic membrane of the chicken embryo; Rous cells produced in vitro transferred onto an uninfected chorio-allantoic membrane give rise to similar small tumors by their multiplication. Therefore the consequence of a Rous virus infection in vitro and in the animal appears to be similar. This conclusion suggests that the Rous cell observable in vitro is similar to the first product of virus infection in the animal. This conclusion is strengthened by the similarity in host range of the virus in vitro and in the animal, as already discussed.

However, the question whether the Rous cell, defined as the first transformed cell, either in vitro or in the animal, is a cancer cell, still remains unanswered. Between the first infection of a cell by the virus and the development of the fatal sarcoma in the animal many cellular divisions take place. As in other virus-induced tumors, the primary growth might be of benign nature and might become transformed into the malignant form at a later stage, by a second event, for instance, a genetic change of some cells.

A considerable amount of work will have to be done before these questions are answered; until then the real nature of the primary Rous cells and their relation to the final sarcoma will be doubtful. However, since, under favorable circumstances, the formation of the sarcoma in the chicken follows promptly the infection with the virus, the primary Rous cells represent at least something very near to the ultimate cancer cell.

In conclusion, I wish to emphasize that the system constituted by the Rous virus and chicken-embryo cells is especially suited for in vitro experimental work of oncological nature. This system has two outstanding properties: (1) the ease with which the normal cells can be transformed into Rous cells; (2) the opportunity that it offers, unique in the cancer field, to compare the properties of a normal cell type with those of its direct pathological derivative.