Oncogenic Viruses

P Lambert, University of Wisconsin, Madison, WI, USA

ª 2009 Elsevier Inc. All rights reserved.

Introduction Human Tumor Viruses

General Principles of Virally Induced Cancers Conclusion
RNA Tumor Viruses Further Reading
DNA Tumor Viruses

Glossary provirus The double-stranded DNA replicative

apoptosis A form of cell death in which the cell is intermediate of a retroviral genome found integrated into
argued to actively commit suicide. Apoptosis is the host chromosome.
characterized by the nucleolytic digestion of the cellular transformation Morphological alteration or altered
chromatin at points between nucleosomal units, the growth properties of a cell in tissue culture that are thought
breakdown of the nucleus, and blebbing at the cell to reflect oncogenic properties; that is, contact-uninhibited
surface. growth, anchorage-independent growth, and
immortalization The capacity of a cell in tissue culture immortalization.
to grow ad infinitum. Normally, somatic cells have a tumor suppressor gene A gene whose loss of function
finite life span, at the end of which the cells senesce. is associated with the tumorigenic phenotype.
Immortalization requires that the cells override the Commonly, such genes when reintroduced into a tumor
mechanisms that normally cause senescence. cell inhibit its growth or tumorigenic phenotype.
oncogene A gene that encodes a factor that can confer
a transformed phenotype onto cells in tissue culture or a
tumorigenic phenotype in vivo.

Abbreviations HPV human papillomaviruse

ATL T cell leukemia/lymphoma HSV herpes simplex virus
BL Burkitt’s lymphoma HTLV-1 human T cell lymphotropic virus I
DLG Drosophila disk large IL-2 interleukin-2
EBV Epstein–Barr virus LMP1 latent membrane protein 1
HBV hepatitis B virus RSV Rous sarcoma virus
HCC hepatocellular carcinoma VLP virus-like particle
HNSCC head and neck squamous cell carcinomas WHV woodchuck hepatitis virus

Introduction insertionally mutagenic retroviruses) or the introduction

of viral genes that directly contribute to oncogenesis
Oncogenesis is believed to be a multifactorial process in (e.g., the human papillomaviruses (HPVs) that cause cervi-
which cells must acquire multiple aberrant properties that cal cancer). In some cases, it is not apparent that a
permit them to grow in an unregulated manner, to invade tumorigenic virus causes any genetic or epigenetic change
tissue, and to metastasize – that is, to establish residence in to a cell that directly contributes to oncogenesis. Here, it is
new locations within the organism. These properties are possible that the host response to the infection is the con-
thought to be conferred mostly through inheritable changes tributing cause of virally induced cancer (e.g., human
to a cell, which occur from both genetic and epigenetic hepatitis B virus (HBV)). In this article, we describe the
events. The contributions of tumorigenic viruses to cancer current understanding of how various viruses contribute to
mostly reflect genetic alterations to the cell. These altera- cancer. Emphasis is placed on describing the role of viruses
tions include the alteration of cellular genes (e.g., the in human cancers. We also describe the important insights

421
422 Viruses | Oncogenic Viruses

gained from the study of animal viruses, which have served RNA Tumor Viruses
historically as important models for the study of tumor
viruses. The field of tumor virology initially arose from the study of
retroviruses that cause tumors in birds and mammals other
than humans. The oncogenic retroviruses can be divided
into three classes: (1) transducing retroviruses, (2) insertion-
General Principles of Virally Induced ally mutagenic retroviruses, and (3) true oncoretroviruses.
Cancers Transducing retroviruses are ones that acquire genetic
information from the host genome. RSV is one example.
Infectious agents, including viruses, bacteria, and parasites, Most other examples also were isolated from avian hosts,
are thought to be the etiologic agents in approximately although others have been isolated from feline, murine, and
20% of human cancers. Human oncogenic viruses include simian hosts. The cellular genetic information being trans-
hepatitis B and hepatitis C viruses (associated with hepato- duced is a cellular proto-oncogene which, as present in
cellular carcinoma (HCC)), Epstein–Barr virus (EBV; and expressed from the transducing retrovirus, has onco-
associated with B cell lymphomas and nasopharyngeal and genic activity, that is, it contributes to tumor formation
gastric carcinomas), HPVs (associated with cervical carci- in vivo and cellular transformation in tissue culture. The
noma, other anogenital cancers, and a subset of head and study of transducing retroviruses led to the discovery of
neck cancers), human T cell lymphotropic virus I (HTLV- more than a dozen oncogenes, including src, ras, and myc. It
1; associated with adult T cell lymphomas), and human was not recognized until the 1970s that these viral onco-
herpesvirus type 8 (associated with Kaposi’s sarcoma and genes actually had been picked up by the virus from the
primary effusion lymphomas). The initial recognition that host organism through genetic recombination, which arises
viruses cause cancer arose from studies of animal viruses. as a by-product of reverse transcription. This led to the
In the early twentieth century, transmittable agents were recognition that animals encode oncogenes. The cellular
demonstrated to cause tumors in chickens and rabbits. The copies of these genes were initially referred to as proto-
respective agents were later identified to be Rous sarcoma oncogenes because it was recognized that the viral copies
virus (RSV), the first-studied RNA tumor virus, and Shope of these genes had undergone genetic changes that led to
papillomavirus, the first-studied DNA tumor virus. Much the increased expression and/or activity of the encoded
of our understanding of viral oncogenesis derived initially gene product. These changes led to the oncogenic nature
from the study of such animal viruses. In recent years, much of the virally transduced gene. In contrast, the cellular
attention has been focused on the study of human tumor homologues do not contribute to cancer in their wild-
viruses as etiological roles of such viruses in human cancers type state. For example, the virally encoded src and ras
have been elucidated. The study of virally induced cancers genes have amino acid substitutions at positions in these
has provided many basic insights into cancer, the most proteins that lead to their constitutive activity in signal
important of which is the identification and functional transduction pathways, and the virally encoded myc gene is
characterization of many oncogenes and tumor suppressor missing regions of the cellular gene that normally lead to a
genes. From these collective studies, several generalizable tight regulation of its expression at the transcriptional and
principles of virally induced cancers can be proffered. First, post-transcriptional levels. No examples of transducing
the onset of virally associated cancers is characterized by retroviruses have been found for human tumors; rather,
long latent periods following initial infection. This suggests most such viruses were identified from the evaluation of
that the viruses alone are not sufficient to cause cancer. laboratory stocks of retroviruses that infect laboratory ani-
Also, given this long latency, oncogenic viruses must be mal species. Generally, these transducing, oncogenic
able to persist in the host for long periods of time. Second, retroviruses are replication defective because of the dis-
the cancers induced by viruses do not constitute a natural ruption of one or more essential viral gene by the insertion
part of the viral life cycle. Rather, virally associated cancers of the cellular proto-oncogene. The rare exception is RSV,
appear to be dead-end streets for viruses – by-products of which is a replication-competent virus in which the trans-
the natural infection that provide no advantage to the virus duced cellular src gene did not disrupt the function of any
evidenced by the fact that it is uncommon for progeny viral gene. How a transducing oncogenic retrovirus arises
virus to be produced in the associated tumors. The third remains unclear. It likely results from a provirus that
principle is that viral properties that contribute to the integrates close to a cellular proto-oncogene. This could
induction of tumors often play critical roles in the life result in the formation of a chimeric RNA of viral and
cycle of the virus. In the following sections, we describe cellular origin, which, when packaged into progeny retro-
the basic properties of RNA and DNA tumor viruses, and virus and on reverse transcription, undergoes
the major insights gained regarding how they contribute to recombination with a copackaged, intact viral transcript
cancer. to produce the actual transducing viral genome.
 Viruses | Oncogenic Viruses 423

Insertionally mutagenic retroviruses are ones that, as a hamsters. Several viral gene products were found expressed
consequence of insertion of the proviral genome into the in these transformed/tumorigenic cells and were identified
host chromosomes, alter the structure and/or expression of because hamsters made antibodies against them. Most nota-
a cellular proto-oncogene or tumor suppressor gene. These ble of these was the SV40 large tumor antigen (T Ag). This
retroviruses are also known as the weakly oncogenic retro- protein is one of the most potent oncogenic factors known.
viruses since they only rarely cause cancers in laboratory Its oncogenic activity is multifactorial, largely arising from
animals, reflective of the low likelihood that the provirus its ability to bind and inactivate the cellular tumor suppres-
insertion occurs at or near a cellular proto-oncogene or sor proteins p53 and pRb. p53 is now recognized to be
tumor suppressor gene. There are several mechanisms by functionally disrupted in most human cancers. The protein
which insertion of provirus can lead to the activation of a was originally identified by virtue of its association with
cellular proto-oncogene. First, the virus may integrate SV40 T Ag.
upstream of the cellular gene in the same ‘sense’ orientation Initial cloning and functional analysis of p53 led to its
such that read-through transcription from the viral promo- designation not as a tumor suppressor gene but as an onco-
ter leads to the generation of chimeric virus/cell mRNAs gene. This was due to the fact that the initial p53 clones were
that can encode the cellular proto-oncogene. Second, the isolated from established cell lines. In these cell lines, the p53
viral transcriptional enhancer can upregulate the activity of gene had undergone mutations that cause the encoded gene
the cellular gene’s own transcriptional promoter. Finally, product to possess dominant-negative activity; that is, it
the virus can disrupt the integrity of the cellular mRNA, could inactivate the normal p53 protein when put back
leading to the loss of negative regulatory elements such as into cells. It was not until the 1980s that the wild-type p53
mRNA instability elements in the mRNA encoding the gene was recognized to be a tumor suppressor gene. p53 is
cellular proto-oncogene. Examples of each of these three thought to play an important role in orchestrating the cel-
cases exist in the literature. Many of the insertionally muta- lular responses to certain forms of stress, including insult
genic retroviruses cause lymphomas/leukemia in their from DNA-damaging agents. Activation of p53 leads either
natural host. In these tumors, the provirus is found nearby to growth arrest, which is believed to allow cells time to
to the c-myc locus, leading to increased expression of the repair damaged DNA, or to apoptosis (programmed cell
c-myc gene product. Insertionally mutagenic retroviruses death). Without functional p53 cells have a greater propen-
can also integrate within alleles of tumor suppressor genes sity for undergoing genetic alterations, some of which can
and disrupt their expression, thereby leading to cancer. The contribute to tumorigenesis. It is now understood that when
identification of such events is much less frequent than SV40 T Ag binds the wild-type p53 protein, it inactivates
those that lead to the activation of cellular proto-oncogenes, the latter’s function. SV40 T Ag also can bind the gene
presumably because of haplosufficiency. product of the retinoblastoma tumor susceptibility locus
True oncoretroviruses are ones that, through virally (Rb). Rb and the related ‘pocket’ proteins, p107 and p130,
encoded activities, alter the host cell or the host cell envir- are critical regulators of the cell cycle. In the G1 stage of the
onment, thereby making cells more susceptible to cancer cell cycle, pRb binds to and inactivates a family of transcrip-
development. There are but a few cases of true oncoretro- tion factors, E2Fs, that regulate the activity of cellular genes
viruses. One is HTLV-1, which causes T cell lymphomas involved in DNA synthesis and cell cycle control. Normally,
in humans. The role of HTLV-1 in human tumors is pRb’s activity is regulated by cyclin-associated kinases that
addressed in detail later. phosphorylate pRb, leading to its dissociation from E2Fs. T
Ag upon binding to pRb or its relatives p107 and p130
can also cause this dissociation and thereby lead to cell
DNA Tumor Viruses cycle deregulation. Adenoviruses cause respiratory diseases
in humans but are not associated with any human cancers.
Whereas the study of RNA tumor viruses was largely However, these viruses can efficiently transform cells in
responsible for the discovery of oncogenes, the study of tissue culture. The transforming activities of adenoviruses
DNA tumor viruses led to the discovery and elucidation were mapped to a region of the viral genome that encodes
of the function of cellular tumor suppressor genes, most sets of gene products, referred to as E1A and E1B, normally
notably p53 and pRb. Much of the initial knowledge about expressed early after infection. The E1A gene products were
DNA tumor viruses came from the study of SV40, a simian demonstrated to bind a set of cellular factors with different
virus, and human adenoviruses. Neither of these are molecular weights. The ability of E1A to transform cells
thought to cause tumors in their natural host; however, correlated with its ability to bind a subset of these associated
both can transform cells in tissue culture. Recently, SV40 factors. E1A was also known to activate transcription of viral
has been found in the human population and in certain promoters, including that for the viral E2 gene. E1A’s reg-
human cancers; however, an etiologic role of SV40 in ulation of the E2 promoter mapped to a promoter proximal
human cancer is not yet established. SV40 efficiently trans- DNA element that bound a family of cellular factor, appro-
forms cells in tissue culture and can cause tumors in priately labeled E2Fs. E1A’s regulation of the E2 promoter
424 Viruses | Oncogenic Viruses

through E2Fs correlated with its ability to bind a subset of often too late for clinical intervention. Unfortunately,
cellular factors including ones of molecular weight p105, the survival rate for individuals diagnosed with HCC is
p107, and p130. One of these factors was later recognized to low. The association of HCC and HBV was determined
be pRb (p105) and the others were demonstrated to be pRb- from worldwide epidemiological studies. In those areas of
like proteins. The E2F factor is the same as the aforemen- the world where a higher incidence of HCC was observed
tioned E2F that regulates the expression of cellular genes (e.g., sub-Saharan Africa and Southeast Asia), there was a
involved in DNA synthesis and cell cycle regulation. Thus, higher rate of HBV. Although HBV-induced HCC is
it was the analysis of E1A that led to the initial understand- relatively rare in the United States and Europe where
ing of the function of pRb as a regulator of transcription HBV infection rates are low, it is a leading cause of cancer
factors, the appreciation that multiple proteins (p107 and in those areas where HBV infection rates are high. This
p130) share functions with pRb, and the discovery of the etiological association of HBV with HCC was clearly
E2F family of transcription factors. The E1B region, which demonstrated in a study performed on men in Taiwan
also contributes to transformation by adenoviruses, encodes in which 191 of 194 of those studied who died of HCC
a factor (E1B55kd) that binds and inactivates p53, like tested positive for the HBsAg surface antigen of the virus.
SV40 T Ag. Another E1B gene product (E1B19kd) can Studies in animal systems have also demonstrated a cor-
directly inhibit apoptosis, a process that is induced by p53. relation between infection with HBV and the occurrence
Thus, both SV40 and adenoviruses have evolved the ability of HCC. For example, 100% of woodchucks infected
to regulate the pRb and p53 tumor suppressor genes. The with woodchuck hepatitis virus (WHV) develop HCC.
importance of these capacities in the context of the viral life There is a long latency period of at least 30 years between
cycle is not completely understood but is thought to relate to infection by HBV and the development of HCC in
the need for the virus to induce a DNA synthesis-competent humans. During this time, the virus integrates into
environment in the infected host cell, thereby allowing for the host genome and viral protein expression is ablated.
production of progeny virus. Inactivation of pRb is likely to Infection by HBV can lead to either acute or chronic
be critical for this process. However, why inactivate p53? hepatitis. Acute infection involves symptoms such as
Normally, it is not active in cells and only becomes active fever, fatigue, anorexia, and nausea, which are soon fol-
under conditions of stress. One theory is that the aberrant lowed by jaundice. In adults, the disease is usually cleared
induction of DNA synthesis by these viruses leads to the by an immune system response to viral surface antigen. In
induction of p53, which, unless overcome, would inhibit 1–8% of infected adults, however, infection persists and
viral DNA replication. Recently, multiple links in the path- the virus undergoes active replication. Chronic infection
ways that regulate pRb and p53 activity have been made, rates are much higher (approximately 90%) in the case
providing a potential understanding for why inactivation of of perinatal and early childhood infections. It is these
pRb might trigger activation of p53. The insights gained chronically or persistently infected patients who are at
from the study of SV40 and adenovirus have shed light on highest risk for developing HBV-associated HCC.
how papillomaviruses cause human cancers. Persistently infected individuals may be asymptomatic
for years, although the HBsAg surface antigen can be
detected during this time. In other patients, chronic hepa-
Human Tumor Viruses titis B occurs, with symptoms including fatigue, anxiety,
and anorexia, associated with hepatocellular necrosis and
In the following sections, we provide brief descriptions of inflammation. A fraction of those patients with chronic
four human tumor viruses, the evidence for their role in hepatitis will develop cirrhosis of the liver. Cirrhosis can
human cancers, and the current understanding of their then lead to HCC. How HBV causes HCC is a matter of
mechanism of action in carcinogenesis. debate, although there are two main pathways by which
HBV might be acting: directly through a viral protein or
integration or indirectly, as a consequence of chronic
HBV
liver damage due to viral infection and host immune
Infection by HBV is strongly associated with the devel- responses to it. HBV may directly induce HCC either in
opment of HCC, one of the ten most common carcinomas trans, by an activity of one of its gene products early on in
in the world. This oncogenic association categorizes HBV cancer development, or by a cis, as a result of viral inte-
as a tumor virus. HBV is a hepatotropic, circular, partially gration (i.e., akin to insertionally mutagenic retroviruses).
duplex DNA virus of 3.2 kb and replicates via an RNA For HBV to cause oncogenesis through a trans-acting
intermediate and reverse transcription. Infection by HBV mechanism, it must be actively undergoing transcription.
can lead to liver damage and such pathological diseases Since most advanced HCCs display no viral gene expres-
as chronic hepatic insufficiency and cirrhosis, leading to sion, this putative pathway must act primarily at the
HCC. HCC is a difficult cancer to treat because symp- initial stages of cancer cell growth while viral genes
toms occur late in the development of the carcinoma, are still being expressed. There are multiple viral gene
 Viruses | Oncogenic Viruses 425

products that have been argued to play a role in promot- fact that most diseases of the liver leading to liver damage,
ing uncontrolled cell growth. It is common to find HBV such as HCV infection, alcoholic cirrhosis, and Wilson’s
genomes integrated into the host chromosome in HCC. disease, are associated with an increased risk for the
These integrations commonly lead to the retention of a development of HCC. The Large S (L) protein product
portion of the viral genome, including the X and preS/S of HBV, when expressed, is toxic to cells and can lead to
regions. The X gene product of HBV has transcriptional liver damage. Transgenic mice engineered to overexpress
activation activity. The activity of X in the context of the the HBV L protein display hepatocellular necrosis and
in vivo viral life cycle is unclear, although X has been regeneration, which eventually leads to HCC. Although
shown to activate many promoters in reporter gene this result is potentially informative, it should be inter-
assays, including the herpes simplex virus (HSV) tk promo- preted carefully because the levels of L protein in these
ter, the HIV long terminal repeat (LTR), and the HBV mice are much higher than those in an HBV infection.
Enl. Cellular promoters activated by X include those HBV-induced liver damage alternatively could be the
for _-globulin, MHC class I, c-myc, and RNA polymerase result of host immune responses to viral gene products.
III. It is possible that X is activating cellular oncogenes, Although the association of HBV infection with the
leading to uncontrolled cell growth. X has also been development of HCC is high, the mechanism by which
shown to inhibit nucleotide excision repair, leading to the virus causes cancer clearly is still a matter of debate.
accumulated mutations. This could cause an increased It is possible that a combination of pathways described
mutation rate induced by chemical carcinogens such as above is contributing to the oncogenic activity of HBV.
aflatoxin B1. X has also been found to bind to the p53 Further clarification of this effect will pave the way for
tumor suppressor protein and may therefore interfere more effective treatments of HCC associated with HBV.
with the latter’s role in DNA damage response. Some Meanwhile, efforts are under way to eradicate HBV
mice transgenic for X driven by the X promoter develop infections through mass immunizations.
HCC. However, this effect is not seen using heterologous
promoters or subgenomic or genomic HBV DNAs con-
EBV
taining the X region. Further support that X plays an
active role in tumor development stems from the fact In the 1950s, while working as a missionary doctor in
that avian hepatitis viruses that lack X do not cause Africa, Denis Burkitt described the childhood tumor
HCC. The HBV truncated major S protein may also play Burkitt’s lymphoma (endemic BL) and postulated that an
a direct role in HCC formation for some of the same infectious agent could possibly act as the etiologic agent of
reasons as X does because it too possesses transcriptional BL. In 1961, a collaboration among Denis Burkitt, Tony
activation activity. The pre-S protein product of HBV has Epstein, and Yvonne Barr resulted in the establishment of
also been shown to activate transcription of TGF-_, an BL-derived cell lines in culture. Electron microscopy
effect which can lead to uncontrolled cell growth. examinations of the cell lines showed herpesvirus-like
HBV alternatively may lead to cancer via a cis-acting particles that are biologically and antigenically dissimilar
mechanism, whereby the integration of the virus into to other human herpesviruses. Subsequent seroepidemio-
the host genome can cause interruption of a tumor sup- logical studies indicated that the candidate human tumor
pressor gene or activation of an oncogene. An example of virus named EBV, recently renamed human herpesvirus 4,
insertional activation is observed in WHV-induced HCC. presents ubiquitously throughout the human populations
Using Southern blotting analysis, WHV has been found to in two different types and is persistently infectious. The
be integrated near to the N-myc loci in 40% of HCCs. strong association between the virus and BL, its ability to
Using the more broad-scanning technique of pulse-field transform primary human B lymphocytes into permanent
gel analysis, an even higher percentage of tumors have lymphoblastoid cells, and a predominance of a single gly-
been found to be integrated near N-myc. However, inser- coprotein in the viral outer surface further differentiate
tional activation by HBV has not been regularly EBV from other herpes members. All three subtypes of
demonstrated in human HCCs using either technique. BLs – endemic, sporadic, and AIDS-related – are thought
However, this does not mean that this phenomenon is to be associated with EBV infection. Early onset endemic
not occurring because current techniques allow detection BL, the most common childhood cancer in equatorial
only within a few kilobases of a gene of interest. Africa, occurs with an unusually high incidence of 5–10
Another major school of thought is that HBV can lead cases per 100 000 individuals per year. Virtually all of the
to HCC via an indirect pathway. Infection of the liver malignant cells of every analyzed endemic BL tumor are
by HBV can cause liver damage, which leads to a chronic EBV genome-positive. The sporadic BL, more commonly
proliferative response in the liver to replace lost cells. found in the United States and other Western countries,
The resulting increase in DNA synthesis increases the has a 50- to 100-fold lower incidence in comparison to the
probability of mutations occurring, leading to uncon- endemic BL with about 15–25% EBV genome-positive
trolled cell growth. This theory is supported by the tumors. Also, sporadic BL tumors have a slightly later age
426 Viruses | Oncogenic Viruses

of onset, a different pattern of presentation, and frequently an extracellular ligand could bind. Rather, LMP1’s signal
involve the abdomen and bone marrow instead of the jaw, transduction activity appears to be constitutively active.
as in the cases of the endemic BL. AIDS-related BL devel- In the past, LMP1 has been argued to cause the immorta-
ops in approximately 10% of AIDS patients in Western lization of B lymphocytes in part by inducing expression
societies and reflects both early and late onset B cell of the cellular antiapoptotic factor Bcl-2; however, it is
malignancies. However, even with such a high incidence unclear whether the increased levels of Bcl-2 found in
of AIDS-related BL tumors, only 30–40% of these are EBV some EBV-infected cells are a direct consequence of
genome-positive. Remarkably, all known forms of BL have LMP1 action or an acquired phenotype that correlates
the t(8:14), t(2:8), and t(8:22) chromosomal translocations with improved survival of those cells in tissue culture.
that are likely deleterious to B lymphocyte development. Another EBV-associated gene, EBNA2, has been shown
In addition to BL, EBV-associated malignancies also to be critical for EBV’s immortalization of B lymphocytes;
include immunoblastic B cell lymphomas in immunosup- however, its role is likely indirect through its ability to
pressed individuals, nasopharyngeal carcinoma, a subset of activate the transcription of LMP1.
gastric carcinoma, Hodgkin’s disease, and EBV-associated With the recent advances in understanding the role of
T cell lymphomas. EBNA1 in the persistence of the EBV genome in human
Acute infection of EBV causes mononucleosis. How cells and the role of LMP1 in EBV’s immortalization of B
EBV initially infects an individual is not clear, but it may lymphocytes, scientists now have two relevant targets for
do so through infection of epithelial cells lining the oral developing novel antiviral therapies that could reduce the
cavity. Usually, EBV infections become persistently incidence of EBV-associated malignancy in the future.
latent, in which the viruses induce B lymphocyte prolif-
eration concurrently with viral DNA replication. Stable
retention of EBV episomes is mandatory for the continual
HPVs
growth of EBV-positive BL cells. The viral protein
EBNA1 is critical for the maintenance of the EBV repli- Papillomaviruses are the causative agents of warts. A
con in infected cells and for the continued growth of BL subset of HPVs that are sexually transmitted cause warts
cells. EBNA1 contributes both to the replication of the of the anogenital tract. After decades of latency, a small
viral genome and to its efficient inheritance during cell fraction of these HPV-induced anogenital lesions pro-
division. gress to carcinomas. The cancers most highly associated
How EBV contributes to human cancer is not well with infection by these so-called high-risk HPVs are
understood. In tissue culture, EBV can efficiently induce cervical cancer. Interestingly, cervical cancer was pre-
the immortalization of B lymphocytes, a property that is viously thought to be casually associated with infection
likely to be related to its oncogenic properties in vivo. The by another sexually transmitted virus, HSV; however, this
EBV genome is a 172-kbp linear, double-stranded DNA. association was dismissed after it became clear that HSV
In infected cells, the EBV genome is maintained as a was not present in the cancers.
circular episome. Although the EBV genome can encode In contrast, more than 99% of all cervical cancers
many genes, the fact that in EBV-immortalized B cells contain papillomavirus DNA. The presence of HPV
only a handful of viral genes are expressed helped DNA in cervical cancers and the strong epidemiological
researchers identify the genes relevant for immortaliza- association of HPV infection with cervical cancer have
tion. One viral gene, latent membrane protein 1 (LMP1), led the National Cancer Institute to conclude that HPV is
has been demonstrated to be critical for EBV’s immorta- the main etiologic agent of cervical cancer in humans.
lization of B lymphocytes and in transgenic mice is More recently, a subset of head and neck squamous cell
sufficient to induce lymphomas. As its name implies, carcinomas (HNSCC) also have been associated with the
LMP1 is an integral membrane protein found in the same type of high-risk HPVs. About 20–30% of HNSCC,
cytoplasmic membrane of latently infected cells. LMP1 particularly those found at the oropharyngeal sites, test
has characteristics of a cell surface receptor; it displays a positive for HPV DNA, and epidemiological and mole-
rapid turnover that is associated with endocytosis and it cular characteristics of these cancers suggest that the
signals to NF-B through association with TRAFFs and mechanism of carcinogenesis is distinct from that in the
TRADDs. This signaling capacity of EBV appears to be HPV-negative HNSCC.
required for its capacity to induce B lymphocyte prolif- Papillomaviruses are nonenveloped icosohedral viruses
eration. In this regard, LMP1 is functionally analogous to that contain a double-stranded DNA genome of about
CD40, a cellular receptor normally expressed in B lym- 8000 bp that encodes early and late genes. These viruses
phocytes, which can induce B lymphocyte proliferation infect the poorly differentiated basal cells within the epi-
upon stimulation by its ligand. Unlike CD40, LMP1 is not dermis of the skin and express the early viral genes.
known to have an associated ligand; consistent with this, Progeny viruses are only produced when the infected
LMP1 lacks a prominent extracellular domain to which basal cell differentiates; this triggers amplification of the
 Viruses | Oncogenic Viruses 427

viral genome and expression of the viral late genes that cells. An exception is that when the CR1 and CR2 domains
encode the capsid proteins. of E7 are mutated in the context of the whole papilloma-
In the majority of cervical cancers, the viral DNA is virus genome, cells transfected with the genome still
found integrated into the cellular genome and no progeny undergo transformation. This suggests that E7 performs
virus is produced. The viral integration event almost always functions other than pRb inactivation. This is further sup-
leads to a disruption in the E1 and E2 open reading frames. ported by recent findings that E7 can bind to and inactivate
These open reading frames encode proteins involved in p21 and p27, inhibitors of cyclin-associated kinases that
viral transcription and replication and their disruption may play critical roles in the control of the cell cycle. In fact,
lead to derepression of viral transcriptional promoters as recent in vivo studies have confirmed that even though the
well as loss of viral episomal replication activity. The E6 inactivation of pRb by E7 is necessary and sufficient for
and E7 open reading frames, on the other hand, are consis- some E7-mediated phenotypes in mice (short-term skin
tently intact in cancers. Integration disrupts an mRNA phenotypes), longer term carcinogenic phenotypes both in
instability element, which allows increased expression of the cervix and in the head and neck of the mouse cannot be
E6 and E7. Integration is also reported in HPV-positive attributed merely to E7-mediated loss of pRb. Prime tar-
HNSCC; however, even in cases where the genome remains gets, other than pRb, that may be important for E7 to
extrachromosomal, E6 and E7 expression is reported. E6 and mediate its oncogenic activities include p21 (since the
E7 are directly implicated in oncogenesis. E7 can cooperate ability of E7 to bind p21 has been shown to be critical to
with E6 or the activated ras oncogene to transform fibro- its transforming abilities), and also the pRb-related proteins
blasts in tissue culture, and E6 and E7 alone or together can p107 and p130. The ability of the virus to inactivate pRb,
immortalize epithelial cells in tissue culture. The expression p107, and p130 has been shown to be critical for the life
of E6 and E7 was shown to be required for the continued cycle of the virus.
proliferation of HPV-positive cervical cancer cell lines; thus, The discovery of the ability of E6 to inactivate the
E6 and E7 are thought to be required for the onset as well as tumor suppressor, p53, was prompted by the oncogenic
for persistence of cervical cancers. When the expression of similarities between papillomaviruses and SV40 and ade-
E6 or E7 is directed to the skin of transgenic mice, the novirus. E6 was found to form a tertiary complex with p53
mice exhibit thickening of the skin and spontaneously and a cellular ubiquitin ligase, E6-AP. This leads to the
develop skin tumors. When treated persistently with exo- ubiquitination of p53 which is then targeted for degrada-
genous estrogen, mice transgenic for both E6 and E7 tion. The p53 tumor suppressor is involved in the G1 to S
develop cervical cancers (whereas nontransgenic animals phase transition of the cell cycle and is induced in response
do not). Furthermore, these mice are a lot more susceptible to cellular stresses such as DNA damage. Although the
to carcinogenesis by an oral carcinogen than nontransgenic importance of p53 in other cancers suggests that it is
mice. important in papillomavirus-associated cancers, inactiva-
Even though E6 and E7 are well accepted to be the tion of p53 cannot fully substitute for the activities of
main oncogenes for HPV, more recent evidence impli- E6. For example, E6 was found to abrogate differentiation
cates the E5 viral protein as a possible contributor to of skin epithelial cells in tissue culture. However, when
carcinogenesis, at least in the cancers where its expression the p53 in these cells was inactivated by overexpressing a
is retained. E5 is the main transforming oncogene of dominant-negative form of p53, the cells could still differ-
bovine papillomavirus; however, in HPV-induced carci- entiate. Also, p53-null mice differ from E6 transgenic
nogenesis its role had been largely unclear. A more recent animals in that they do not display thickened skin or
study has shown that a large fraction of cervical cancers develop skin tumors. The alpha-helix binding domain of
still express E5. Furthermore, transgenic mice expressing E6, through which E6 associates with E6AP, has been
E5 are more susceptible to skin carcinogenesis, similar to shown to be important for E6 oncogenic activities in
transgenic animals expressing E6 or E7 suggesting in vivo mice, suggesting that the inactivation of p53 contributes
carcinogenic roles for E5. to this process. However, conditional deletion of p53 in the
As is the case with SV40 and adenoviruses, papilloma- cervical tissues has revealed that the loss of p53 in these
viruses have the ability to inactivate tumor suppressors. E7 tissues is a lot more carcinogenic than the expression of E6,
binds to and induces the degradation of the tumor suppres- suggesting that E6 is relatively inefficient at disrupting p53
sor, pRb. The inactivation of pRb allows E2F to be in vivo.
constitutively active, which leads to increased proliferation E6 has been shown to bind a variety of protein part-
in E7-expressing cells. The regions of E7 required for its ners, including E6-BP, paxillin, E6-TP, bak, and the
inactivation of pRb have been mapped to conserved human homologue of Drosophila disk large (DLG) tumor
regions (CR)1 and 2, so named because of their similarity suppressor gene. In fact, the ability of E6 to associate with
to pRb-binding regions in the adenovirus oncoprotein E1A. proteins such as Dlg through its PDZ-binding domain was
The CR1 and CR2 domains of E7 are both necessary for shown to correlate with E6 ability to induce aberrant
E7’s induction of tumors in animals and transformation of DNA synthesis and contribute to carcinogenesis. The
428 Viruses | Oncogenic Viruses

binding partner responsible for those phenotypes has not changes. Therefore, the immortalized T cells can progress
been identified but candidates include Dlg, Scribble, and to the second stage of transformation, which has a typical
MUPP1, all of which have been reported to bind E6 property of IL-2-independent growth. Recent evidence
through the PDZ domain. suggests that HTLV-1 not only induces the proliferation
In addition, E6 is known to activate telomerase, the of host T cells but also protects the infected cells from
enzyme that adds DNA to the ends of telomeres. undergoing apoptosis.
Mutational analysis of E6 has revealed that these novel E6 The HTLV-1 provirus genome is 9032 bp long and
activities can be separated from E6’s ability to degrade p53. contains gag, pol, and env genes that encode the viral matrix
Thus, these aforementioned proteins may be involved in (capsid and nucleocapsid proteins), enzymes (reverse tran-
unknown E6 pathways. Ongoing research of E6 and E7 is scriptase, integrase, and protease), and envelope protein
aimed at further elucidating their roles in carcinogenesis. (surface glycoprotein and transmembrane protein), respec-
HPV infection is the most prevalent sexually transmitted tively. The HTLV-1 genome contains unique regulatory
disease. genes, the rex and tax genes (analogous to the HIV rev and tat
During the past half century, use of Pap smears have genes), at the 39 end of the genome, which are not common
reduced the incidence of frank cervical cancer through in other retroviruses. Rex is a 27-kDa nuclear phosphopro-
early detection of preinvasive lesions. Due to the papillo- tein that regulates viral RNA processing. Rex enhances the
maviruses’ intimate association with cervical cancer, expression of single-spliced mRNA and unspliced viral
much effort has been put forth to prevent these infections. genomic RNAs encoding the gag-pol and env gene products
In 2006, the FDA approved the first prophylactic vaccine but reduces the expression of double-spliced tax/rex mRNA.
aimed to protect against HPV infection. This is a quad- At the early stage of viral gene expression in the viral life
rivalent vaccine that protects against four genotypes of cycle, double-spliced mRNAs for Tax and Rex proteins are
HPVs (6 and 11, which are associated with 90% of genital mainly produced. Newly synthesized Tax protein transacti-
warts, and 16 and 18, which are associated with 70% of vates the transcription of HTLV-1. However, accumulated
cervical cancers). This vaccine uses highly purified virus- Rex protein enhances the accumulation of unspliced and
like particles (VLPs) to induce a protective immune single-spliced viral RNAs and suppresses the accumulation
response in vaccine recipients. VLPs are ordered struc- of tax/rex mRNA. Therefore, by the function of Tax and
tures that resemble the natural papillomavirus capsid. Rev regulatory proteins, viral expression is regulated both
They can be formed when the papillomavirus major cap- positively and negatively, and it helps the transient expres-
sid protein L1 is expressed at high concentrations. sion of HTLV-1. Transient expression of the viral genes
Clinical trials have shown this vaccine to be highly effec- may be one of the mechanisms to escape from immune
tive against infection by the virus and also wart and surveillance of the host. Tax protein is a 40-kDa nuclear
cancer formation. A similar vaccine developed by phosphoprotein and is considered as an oncogenic viral
GlaxoSmithKline is anticipated to become available soon. protein that transactivates the transcription of HTLV-1
and also binds to cellular transcription factors or other
cytoplasmic cellular molecules involved in the fundamental
HTLV-1
cell function such as IL-2, IL-2R, and c-fos and the para-
HTLV-1 is an oncogenic retrovirus associated with a vari- thyroid hormone-related peptide. HTLV-1 has no
ety of human diseases, including adult T cell leukemia/ preferential site of integration, which could explain why a
lymphoma (ATL), myelopathy, uveitis, and arthropathy. transformation process is initiated. The oncogenic property
The epidemiology of ATL suggests that cumulative of Tax protein is well documented in the development of
genetic defects may be responsible for the acute T cell ATL. Indirect evidence showed that the 39 end of the
malignancy in a given T cell clone. In HTLV-1 infection, HTLV-1 genome (tax gene) is necessary and sufficient for
the time-dependent emergence of infected T cell clones is cell immortalization. In addition, cotransfection of Tax with
well studied, and it has been shown that the uncontrolled Ras can also induce transformation of primary rat embryo
growth of a single clone causes ATL. In vitro, T cell fibroblasts. Finally, Tax transgenic mice develop several
immortalization by HTLV-1 occurs within a few months pathologies, including leukemia, mesenchymal tumor, and
of culture. HTLV-1 transforms human T cells in vivo and neurofibromas.
in cell culture. The development of T cell transformation
can be divided into two stages. In the early stage, the virus
induces interleukin-2 (IL-2)-dependent T cell prolifera- Conclusion
tion, which mimics the action of antigens. In the normal T
cells, antigen-stimulated T cells cease growth after a few It is clear that tumorigenic viruses contribute to cancer
weeks. However, HTLV-1-infected T cells show infinite through multiple distinct mechanisms. In cases in which
proliferation (i.e., immortalization). The uncontrolled cell viruses are thought to encode genes that directly contribute
proliferation is believed to facilitate secondary genetic to cancer, why is this so? In these cases (EBV, HPVs, and
 Viruses | Oncogenic Viruses 429

HTLV-1) it is not obvious that the cancers contribute contributes to carcinogenesis. These include the human
positively to the viral infections; rather, the cancers repre- homologue for the DLG tumor suppressor gene, a target
sent dead-end streets. For example, in HPV-induced for the adenovirus E4ORF1 and HPV E6 gene products,
cervical cancers, the virus is commonly integrated, with which is thought to alter -catenin-mediated signal trans-
the integration event causing disruption of the viral gen- duction through its interaction with the gene product of
ome, and the cells, being poorly differentiated, cannot the adenomatous polyposis coli gene, APC, a tumor sup-
support progeny virus production. Recent evidence sug- pressor important in colon cancer. Another example is
gests that papillomaviruses use their oncogenes during the XRCC1, a putative target for the HPV E6 oncogene,
normal viral life cycle within a wart to reprogram differ- which functions in DNA repair. It is reasonable to predict
entiating epithelial cells to support the late stages of the that the continued study of tumor viruses will shed new
viral life cycle in which the viral genome is amplified. This insights into the carcinogenic process.
occurs within a compartment of the epithelium in which
normally no DNA synthesis is thought to occur. The E7 See also: Antiviral Agents; Hepatitis Viruses;
oncogene is thought to trigger the infected, differentiating Herpesviruses; Polyomaviruses and Papillomaviruses;
epithelial cell to express cellular genes that are necessary to Retroviruses; Vaccines, Viral; Virus Infection
provide a DNA synthesis-competent environment.
Another example is provided by EBV in which LMP1
causes B cells to remain in a proliferating state, thereby Further Reading
providing a means by which the infected cell population is Bajaj BG, Murakami M, and Robertson ES (2007) Molecular biology of
expanded and the viral episomal genome is allowed to EBV in relationship to AIDS-associated oncogenesis. Cancer
persist within the organism. In the case of other viruses, Treatment and Research 133: 141–162.
Cesarman E and Mesri EA (2007) Kaposi sarcoma-associated
the carcinogenic event appears not to be due to the virus herpesvirus and other viruses in human lymphomagenesis. Current
encoding an oncogene but rather to the virus altering the Topics in Microbiology and Immunology 312: 263–287.
expression of cellular genes that contribute to cancer or Greene W, Kuhne K, Ye F, et al. (2007) Molecular biology of KSHV in
relation to AIDS-associated oncogenesis. Cancer Treatment and
protect cells from becoming tumorigenic. The most clear Research 133: 69–127.
examples of this are the insertionally mutagenic retro- Hammerschmidt W and Sugden B (2004) Epstein-Barr virus sustains
viruses in which the provirus integration event can cause Burkitt’s lymphomas and Hodgkin’s disease. Trends in Molecular
Medicine 10(7): 331–336.
increased expression of a cellular proto-oncogene. The Koike K (2007) Hepatitis C virus contributes to hepatocarcinogenesis by
most enigmatic tumor viruses are those in which no direct modulating metabolic and intracellular signaling pathways. Journal of
mechanism of tumorigenesis has been clearly implicated. Gastroenterology and Hepatology 22(supplement 1): S108–S111.
Longworth MS and Laimins LA (2004) Pathogenesis of human
The prime example is human HBV in which, while regions papillomaviruses in differentiating epithelia. Microbiology and
of the viral genome are often retained, the viral genes are Molecular Biology Reviews 68(2): 362–372.
not expressed commonly in the cancers and the viral Lupberger J and Hildt E (2007) Hepatitis B virus-induced oncogenesis.
World Journal of Gastroenterology 13(1): 74–81.
genome is not thought to cause deregulation of any cellular Münger K (2002) The role of human papillomaviruses in human cancers.
gene through the integration events. In this case, it may Frontiers in Bioscience 7: d641–d649.
well be the host’s own immune response to the infection Rezk SA and Weiss LM (2007) Epstein-Barr virus-associated
lymphoproliferative disorders. Human Pathology 38(9): 1293–1304.
that causes the virally induced cancers to occur. Schiffman M, Castle PE, Jeronima J, Rodriguez AC, and Wacholder S
The study of tumor viruses during the past century has (2007) Human papillomavirus and cervical cancer. Lancet
provided many insights about cancer. Their study led to 370(9590): 890–907.
Shah KV (2007) SV40 and human cancer: A review of recent data.
the discovery of oncogenes and tumor suppressor genes. International Journal of Cancer 120(2): 215–223.
Current studies continue to shed new insight into the Stanley M (2007) Prophylactic HPV vaccines. Journal of Clinical
process of carcinogenesis. For example, recent studies Pathology 60(9): 961–965.
Yasunaga J and Matsuoka M (2007) Leukaemogenic mechanism of
on adenoviruses and HPV have led to the identification human T-cell leukaemia virus type I. Reviews in Medical Virology
of novel cellular targets, the deregulation of which likely 17(5): 301–311.