It has long been suspected that various forms of cancer, particularly certain lymphomas and
leukemias, are caused or at least ‘‘co-caused’’ by transmissible viruses. This theory has had its ups
and downs during the first half of this century, and it was not generally accepted until the 1950s
that viruses can cause malignant tumors in animals. The known carcinogenic effects of certain
chemicals, irradiation, chronic irritation, and hormones did not fit with the idea of an infectious
origin of cancer. In early experiments, the basic assay to determine whether cancer could
be induced by a transmissible agent involved transmititng malignant disease by inoculation of
filtered extracts prepared from diseased tissues. If the disease occurred in animals inoculated
with such filtrates, it was assumed to be caused by a virus. In 1908, Ellermann and Bang137
transmitted chicken leukemia by cell-free, filtered extracts and thus were among the first to
demonstrate the viral etiology of this disease. In 1911, Rous induced sarcomas in chickens by
filtrates obtained by passing tumor extracts through filters that were impermeable to cells
and bacteria. These findings remained dormant for two decades until Shope showed, in 1933,
that the common cutaneous papillomas of wild rabbits in Kansas and Iowa were caused by a
filterable agent.139 It was later found that when these tumors were transplanted subcutaneously
they became invasive squamous cell carcinomas.
In 1934, Lucke´ observed that kidney carcinomas commonly found in frogs in New
England lakes could be transmitted by lyophilized cell-free extracts.141Twoyears later, Bittner
demonstrated the transmission of mouse mammary carcinoma through the milk of mothers to
offspring.142 This was the first documented example of transmission of a tumor-inducing virus
from one generation to another. Drawing on the experiments of Bittner, Gross postulated that
mouse leukemia was also caused by a virus and that occurrence of the disease in successive generations of mice was due to transmission of virus from parents to offspring. The proof of this hypothesis eluded Gross for a number of years until he was prompted, by evidence based on transmission of Coxsackie viruses to newborn mice, to attempt inoculation of mice less than
48 hours old. Using this approach, he successfully transmitted mouse leukemia by injecting
filtered extracts preparedfrom borgans of inbred AK or C58 mice, which have a high incidence
of ‘‘spontaneous’’ leukemia, or from embryos of these mice, into newborn C3H mice, which have
a very low incidence of leukemia. These experiments demonstrated for the first time that mouse
leukemia is caused by a virus and that the virus is transmitted in its latent form through embryos.
This led to the isolation of a mouse leukemia virus.The isolated virus was also found to induce
Leukemias and lymphomas in inbred strains of mice. Electron-microscopic studies145 showed that
the mouse leukemia virus is spheroid, has a diameter of about 100 nm, and contains a dense,
centrally located ‘‘nucleus’’ separated from the external envelope by a clear circular zone. The
Gross mouse leukemia virus was classified as a type C virus, a term now used to describe a wide
variety of RNA-containing oncogenic viruses of similar morphology.
The RNA oncoviruses have been classified by morphological criteria. Intracytoplasmic type A
particles were initially observed in early embryos of mice and in certain murine tumors. These A
particles are noninfectious, bud into intracellular membranes rather than through the plasma
membrane, and thus stay within the cell. They have an active reverse transcriptase and exist as
a proviral form in chromosomal DNA. Type B viruses have spikes on their outer envelope, bud
from cells, and have been identified primarily in murine species, mouse mammary tumor virus
(MMTV) being an example. Type C viruses have been found widely distributed among birds and
mammals, can induce leukemias, sarcomas, and other tumors in various species, and have certain
gene sequences that are homologous to ‘‘transforming’’ sequences isolated from various
human tumors (see below). Another subgroup, type D RNA oncoviruses, has been isolated
from primate species but their oncogenic potential is not well established. The subtypes of
RNA tumor viruses, known as Retroviridiae, share a genetically related genome containing a
gag-pol-env gene sequence coding for virus internal structural proteins, the special type of
RNA-directed DNA polymerase called reverse transcriptase and viral envelope proteins, respectively.
Thus, they most likely share a common evolutionary heritage.146 However, distinct subclasses of retrovirus evolution, based on pol gene sequence homologies, have been found; one major pathway gives rise to mammalian type C viruses and a second to A, B, D, and avian type C oncoviruses.146 A more recent addition to the retrovirus classification is the human Tcell leukemia virus (HTLV), isolated from patients with certain forms of adult T-cell leukemias (discussed later). The pol gene of HTLV
appears to have evolved from a progenitor common to the types A, B, D, and avian C oncoviruses
rather than from the mammalian C type.146 If true, this would be unusual because most mammalian type C viruses share antigenic determinants among several gag, pol, and env gene products, suggesting a common progenitor for this subclass of retroviruses.
