SV40-Mediated Oncogenesis

Maurizio Bocchetta and Michele Carbone

Simian virus 40 (SV40) was first isolated in 1958 among other simian viruses
from contaminated polio vaccine preparations, which were inadvertently
administered to millions of people in different countries from 1954 to 1963.
Soon after SV40 was introduced to the scientific com¬munity (1) its capabilities
to induce different forms of cancer in exper¬imental animals were recognized
(2,3). However, epidemiology failed to establish a conclusive link between the
administration of SV40-contaminated polio vaccines to humans and the development
of cancer (4–8). Because epidemiology was inconclusive, SV40 has been
consid¬ered for many years to be harmless to humans. From the 1970s, throughout
the 1980s, and until recently, SV40 has been utilized mainly as a tool to
understand key molecular processes such as DNA replica¬tion, splicing, and
translation in mammalian cells. It has also been widely used to uncover the
process of the cell cycle control because of the interaction of its major
oncogenic protein products with critical tumor suppressor gene pathways of the
cell. Indeed, the SV40 onco-genes have probably been the most commonly used
tools to experi¬mentally immortalize or transform rodent and human cells, mainly
fibroblasts. Occasional screening of human tumors suggested that SV40 could
participate in the development of human cancer (9–15).

The interest concerning the association of SV40 with certain human malignancies
(specifically, malignant mesothelioma, tumors of the brain and bone, and
non-Hodgkin’s lymphoma) and the possible causative role of SV40 in the onset of
these forms of cancer has vigor¬ously resurfaced during the past decade
(reviewed in refs. 16 and 17). This has been caused by the development of new
molecular techniques that now allow investigators to better study the presence
and the bio¬logic effects of viruses in infected cells. The wealth of
experimental data accumulated over the recent past conclusively associates SV40
with human tumors, especially with malignant mesothelioma (16,18,19). A recent
meeting of the National Academy of Science, Institute of Medi¬cine (IOM)
concluded that SV40 is a “strong” carcinogen, and that there is “moderate
evidence” that SV40 causes some human tumors (20). This was emphasized by the
recognition that previous epidemiologic

studies were flawed and could not provide any conclusive indication regarding
the potential oncogenicity of SV40-contaminated poliovac-cines. The IOM has also
concluded that the experimental evidence available so far suggests that SV40 may
be transmitted among humans, and that SV40 can cause cancer in humans under
natural conditions.

This chapter reviews the virology of SV40, discusses the association of SV40
with human malignant mesothelioma, and describes the inter¬actions that SV40
establishes with human mesothelial cells, since these cells are uniquely
susceptible to SV40 transformation and immortal¬ization (21).

SV40 Genomic Organization, Gene Transcription, and Cycle of
Infection

The genome of SV40 is a small, circular double-stranded DNA mole¬cule. The
genome of SV40 strain 776 (also called reference strain, or wild-type SV40) is
composed of 5243 base pairs (bp). Different strains of SV40 exist, all sharing a
very high level of DNA sequence conser¬vation, with the exception of the
transcriptional enhancer region, the very C-terminal portion of the SV40 major
oncoprotein (22), and the intron of the early transcripts (23). Aside from
differences in the enhancer region, the following description of the genomic
organization applies to all SV40 strains. At least six virally encoded protein
products are translated in permissive host cells through alternative splicing
and translation of overlapping reading frames of the SV40 messenger RNAs
(mRNAs). The SV40 genome is organized in three regions: a regulatory region
(that includes the viral origin of replication and a bidirectional promoter), a
region including the early genes, and a region comprising the late genes. The
early and late genes extend in opposite directions with respects to the
regulatory regions (Fig. 3.1). The denominations

“early” and “late” reflect the order of their transcription/synthesis
in the host cell after SV40 infection. SV40 enters the cells after interaction
with its receptor [major histocompatibility complex (MHC) class I] (24, 25) on
the plasma membrane, and it is internalized within the cytoplasm through a
specific endocytosis pathway (caveole) (26). It is then traf¬ficked into the
nucleus by means of the interaction of the importin a2/b heterodimer with an
SV40 capsid protein (VP-3) (27). It is still unclear where the viral genome is
released from its protein envelope. Once within the nucleus, the SV40 early
genes are transcribed. The early mRNAs (the precise transcription initiation
sites of both the early and late mRNAs vary over a number of positions) contain
a 347 bases intron that can be alternatively spliced giving rise to two classes
of mRNA (28). Translation of these two types of mRNAs produce two proteins: the
large-tumor antigen (or Tag), and the small-tumor antigen (or tag). Overall, the
early SV40 mRNAs represent a small fraction of the total RNAs in infected cells,
and early genes mRNAs and protein products can be detected in freshly infected
cells using only very sensitive methods (29). The ratio between Tag-encoding
mRNAs and tag-encoding mRNAs varies in different cell types, and it has been
studied only in vitro. In HeLa cell extracts the Tag:tag mRNA ratio is about
100:1 (30). To summarize, the SV40 early genes exert their function even though
they are synthesized at very low levels in infected cells.

The SV40 Tag and tag interact with a number of cellular proteins (these
activities are discussed below), and the end result of these con¬certed
activities is driving the host cell into the S phase so that the viral genome
can be replicated. Tag is required for the replication of the SV40 circular
chromosome. Tag binds as a double hexamer to the SV40 origin of replication (31)
where it interacts with the host’s DNA polymerase a-primase to initiate DNA
replication (32). Tag also has DNA helicase activities that play an important
role in the SV40 chromosome replica¬tion process (33).

As previously stated, the SV40 regulatory region contains a bidirec¬tional
promoter. This means that the regulatory region can promote transcription of
both the early and late genes. During the early stages of infection, however,
transcription of the late genes does not take place because of the binding of
transcriptional repressors at sites located in the proximity of the late mRNAs
transcription initiation site. These transcriptional repressors belong to the
steroid-thyroid hormone recep¬tor superfamily (34,35). During SV40 DNA
replication these repressors are progressively titrated-off from the late
promoter, so that transcrip¬tion of the late genes can take place. The binding
of Tag to the regula¬tory region also enhances the latter process, since Tag
represses the transcription of its own mRNA and promotes transcription of the
late gene mRNAs (36). Efficient synthesis of the late gene mRNAs marks the
beginning of the late SV40 cycle of infection during which large amounts of the
viral capsid proteins are produced. The late mRNAs also arise from alternative
splicing of the same family of transcripts. Two major classes of late mRNAs are
produced: the 16S and 19S mRNAs [the classification derives from the migration
of these mole¬cules in sucrose gradients (29)]. The 16S mRNAs code for VP-1,
which

is the major SV40 capsid protein, while the 19S mRNAs code for the agnoprotein,
VP-2 and VP-3. VP-2 and -3 are less abundant capsid pro¬teins but play an
essential role in the SV40 packaging process (37,38). VP-3 has also been shown
to interact with basal cellular transcriptional factors that repress
transcription of the early genes and enhance the completion of the SV40
infection cycle (39). The function of the small agnoprotein is still unclear.
Its perinuclear localization suggests that it may participate in nuclear
trafficking of SV40 capsid proteins (40, 41). Recent studies have indicated that
during JCV infection ( JCV is a human polyomavirus closely related to SV40) the
agnoprotein may interact (directly or indirectly) with Tag and contribute to the
tran¬scription regulation of JCV (42). Furthermore, JCV agnoprotein can inhibit
cell cycle progression by binding to cellular p53 and thus increasing the
expression of the cyclin-dependent protein kynase (CDK) inhibitor p21WA F (41).
Whether the SV40 agnoprotein has similar biologic activities has not been
investigated.

The late phase of SV40 infection is characterized by a massive pro¬duction of
capsid proteins that accumulate in the nuclei of infected cells. A large number
of viral particles are assembled, and the host cell is eventually lysed, with
consequent release in the extracellular envi¬ronment of infectious SV40
particles (Fig. 3.2). Therefore, SV40 manip¬ulates the host’s cell cycle to
ensure replication of its own DNA genome. Malignant transformation of the host
is not required for

SV40’s life cycle, and the most common outcome of SV40 infection is the
lysis of the infected host. However, SV40 uses two extraordinarily powerful
oncoproteins to undermine the host’s cell cycle checkpoints, and, if anything
goes wrong with the SV40 lytic pattern of infection, any mammalian cell
containing SV40 may undergo malignant trans¬formation. The entity of such risk
varies between different cell types. As a whole, in vitro SV40-infected human
cells from different tissues display a mixture of cytopathic and transformed
phenotypes. This characteristic pattern of infection of human cells by SV40 led
some investigators to call SV40 infection of human cells “semipermissive” (43).

Susceptibility to SV40 Infection and SV40-Mediated
Transformation

The in vitro outcome of SV40 infection critically depends on the species and
cell type of the host. Traditionally, cells are classified as permis¬sive,
nonpermissive, and semipermissive to SV40 infection (reviewed in ref. 16).
Prototypes of permissive cells are those derived from African green monkey
kidneys. These cells are uniformly infected by SV40, synthesize large amounts of
viral particles, and display a typical pathologic morphology after SV40
infection characterized by large perinuclear vacuoles (Fig. 3.2) when the SV40
titer reaches about 107 median tissue culture infective dose (TCID)50/mL in the
tissue culture medium (44). SV40-infected African green monkey kidney cells
invari¬ably undergo cell lysis in vitro, and SV40-mediated malignant cell
transformation in these populations, although theoretically possible, must be an
extremely rare event. During the past decade we have infected a substantially
large number of African green monkey kidney cells with different strains of SV40
and we have never observed cell survival after SV40 infection. Cell lysis, of
course, prevents malignant transformation.

Nonpermissive cells, such as rodent cells, on the other hand, do not allow the
replication of the SV40 genome. In rodent cells the SV40 Tag does not properly
interact with the host’s DNA polymerase a primase, and thus it is unable to
initiate the replication of the SV40 chromosome (45). Nevertheless, the SV40
oncoproteins are still capable of driving the host cell into the S phase and
eventually into mitosis, but the outcome of this process is a sort of abortive
transformation, since the SV40 DNA cannot replicate and is not propagated in the
dividing cells. Therefore, SV40 can transform rodent cells only after
integration of its genome in the host’s chromosomes in such a way that the
integrity of the SV40 early genes and their expression are preserved.
Integration is rather infrequent. For example, the average rate of cell
transformation of mouse fibroblasts after SV40 infection is about 10-7 (46).
This rela¬tively low frequency of transformation mainly reflects the infrequency
of proper integration of the SV40 genome into the host’s genome, and does not
imply that SV40 is a poor oncogenic factor in rodent systems. In fact, nearly
100% of artificially engineered transgenic mice expressing the SV40 oncoproteins
develop tumors. If the SV40 regulatory region drives the expression of the SV40
oncoprotein in these animals, mice develop brain tumors (47). However,
tissue-specific expression of the SV40 early genes in transgenic mice has led to
the development of in vivo tumor models for virtually all tissues (48–50).
Altogether, these experimental evidences demonstrate that the SV40 tumor
antigens are exceedingly potent cancer-inducing agents in rodents.

Figure 3.3. Electron microscopy displaying complete SV40 viral particles
accu¬mulating in an infected human fibroblasts. Original magnification: 20,000¥.

Human cells are traditionally described as semipermissive to SV40 infection
(43). This term has been used to emphasize different features of human cells
infected with SV40, and it reflects the variance in sus¬ceptibility to SV40
infection of human cells derived from different tissues (21,51–54). SV40 grows
efficiently in some human cells, such as newborn kidney cells (55) and
spongioblasts (54), but it grows poorly in cells from other tissues (56).
Despite these differences, a unifying feature of human cells exposed to SV40 in
vitro is that only a fraction of the cell population supports SV40 infection at
any given time, while a substantial percentage of exposed cells is apparently
unaffected (reviewed in ref. 16). The molecular basis of this situation is still
rather undefined. Early studies indicated that SV40 enters all human cells after
exposure, but that the cellular environment plays a pivotal role in determining
whether SV40 will produce a productive infection or not (57,58). Therefore, only
a fraction of human cells exposed to SV40 in vitro express the SV40 early genes,
replicate SV40 DNA, produce viral particles, and undergo SV40-mediated cell
lysis at any given time. Accordingly, only a small percentage of human
fibroblasts exposed to SV40 expresses Tag 48 hours after infection, actively
replicates the SV40 chromosome, and assembles viral particles (Fig. 3.3), an
event that

eventually leads to SV40-induced cell lysis, with consequent release
of SV40 viral particles in the medium (Fig. 3.4). This equilibrium seems to be
maintained in human fibroblast cultures over several weeks after SV40
introduction into the cell population (21). SV40 does not trans¬form human
fibroblasts because the viral cytotoxic effects are predom¬inant, and SV40 lyses
the cells that support its transcription and replication. The notion that
SV40-mediated cell death represents the reason why human fibroblasts are not
permanently transformed after SV40 infection is supported by transfection
experiments. Introduction into human fibroblasts of SV40 genomes mutated in
their origin of replication gives rise to the production of malignantly
transformed clones. In the latter scenario, the rodent nonpermissibility to SV40
is artificially reproduced in human cells because of mutations in the SV40
origin of replication that do not allow replication of the SV40 chromo¬some. In
these conditions the frequency of cell transformation of human fibroblasts
closely approaches that of mice fibroblasts (rate of transformation of about
10-7) (21). Because of the difficulty of trans¬forming human cells by natural
infection, most of the studies concern-

ing SV40-mediated transformation of human cells have been con¬ducted
using transfection of replication defective SV40 genomes. Since stable
transfection requires, similarly to the nonpermissive host, inte¬gration into
the host’s genome, it was erroneously assumed that SV40-mediated transformation
of human cells required the integration of the SV40 genome in the human cellular
DNA (Fig. 3.5).

The SV40 Oncoproteins

As mentioned above, the SV40 genome is rather small. For this reason SV40 must
rely mainly on cellular genes to complete its life cycle. Nevertheless, SV40
needs to manipulate the host’s cell cycle to ensure

the replication of its DNA. Despite its intrinsic “economy” limitations, SV40
has evolved effective subversive proteins that target both nuclear and
cytoplasmic activities. Deregulation of nuclear activities is achieved by Tag,
while deregulation of cytoplasmic activities is left to the SV40 tag. Tag is
probably one of the most multifunctional proteins in nature, since it
participates in DNA, RNA, and protein–protein inter¬actions to ensure SV40
replication and transcription. Tag also engages in a number of interactions with
host’s cell proteins aimed at the deregulation of the host cell cycle. Tag binds
and inactivates the two major tumor suppressor pathways of the cell, p53 (59,60)
and pRb protein family (59,61,62). Through these interactions Tag
simultane¬ously knocks out the two most critical cellular networks controlling
G1/S transition and possibly G2/M checkpoints (63–65). Through the inhibition of
p53, Tag also impairs the major cellular control of genomic stability and
apoptosis program (reviewed in ref. 62). Furthermore, the SV40 Tag binds and
affects the functions of a number of cellular pro¬teins all involved in protein
folding, transcriptional regulation, cell cycle progression, and stability of
the genome. So far, the known cel¬lular proteins that interact with Tag include
(besides p53 and pRb protein family) the transcriptional coactivator p300 and
its closely related p400 (66,67), the mammalian homologue of heat shock protein
(Hsc)70 (reviewed in ref. 68), the DNA binding protein kin17 (69), TATA binding
protein (TPB) and TFII-D complexes (70), cyclin A/p33CDK2 complexes (71), and
possibly others. It is reasonable to anticipate that the list of cellular
proteins interacting with the SV40 Tag will expand with further investigations.
Probably, we still have an insufficient inter-pretation of the net result of all
these interactions. However, the known outcomes of SV40 introduction in a host
cell include induction of the insulin-like growth factor (IGF)-I and its
receptor (72,73), induction of cyclin A (74) and the cyclin-dependent kinase
cdc2 (75), and promotion of chromosomal instability (76,77). More recent studies
(see below) have indicated that Tag participates (at least in some human cell
systems) in the induction of met signaling (78), telomerase activation (79),
RASSF1A repression (80), and Notch-1 induction (81). To summa¬rize, Tag affects
nuclear functions through transcriptional transactiva-tion, protein inhibition,
and direct binding of Tag to the host DNA. The outcomes are aimed both at the
inhibition of cell cycle checkpoints (p53 and pRbs interactions) and at the
induction of proteins involved in the promotion of cell cycle (cyclin A, cdc2).
At the same time, Tag indirectly provides critical survival signals to the cell
through the IGF-I pathway, which may play an essential role for circumventing
apoptosis during the early phases of SV40 infection.

Tag has a J-domain in its N-terminal region that mediates its binding to Hsc-70.
Mutational analyses have revealed that the majority of Tag functions require an
intact J-domain, suggesting that Hsc-70 may play a critical role in the
execution of Tag activities (82). In mouse cells, the minimal fragment of Tag
sufficient for cell transformation includes the J-domain and a region (amino
acids 102–115) responsible for Tag binding to pRb protein family (82,83). On the
other hand, mutant Tag molecules missing the N-terminal J-domain are virtually
inactive.

These observations have suggested that the SV40 Tag may function essentially as
a molecular chaperone through its interaction with Hsc-70 mediated by the Tag
J-domain (82,84). It is possible that Tag may interfere with a number of
cellular protein functions by modifying their conformation, thus operating as a
sort of nuclear chaperone specific for a number of substrates. This would
explain why Tag profoundly affects many cellular pathways though being expressed
at very low levels in early-infected cells (29).

The functions of Tag are implemented in the cytoplasm by the SV40 tag. The
latter is a small protein consisting of 174 amino acids. The first 82 residues
of tag are identical to those of Tag; therefore, both SV40 oncoproteins share
the same J-domain (reviewed in ref. 68). Down¬stream from the J-domain there are
sequences that mediate tag inter¬action with cellular protein phosphatase 2A
(PP2A) (85). Functions of the SV40 tag include transactivation (and
transrepression) of cellular and viral promoters and inhibition of PP2A
(reviewed in ref. 86). The latter represents the best known function of tag.
PP2A dephosphory-lates a number of cellular proteins including components of the
mitogen activated protein kinases (MAPKs). The latter proteins are involved in
the phosphorylation cascade that mediates the signal trans-duction pathway
common to several growth factor receptors. PP2A exists as a heterotrimer of B,
A, and C subunits. The SV40 tag binds PP2A trimer and displaces the B subunit,
with consequent reduction of PP2A activity (86). Thus, through its inhibition of
PP2A, tag indirectly reinforces mitogenic stimuli by intensifying MAPK
signaling, an event that leads to activating protein (AP)-1 induction (87), and
transcrip-tional induction of the key cell cycle regulators cyclin D1 (88). The
resulting effect of the SV40 tag is to amplify the activities of Tag, especially
during the first phase of SV40 infection (reviewed in ref. 86). Nevertheless,
tag is capable of inducing S-phase entry indepen¬dently from Tag in certain cell
systems and, more importantly, its func¬tions are required for SV40-mediated
transformation of human cells (86).

SV40 and Apoptosis

The mechanisms through which SV40-containing cells escape apopto-sis are still
controversial. In theory, Tag should possess both pro- and antiapoptotic
activities. Tag binds and inhibits p53; thus it should sup¬press p53-dependent
apoptosis. On the other hand, binding of Tag to pRb causes the release of E2F-1,
which is a potent inducer of apoptosis (89,90). However, the experimental data
gathered so far complicate this probably oversimplistic model of Tag interaction
with proteins that reg¬ulate apoptosis. In fact, Tag promotes p53-dependent
apoptosis in rat embryonic fibroblasts exposed to genotoxic chemicals (91). A
Bcl2-like domain has been identified in Tag, and this domain works in protecting
certain cells from apoptosis (92). These observations suggest that Tag may
affect the intrinsic pathway of apoptosis directly by means of its Bcl2-like
domain. Yet, the actual significance of this domain is rather

obscure. As an example, Tag-induced enhancement of p53-dependent apoptosis
induced in rat embryo fibroblasts after exposure to genotoxic chemicals appears
completely unaffected by mutations in the Bcl2-like domain (91). We have
recently shown that SV40 infection induces Notch-1 (81). Notch-1 is a highly
pleiotropic protein that regulates crit¬ical cell fate decisions during
development and differentiation in a wide spectrum of Methazoan (93). Among
other functions, Notch-1 acts as an antiapoptotic protein in murine
erythroleukemia cells (94). Thus, it is possible that activation of Notch-1
signaling may participate in preven¬tion of apoptosis in SV40-containing cells,
and this hypothesis is now being tested in our laboratory.

Tag seems to protect liver cells apoptosis mediated by Fas. However, in this
system Tag apparently does not affect the expression levels of Fas itself;
rather, it enhances a protective mechanism involving the protein kinase C
signaling pathway (95), thus implicating tag in this process. One study supports
this interpretation and indicates that expression of the tag alone in transgenic
mice is sufficient to confer liver cells resistance to CD95-mediated apoptosis
(96). Furthermore, tag pro¬tects from apoptosis rat embryo fibroblasts harboring
mutations in the Tag (97). Indeed, the SV40 tag is potentially an antiapoptotic
protein. By means of its inhibition of PP2A, tag should reinforce survival
signals acting through the protein kinase B/Akt signaling pathway. Akt
phos-phorylates and inhibits a number of components of the cell death machinery,
including BAD, pro-caspase 9, and MDM2 (reviewed in ref. 98). Nevertheless, the
outcomes of tag functions on apoptosis are rather controversial as well. In
fact, tag induces apoptosis in human osteosar-coma cells through a
p53-independent mechanism (99).

Overall, it is probably reductive to clearly identify pro- or antiapo-ptotic
functions of any of the SV40 oncoproteins. All cellular oncogenes share
activities that induce cell proliferation and apoptosis. It is now recognized
that the proliferative and apoptotic pathways are coupled, and that the role
played by survival factors in completing proliferation versus programmed cell
death choices can be critical (100). In this light, induction of the IGF and
hepatocyte growth factor (HGF) pathways by SV40 may be the key in understanding
the overall avoidance of apo-ptosis in SV40-infected cells (which express the
entire SV40 genome and not just portions of the SV40 genome, as most of the
systems described above). The recent finding that Tag translocates the IGF-I
receptor substrate in the nucleus (101) further underscores the inter¬twined
relation between SV40 and survival signals acting through the IGF pathway.

SV40 Transformation of Human Cells

Human cells from diverse tissues have been transformed using SV40 or the SV40
oncoproteins (59). Among cells of different tissue origin, fibroblasts are the
most thoroughly investigated human cell type for SV40-mediated transformation in
vitro. Most of the studies have been performed using replication-defective SV40
mutants to avoid SV40induced cell death. In these stable transfections, SV40
increases the fibroblasts’ proliferative potential of 20 to 30 cell doublings,
without conferring tumorigenicity in immunocompromised mice. Then the
transformed population enters the so-called crisis characterized by cell growth
arrest and senescence that ultimately causes apoptosis in most of the cell
population (102). Occasional cells escape senescence and give rise to immortal
clones (103,104). In conclusion, in the conditions described above (transfection
of the SV40 early region), SV40 inhibi¬tion of the p53 and pRb pathways is
insufficient to prevent crisis.

The discovery of telomerase and its role in cellular immortalization has
provided a better understanding of the process of malignant trans¬formation of
the cell (reviewed in ref. 105). Telomerase is an RNA-protein complex that
ensures the maintenance of the length of chromosome ends (telomeres). One of the
protein components of telomerase (TERT) is a reverse transcriptase that uses the
RNA com¬ponent of the telomerase complex as a template to add tandem repeats of
a short sequence (TTAGGG in mammals) at the chromosome ends (105). Human somatic
cells do not express TERT. In the absence of telomerase activity the ends of
chromosomes shorten at every cycle of cell division because of the intrinsic
mechanism of DNA replication. The cell has evolved a sensitive network
controlling the length of telomeres. This is necessary because progressive
shortening of telo-meres eventually leads to dramatic chromosomal instability.
Therefore, telomeres function as a sort of “hourglass” for cell division: when
telomeres approach a critical size, the cell deploys an irreversible pathway
that leads to cell cycle arrest, senescence, and apoptosis (102,106). As a
consequence, all tumor cells need to induce some mech¬anism to ensure the
maintenance of telomere length in order to prolif¬erate indefinitely. The
experimental evidence available indicates that the majority of human tumors
achieve de novo expression of TERT as a result of mutational events. Human
tumors in which TERT is not expressed have developed an alternative process of
telomeres elonga¬tion through recombination, termed ALT, characterized by
unusually long telomeres (107). A recent study has shown that SV40 is unable to
induce telomerase activity in human fibroblasts (79). Therefore, immor¬tal
fibroblasts can arise only after sporadic activation of telomerase due to
chromosomal rearrangements or mutations. This explains the infrequency of
malignant transformation of fibroblasts after exposure to SV40. A similar
situation is repeated in most human cell systems. For example, mammary gland
epithelial cells transformed by SV40 undergo senescence, which is not
circumvented by the introduction of oncogenic H-ras. However, sporadic cells
escape crisis and form tumorigenic lines in mice (108). In synthesis, these
studies have shown that the introduction of the SV40 oncoproteins, in
conjunction with the expression of a chronic mitogenic stimulus represented by
oncogenic H-ras, is insufficient for malignant transformation of human cells. In
these settings, malignant transformation still requires random muta-genic events
that affect an undefined number of genetic elements. Recent studies have
demonstrated that in the conditions described above, activation of TERT is the
critical requirement for malignant

transition. Indeed, the minimal genetic elements for normal human cells to
become transplantable tumors in experimental animals are rep¬resented by an
oncogenic H-ras allele, the SV40 early region and active telomerase (109). These
apparently limited requirements have led some researchers to the conclusion that
the transition to a fully malignant state is achieved in human cells through
activation of telomerase, the induction of a chronic mitogenic stimulus,
inhibition of both p53 and pRb pathways, and inactivation of PP2A. The SV40
early gene prod¬ucts provide the latter three functions (110).

This recapitulation of the carcinogenetic process in human systems is probably
oversimplified. The SV40 oncoproteins do not simply inhibit p53, pRb protein
family, and PP2A, but participate in a number of interactions that profoundly
affect the entire cell fate program of the host. In other terms, researchers
tend to underestimate the extent of multifunctionality and the complexity of
operation of the SV40 tumor antigens. A good example of this is provided by the
interaction of SV40 with the Notch signaling pathway (see below). Immortal,
SV40-transformed human mesothelial cells are completely growth-arrested if
activation of Notch is impaired through chemical inhibition (81), implicating
Notch-1 as another essential player in the process of malig-nant transformation
of at least some human cell types. Accordingly, interference with Notch-1 using
either chemical inhibition or antisense technology suppresses the malignant
phenotype of human fibroblasts expressing TERT, oncogenic ras, and SV40
oncoproteins (111). This evi¬dence suggests that our interpretation of the
molecular circuitry of cancer is still rather incomplete, and additional genetic
elements may soon be included among those required for the malignant
trans¬formation of human cells.

SV40 and Malignant Mesothelioma

SV40 is highly oncogenic in the hamster, where it induces tumors of
different types. The pattern of tumor formation is dependent on the site of
administration. When injected intracardially, so that it can spread to all
organs, SV40 induces malignant mesothelioma (MM) in about 60% of animals after a
latency of 3 months (112). The remaining 40% of ham¬sters develop lymphomas,
osteosarcomas, and myxomas (112). The induction of mesotheliomas in these
animals is strictly dependent on the expression of tag, since hamsters injected
with SV40 mutants unable to synthesize tag developed lymphomas (the relevance of
tag in the process of mesothelial cells transformation is further discussed
below). If SV40 is injected directly into the pleural space, 100% of ham¬sters
develop MM (112). Intracranial injection of SV40 causes brain tumors, while
subcutaneous injection causes lymphomas and sarcomas at the site of injection
(reviewed in ref. 16). This pattern of tumor for¬mation suggests that the
hamster mesothelium is particularly suscep¬tible to SV40-mediated oncogenesis.

In the past decade, SV40 has been detected by different laboratories worldwide
in a number of human tumors using a variety of techniques

pointed at the detection of SV40 DNA, RNA, and oncoproteins in the tumor
specimens (16,17). The wealth of evidence linking SV40 to spe¬cific human
cancers is such that now three different panels of scientists have conclusively
linked SV40 to human cancer (18–20). Strikingly, the panel of human tumors in
which SV40 has been detected perfectly matches that induced by SV40 in hamsters,
suggesting that SV40 may play a causative role in the onset of these
malignancies (Fig. 3.4). Malignant mesothelioma represents the human tumor in
which SV40 association and putative oncogenicity have been more extensively
studied. The overall consensus is that 50% to 60% of MMs in the United States
contain SV40. Geographic differences in the prevalence of SV40 in MM have been
described, and the possibility that such variance may originate from differences
in SV40 contamination of poliovaccine preparations used in various countries has
been proposed (113–118). The association of SV40 with MM is highly specific,
since SV40-positive MM specimens contain SV40 only in the tumor cells, while the
surrounding stromal tissue is SV40-free (119,120). SV40 is biologically active
in MM, because its major oncoprotein Tag has been demon¬strated to bind cellular
p53 (121) and pRb protein family members (122). The functions of the SV40
oncoproteins are required for the main¬tenance of the malignant phenotype in MM,
since targeting Tag through antisense techniques causes growth arrest and
apoptosis in SV40-positive MM cell lines (123). Moreover, human mesothelial
cells are uniquely susceptible to SV40-mediated transformation and
immor¬talization in vitro (see below), and evidence of co-carcinogenicity
between asbestos and SV40 has been described (21). All this evidence strongly
implicates SV40 as a causative agent in MM.

In theory, SV40 may be implicated in the origin of more MMs than those in which
it is detected. Some investigators have proposed that SV40 may contribute to
cancer formation according to a “hit-and-run” mechanism (reviewed in ref. 124).
According to this model, SV40 may be required for the initial stages of tumor
formation. During prolifera¬tion of the tumor a fraction of the cell population
may acquire a number of mutations so that the functions of the SV40 oncoproteins
may become disposable. In such a scenario it is conceivable that the
SV40-containing cancer cells may be counterselected, since the SV40
onco-proteins are immunogenic, and because SV40 (in these settings) would
represent just a metabolic burden. Furthermore, SV40 DNA does not preferentially
integrate into the genome of human cancer cells (22), a situation that mirrors
that of Epstein-Barr virus (EBV)-mediated car-cinogenicity (125). In conclusion,
the hit-and-run model could poten¬tially take place in certain cases, and some
in vitro experimental data have been produced supporting it (126–130). Some in
vivo SV40-driven tumor models also support the hit-and-run mechanism.
Trans-genic mice expressing Tag under the control of inducible promoters display
proliferative disorders and tumors that are dependent on Tag, since suppression
of Tag expression causes reversion of the malignant phenotype. However, these
tumors become Tag-independent (or par¬tially Tag-independent) if the expression
of Tag is silenced after longer periods of time (48,131). In spite of these
experimental models, there is

insufficient knowledge on the extent of the hit-and-run process in human cancer,
or whether it occurs at all in natural conditions.

Interaction of SV40 with Human Mesothelial Cells in Vitro

In the past 10 years we have studied the molecular effects produced
by infection of primary human mesothelial (HM) cells with live SV40 virus. These
studies are critically important to understand the patho-genesis of MM, since
SV40 is present in MM as a complete virus and not as a nonreplicating,
molecularly engineered, transfected plasmid. We found that the SV40 pattern of
infection of HM cells is substantially different from the traditional
semipermissive SV40 infection of human cells. SV40-infected HM cells express the
SV40 early genes, replicate SV40 chromosomes, and synthesize capsid proteins and
complete viral particles. However, the majority of infected cells do not undergo
SV40-mediated cell lysis, because HM cells synthesize amounts of SV40
com¬patible with cell survival (21). Survival in the presence of potentially
oncogenic SV40 tumor antigens causes a very high rate of cell trans¬formation in
SV40-infected HM. A few weeks after the introduction of SV40 in the population,
three-dimensional foci of HM cells that have lost contact-inhibition arise. The
frequency of focus formation in HM cells is on average 0.2 ¥ 10-4 in different
primary cultures (21). Nearly 100% of these foci can be established in culture
as cell lines that display a completely transformed phenotype in vitro and are
immortal. These results are unique under several perspectives. No other cells
have been described to acquire a completely transformed phenotype with such high
frequency after SV40 infection (about 1 of 5000 infected HM cells). Moreover,
less than 5% of foci developing after SV40 transformation of cells of different
origin can be successfully established as cell lines in culture (reviewed in
ref. 132). Instead, about 90% of SV40-transformed HM cell lines appear immortal
from the start because they never develop a crisis. Recent data have provided a
rationale for this unprece¬dented result. We found that SV40 induces telomerase
activity in HM cells as early as 72 hours after infection, and this activity
increases with cell passage in vitro (79). This indicated that SV40-transformed
HM cells do not need additional mutational events to become immortal, since
telomerase activation is a process intimately connected with SV40 infection.
SV40-mediated induction of telomerase activity may be specific for HM cells,
because telomerase activity was not detected in SV40-infected primary
fibroblasts (79). Besides its role in cell immor¬talization, induction of
telomerase by SV40 may be a process connected with focus formation, since TERT
has been shown to possess trans¬forming activities (106). SV40 also specifically
induces HGF and pro¬motes HGF receptor (met) phosphorylation in HM cells, and
these cellular effects appear to be a consequence of Tag interaction with pRb
(78). SV40-dependent activation of the met pathway in HM cells has several
implications in the process of HM malignant transformation. Since the
proto-oncogene met is upstream from ras, chronic activation

of the met pathway may reproduce, at least in part, the consequences of a
constitutively active ras allele. Collectively, SV40 infection of HM cells
provides a human cell with telomerase activation, met, and IGF pathways
activation, and (obviously) with the expression of the SV40 oncoproteins. Taking
into account the minimal genetic requirements for oncogenic transformation of
human cells (109), SV40 appears to be a complete carcinogen for HM cells, with
the immune system acting as the ultimate fail-safe for MM development after SV40
infection of the human mesothelium. However, the in vitro data would not support
the latter interpretation, since (on average) 1 out of 5000 HM cells will give
rise to an immortal cell line after SV40 infection (21). This entails that met
and IGF pathways activation do not fully complement an oncogenic ras allele, and
that the event of focus formation in HM is determined by any mutational
occurrence that reproduces the cel¬lular outcomes of oncogenic ras expression.
Taking into account the extremely high frequency of malignant transformation of
HM cells after SV40 infection, the latter situation must be achieved fairly
fre¬quently. SV40 induction of the met signaling pathway in HM cells has
additional implications in the pathogenesis of MM. SV40-containing HM cells
produce increased amounts of HGF, implying that SV40 can operate also as a
landscaper factor in MM, and that SV40 may exert its oncogenic properties even
though being expressed only in a fraction of the tumor cells population.

The study of the interactions between SV40 and HM cells in vitro provides a
number of mechanistic explanations concerning typical phe-notypes characteristic
of SV40-positive MMs. In the latter tumors the suppressor gene RASSF1A is
commonly silenced through methylation of its promoter. The latter instance
appears directly mediated by SV40, because SV40 specifically promotes RASSF1A
promoter methylation during the process of SV40 HM cell transformation (80).
Moreover, in SV40-positive MM specimens the expression of the Notch-1 gene is
specifically induced, while SV40-negative MMs do not display upreg-ulation of
Notch-1 (81). As previously indicated, Notch genes are pro¬teins regulating
crucial cell fate decision during development and differentiation. Such
regulation is achieved through Notch-dependent control of cell proliferation and
survival (reviewed in ref. 93). Deregu¬lated or aberrant expression of Notch
proteins is currently being studied in a number of human tumors (133), and
appears to critically sustain the malignant phenotype (111). The characteristic
Notch-1 over-expression in SV40-positive MMs is also mirrored and
mechanistically explained in the SV40-HM in vitro system. SV40 induces Notch-1
expression early after HM infection. This induction is maintained through the
process of SV40-dependent transformation of HM (and in cell lines derived from
SV40-positive MMs), and Notch-1 activation is necessary for the growth of
SV40-containing transformed HMs (81).

As mentioned above, SV40 preferentially induces mesotheliomas in hamsters.
However, SV40 mutants unable to produce a functional tag only occasionally
induce mesotheliomas in these animals, while, in similar settings, the
development of tumors of other origin appear to

be only partially dependent on the function of tag (112). This in vivo evidence
indicates that tag functions must play critical roles in the hamster
mesothelium. The interaction between hamster mesothelial cells and SV40 has been
insufficiently studied so far. However, in human mesothelial cells tag is
required for telomerase induction (79), Notch-1 induction (81), and focus
formation in vitro (21). These data underscore the pivotal role of the SV40 tag
in the process of HM cell transformation in vitro.

Interactions of live virus with HM cells has also evidenced that gene dosage or
expression levels of the SV40 oncoproteins can produce effects that, ultimately,
are qualitatively different from transfection experiments. In fact, SV40 induces
the Notch-1 signaling pathway in HM cells only as a replication competent virus,
and not as a replication-defective mutant. This situation can be circumvented if
the SV40 oncoproteins are artificially expressed in HM cells under the control
of a strong promoter (such as the cytomegalovirus promoter).

Crossing the Species Boundaries: SV40, a Human
Virus?

SV40 shares with its closest homologues Jamestown Canyon virus (JCV)
and BKV the property of exerting powerful oncogenic activities when it is
introduced in other species; JCV and BKV are poly-omaviruses that commonly
infect humans (134,135). Their putative implication in certain human cancers has
been proposed (136–139). More often JCV and BKV produce asymptomatic infections
and latently persist in human tissues, but they can be reactivated in
immu-nocompromised individuals and consequently cause lethal diseases. Likewise,
SV40 establishes in monkeys a similar life cycle to that of polyomaviruses in
humans. However, JCV is highly oncogenic in owl and squirrel monkeys (140), BKV
causes different types of cancers in hamsters (141) and rats (142), and SV40 is
highly oncogenic in hamsters and other rodents, and it is associated with
different human cancers in which the experimental data available indicate that
SV40 plays a pathogenic role (reviewed in ref. 16). We are not aware of any
evidence that human polyomaviruses have propagated in other species under
natural conditions. On the other hand, SV40 has been massively introduced into
the human population during the early poliomyelitis vaccination program.
Furthermore, SV40 seems to be capable to infect humans under natural or
pseudonatural conditions. Up to 10% of individuals working in close contact with
monkeys develop SV40-specific antibodies (143), and different percentages of
laboratory personnel working with SV40, with monkeys, or with monkey cells
display immunologic evidences of previous SV40 infec¬tions (reviewed in ref.
22). In conclusion, there is evidence that SV40 has entered the human population
in different circumstances. The issue of possible human-to-human transmission of
SV40 infections has been poorly investigated to date. However, SV40-specific
antibodies are

detectable in children born many years after the administration of
SV40-contaminated poliovaccines, and the prevalence of SV40-specific antibodies
appears to increase with age (22). In some of the children possessing
SV40-specific antibodies, SV40 DNA has been detected in tissue specimens using
polymerase chain reaction (22). Furthermore, SV40 has been recently retrieved
from sewage waters of Calcutta (India) where monkeys are absent (144). These
findings strongly sug¬gest that SV40 is being propagated in the human
population, and that human-to-human transmission of SV40 is taking place at
least in some areas of the world. More studies are needed to assess the extent
of SV40 diffusion in human. However, the evidence available so far indicates
that SV40 has permanently crossed the barrier between monkeys and humans, and
that SV40 is now also a human virus.

Conclusion

The studies of SV40-mediated malignant transformation of cells of
dif¬ferent origin, including human cells, underscore the extraordinarily
powerful oncogenic potential of SV40. Many aspects of the SV40 natural
distribution, prevalence, and life cycle are still largely unknown. The
incidence of the human malignancies associated with SV40 is constantly rising,
especially mesothelioma. Malignant mesothe-lioma was virtually unknown before
the second half of the 20th century (145). The increase of mesothelioma cases
from practically zero to the current 2000 to 3000 new cases per year in the
United States alone has been attributed to the widespread use of asbestos during
the first half of the 20th century. However, we also know that between 1954 and
1963 about 32 million individuals were injected with various amounts of
infectious SV40 in the United States alone (145). Human mesothelial cells are
uniquely susceptible to SV40-mediated transformation (21). SV40 is specifically
linked to a substantial percentage of mesothelioma cases (16), and there is
evidence that asbestos and SV40 can act as co-carcinogens (21). All the above
facts indicate that SV40 is at least co-responsible for the rise of the
incidence of mesothelioma. The evidence indicating that SV40 is presently
propagating in the human population implies that human contact with this virus
is not a mere accident circumscribed in time, but that SV40 represents a hazard
for human health for some time to come. Therefore, there is an evident need for
a broad and detailed study of SV40-mediated carcinogenesis in order to prevent
and treat SV40-associated human diseases, including mesothelioma.

Acknowledgments

This work was supported by National Institutes of Health (NIH) grants
RO1 CA092657, Riviera Country Club CRFA to Michele Carbone, and by NIH R21
CA91122 to Maurizio Bocchetta.