Cancer Therapy Vol 2, 85-98, 2004

Center for Neurovirology and Cancer Biology, Laboratory of Molecular Neurovirology, Temple University, College of Science and Technology, 1900 North 12th Street, 015-96, Room 204A, Philadelphia, PA 19122

__________________________________________________________________________________

*Correspondence: Mahmut Safak, Laboratory of Molecular Neurovirology, Center for Neurovirology and Cancer Biology, College of Science and Technology, Temple University, 1900 N. 12th St., 015-96, Room 204A, Philadelphia, PA 19122. Phone: (215) 204-6340. Fax: (215) 204-0679. E-mail: msafak@temple.edu

Key Words: Polyomaviruses, JCV, BKV, SV40, T antigen

Abbreviations: BK virus, (BKV); central nervous system, (CNS); CREB-binding protein, (CBP); human immunodeficiency virus, (HIV); insulin-like growth factor receptor, (IGF-IR); JC virus, (JCV); Jun N-terminal kinase, (JNK); Kaposi's sarcoma, (KS); myelin basic protein, (MBP); nuclear localization signal, (NLS); polymerase a -primase, (Pola ); progressive multifocal encephalopathy, (PML); proteolipid protein, (PLP); simian virus 40, (SV40)

The genome structure and gene products of polyomaviruses have been under intense investigation in recent years for several reasons. First, their small, circular genomes serve as miniature model systems to study many aspects of DNA structure for more complex eukaryotic genomes. Second, their oncogenic proteins can transform cells under certain conditions in both tissue culture and experimental animals in a manner resembling malignancies seen in humans. Particularly, recent findings regarding the detection of the genomes of both human (JCV and BKV) and simian virus 40 (SV40) polyomaviruses in a variety of human tumors suggest that this group of viruses may play a role in the induction of certain human tumors, although controversy still remains as to whether these viruses indeed induce such tumors. Such observations have led to investigators to further study the mechanisms of tumor induction by these viruses. In this short review, we focused our attention to recent developments with respect to polyomavirus-induced tumors in experimental animals and the detection of viral genomes in a variety of human malignancies.

JCV is a small human DNA virus with a double-stranded, covalently linked circular genome, 5130 base pair in size. It is classified in the Papovaviridae family within the polyomavirus genus . JCV genome is composed of bidirectional regulatory elements and coding regions (Figure 1). The regulatory region contains the origin of DNA replication and promoter/enhancer elements for viral early and late genes. The coding regions can be divided

Figure 1. Genomic organization of JCV. JCV genome is composed of regulatory and coding regions. The regulatory region contains the origin of DNA replication and promoter/enhancer elements. The coding regions are divided into an early and late region. The early region encodes regulatory proteins, small and large T antigen. The late coding region encodes viral structural proteins (VP-1, VP-2 and VP-3) and a short regulatory peptide, agnoprotein.

into early and late regions. The early coding region primarily encodes two regulatory proteins, small and large T antigen although recent findings indicate that this region also encodes three additional small peptides called T's . The late coding region encodes structural capsid proteins (VP-1, VP-2 and VP-3) and a small regulatory agnoprotein. Structural and antigenic studies demonstrated that JCV is related to other polyomaviruses such as human BK virus, and a primate virus, simian virus 40 in the genus. Serological data indicate that, unlike SV40, JCV and BKV share the property of hemagglutination of human type O erythrocytes. It should also be noted here that there is lack of convincing sera conversion data for wide infectivity of SV40 in human population as seen for JCV and BKV. Seroepidomological data shows that overwhelming majority of the world's population is infected by JCV and the virus establishes a persistent infection in the kidneys (latent infection) after a subclinical primary infection. Recent reports indicate that peripheral blood B lymphocytes, hematopoietic progenitor cells, and tonsillar stromal cells could also harbor JCV. These sites, therefore, can be considered additional potential sites for JCV infection and latency .

JCV was first isolated from brain tissue of a PML patient by Padgett et al, in 1971. The brain tissue was used as a source of inoculum to infect primary cultures derived from human fetal brain and the virus was successfully isolated from long-term cultures mainly consisting of glial cells . This was the first direct evidence suggesting that a neurotropic viral agent was associated with the occurrence of PML. Shortly after its isolation, the oncogenic potential of the virus was tested both in tissue culture and experimental animals. Particularly, recent findings regarding the detection of JCV genome in a variety of human tumors indicate that JCV may be associated with the induction of human tumors.

JCV is a neurotropic virus that lytically infects oligodendrocytes in the central nerves system and causes a neurodegenerative disease of the white matter in the human brain, progressive multifocal encephalopathy (PML). The disease develops mostly in patients with underlying immunosuppressive conditions, including Hodgkin's lymphoma, lymphoproliferative diseases, and AIDS . In a small number of cases, however, PML was also found to affect individuals with no underlying disease . While PML was previously considered a rare complication of middle-aged and elderly patients with lymphoproliferative diseases, due to the AIDS epidemic in recent years, it is now a commonly encountered disease of the CNS in patients of different age groups. This suggests that human immunodeficiency virus (HIV) infection may directly or indirectly participate in this process. Recent estimates indicate that the incidence of PML in HIV-seropositive patients reached up to 5%, compared to that 0.8% before the AIDS epidemic .

Following its isolation, JCV has not only been shown to cause a variety of tumors in experimental animals but also shown to have the ability to induce neoplastic cell transformation in tissue culture. Since JCV induced tumors arise in tissues of neural origin , tissue-specific expression of JCV regulatory region is thought to play a major role in this process. Inoculation of JCV into several experimental animal models, including hamsters, nonhuman primates, and transgenic mice, resulted in variety of tumors depending on the animal type, age and site of inoculation. For instance, more than 80% of newborn Syrian hamsters when inoculated intracerebrally and subcutaneously with the Mad-1 strain of JCV developed glioblastomas, neuroblastomas and medullablastomas . Even the presence of an entire biologically active JCV genome was demonstrated when cells from these tumors were co-cultivated with permissive glial cells . JCV was also inoculated intraoccularly into newborn hamsters and resulted in abdominal neuroblastomas developing in several locations of the body .

Unlike the other members of the polyomavirus family (BKV and SV40), JCV is the only polyomavirus shown to induce tumors in nonhuman primates, such as monkeys. When owl squirrel monkeys were inoculated with live JCV subcutanously, intraperitoneally, and intracerebraly , the animals developed tumors at different time intervals. One owl monkey developed a malignant cerebral tumor similar to an astrocytoma seen in humans after 16 months of inoculation. Another one developed a malignant neuroblastoma 25 months after inoculation. Further analysis of the tumors for the expression of JCV large tumor antigen which is the main viral regulatory protein involved in tumor induction revealed both the presence of large T antigen and complex formation with tumor suppressor protein p53 .

Mechanistically, the tumorigenic potential of JCV T antigen appears to be, at least in part, mediated by its interaction with tumor suppresser genes including p53 and retinoblastoma gene products, pRb and p130. Upon binding, T antigen appears to interfere with the cell cycle progression properties of these proteins. Coimmunoprecipitation assays using cellular extracts from JCV-transformed glial cells show T antigen complex formation with pRb, p53 and p107 . A report by Rencic et al, also suggests a role for T antigen in the induction of oligoastrocytomas in an immunocompetent patient. JC virus large T antigen has also been shown to interact with cellular and viral proteins including YB-1, Pura , JCV agnoprotein, and insulin receptor substrate 1 (IRS-1) . IRS-1 is the major signaling molecule for the type I insulin-like growth factor receptor (IGF-IR) . In addition, recent reports also indicate a possible communication between JCV T antigen and the Wnt signaling pathway in induction of tumor formation because T antigen expressing cells express higher levels of b -catenin and its partner LEF-1 .

Our group also described the formation of different tumors in tissues that derived from neural origin in transgenic mice models . JCV early coding region, driven by its own promoter, was utilized to create these transgenic animal models. Histological and histochemical analysis of the tumor masses demonstrated the expression of JCV large T antigen in tumors versus control tissues. In contrast to previous observations by Small et al , transgenic animals created with the early region of JCV archetype strain did not show any sign of hypomyelination in the central nervous system which was a feature observed in transgenic mouse models. On the contrary, cerebellar tumors that resemble human medullablastomas appeared in the transgenic animals . In another line of transgenic mouse, half of the animals developed large, solid masses within the base of the skull by one year of age. Histological evaluation of the tumors by location and by histochemical studies demonstrated that these tumors arose from the pituitary gland . Figure 2 exemplifies a variety of tumors induced by JCV in an experimental animal model system.

In addition to the evaluation of tumorogenic activity of JCV in mice, transgenic mice were also used to study the process of the acute demyelination occurring in PML-affected brain tissue. Some of the offsprings of a transgenic mouse created with the regulatory and coding sequences of JCV T-Ag exhibited mild to severe tremor phenotypes with hypo and dysmyelination occurring in the central nervous system (CNS). In addition, dysmyelination was further characterized in transgenic animals by Trapp et al, by examining the expression of the JCV and myelin-specific genes. Initial examination of brain tissue from transgenic mice revealed relatively low expression levels of proteolipid protein (PLP), myelin basic protein

Figure 2. JCV transgenic animal models. Transgenic mice containing the full sequence of the JCV genome (archetype), develop primitive neuroectodermal tumors in the brain, characterized by numerous packed cells with an elongated nuclei and scanted cytoplasm (Panel A, Hematoxilin & Eosin). Immunohistochemistry against the early gene product T-antigen, demonstrates the nuclear localization of the protein (Panel B), and in the same group of cells there is intense immunoreactivity for p53 (Panel C). Transgenic animals containing only the early sequence of JCV, develop a variety of neural-origin tumors, including adrenal neuroblastomas, characterized by rounded homogeneous cells with a perinuclear halo of cytoplasm (Panel D, Hematoxilin & Eosin), which also express nuclear T-antigen when tested by immunohistochemistry (Panel E). In the same cellular compartment there is strong immunoreactivity for p53 (Panel F). Another tumor developed by a line of JCV early transgenic mice is pituitary adenomas, characterized by numerous pleomorphic cells of different sizes and abundant eosinophilic cytoplasm (Panel G, Hematoxilin & Eosin). The neoplastic cells demonstrate intense nuclear positivity for T-antigen (Panel H), as well as p53 (Panel I). All panels original magnification x1000.

(MBP) and myelin associated glycoprotein which collectively make up the axonal myelin sheet although the mRNA message levels for those proteins appeared to be normal. The mechanism by which T antigen plays a critical role in the reduction of these respective protein levels in the brains of transgenic mice remains unknown, however, it is suggested that T antigen may alter the expression levels of both proteolipid and myelin basic protein at the protein levels or may inhibit the maturation process of oligodentrocytes thereby altering the level of myelin around the axons.

In recent years, a widespread detection of JCV genome in variety of human tumors raised the possibility that JCV may induce tumors in humans. In fact, Richardson, who first described PML in 1961 , reported the incidental detection of an oligodendroglioma in a patient with concomitant occurrences of chronic lymphatic leukemia and PML. Following this report, concomitant occurrences of PML with different human tumors was described in several more cases. Sima et al, reported the association of PML with multiple astrocytomas in 1983 . Similarly, Casteigne et al, (1974) described a case where a patient with long history of immunodeficiency syndrome, in addition to PML, showed numerous foci of anaplastic astrocytes. Microscopic analysis of the demyelinating lesions demonstrated the presence of viral particles in both oligodendrocytes and astrocytes within PML foci, but not in the neoplastic astrocytes . A more recent report by Shintaku and colleagues showed dysplastic ganglian-like cells in a patient with PML . A large number of dysplastic or dysmorphic ganglian-like cells were found in the cerebral cortex that showed properties of neurons. Expression of JCV large T antigen was demonstrated in the infected neurons, however, the late gene products were not.

In addition to the cases described above, JCV genome has also been detected in human brain tumors in the absence of PML lesions. Boldorini et al, reported the detection of JCV DNA in the brain tumors of an immunocompetent patient with a pleomorphic xantoastrocytoma . An earlier study by Rencic et al, demonstrated the presence of JCV viral DNA and expression of large T antigen in tumor tissue from an immunocompetent HIV-negative patient with oligoastrocytoma . These two cases presented the experimental evidence for a possible association of JCV in brain tumors of immunocompetent non-PML patients. Such findings further prompted the attempts to establish the association of JCV with different types of brain tumors in humans. In fact, Del Valle et al, recently analyzed multiple brain tumors for the detection of JCV genome and showed that 62.5% of oligoastrocytomas, 83.3% of ependymomas, 80% of pilocytic astrocytomas, 57.1% of oligodendrogliomas, 76.9% of astrocytomas and 66% of anaplastic oligodendrogliomas contained JCV early gene sequence. Figure 3 illustrates the detection of JCV early oncogenic protein, large T antigen, and cellular tumor suppressor protein, p53, in a variety of human tumors JCV genomic DNA has also been shown to be present in tumor tissue which is not neural origin. Recent reports indicate that the JCV genome was detected in.

Figure 3. Detection of JCV proteins in human brain tumors. Expression of JCV early protein has been found in a wide variety of brain neoplasms, including low grade glial tumors, such as oligodendrogliomas (Panel A, Hematoxilin & Eosin), characterized by homogeneous cells with a clear halo surrounding their nuclei. Immunohistochemistry from T-antigen is positive in the nuclei of the majority of the neoplastic cells (Panel B), where the cell cycle regulator protein p53 is also found (Panel C). High-grade glial tumors such as glioblastoma multiforme (Panel D) characterized by extensive areas of necrosis and pleomorphic, atypical cells expressing T-antigen in their nuclei (Panel E). p53 is also present in the nuclei of the neoplastic cells (Panel F). Tumors of neural origin, such as medulloblastomas, characterized by numerous sheaths of homogeneous cells, with scanted cytoplasm (Panel G, Hematoxilin & Eosin), demonstrate nuclear expression of the early JCV protein, T-antigen (Panel H), and also nuclear immunoreactivity for p53 (Panel I). All panels original magnification x1000.

gastrointestinal tract and solid non-neural tumors including colorectal cancers . It should be however noted here that such studies explored the possibility of whether JCV genome or its expressed proteins could be detected by certain molecular biology techniques but does not provide information about the mechanism by which JCV could possibly induce tumors in humans

Another human polyomavirus which is classified within the Papovaviridae family is BK virus. This virus was first isolated in 1971 from the urine of a renal allograft recipient who developed ureteric stenosis . Like JCV and SV40, the BKV early and late genomes code for six viral proteins, two from the early genome and four from the late genome. Early proteins are nonstructural regulatory proteins (small t and large T antigens), of which large T antigen is involved in regulation of the viral DNA replication and late gene expression. The function of small t antigen in this regard is not known. The viral late genome, in addition to encoding the structural proteins VP-1, VP2 and VP3, also encodes a small regulatory peptide, agnoprotein, whose function largely remains unclear in the viral lytic cycle. Recent evidence from JCV virus agnoprotein work, however, suggests that it plays a role in viral DNA replication, transcription , and cell cycle regulation .

Like JCV, BKV has also a worldwide distribution in the human population. Primary infection by BKV takes place during early childhood and is subclinical although a mild respiratory illness or urinary track disease may occur . Little is known about the route of BKV transmission although induction of upper respiratory disease by BKV and detection of latent BKV DNA in tonsils suggests a possible oral or respiratory route of transmission . During primary infection, viremia occurs and the virus spreads to a number of organs in the infected individuals including kidneys, bladder, prostate, uterine cervix, lips and tongue where it remains in a latent state. Reactivation of the virus from latent state is mostly associated with the immunocompromised state of individuals. Reactivated virus was detected in the urine of renal and bone marrow transplant recipients undergoing immunosuppressive therapy as well as in the urine of pregnant women . Upon reactivation, BKV may cause interstitial nephritis and ureteral obstruction in patients receiving renal transplants, and in some cases, it can cause viral-infection-induced transplant dysfunction and graft rejection . In addition, an association between hemorrhagic cystitis and BKV was shown in bone marrow transplant recipients .

Like JCV, the oncogenic potential of BKV has been tested in experimental animals including young and newborn mice, rats, and hamsters by inoculation of live virus. . The type of tumors induced by BKV was strictly dependent on the route of inoculation. It was observed that BKV is weakly oncogenic when inoculated subcutaneously but induced tumors in high proportions when inoculated intracerebrally or intravenously . Tumors induced by BKV belong to a variety of histotypes including ependymoma, neuroblastoma, pineal gland tumors, fibrosarcoma, osteosarcoma and tumors of pancreatic islets . Rats inoculated with BKV developed fibrosarcoma, liposarcoma, osteosarcoma, nephroblastoma, and glioma. Mice, however, developed only choroids plexus papilloma in a similar setting .

Transgenic mice were also used to test the oncogenicity of BKV large T antigen (T-Ag). Transgenic mice with BKV T-Ag developed renal tumors, hepatocellular carcinoma, and lymphoproliferative disease . In such studies, there appears to be differences among the strains of BKV in terms of oncogenicity. For example, Gardner's BKV strain seems to be more potent to induce tumors in transgenic mice than other isolates such as MM, BKV-IR or RF .

The mechanism by which BKV causes tumors in experimental animals and cell transformation in tissue culture remains elusive. It was shown that like JCV and SV40 T-Ag, BKV T-Ag interacts with several key cell cycle regulatory proteins, including tumor suppressor proteins p53 and the family members of retinoblastoma proteins, pRb105 and Rb130. BKV T-Ag perhaps inactivates the function of these proteins and thereby contributes to the cell transformation . It was recently shown that the complex formation of SV40 T-Ag with mouse p53 completely blocks the transactivation function of p53 protein . Due to the high homology between BKV T-Ag and SV40 T-Ag, a similar mechanism may hold for the BKV T-Ag as well.

It is proposed that BKV T-Ag may also transform cells through a "hit and run" mechanism. In a study by Brunner et al, it was observed that although transfection of BKV DNA into human cells resulted in a transformed phenotype, viral DNA was absent in most of the clones. This suggested that transformed cells no longer require the expression of T-Ag after a certain stages in the transformation process.

BKV T-Ag was also shown to induce a number of structural chromosomal alterations characterized by breaks, gabs, dicentric and ring chromosomes, deletions, duplications and translocations . While the molecular mechanism of this clastogenic effect of BKV on host DNA is unknown, it is thought to reside in its ability to bind to topoisomerase I or in its helicase activity in which it may induce chromosome damage when unwinding the strands of cellular DNA. Moreover, since BKV binds to tumor suppressor protein p53 and inactivates its function, this may lead to survival of DNA-damaged cells and increase their probability to transform and acquire immortality. As a result, the clastogenic and mutagenic activities of BKV may disturb the crucial function of the genes that are important for the maintenance of genomic stability such as oncogenes, tumor suppressor genes and DNA repair genes.

Detection of BKV DNA in a variety of human tumors and tumor cell lines during the late 1970's prompted researchers to further investigate the possible association of BKV with the induction of a variety of human tumors . Since BKV exhibits a specific oncogenic tropism for the ependymal tissue, endocrine pancreas, and osteosarcomas in rodents , investigators primarily focused on the characterization of such tumors in humans for the detection of BKV genome. Southern blot hybridization studies showed that 4 out of 9 (44%) human tumors of the pancreatic islets and 19 out of 74 (26%) of human brain tumors contained BKV DNA in a free, episomal state . BKV was even rescued from some of the tumors by transfection of human embryonic fibroblasts with tumor DNA.

The detection of BKV DNA was also reported by Dorries et al, in 46% of brain tumors of the most common histotypes . In this particular study, BKV DNA was found to be integrated into the chromosomal DNA. Human tumors associated with immunocompromised conditions were also analyzed by Southern blotting and it was shown that BKV DNA was associated with Kaposi's sarcoma with low frequencies (20%) .

Recently, tumor cell lines, normal and neoplastic human tissues were investigated for the detection of BKV by PCR methods utilizing specific primers for the early region of BKV DNA. The nucleotide sequence analysis of PCR products from these studies revealed the presence of BKV specific sequences in several brain tumor samples: one osteocarcinoma, two glioblastoma cell lines, one normal brain tissue and one normal bone tissue specimen . Even the expression of the early region of the BKV was demonstrated by Northern blotting of RT-PCR studies in some of the samples in those studies. The presence of BKV DNA was also investigated in several different tumors including urinary track tumors, in carcinomas of the uterine cervix, vulva, lips and tongue . However, data obtained from such studies were inconclusive because the percentage of positive samples in these neoplastic tissues of the urinary and genital tracks and of the oral cavity were similar to that detected in the corresponding normal tissues . BKV DNA was shown to be present in Kaposi's sarcoma (KS) cases in high percentages suggesting that BKV may be an important co-factor in KS .

SV40 is the most extensively studied polyomavirus. Its small genome size was exploited as a model system to study transcription and replication for more complex eukaryotic systems. Characteristic cytopathic vacuolization effects caused by SV40 in African green monkey cells led to the recognition and isolation of the virus in 1960 by Sweet and Hilleman, . Apparently SV40 was introduced into the human population through widespread use of contaminated poliovaccines. Contamination occurred during the vaccine preparation process because the early poliovaccines were prepared in primary cultures of kidney cells derived from rhesus monkeys, which are often naturally infected with SV40. As described above, SV40 genome is very similar to the other polyomaviruses, BKV and JCV, in structure containing regulatory and coding regions. Coding regions encode regulatory (small t and large T antigens, agnoprotein) and structural capsid proteins (VP-1, VP-2 and VP-3). The regulatory region of SV40, like JCV and BKV, contains the origin of DNA replication and promoter/enhancer elements which are targets for transcription factors. SV40's genome shows significant sequence homology to BKV and JCV at the coding regions, however, more divergent sequences lie within its regulatory region.

Following its discovery, SV40 was tested for its ability to induce tumors in experimental animals and to transform a variety of cell types from different species in tissue culture. Particularly, studies with Syrian hamsters showed the ability of SV40 to induce a variety of tumors in experimental animals . Such observations raised a question whether SV40 is involved in human carcinogenesis because SV40 was shown to establish infections in humans . Injection of SV40 DNA into hamsters resulted in a variety of tumors depending on the site of injection. For example, injection of SV40-infected rhesus monkey kidney cells into newborn hamsters induced sarcomas at the site of inoculation . Intravenous injection of SV40 into weanling hamsters resulted in lymphocytic leukemia, soft tissue sarcoma, osteosarcoma and lymphoma . Intracranial injection of SV40 into both newborn hamsters and Mastomys produced ependymomas . Mesontheliomas were induced upon injection of SV40 into the intrapleural region of weanling hamsters .

A variety of cell types have been used to characterize the transforming properties of SV40 including humans, hamsters, mice, rats, guinea pigs and cattle . It turned out that not every cell is permissive to infection of SV40. Monkey cells are considered to be permissive to SV40 infection. Mouse cells are nonpermissive, and human cells are considered to be "semipermissive" to SV40 infection. It was observed that in nonpermissive cells, the viral genome is often found to be integrated into the host genome and the integration is not directed to any specific site . The cellular transformation and immortalization are the consequence of nonlytic infection of the host cells. Viral oncogenic proteins are generally expressed continuously during that period perhaps to maintain the cells in the transformed state. The exact mechanism of cell transformation and immortalization is unknown. However, it appears that viral onco-protein, T-Ag, targets primarily tumor suppressor and key cell cycle regulator proteins, such as p53 and pRb, which inactivates their function and results in deregulation of cell cycle progression.

SV40 T-Ag is a multifunctional oncoprotein that possesses several defined functional domains and has been shown to play a critical role in cell transformation and tumor induction . Figure 4 schematically illustrates different functional domains of SV40 large T antigen. The amino terminus of the T-Ag contains two distinct domains important in cell transformation. The far amino terminus of T-Ag includes the J domain involved in proper folding of protein complexes. This region shares 82 amino acid residues with small t antigen. The second region of the amino terminus of T-Ag mediates the binding to pRb and the pRb family members p107 and p130 . Although the function of p107 and p130 in cell cycle regulation remains unclear, the mechanism of action of tumor suppressor protein pRb at the G1 checkpoint has been well demonstrated. It forms an inactive complex with a transcription factor E2F and arrests cells at the G1 phase of cell cycle. When specific cyclin dependent kinases phosphorylate Rb, it releases transcription factor E2F which in turn transactivates S phase specific gene promoters and causes the cell to progress into S phase. When bound to Rb, T-Ag inactivates the regulatory function of pRb which allows unscheduled S-phase entry thereby establishing favorable conditions for cellular transformation . T-Ag also targets another tumor suppressor protein, p53, which plays a critical role in cell cycle progression at the G1 checkpoint and induces apoptosis when overexpressed in cultured cells . A possible mechanism by which p53 regulates the genomic stability is through the induction of apoptosis in DNA damaged cells before potentially oncogenic events deregulate cell cycle progression. p53 is found mutated or lost in up to 50% of all human cancers which emphasizes the importance of its functional loss in carcinogenesis (. SV40 T-Ag possesses two p53 binding sites near its carboxy-terminal end. By binding to p53 at these sites, T-Ag inhibits p53-mediated activities including arresting cells that have mild DNA damage in G1 or G2/M phases of the cell cycle for DNA repair and eliminating the cells that has extensive DNA damage by apoptosis. Under these circumstances, the cells with damaged DNA go through the cells cycle stages without DNA repair which results in accumulation of cellular mutation and increased genomic instability that can lead to cancer.

T-Ag, in addition to targeting cellular tumor suppressor proteins, also targets nuclear acetylases including CREB-binding protein (CBP), P/CAF and p300. These regulatory proteins function as cofactors and play important roles in transcription and posttranslational modification of cellular tumor suppressor proteins. T-Ag interacts with these proteins through multiple regions and inactivates their important cellular functions. This is also thought to contribute to deregulation of cell cycle progression.

Small t antigen of SV40, which is produced by alternative splicing of early transcripts, was shown to form complexes with the regulatory subunit of PP2A. This association appears to inhibit the function PP2A which inturn leads to more phosphorylated and increased kinase activity of several cellular kinases including MAP kinase and its kinase ERK, Jun N-terminal kinase (JNK) and a key ion transporter, the Na/H antiporter .

Figure 4. Schematic representation of functional domains of SV40 large T-antigen. Approximate minimal regions of T-antigen that retain binding activity to polymerase a -primase (Pola ), tumor suppressor proteins Rb and p53, human heat shock protein 70 (hsc70) and coactivators p300 and CBP are illustrated. DNA binding domain, ATPase activity domain, nuclear localization signal (NLS) domain, helicase domain, host range domain, Zn finger domain, and J domain are also depicted.

It is also believed that small t antigen antagonizes T-Ag-induced cellular apoptosis and thereby contributes to more efficient transformation of rat embryo fibroblasts . Transgenic animals created with a small t antigen deletion mutant of SV40 genome consistently developed tumors in highly mitotic tissues relative to wild-type virus indicating that small t antigen contributes to large T-Ag-mediated transformation of resting cells.

The detection of SV40 in a metastatic melanoma patient by Soriano et al, in 1974 was the first observation that links the association of SV40 with human cancers. The virus was isolated from a lung metastasis and viral T-Ag and capsid proteins were detected in lung, liver and muscle metastasis but not in normal tissue. Since then, numerous reports have been published regarding a possible link between SV40 and human tumors. SV40 genome and the expression of T-Ag were detected by PCR, DNA hybridization, DNA sequencing and immunofluorescence techniques in a variety of human tumors and nontumor tissues including mesotheliomas , brain tumors , and other human tumors and nontumor tissues including osteosarcomas, AIDS-related lymphomas, peripheral blood cells, kidney tissue from pediatric renal transplant patients and non-Hodgkin's lymphomas .

A large number of reports have described the association of SV40 with malignant mesothelioma and yet asbestos, an environmental carcinogen is believed to be the predominant cause of mesotheliomas. Development of malignant mesotheliomas (up to 20%) in patients with no known asbestos exposure raised a controversial case of whether asbestos can be considered as the only causative agent of fatal mesotheliomas or there are other factors or co-factor, such as SV40, that play a role in the development of such tumors. Many studies have repeatedly linked the association of SV40 with mesothelioma. A recent multi-laboratory study by Testa et al, confirmed the presence of SV40 sequences in frozen mesothelioma samples by PCR, DNA hybridization and/or DNA sequencing . The complex formation between T-Ag with p53 and T-Ag with retinoblastoma family members, including pRb, p107 and p130, was also demonstrated by co-immunoprecipitation assays . Some studies suggested that a relatively higher susceptibility of mesothelial cells to SV40 infection maybe a part of the determining factor in development of mesotheliomas. Bochetta et al, compared the rate of transformation of SV40-infected mesothelial cells with that of human fibroblasts in a tissue culture system and the results were striking . Mesothelial cells were found to be 1000 times more susceptible to transformation upon SV40 infection than human fibroblast cells. This may partially offer an explanation for the relationship between SV40 and human mesotheliomas.

There are also now a number of studies describing the association of SV40 with human brain tumors. Experimental animal studies showed that SV40 is oncogenic in neural tissues when injected, for example, into the newborn hamsters and SV40 was shown to be capable of transforming primary human astrocytes in culture . SV40 genome and its gene products were detected by PCR or Western blotting in a variety of brain tumors including glioblastomas, gliomas, gliosarcomas, medullablastomas, meningiomas, pituitary adenomas and oligodendromas . Even a complex formation of T-Ag with p53 and T-Ag with pRb was demonstrated by co-immunoprecipitation assays suggesting that T-Ag targets common pathways in different tumors.

We have briefly reviewed recent developments regarding the tumor inducing aspects of polyomaviruses JCV, BKV and SV40. We have learned much about the molecular mechanisms underlying the cell transformation process induced by the oncogenic protein of each virus, large T antigen. However, many questions still remain unanswered as to how large T antigen can perturb the normal cell cycle progression and eventually cause cell transformation and immortalization. Further research is required to understand the molecular mechanism(s) of cell transformation, and polyomaviruses offer an excellent model system to study many aspects of this process. This in turn may help us to understand the foundation of human cancers.

We would like to thank past and present members of the Center for Neurovirology and Cancer Biology for their insightful discussion and sharing of ideas. We particularly appreciate Jessica Otte's operational efforts in our laboratory. She is the technical manager of the Center for Neurovirology and Cancer Biology. This work was supported by National Institutes of Health grants to M. S. and K. K.

Dr. Mahmut Safak