Chapter 43: Human Cancer Viruses

Viruses are etiologic factors in the development of several types of human tumors, including two of great significance worldwide—cervical cancer and liver cancer. At least 15–20% of all human tumors worldwide have a viral cause. The viruses that have been strongly associated with human cancers are listed in Table 43-1. They include human papillomaviruses (HPVs), Epstein-Barr virus (EBV), human herpesvirus 8, hepatitis B virus, hepatitis C virus, and two human retroviruses plus several candidate human cancer viruses. New cancer-associated viruses are being discovered by the use of molecular techniques. Many viruses can cause tumors in animals, either as a consequence of natural infection or after experimental inoculation.

Animal viruses are studied to learn how a limited amount of genetic information (one or a few viral genes) can profoundly alter the growth behavior of cells, ultimately converting a normal cell into a neoplastic one. Such studies reveal insights into growth regulation in normal cells. Tumor viruses are agents that can produce tumors when they infect appropriate animals. Many studies are done using cultured animal cells rather than intact animals, because it is possible to analyze events at cellular and subcellular levels. In such cultured cells, tumor viruses can cause “transformation.” However, animal studies are essential to studying many steps in carcinogenesis, including complex interactions between virus and host and host responses to tumor formation.

Studies with RNA tumor viruses revealed the involvement of cellular oncogenes in neoplasia; DNA tumor viruses established a role for cellular tumor suppressor genes. These discoveries revolutionized cancer biology and provided the conceptual framework for the molecular basis of carcinogenesis.

Tenets of viral carcinogenesis are summarized in Table 43-2.

Tumor Viruses Are of Different Types

Like other viruses, tumor viruses are classified among different virus families according to the nucleic acid of their genome and the biophysical characteristics of their virions. Most recognized tumor viruses either have a DNA genome or generate a DNA provirus after infection of cells (hepatitis C virus is an exception).

DNA tumor viruses are classified among the papilloma-, polyoma-, adeno-, herpes-, hepadna-, and poxvirus groups. DNA tumor viruses encode viral oncoproteins that are important for viral replication but also affect cellular growth control pathways.

Most RNA tumor viruses belong to the retrovirus family. Retroviruses carry an RNA-directed polymerase (reverse transcriptase) that constructs a DNA copy of the RNA genome of the virus. The DNA copy (provirus) becomes integrated into the DNA of the infected host cell, and it is from this integrated DNA copy that all proteins of the virus are translated.

RNA tumor viruses are of two general types with respect to tumor induction. The highly oncogenic (direct-transforming) viruses carry an oncogene of cellular origin. The weakly oncogenic (slowly transforming) viruses do not contain an oncogene and induce leukemias after long incubation periods by indirect mechanisms. The two known cancer-causing retroviruses in humans act indirectly. Hepatitis C virus, a flavivirus, does not generate a provirus and appears to induce cancer indirectly.

Multistep Carcinogenesis

Carcinogenesis is a multistep process; that is, multiple genetic changes must occur to convert a normal cell into a malignant one. Intermediate stages have been identified and designated by terms such as “immortalized,” “hyperplastic,” and “preneoplastic.” Tumors usually develop slowly over a long period of time. The natural history of human and animal cancers suggests a multistep process of cellular evolution, probably involving cellular genetic instability and repeated selection of rare cells with some selective growth advantage. The number of mutations underlying this process is estimated to range from five to eight. Observations suggest that activation of multiple cellular oncogenes and inactivation of tumor suppressor genes are involved in the evolution of tumors whether or not a virus is involved.

It appears that a tumor virus usually acts as a cofactor, providing only some of the steps required to generate malignant cells. Viruses are necessary—but not sufficient—for development of tumors with a viral etiology. Viruses often act as initiators of the neoplastic process and may do so by different mechanisms.

Cellular Oncogenes

“Oncogene” is the general term given to genes that are involved in cancer causation. Normal versions of these transforming genes are present in normal cells and have been designated proto-oncogenes.

The discovery of cellular oncogenes came from studies with acutely transforming retroviruses. It was found that normal cells contained highly related (but not identical) copies of various retrovirus transforming genes; cellular sequences had been captured and incorporated into the retrovirus genomes. Transduction of the cellular genes was probably an accident, as the presence of the cellular sequences is of no benefit to the viruses. Many other known cellular oncogenes that have not been segregated into retrovirus vectors have been detected using molecular methods.

Cellular oncogenes are partly responsible for the molecular basis of human cancer. They represent individual components of complicated pathways responsible for regulating cell proliferation, division, and differentiation and for maintaining the integrity of the genome. Incorrect expression of any component might interrupt that regulation, resulting in uncontrolled growth of cells (cancer). Examples exist of tyrosine-specific protein kinases (eg, src), growth factors (sis is similar to human platelet-derived growth factor, a potent mitogen for cells of connective tissue origin), mutated growth factor receptors (erb-B is a truncated epidermal growth factor receptor), GTP-binding proteins (Ha-ras), and nuclear transcription factors (myc, jun).

The molecular mechanisms responsible for activating a benign proto-oncogene and converting it into a cancer gene vary—but all involve genetic damage. The gene may be overexpressed, and a dosage effect of the overproduced oncogene product may be important in cellular growth changes. These mechanisms might result in constitutive activity (loss of normal regulation), so that the gene is expressed at the wrong time during the cell cycle or in inappropriate tissue types. Mutations might alter the carefully regulated interaction of a proto-oncogene protein with other proteins or nucleic acids. Insertion of a retroviral promoter adjacent to a cellular oncogene may result in enhanced expression of that gene (ie, “promoter-insertion oncogenesis”). Expression of a cellular gene also may be increased through the action of nearby viral “enhancer” sequences.

Tumor Suppressor Genes

A second class of human cancer genes is involved in tumor development. These are the negative regulators of cell growth, tumor suppressor genes. They were identified because they form complexes with oncoproteins of certain DNA tumor viruses. The inactivation or functional loss of both alleles of such a gene is required for tumor formation—in contrast to the activation that occurs with cellular oncogenes. The prototype of this inhibitory class of genes is the retinoblastoma (Rb) gene. The Rb protein inhibits entry of cells into S phase by binding to key transcription factors that regulate expression of S-phase genes. The function of normal Rb protein is regulated by phosphorylation. The loss of Rb gene function is causally related to the development of retinoblastoma—a rare ocular tumor of children—and other human tumors.

Another crucial tumor suppressor gene is the p53 gene. It also blocks cell cycle progression; p53 acts as a transcription factor and regulates the synthesis of a protein that inhibits the function of certain cell cycle kinases. It also causes cells with DNA damage to undergo apoptosis. The loss of p53 function allows cells with damaged DNA to progress through the cell cycle, leading to the eventual accumulation of genetic mutations. The p53 gene is mutated in over half of all human cancers.

Persistent Infections

The pathogenesis of a viral infection and the response of the host are integral to understanding how cancer might arise from that background. The known tumor viruses establish long-term persistent infections in humans. Because of differences in individual genetic susceptibilities and host immune responses, levels of virus replication and tissue tropisms may vary among persons. Even though very few cells in the host may be infected at any given time, the chronicity of infection presents the long-term opportunity for a rare event to occur that allows survival of a cell with growth control mechanisms that are virus-modified.

Host Immune Responses

Viruses that establish persistent infections must avoid detection and recognition by the immune system that would eliminate the infection. Different viral evasion strategies have been identified, including restricted expression of viral genes that makes infected cells nearly invisible to the host (EBV in B cells), infection of sites relatively inaccessible to immune responses (HPV in the epidermis), mutation of viral antigens that allows escape from antibody and T-cell recognition (human immunodeficiency virus [HIV]), modulation of host major histocompatibility complex class I molecules in infected cells (adenovirus, cytomegalovirus), inhibition of antigen processing (EBV), and infection and suppression of essential immune cells (HIV).

It is believed that host immune surveillance mechanisms usually eliminate the rare neoplastic cells that may arise in normal individuals infected with cancer viruses. However, if the host is immunosuppressed, cancer cells are more likely to proliferate and escape host immune control. Immunosuppressed organ transplant recipients and HIV-infected individuals are at increased risk of EBV-associated lymphomas and of HPV-related diseases. It is possible that variations in individual immune responses may contribute to susceptibility to virus-induced tumors in normal hosts.

Mechanisms of Action by Human Cancer Viruses

Tumor viruses mediate changes in cell behavior by means of a limited amount of genetic information. There are two general patterns by which this is accomplished: The tumor virus introduces a new “transforming gene” into the cell (direct-acting), or the virus alters the expression of a preexisting cellular gene or genes (indirect-acting). In either case, the cell loses control of normal regulation of growth processes. DNA repair pathways are frequently affected, leading to genetic instability and a mutagenic phenotype.

Viruses usually do not behave as complete carcinogens. In addition to changes mediated by viral functions, other alterations are necessary to disable the multiple regulatory pathways and checkpoints in normal cells to allow a cell to become completely transformed. There is no single mode of transformation underlying viral carcinogenesis. At the molecular level, oncogenic mechanisms by human tumor viruses are very diverse.

Cellular transformation may be defined as a stable, heritable change in the growth control of cells in culture. No set of characteristics invariably distinguishes transformed cells from their normal counterparts. In practice, transformation is recognized by the cells’ acquisition of some growth property not exhibited by the parental cell type. Transformation to a malignant phenotype is recognized by tumor formation when transformed cells are injected into appropriate test animals.

Indirect-acting tumor viruses are not able to transform cells in culture.

Cell Susceptibility to Viral Infections and Transformation

At the cellular level, host cells are either permissive or nonpermissive for replication of a given virus. Permissive cells support viral growth and production of progeny virus; nonpermissive cells do not. Especially with the DNA viruses, permissive cells are often killed by virus replication and are not transformed unless the viral replicative cycle that results in death of the host cell is blocked in some way; nonpermissive cells may be transformed. However, there are situations in which DNA virus replication does not lyse the host cell and such cells may be transformed. Nevertheless, transformation is a rare event. A characteristic property of RNA tumor viruses is that they are not lethal for the cells in which they replicate. Cells that are permissive for one virus may be nonpermissive for another.

Not all cells from the natural host species are susceptible to viral replication or transformation or both. Most tumor viruses exhibit marked tissue specificity, a property that probably reflects the variable presence of surface receptors for the virus, the ability of the virus to cause disseminated versus local infections, or intracellular factors necessary for viral gene expression.

Some viruses are associated with a single tumor type, whereas others are linked to multiple tumor types. These differences reflect the tissue tropisms of the viruses.

Retention of Tumor Virus Nucleic Acid in a Host Cell

The stable genetic change from a normal to a neoplastic cell generally requires the retention of viral genes in the cell. Often, but not always, this is accomplished by the integration of certain viral genes into the host cell genome. With DNA tumor viruses, a portion of the DNA of the viral genome may become integrated into the host cell chromosome. Sometimes, episomal copies of the viral genome are maintained in tumor cells. With retroviruses, the proviral DNA copy of the viral RNA is integrated in the host cell DNA. Genome RNA copies of hepatitis C virus that are not integrated are maintained in tumor cells.

In some viral systems, virus-transformed cells may release growth factors that affect the phenotype of neighboring uninfected cells, thereby contributing to tumor formation. It is also possible that as tumor cells collect genetic mutations during tumor growth, the need for the viral genes that drove tumor initiation may become unnecessary and viral markers will be lost from some cells.

HEPATITIS B VIRUS

Hepatitis B virus (see Chapter 35), a member of the Hepadnaviridae family, is characterized by 42-nm spherical virions with a circular genome of double-stranded DNA (3.2 kbp). One strand of the DNA is incomplete and variable in length. Studies of the virus are hampered because it has not been grown in cell culture.

In addition to causing hepatitis, hepatitis B virus is a risk factor in the development of liver cancer in humans. Epidemiologic and laboratory studies have proved persistent infection with hepatitis B virus to be an important cause of chronic liver disease and the development of hepatocellular carcinoma. Hepatitis B virus infections occurring in adults are usually resolved, but primary infections in neonates and young children tend to become chronic in up to 90% of cases. It is these persistent hepatitis B virus infections established early in life that carry the highest risk of hepatocellular carcinoma later in life. The mechanism of oncogenesis remains obscure. Persistent viral infection leads to necrosis, inflammation, and liver regeneration that, over time, results in cirrhosis; hepatocellular carcinoma usually arises out of this background. The hepatitis B virus transactivator protein, X protein, is a potential viral oncoprotein. A dietary carcinogen, aflatoxin, may be a cofactor for hepatocellular carcinoma, especially in Africa and China.

The advent of an effective hepatitis B vaccine for the prevention of primary infection raised the possibility of prevention of hepatocellular carcinoma, particularly in areas of the world where infection with hepatitis B virus is hyperendemic (eg, Africa, China, Southeast Asia). Twenty years after the initiation of a universal hepatitis B vaccination program in Taiwan, chronic hepatitis B virus infection rates and liver cancer incidence rates were markedly reduced.

Woodchucks are an excellent model for hepatitis B virus infections of humans. A similar virus, woodchuck hepatitis virus, establishes chronic infections in both newborn and adult woodchucks, many of which develop hepatocellular carcinomas within a 3-year period.

HEPATITIS C VIRUS

Hepatitis C virus (see Chapter 35), a member of the Flaviviridae family, contains a genome of single-stranded RNA 9.4 kb in size. It appears that the majority of infections become persistent, even in adults. Chronic infection with hepatitis C virus leads to chronic inflammation and cirrhosis, and is also considered to be a causative factor in hepatocellular carcinoma. The development of hepatocellular carcinoma is likely mediated by a combination of virus and host specific mechanisms. There are over 170 million chronic carriers of hepatitis C virus, with 1-5% of these going on to develop hepatocellular carcinoma.

There are currently over 350 million people worldwide persistently infected with hepatitis B virus leading to over 500,000 deaths from hepatocellular cancer annually.

RETROVIRUSES

Retroviruses contain an RNA genome and an RNA-directed DNA polymerase (reverse transcriptase). RNA tumor viruses in this family mainly cause tumors of the reticuloendothelial and hematopoietic systems (leukemias, lymphomas) or of connective tissue (sarcomas).

Important properties of the retroviruses are listed in Table 43-3.

Structure and Composition

The retrovirus genome consists of two identical subunits of single-stranded, positive-sense RNA, each 7–11 kb in size. The reverse transcriptase contained in virus particles is essential for viral replication.

Retrovirus particles contain the helical ribonucleoprotein within an icosahedral capsid that is surrounded by an outer membrane (envelope) containing glycoprotein and lipid. Type- or subgroup-specific antigens are associated with the glycoproteins in the viral envelope, which are encoded by the env gene; group-specific antigens are associated with the virion core, which are encoded by the gag gene.

Three morphologic classes of extracellular retrovirus particles—as well as an intracellular form—are recognized, based on electron microscopy. They reflect slightly different processes of morphogenesis by different retroviruses. Examples of each are shown in Figure 43-1.

Type A particles occur only intracellularly and appear to be noninfectious. Intracytoplasmic type A particles, 75 nm in diameter, are precursors of extracellular type B viruses, whereas intracisternal type A particles, 60–90 nm in diameter, are unknown entities. Type B viruses are 100–130 nm in diameter and contain an eccentric nucleoid. The prototype of this group is the mouse mammary tumor virus, which occurs in “high mammary cancer” strains of inbred mice and is found in particularly large amounts in lactating mammary tissue and milk. It is readily transferred to suckling mice, in whom the incidence of subsequent development of adenocarcinoma of the breast is high. The type C viruses represent the largest group of retroviruses. The particles are 90–110 nm in diameter, and the electron-dense nucleoids are centrally located. The type C viruses may exist as exogenous or endogenous entities (see as follows). The lentiviruses are also type C viruses. Finally, the type D retroviruses are poorly characterized. The particles are 100–120 nm in diameter, contain an eccentric nucleoid, and exhibit surface spikes shorter than those on type B particles.

Classification

A. Genera

The Retroviridae family is divided into seven genera: Alpharetrovirus (which contains avian leukosis and sarcoma viruses), Betaretrovirus (mouse mammary tumor virus), Gammaretrovirus (mammalian leukemia and sarcoma viruses), Deltaretrovirus (human T-lymphotropic viruses [HTLV] and bovine leukemia virus), Epsilonretrovirus (fish viruses), Spumavirus (which contains viruses able to cause “foamy” degeneration of inoculated cells but which are not associated with any known disease process), and Lentivirus (which encompasses agents able to cause chronic infections with slowly progressive neurologic impairment, including HIV; see Chapter 44).

Retroviruses can be organized in various ways depending on their morphologic, biologic, and genetic properties. Differences in genome sequences and natural host range are frequently used, but antigenic properties are not. Retroviruses may be grouped morphologically (types B, C, and D); the vast majority of isolates display type C characteristics.

B. Host of Origin

Retroviruses have been isolated from virtually all vertebrate species. Natural infections by a given virus are usually limited to a single species, though infections across species barriers may occur. Group-specific antigenic determinants on the major internal (core) protein are shared by viruses from the same host species. All mammalian viruses are more closely related to one another than to those from avian species.

The RNA tumor viruses most widely studied experimentally are the sarcoma viruses of chickens and mice and the leukemia viruses of mice, cats, chickens, and humans.

C. Exogenous or Endogenous

Exogenous retroviruses are spread horizontally and behave as typical infectious agents. They initiate infection and transformation only after contact. In contrast to endogenous viruses, which are found in all cells of all individuals of a given species, gene sequences of exogenous viruses are found only in infected cells. The pathogenic retroviruses all appear to be exogenous viruses.

Retroviruses may also be transmitted vertically through the germ line. Viral genetic information that is a constant part of the genetic constitution of an organism is designated as “endogenous.” An integrated retroviral provirus behaves like a cluster of cellular genes and is subject to regulatory control by the cell. This cellular control usually results in partial or complete repression of viral gene expression. Its location in the cellular genome and the presence of appropriate cellular transcription factors determine to a great extent whether (and when) viral expression will be activated. It is not uncommon for normal cells to maintain the endogenous viral infection in a quiescent form for extended periods of time.

Many vertebrates, including humans, possess multiple copies of endogenous RNA viral sequences. The endogenous viral sequences are of no apparent benefit to the animal. However, endogenous proviruses of mammary tumor virus carried by inbred strains of mice express superantigen activities that influence the T-cell repertoires of the animals.

Endogenous viruses are usually not pathogenic for their host animals. They do not produce any disease and cannot transform cells in culture. (There are examples of disease caused by replication of endogenous viruses in inbred strains of mice.)

Important features of endogenous viruses are as follows: (1) DNA copies of RNA tumor virus genomes are covalently linked to cellular DNA and are present in all somatic and germ cells in the host; (2) endogenous viral genomes are transmitted genetically from parent to offspring; (3) the integrated state subjects the endogenous viral genomes to host genetic control; and (4) the endogenous virus may be induced to replicate either spontaneously or by treatment with extrinsic (chemical) factors.

D. Host Range

The presence or absence of an appropriate cell surface receptor is a major determinant of the host range of a retrovirus. Infection is initiated by an interaction between the viral envelope glycoprotein and a cell surface receptor. Ecotropic viruses infect and replicate only in cells from animals of the original host species. Amphotropic viruses exhibit a broad host range (able to infect cells not only of the natural host but of heterologous species as well) because they recognize a receptor that is widely distributed. Xenotropic viruses can replicate in some heterologous (foreign) cells but not in cultured cells from the natural host. Many endogenous viruses have xenotropic host ranges.

E. Genetic Content

Retroviruses have a simple genetic content, but there is some variation in the number and type of genes contained. The genetic makeup of a virus influences its biologic properties. Genomic structure is a useful way of categorizing RNA tumor viruses (Figure 43-2).

The standard leukemia viruses (Alpharetrovirus and Gammaretrovirus) contain genes required for viral replication: gag, which encodes the core proteins (group-specific antigens); pro, which encodes a protease enzyme; pol, which encodes the reverse transcriptase enzyme (polymerase); and env, which encodes the glycoproteins that form projections on the envelope of the particle. The gene order in all retroviruses is 5′-gag-pro-pol-env-3′.

Some viruses, exemplified by the human retroviruses (Deltaretrovirus and Lentivirus), contain additional genes downstream from the env gene. One is a transactivating regulatory gene (tax or tat) that encodes a nonstructural protein that alters the transcription or translational efficiency of other viral genes. The lentiviruses, including HIV, have a more complex genome and contain several additional accessory genes (see Chapter 44).

Retroviruses with either of these two genomic structures will be replication-competent (in appropriate cells). Because they lack a transforming (onc) gene, they cannot transform cells in tissue culture. However, they may have the ability to transform precursor cells in blood-forming tissues in vivo.

The directly transforming retroviruses carry an onc gene. The transforming genes carried by various RNA tumor viruses represent cellular genes that have been appropriated by those viruses at some time in the distant past and incorporated into their genomes (see Figure 43-2).

Such viruses are highly oncogenic in appropriate host animals and can transform cells in culture. With very few exceptions, the addition of the cellular DNA results in the loss of portions of the viral genome. Consequently, the sarcoma viruses usually are replication-defective; progeny virus is produced only in the presence of helper viruses. The helper viruses are generally other retroviruses (leukemia viruses), which may recombine in various ways with the defective viruses. These defective transforming retroviruses have been the source of many of the recognized cellular oncogenes.

F. Oncogenic Potential

The retroviruses that contain oncogenes are highly oncogenic. They are sometimes referred to as “acute transforming” agents because they induce tumors in vivo after very short latent periods and rapidly induce morphologic transformation of cells in vitro. The viruses that do not carry an oncogene have a much lower oncogenic potential. Disease (usually of blood cells) appears after a long latent period (ie, “slow-transforming”); cultured cells are not transformed.

Briefly, neoplastic transformation by retroviruses is the result of a cellular gene that is normally expressed at low, carefully regulated levels becoming activated and expressed constitutively. In the case of the acute transforming viruses, a cellular gene has been inserted by recombination into the viral genome and is expressed as a viral gene under the control of the viral promoter. In the case of the slow-transforming leukemia viruses, the viral promoter or enhancer element is inserted adjacent to or near the cellular gene in the cellular chromosome.

Replication of Retroviruses

A schematic outline of a typical retrovirus replication cycle, represented by HTLV, is shown in Figure 43-3. The pol gene encodes the unique polymerase (reverse transcriptase) protein that has four enzymatic activities (protease, polymerase, RNase H, and integrase). After virus particles have adsorbed to and penetrated host cells, the viral RNA serves as the template for the synthesis of viral DNA through the action of the viral enzyme reverse transcriptase, functioning as an RNA-dependent DNA polymerase. By a complex process, sequences from both ends of the viral RNA become duplicated, forming the long terminal repeat located at each end of the viral DNA (Figure 43-4). Long terminal repeats are present only in viral DNA. The newly formed viral DNA becomes integrated into the host cell DNA as a provirus. The structure of the provirus is constant, but its integration into host cell genomes can occur at different sites. The very precise orientation of the provirus after integration is achieved by specific sequences at the ends of both long terminal repeats.

Progeny viral genomes may then be transcribed from the provirus DNA into viral RNA. The U3 sequence in the long terminal repeat contains both a promoter and an enhancer. The enhancer may help confer tissue specificity on viral expression. The proviral DNA is transcribed by the host enzyme, RNA polymerase II. Full-length transcripts (capped, polyadenylated) serve as genomic RNA for encapsidation in progeny virions. Some transcripts are spliced, and the subgenomic messenger RNAs (mRNAs) are translated to produce viral precursor proteins that are modified and cleaved to form the final protein products.

If the virus happens to contain a transforming gene, the oncogene plays no role in replication. This is in marked contrast to the DNA tumor viruses, in which the transforming genes are also essential viral replication genes.

Virus particles assemble and emerge from infected host cells by budding from plasma membranes. The viral protease then cleaves the Gag and Pol proteins from the precursor polyprotein, producing a mature infectious virion prepared for reverse transcription when the next cell is infected.

A salient feature of retroviruses is that they are not cytolytic; that is, they do not kill the cells in which they replicate. The exceptions are the lentiviruses, which may be cytolytic (see Chapter 44). The provirus remains integrated within the cellular DNA for the life of the cell. There is no known way to cure a cell of a chronic retrovirus infection.

Human Retroviruses

A. Human T-Lymphotropic Viruses

Only a few retroviruses are linked to human tumors. The human T-lymphotropic virus (HTLV) group of retroviruses has probably existed in humans for thousands of years. HTLV-1 has been established as the causative agent of adult T-cell leukemia-lymphomas (ATL) as well as a nervous system degenerative disorder called tropical spastic paraparesis. It does not carry an oncogene. Three related human viruses, HTLV-2, HTLV-3, and HTLV-4, have been isolated but have not been conclusively associated with a specific disease. HTLV-1 and HTLV-2 share about 65% sequence homology and display significant serologic cross-reactivity.

The human lymphotropic viruses have a marked affinity for mature T cells. HTLV-1 is expressed at very low levels in infected individuals. It appears that the viral promoter-enhancer sequences in the long terminal repeat may be responsive to signals associated with the activation and proliferation of T cells. If so, the replication of the viruses may be linked to the replication of the host cells—a strategy that would ensure efficient propagation of the virus.

The human retroviruses are transregulating (see Figure 43-2). They carry a gene, tax, whose product alters the expression of other viral genes. Transactivating regulatory genes are believed to be necessary for viral replication in vivo and may contribute to oncogenesis by also modulating cellular genes that regulate cell growth.

There are several genetic subtypes of HTLV-1, with the major ones being subtypes A, B, and C (these do not represent distinct serotypes).

The virus is distributed worldwide, with an estimated 20 million infected individuals. Clusters of HTLV-associated disease are found in certain geographic areas (southern Japan, Melanesia, the Caribbean, Central and South America, and parts of Africa) (Figure 43-5). Although less than 1% of people worldwide have HTLV-1 antibody, up to 5% of the population in endemic areas may be seropositive.

ATL is poorly responsive to therapy. The 5-year survival rate for patients with this cancer is less than 5%.

Transmission of HTLV-1 seems to involve cell-associated virus. Mother-to-child transmission via breastfeeding is an important mode. Efficiency of transmission from infected mother to child is estimated at 15–25%. Such early-life infections are associated with the greatest risk of ATL. Blood transfusion is an effective means of transmission, as are sharing blood-contaminated needles (drug abusers) and sexual intercourse.

Seroepidemiology has linked infection with HTLV-1 to a syndrome called HTLV-1–associated myelopathy/tropical spastic paraparesis (HAM/TSP). The primary clinical feature is development of progressive weakness of the legs and lower body. The patient’s mental faculties remain intact. HAM/TSP is described as being of the same magnitude and importance in the tropics as is multiple sclerosis in Western countries. Other HTLV-1–associated diseases include uveitis and infective dermatitis.

B. Human Immunodeficiency Viruses

A group of human retroviruses has been established as the cause of acquired immune deficiency syndrome (AIDS) (see Chapter 44). The HIV are cytolytic and nontransforming and are classified as lentiviruses. However, AIDS patients are at elevated risk of several types of cancer because of the immune suppression associated with HIV infection. These cancers include cervical cancer, Kaposi sarcoma, lymphomas, head and neck cancer, liver cancer, and oral cancer.

C. Other

The simian foamy viruses from the Spumavirus genus are highly prevalent in captive nonhuman primates. Humans occupationally exposed to the primates can be infected with foamy viruses, but these infections have not resulted in any recognized disease.

Fundamental differences exist between the oncogenes of DNA and RNA tumor viruses. The transforming genes carried by DNA tumor viruses encode functions required for viral replication and do not have normal homologs in cells. In contrast, retroviruses either carry transduced cellular oncogenes that have no role in viral replication or they act through indirect mechanisms. The DNA virus transforming proteins complex with normal cell proteins and alter their function. To understand the mechanism of action of DNA virus transforming proteins, it is important to identify the cellular targets with which they interact. Examples of such interactions are shown in Table 43-4.

POLYOMAVIRUSES

Important properties of polyomaviruses are listed in Table 43-5.

Classification

The Polyomaviridae family contains a single genus designated Polyomavirus, formerly part of the Papovaviridae family (which no longer exists). Polyomaviruses are small viruses (diameter 45 nm) that possess a circular genome of double-stranded DNA (5 kbp; molecular weight 3 × 10⁶) enclosed within a nonenveloped capsid exhibiting icosahedral symmetry (Figure 43-6). Cellular histones are used to condense viral DNA inside virus particles.

Polyomaviruses are simple DNA-containing viruses that possess a limited amount of genetic information (six or seven genes). Multiple species have been identified, including the tumor virus SV40 and others known to infect humans (BK, JC, KI, WU, MCV, HPyV6, HPyV7, HPyV10, and TSV). Many species of mammals and some birds have been found to carry their own species of polyomavirus.

Polyomavirus Replication

The polyomavirus genome contains “early” and “late” regions (Figure 43-7). The early region is expressed soon after infection of cells; it contains genes that code for early proteins—for example, the SV40 large tumor (T) antigen, which is necessary for the replication of viral DNA in permissive cells, and the small tumor (t) antigen. The murine polyoma virus genome encodes three early proteins (small, middle, and large T antigens). One or two of the T antigens are the only viral gene products required for transformation of cells. Usually, the transforming proteins must be continuously synthesized for cells to stay transformed. The late region consists of genes that code for the synthesis of coat protein; they have no role in transformation and usually are not expressed in transformed cells.

SV40 T antigen interacts with the cellular tumor suppressor gene products, p53 and pRb family members (see Table 43-4). Interactions of T antigen with the cellular proteins are important in the replicative cycle of the virus. This complex formation functionally inactivates the growth-inhibitory properties of pRb and p53, allowing cells to enter S phase so that viral DNA may be replicated. Likewise, functional inactivation of the cellular proteins by T antigen binding is central to the virus-mediated transformation process. As p53 senses DNA damage and either blocks cell cycle progression or initiates apoptosis, abolishing its function would lead to accumulation of T antigen-expressing cells with genomic mutations that might promote tumorigenic growth.

Pathogenesis and Pathology

The human polyomaviruses BK and JC are widely distributed in human populations, as evidenced by the presence of specific antibody in 70–80% of adult sera. Infection usually occurs during early childhood. Both viruses may persist in the kidneys and lymphoid tissues of healthy individuals after primary infection and may reactivate when the host’s immune response is impaired, for example, by renal transplantation, during pregnancy, or with increasing age. Viral reactivation and shedding in urine are asymptomatic in immunocompetent persons. The viruses are most commonly isolated from immunocompromised patients, in whom disease may occur. BK virus causes hemorrhagic cystitis in bone marrow transplant recipients. It is the cause of polyomavirus-associated nephropathy in renal transplant recipients, a serious disease that occurs in up to 5% of recipients and that results in graft failure in up to 50% of those affected patients. JC virus is the cause of progressive multifocal leukoencephalopathy, a fatal brain disease that occurs in some immunocompromised persons, especially those with depressed cell-mediated immunity resulting from immunosuppressive therapies or infection by HIV. Progressive multifocal leukoencephalopathy affects about 5% of AIDS patients. BK virus and JC virus are antigenically distinct, but both encode a T antigen that is related to SV40 T antigen. These human viruses can transform rodent cells and induce tumors in newborn hamsters. JC virus has been associated with human brain tumors, but an etiologic role is not yet established.

KI and WU viruses were discovered in 2007 in nasopharyngeal aspirates from children with respiratory infections. Merkel cell polyomavirus was identified in 2008 in Merkel cell carcinomas, rare skin tumors of neuroendocrine origin. Seroprevalence studies suggest that KI, WU, and Merkel cell virus infections are common and widespread and occur in childhood. HPyV6, HPyV7, and HPyV10 appear to be constituents of human skin. Trichodysplasia spinulosa–associated polyomavirus (TSV) was discovered in proliferative skin lesions, HPyV9 was found in the blood of immunosuppressed patients, and HPyV12 was found in liver tissue. Other polyomaviruses have been found in human stool, including MWPyV, MXPyV, and STL polyomavirus. Because of their recent discoveries, information on disease associations is limited.

Merkel cell virus appears to be etiologically important in a large fraction of Merkel cell carcinomas. In many of the tumors characterized, Merkel cell virus DNA is clonally integrated in tumor cells, oncogene expression is required for cell growth, and the integrated viral genomes have mutations in the T antigen gene that prevent viral DNA replication.

SV40 replicates in certain types of monkey and human cells; it is highly tumorigenic in experimentally inoculated hamsters and in transgenic mice and can transform many types of cells in culture. Tumor induction in the natural host—the rhesus monkey—is rarely observed. SV40 may cause a progressive multifocal leukoencephalopathy-like disease in rhesus monkeys.

SV40 contaminated early lots of live and killed poliovirus vaccines that had been grown in monkey cells unknowingly infected with SV40. Millions of people worldwide received such SV40-contaminated vaccines between 1955 and 1963. SV40 is detected in humans today, including in individuals too young to have been exposed via vaccination. Evidence suggests that it (and other polyomaviruses) may be transmitted by the fecal–oral route in humans. The prevalence of SV40 infections in humans appears to be low.

SV40 DNA has been detected in selected types of human tumors, including brain tumors, mesotheliomas, bone tumors, and lymphomas. The role SV40 may be playing in formation of human cancers is unknown.

The host range for polyomaviruses is often highly restricted. Usually a single species can be infected and only certain cell types within that species. Exceptions are the primate polyomaviruses SV40 and BK virus; SV40 can infect also humans and human cells and BK virus can infect some monkeys and monkey cells. Cell types that fail to support polyomavirus replication may be transformed by a virus.

PAPILLOMAVIRUSES

Important properties of papillomaviruses are listed in Table 43-6.

Classification

The Papillomaviridae family is a very large virus family currently divided into 16 genera, of which five contain members that infect humans (Alpha-, Beta-, Gamma-, Mupa-, and Nupapapillomavirus). The papillomaviruses are former members of the Papovaviridae family. Although papillomaviruses and polyomaviruses share similarities in morphology, nucleic acid composition, and transforming capabilities, differences in genome organization and biology led to their separation into distinct virus families. The papillomaviruses are slightly larger in diameter (55 nm) than the polyomaviruses (45 nm) and contain a larger genome (8 kbp vs 5 kbp). The organization of the papillomavirus genome is more complex (Figure 43-8). There is widespread diversity among papillomaviruses. Because neutralization tests cannot be done since there is no in vitro infectivity assay, papillomavirus isolates are classified using molecular criteria. Virus “types” are at least 10% dissimilar in the sequence of their L1 genes. Almost 200 distinct HPV types have been recovered.

Papillomavirus Replication

Papillomaviruses are highly tropic for epithelial cells of the skin and mucous membranes. Viral nucleic acid can be found in basal stem cells, but late gene expression (capsid proteins) is restricted to the uppermost layer of differentiated keratinocytes (Figure 43-9). Stages in the viral replicative cycle are dependent on specific factors that are present in sequential differentiated states of epithelial cells. This strong dependence of viral replication on the differentiated state of the host cell is responsible for the difficulties in propagating papillomaviruses in vitro.

Pathogenesis and Pathology

Transmission of viral infections occurs by close contact. Viral particles are released from the surface of papillomatous lesions. It is likely that microabrasions allow infection of proliferating basal layer cells at other sites or within different hosts.

Papillomaviruses cause infections at cutaneous and mucosal sites, sometimes leading to the development of different kinds of warts, including skin warts, plantar warts, flat warts, anogenital warts, laryngeal papillomas, and several cancers, including those of the cervix, vulva, penis and anus, and a subset of head and neck cancers (Table 43-7). The multiple types of HPV isolates are preferentially associated with certain clinical lesions, though distribution patterns are not absolute. HPV genital infections are sexually transmitted and represent the most common sexually transmitted disease in the United States. Cervical cancer is the second most frequent cancer in women worldwide (about 500,000 new cases annually) and is a major cause of cancer deaths in developing countries.

Based on the relative occurrence of viral DNA in certain cancers, HPV types 16 and 18 are considered to be high cancer risk; about 16 other less common types are somewhat less frequently associated with neoplasms but are also considered high risk. Many HPV types are considered benign.

Integrated copies of viral DNA are usually present in cervical cancer cells, though HPV DNA is generally not integrated (episomal) in noncancerous cells or premalignant lesions. Skin carcinomas appear to harbor HPV genomes in an episomal state. Viral early proteins E6 and E7 are synthesized in cancer tissue. These are HPV-transforming proteins, able to complex with Rb and p53 and other cellular proteins (see Table 43-4).

The behavior of HPV lesions is influenced by immunologic factors. Cell-mediated immunity is important. The majority of HPV infections are cleared and become undetectable within 2–3 years. Development of HPV-associated carcinomas requires persistent infection.

Cervical cancer develops slowly, taking years to decades. It is thought that multiple factors are involved in progression to malignancy; however, persistent infection with a high-risk HPV is a necessary component to the process (Figure 43-10).

Clinical Findings and Epidemiology

An estimated 660 million people worldwide have HPV genital infections, the most common viral infection of the reproductive tract. An estimated 20 million Americans are infected, with about 6 million new infections occurring annually in the United States. The peak incidence of HPV infections occurs in adolescents and young adults younger than 25 years.

HPVs are accepted as the cause of anogenital cancers. Over 99% of cervical cancer cases and over 80% of anal cancer cases are linked to genital infections with HPVs. Papillomaviruses illustrate the concept that natural viral strains may differ in oncogenic potential. Although many different HPV types cause genital infections, HPV-16 or HPV-18 is found most frequently in cervical carcinomas, although some cancers contain DNA from other types, such as HPV type 31 (see Table 43-7). Epidemiologic studies indicate that HPV-16 and HPV-18 are responsible for more than 70% of all cervical cancers, with type 16 being most common. HeLa cells, a widely used tissue culture cell line derived many years ago from a cervical carcinoma, contain HPV-18 DNA.

Anal cancer is associated with high-risk HPV infection. Immunocompromised patients are especially at risk, as are men who have sex with men. Multiple HPV types have been commonly found in the anal canals of HIV-infected men in the latter group. Oropharyngeal cancers, a subset of head and neck squamous cell carcinomas, are also linked to HPV infections, especially by type 16. About 25% of mouth and 35% of throat cancers are associated with HPV. The oral cavities of both HIV-positive and HIV-negative individuals contain an abundant number of different HPV types.

The role of men as carriers of HPV as well as vectors for transmission of infections is well documented; however, most penile HPV infections in men are subclinical and do not result in HPV-associated disease.

Anogenital warts are usually (90%) caused by low-risk HPV types 6 and 11. Laryngeal papillomas in children, also called recurrent respiratory papillomatosis, are caused by HPV-6 and HPV-11, the same viruses that cause benign genital condylomas. The infection is acquired during passage through the birth canal of a mother with genital warts. Though laryngeal papillomas are rare, the growths may obstruct the larynx and must be removed repeatedly by surgical means. About 3000 cases of this disease are diagnosed annually; up to 3% of children will die.

There is a high prevalence of HPV DNA in normal skin from healthy adults. It appears that these asymptomatic HPV infections are acquired early in infancy. A great multiplicity of HPV types are detected in normal skin. Transmission is thought to occur from those in close contact with the child, with a high concordance (about 60%) between types detected in infants and their mothers.

Immunosuppressed patients experience an increased incidence of warts and cancer of the cervix. All HPV-associated cancers occur more frequently in persons with HIV/AIDS.

Prevention and Control

Widespread use of the Papanicolaou (Pap) smear for cervical cancer detection has led to substantial decrease in deaths due to cervical cancer. The cytological test aims to detect precancerous cellular changes in morphology, allowing for lesion removal prior to cancer development. Testing for the presence of HPV types 16 and 18, or for all high-risk HPV types, enhances the detection sensitivity and specificity for precancerous lesions. Laboratory testing is by DNA hybridization or PCR methods. Testing algorithms include reflex testing of abnormal Pap smears, co-testing, or primary HPV screening. Studies have indicated that primary HPV screening algorithms may be the preferred method in most settings. Women younger than 20 years should not be tested due to frequent HPV clearance in initial infections.

Vaccines against HPV are expected to be a cost-effective way to reduce anogenital HPV infections, the incidence of cervical cancer, and the HPV-associated health care burden. A quadrivalent HPV vaccine was approved in the United States in 2006 and a bivalent vaccine in 2007. Both are noninfectious recombinant vaccines containing virus-like particles composed of HPV L1 proteins. The quadrivalent vaccine contains particles derived from HPV types 6, 11, 16, and 18, whereas the bivalent vaccine contains particles from types 16 and 18. Both vaccines are effective at preventing persistent infections by the targeted HPV types and the development of HPV-related genital precancerous lesions. They are not effective against established HPV disease. Adolescent and young adult females were the initial target population for vaccination, with adolescent and young adult males recommended for vaccination with the quadrivalent vaccine in 2011. It is not known how long vaccine-induced immunity lasts, but it appears to extend for at least 10 years.

ADENOVIRUSES

The adenoviruses (see Chapter 32) comprise a large group of agents widely distributed in nature. They are medium-sized, nonenveloped viruses containing a linear genome of double-stranded DNA (26–45 kbp). Replication is species-specific, occurring in cells of the natural hosts. Adenoviruses commonly infect humans, causing mild acute illnesses, mainly of the respiratory and intestinal tracts.

Adenoviruses can transform rodent cells and induce the synthesis of virus-specific early antigens that localize in both the nucleus and the cytoplasm of transformed cells. The E1A early proteins complex with the cellular Rb protein as well as with several other cellular proteins. Other early proteins, E1B and E4ORF1, bind p53 and other cellular signaling proteins (see Table 43-4). The adenoviruses are important models for studying the molecular mechanisms by which DNA tumor viruses usurp cellular growth control processes. Different serotypes of adenoviruses manifest varying degrees of oncogenicity in newborn hamsters. No association of adenoviruses with human neoplasms has been found.

HERPESVIRUSES

These large viruses (diameter 125–200 nm) contain a linear genome of double-stranded DNA (125–240 kbp) and have a capsid with icosahedral symmetry surrounded by an outer lipid-containing envelope. Herpesviruses (see Chapter 33) typically cause acute infections followed by latency and eventual recurrence in each host, including humans.

In humans, herpesviruses have been linked to several specific types of tumors. Herpesvirus EBV causes acute infectious mononucleosis when it infects B lymphocytes of susceptible humans. Normal human lymphocytes have a limited life span in vitro, but EBV can immortalize such lymphocytes into lymphoblast cell lines that grow indefinitely in culture.

EBV is etiologically linked to Burkitt lymphoma, a tumor most commonly found in children in central Africa; to nasopharyngeal carcinoma, more common in Cantonese Chinese and Alaskan Eskimos than other populations; to posttransplant lymphoproliferative disorders in immunocompromised hosts; and to Hodgkin disease. These tumors usually contain EBV viral DNA (both integrated and episomal forms) and viral antigens.

EBV encodes a viral oncogene protein (LMP1) that mimics an activated growth factor receptor. LMP1 is able to transform rodent fibroblasts and is essential for transformation of B lymphocytes (see Table 43-4). Several EBV-encoded nuclear antigens (EBNAs) are necessary for immortalization of B cells; EBNA1 is the only viral protein consistently expressed in Burkitt lymphoma cells. EBV is very successful at avoiding immune elimination; this may be due in part to the function of EBNA1 in inhibition of antigen processing to allow infected cells to escape killing by cytotoxic T lymphocytes.

Malaria may be a cofactor for African Burkitt lymphoma. Most of those tumors also show characteristic chromosomal translocations between the c-myc gene and immunoglobulin loci, leading to the constitutive activation of myc expression. Consumption of salted or dried fish may be a dietary cofactor in EBV-related nasopharyngeal carcinoma. EBV-associated posttransplant lymphoproliferative disorders are more common in highly immunosuppressed transplant patients and can be detected earlier by screening for EBV in the blood.

Kaposi sarcoma–associated herpesvirus, also known as human herpesvirus 8 (KSHV/HHV8), is not as ubiquitous as most other human herpesviruses. It is believed to be the cause of Kaposi sarcoma, primary effusion lymphoma, and multicentric Castleman disease, a lymphoproliferative disorder. KSHV has a number of genes related to cellular regulatory genes that may stimulate cellular proliferation and modify host defense mechanisms.

Some herpesviruses are associated with tumors in lower animals. Marek disease is a highly contagious lymphoproliferative disease of chickens that can be prevented by vaccination with an attenuated strain of Marek disease virus. The prevention of cancer by vaccination in this case establishes the virus as the causative agent and suggests the possibility of a similar approach to prevention of human tumors with a virus as a causative agent. Other examples of herpesvirus-induced tumors in animals include lymphomas of certain types of monkeys and adenocarcinomas of frogs. The simian viruses cause inapparent infections in their natural hosts but induce malignant T-cell lymphomas when transmitted to other species of monkeys.

POXVIRUSES

Poxviruses (see Chapter 34) are large, brick-shaped viruses with a linear genome of double-stranded DNA (130–375 kbp). Yaba virus produces benign tumors (histiocytomas) in its natural host, monkeys. Shope fibroma virus produces fibromas in some rabbits and is able to alter cells in culture. Molluscum contagiosum virus produces small benign growths in humans. Very little is known about the nature of these proliferative diseases.

It is clear that viruses are involved in the genesis of several types of human tumors. Proving a causal relationship between a virus and a given type of cancer is, in general, very difficult.

If a virus is the only etiologic agent of a specific cancer, the geographic distribution of viral infection should coincide with that of the tumor; the presence of viral markers should be higher in cases than in controls; and viral infection should precede the tumor. These criteria can be difficult to establish if other environmental or genetic factors cause some cases of the same type of cancer. Only if the continued expression of a viral function is necessary for maintenance of transformation will viral genes necessarily persist in every tumor cell. If the virus provides an early step in multistep carcinogenesis, the viral genome may be lost as the tumor progresses to more altered stages. Conversely, a virus may be found associated frequently with a tumor but be there simply as a passenger because of an affinity for the cell type.

Tumor viruses are usually not replicating in transformed cells, so it is necessary to use very sensitive methods to search for viral nucleic acids or proteins in cells to detect virus presence. Viral structural proteins are frequently not expressed, but virus-encoded nonstructural proteins may be expressed as markers of virus presence.

Tumor induction in laboratory animals and transformation of human cells in culture are good circumstantial lines of evidence that a virus is tumorigenic, and those systems can provide models for molecular analyses of the transformation process. However, they do not constitute proof that the virus causes a particular human cancer.

The most definitive proof of a causal relationship is decreased tumor incidence by prevention of viral infection. Intervention methods should be effective in reducing the occurrence of the cancer even if the virus is only one of several cofactors.

• Viruses are causative agents for several types of human cancer.

• Cancer viruses are classified in different virus families, include both RNA-containing and DNA-containing agents, and function by different mechanisms of carcinogenesis.

• Human cancer viruses include hepatitis B virus, hepatitis C virus, papillomaviruses, EBV, human herpesvirus 8, HTLVs, and Merkel cell virus. HIV is also considered a cancer agent because of the immune suppression associated with infection by the virus.

• There are effective vaccines against hepatitis B virus and the high-risk HPV types 16 and 18.

• Animal models and cultured cells are used to explore mechanisms of viral carcinogenesis.

• Studies with tumor viruses revealed the roles of cellular oncogenes and tumor suppressor genes in cancer and fostered the recognition of the molecular basis of carcinogenesis.

• Tumor viruses establish persistent infections in hosts, with long latent periods between initial infection and tumor appearance.

• Cancer virus infections are much more common than virus-related tumor formation.