Although the virus, the host, and, in particular, the extent and dynamics of viral replication are quite different from lentiviruses, the oncogenic potential of many simple retroviruses also rests on the fact that there are a large number of infected cells in the host. As mentioned earlier, for infections with a number of simple retroviruses (MLV in mice, ALV in chickens, MMTV in mice, etc.), the large number of infected cells means that the probability of the viral DNA inserting near a proto-oncogene, which is unlikely when considered on the basis of a single cell with one or two viral DNA insertions, becomes quite likely when considered in the context of the millions of infected cells present in the host. Although the activation of a proto-oncogene by retroviral insertion is not usually sufficient to convert a normal cell into a tumor cell, it can be the rate-limiting step in the development of a tumor, by creating a large pool of cells within which subsequent mutations—either spontaneous or induced by additional insertions of viral DNA—are likely to occur. This model requires both that viral DNA can insert at a large number of sites (so that insertion near the target gene is possible) and that an appropriately large number of cells be infected for the activation (or, much more rarely, inactivation) of specific target genes.

As is pointed out in Chapter 8, the effects of retroviral insertions on the expression of host genes also applies to endogenous viruses in evolutionary time. To be selected on an evolutionary time scale, the insertion must benefit the host. The clearest benefit conferred by endogenous proviruses is resistance to infection by certain (often related) exogenous viruses. However, as in the case of exogenous proviruses, endogenous proviruses can also activate or inactivate expression of nearby cellular genes. Indeed, a number of familiar mutations in inbred mice were caused by this kind of insertional inactivation. Over longer periods of evolutionary time, such inactivation events are likely to be lost by selection, but a few cases of apparently selectively favorable activation of cellular genes, such as the expression of amylase in the human salivary gland described in Chapter 8, have occurred.

Clearly, retroviruses not only depend on, and are influenced by, their hosts, but they, in turn, can profoundly affect the genetic makeup of their hosts. Building on the example of retroviruses that naturally acquire cellular oncogenes, major efforts are under way to modify retroviruses deliberately so that they will carry and express genes that have desirable, rather than undesirable, effects on their hosts. Retroviral vectors have become useful tools in the laboratory for genetic alteration of cells in culture and for modifying animals for experimental purposes. The effective application of vector strategies to humans remains to be achieved, but preliminary experiments suggest that modified retroviruses will eventually be used for the treatment of hu man diseases of both genetic and nongenetic origin. Effective use of these strategies must be based on a thorough understanding of the biology of the retrovirus-host interaction. For example, a vector based on a simple retrovirus will only be able to infect dividing cells, severely limiting the range of cell types available for treatment, especially in adults. It is also important to remember that proteins expressed by cells infected by retroviral vectors may be treated as foreign by the host and give rise to undesirable immune responses against them or the transduced cell. Such responses can affect the efficacy of the therapeutic protein or the survival of the infected cell.