Both viruses and carcinogenic chemicals can cause tumors by

General Principles of Virally Induced Cancers

Infectious agents, including viruses, bacteria, and parasites, are thought to be the etiologic agents in approximately 20% of human cancers. Human oncogenic viruses include hepatitis B and hepatitis C viruses (associated with hepatocellular carcinoma (HCC)), Epstein–Barr virus (EBV; associated with B cell lymphomas and nasopharyngeal and gastric carcinomas), HPVs (associated with cervical carcinoma, other anogenital cancers, and a subset of head and neck cancers), human T cell lymphotropic virus I (HTLV-1; associated with adult T cell lymphomas), and human herpesvirus type 8 (associated with Kaposi’s sarcoma and primary effusion lymphomas). The initial recognition that viruses cause cancer arose from studies of animal viruses. In the early twentieth century, transmittable agents were demonstrated to cause tumors in chickens and rabbits. The respective agents were later identified to be Rous sarcoma virus (RSV), the first-studied RNA tumor virus, and Shope papillomavirus, the first-studied DNA tumor virus. Much of our understanding of viral oncogenesis derived initially from the study of such animal viruses. In recent years, much attention has been focused on the study of human tumor viruses as etiological roles of such viruses in human cancers have been elucidated. The study of virally induced cancers has provided many basic insights into cancer, the most important of which is the identification and functional characterization of many oncogenes and tumor suppressor genes. From these collective studies, several generalizable principles of virally induced cancers can be proffered. First, the onset of virally associated cancers is characterized by long latent periods following initial infection. This suggests that the viruses alone are not sufficient to cause cancer. Also, given this long latency, oncogenic viruses must be able to persist in the host for long periods of time. Second, the cancers induced by viruses do not constitute a natural part of the viral life cycle. Rather, virally associated cancers appear to be dead-end streets for viruses – by-products of the natural infection that provide no advantage to the virus evidenced by the fact that it is uncommon for progeny virus to be produced in the associated tumors. The third principle is that viral properties that contribute to the induction of tumors often play critical roles in the life cycle of the virus. In the following sections, we describe the basic properties of RNA and DNA tumor viruses, and the major insights gained regarding how they contribute to cancer.

1. Oncogenic viruses commonly activate cyclin-dependent kinases and inactivate the checkpoint pathways.

Tumor viruses commonly inactivate the pRb and p53 pathways to force cell division and block apoptosis. For instance, two genes (E6 and E7) are essential for replication of the human papillomaviruses (HPVs) that are associated with human cervical carcinomas. The E6 and E7 proteins bind and promote the degradation of p53 and pRb, respectively. The E1A and E1B gene products of adenoviruses inactivate pRb and p53, respectively. In the case of the papillomaviruses, p53 and pRb are inactivated by a single polypeptide, the SV40 T antigen.

Some tumor viruses can also activate the G1 cyclins independently from p53. The herpesvirus associated with Kaposi's sarcoma (KSHV/HHV8) encodes a cyclin D2 homologue that forms a complex with Cdk6. This complex promotes the degradation of the CDK inhibitor, p27KIP1, and thereby lifts the inhibition of cyclin D/Cdk complexes (Ills et al., 1999; Mann et al., 1999). The HPV E7 protein, in addition to inactivating pRb, also inactivates the other major CKIs, including p21WAF1/CIP1 (Funk et al., 1997; Jones et al., 1997).

4 Reprise

Oncogenic viruses are divided into two broad classes, RNA and DNA viruses. Most RNA tumor viruses are retroviruses, which replicate via a DNA provirus generated by their reverse transcriptase. Retroviruses transform cells in three distinct ways. Acute transformation involves two different viruses, a replication-deficient transforming virus that encodes a v-oncogene (a mutated form of a cellular gene) and a nontransforming replication-competent helper virus, which enables replication of the transforming oncovirus. The resulting tumors are polyclonal. Nonacute transformation occurs when an integrated provirus acts as an insertional mutagen, up-regulating a cellular oncogene. The resulting tumors are monoclonal reflecting the inefficiency of the transforming process. Trans-acting transformation is dependent on accessory viral proteins. These trans-acting proteins are essential to viral replication and contribute to the transformation of their host cell by activating cellular proteins that drive proliferation or hinder apoptosis. The resulting tumors are oligoclonal and slow to develop.

Although a number of oncogenic retroviruses are capable of inducing tumors in animals, their relevance to human disease is mainly their role in the discovery of oncogenes. Oncogenic DNA viruses, in contrast, are an etiological factor in a significant portion of human cancers. Recent technological advancements have facilitated the identification of DNA oncogenic viruses and the types of tumors they produce, and these numbers are likely to grow.

Oncogenic DNA viruses use a great variety of transformation mechanisms. Small oncogenic DNA viruses transform via viral oncogenes that do not have cellular homologs. Large oncogenic DNA viral genomes encode oncogenes that are either similar to or distinct from cellular genes. By disabling two major tumor suppressor proteins, p53 and pRB, many oncogenic DNA viruses avoid apoptosis and ensure continued proliferation of their host cell. In addition, tumorigenic DNA viruses often induce cellular genomic instability, either actively through the inhibition of DNA damage repair, or indirectly through increased cellular proliferation.

An additional indirect mechanism of viral-associated oncogenesis is the induction of chronic inflammation. As cells proliferate to replace cells destroyed by chronic inflammation, they may acquire tumorigenic mutations. Inflammation-induced tumorigenesis is a very slow process occurring over the period of multiple decades. Both HCV and HBV induce hepatocellular carcinoma through this mechanism, and cause the majority of human hepatocellular cancers.

Prophylactic vaccines against several important oncogenic DNA viruses (such as HPV and HBV) have been developed, and there have been recent advances in the treatment of HCV infections. The recent explosion of research-driven knowledge about tumor viruses will undoubtedly lead to the prevention or cure of many virus-associated cancers.

2 Viral carcinogens

Oncogenic viruses belong to both major viral groups. Examples of DNA viruses include the infective agent of the human wart, adenoviruses, the Shope rabbit papilloma, and the polyoma virus originally isolated from mice. RNA viruses include the Rous Sarcoma of chickens, the mouse mammary carcinoma virus, the agents of leukaemia in mice and leucosis in fowls, and the recently isolated monkey sarcoma virus. Malignant transformation of cells by viral agents is to some extent a reflection of failure on the part of the virus to establish itself (although the immunologist might consider this failure to be the result of a good host defence mechanism). A well established virus grows rapidly within a cell, and kills it, thus releasing large numbers of viral particles. The carcinogenic virus is able to infect the host cell, but does not replicate rapidly enough to kill it. It appears that in the case of DNA viruses, the viral chromosome merges with the host cellular chromosomes. In some way as yet unknown, it interferes with the normal regulatory processes of the cells, which then escape the control mechanism responsible for maintaining the equilibrium between cell proliferation and cell destruction. As a result of the chromosomal merger, when the host cell divides, the daughter cells contain replicas of the virus and are themselves cancerous.

In the case of the carcinogenic RNA viruses, viral particles may modify messenger RNA, or substitute for a part of, or the whole molecule. In this way, new information for the synthesis of protein may be introduced, leading to changes in cellular metabolism, as well as viral replication. However, at the present state of our knowledge, this theory fails to explain how new genetic information is passed on to the daughter cells in a fast growing tumour. There are indications that in vitro at least, virally infected cells may occasionally revert to the untransformed state, but this appears to be an uncommon happening in vivo, where perhaps the conditions for multiplication of transformed cells are more favourable.

The problem of isolating virus from cancerous cells is an exceedingly difficult one. Recently, however, it has been shown that when virally infected cancer cells are fused with normal cells, the latent virus becomes unmasked and can be recovered from the hybrid cell, in which it is able to multiply. The importance of the technique for the detection of the viral origin of tumours can hardly be overestimated (Watkins and Dulbecco, 1967).

The carcinogenic effects of viruses may be reinforced by cocarcinogenic factors. For example in the induction of the Shope papilloma, painting the skin with tar shortens the incubation period of the virally induced tumour and increases its virulence. Cater (1951) has shown that the growth and spread of Rous sarcoma in chicks is favoured by pre-existing vitamin E deficiency. The cocarcinogenic effect of hormones on viral tumour induction has already been mentioned in connection with the Bittner mouse mammary carcinoma.

ENZOOTIC NASAL TUMOR VIRUS

The enzootic nasal tumor virus (ENTV) infects both sheep (ENTV-1) and goats (ENTV-2), and these (ENTV-1 and ENTV-2) are apparently distinct viruses that share ~95% overall amino acid similarity to Jaagsiekte sheep retrovirus. Less is known regarding the global distribution, pathogenesis, and molecular constitution of enzootic nasal tumor virus than the related virus of sheep. The enzootic nasal tumor virus can transform secretory epithelial cells of the ethmoid turbinate, a restricted region of the nasal cavity, thus infections with this virus frequently result in nasal adenocarcinomas (enzootic nasal adenocarcinoma, enzootic nasal tumor; Fig. 14.7). The condition has been experimentally reproduced in lambs following inoculation with enzootic nasal tumor virus.

Although experimental studies have shown tumor formation can occur as little as 12 weeks after virus infection, the incubation period for natural infections is typically longer. Once clinical signs manifest, they often progress rapidly to death or lead to culling of the affected animal from the herd. Clinically, affected animals may have copious nasal discharge, respiratory distress, open mouth breathing, exophthalmos, and nasal bone/cribriform plate deformities/destruction. Metastasis of the virus-induced nasal tumors has not been reported. As for Jaagsiekte sheep retrovirus, the cellular receptor for enzootic nasal tumor virus is Hyal2; however, enzootic nasal tumor virus targets epithelial cells in the nasal passages, not the lower respiratory tract. The LTR regions of enzootic nasal tumor virus genome has enhancer elements that promote virus transcription in the nasal epithelium of ruminants, but not the lung-specific enhancer elements that are characteristic of Jaagsiekte sheep retrovirus. Like Jaagsiekte sheep retrovirus, the Env protein of enzootic nasal tumor virus has itself been demonstrated to be necessary and sufficient to induce nasal adenocarcinomas.

Although the virus is present in tumor tissue and nasal secretions, enzootic nasal tumor virus, like Jaagsiekte sheep retrovirus, cannot yet be cultivated in tissue culture systems. Similarly, confirmation of enzootic nasal tumor virus infection of small ruminants requires PCR amplification of viral nucleic acid as infected animals do not mount a serologic response.

Human tumor viruses are members of distinct taxonomic groups that are causally linked to malignant diseases, which develop from transformed infected cells. They include the high-risk human papillomaviruses (HPV), a hepadnavirus (hepatitis B virus, HBV), a flavivirus (hepatitis C virus, HCV), a retrovirus (human T-cell leukemia virus type 1, HTLV-1), and two herpesviruses (Epstein–Barr virus, EBV; Kaposi's sarcoma-associated herpesvirus, KSHV). The most frequent virus-induced malignancies originate from epithelia as genital carcinomas (HPV16, 18, etc.), the viral hepatocellular carcinoma (HBV, HCV), or the nasopharyngeal carcinoma (EBV). A second major target of tumor virus-mediated oncogenesis are hematopoetic cells. Lymphomas in immunosuppressed individuals (EBV), Hodgkin's lymphoma (EBV), Burkitt's lymphoma (EBV), multicentric Castleman's disease (KSHV), primary effusion lymphoma (KSHV), and adult T-cell leukemia/lymphoma (HTLV-1) are diseases that develop from the lymphoid lineage. Most virally induced malignancies develop after long periods of persistent infection in a minority of the infected individuals, indicating the requirement of rare additional events in oncogenesis. Since all human tumor viruses are important or necessary for the development of particular cancers, these cancers are, in principle, preventable by protective immunization.

Cancer-Causing Agents

Cancer-causing agents can be categorized into three groups: oncogenic viruses, chemicals, and radiation. All three either independently or in combination drive the carcinogenic cycle.

2.11.1 FOXM1 and HPV16 E7

Tumor viruses contribute to 15–20% of human cancers (Butel, 2000; Carillo-Infante et al., 2007; Damania, 2007; Davey et al., 2011; Dayaram and Marriott, 2008; Fernandez and Esteller, 2010; Hoppe-Seyler and Butz, 1995; Javier and Butel, 2008; Liao, 2006; Martin and Gutkind, 2008; McLaughlin-Drubin and Münger, 2008; Moore and Chang, 2010; Parkin, 2006; Ziegler and Bounaguro, 2009; zur Hausen, 2001a,b).

E7 is a transforming oncoprotein of the high-risk human papillomavirus 16 (HPV16), a small DNA tumor virus belonging to the Papillomaviridae family (Chakrabarti and Krishna, 2003; Doorbar, 2006; Duensing and Münger, 2002; Duensing and Münger, 2004; Duensing et al., 2009; Fehrmann and Laimins, 2003; Finzer et al., 2002; Galloway and McDougall, 1996; Ganguly and Parihar, 2009; Garnett and Duerksen-Hughes, 2006; Ghittoni et al., 2010; Hamid et al., 2009; Jones and Münger, 1996; Klingelhutz and Roman, 2012; Korzeniewski et al., 2011; Lagunas-Martinez et al., 2010; Ledwaba et al., 2004; Longworth and Laimins, 2004; McCance, 2005; McLaughlin-Drubin and Münger, 2009a,b; Moody and Laimins, 2010; Münger, 2002; Münger and Howley, 2002; Münger et al, 2001, 2004; Narisawa-Saito and Kiyono, 2007; Pim and Banks, 2010; Scheffner and Whitaker, 2003; Tan et al., 2012 ; Wise-Draper and Wells, 2008; Yugawa and Kiyono, 2009; zur Hausen, 1996, 2002; Zwerschke and Jansen-Dürr, 2000). HPV16 infects the basal cell layer of mucosal epithelia in the anogenital tract and causes lesions, which have a propensity for carcinogenic progression, so that HPV16 is the causative agent of about half of all human cervix cancers (Baseman and Koutsky, 2005; Burk et al., 2009; Doorbar, 2006; Fehrmann and Laimins, 2003; Feller et al., 2009; Hebner and Laimins, 2006; Jayshree et al., 2009; Kisseljov et al., 2008; Lehoux et al., 2009; Lizano et al., 2009; Longworth and Laimins, 2004; McLaughlin-Drubin and Münger, 2009b, 2012; Moody and Laimins, 2010; Münger, 2002; Münger et al, 2004; Narisawa-Saito and Kiyono, 2007; Roden and Wu, 2006; Schiffman et al., 2007; Stanley et al., 2007; Stubenrauch and Laimins, 1999; Subramanya and Grivas, 2008; Woodman et al., 2007; zur Hausen, 1996, 2002, 2009).

The viral oncoprotein HPV16 E7 inactivates the tumor suppressor RB by disrupting E2F-RB complexes and by triggering the degradation of RB via the ubiquitin-proteasome pathway (Blanchette and Branton, 2009; Chakrabarti and Krishna, 2003; D’Abramo and Archambault, 2011; Duensing and Münger, 2004; Duensing et al., 2009; Felsani et al., 2006; Ganguly and Parihar, 2009; Ghittoni et al., 2010; Hamid et al., 2009; Helt and Galloway, 2003; Jones and Münger, 1996; Klingelhutz and Roman, 2012; Korzeniewski et al., 2011; Lee and Cho, 2002; Lehoux et al., 2009; Liu and Marmorstein, 2006; Lizano et al., 2009; McLaughlin-Drubin and Münger, 2009a,b; Moody and Laimins, 2010; Münger, 2003; Münger and Howley, 2002; Münger et al., 2001, 2004; Pim and Banks, 2010; Randow and Lehner, 2009; Scheffner and Whitaker, 2003; Tan et al., 2012; Yugawa and Kiyono, 2009; Zwerschke and Jansen-Dürr, 2000). The analysis of a RB mutant, which was selectively defective for binding E7, demonstrated that the inactivation of RB is required for nearly all acute in vivo effects of E7 (Balsitis et al., 2005).

E7 binds directly to FOXM1c (Lüscher-Firzlaff et al., 1999) and increases the transcriptional activity of FOXM1c (Lüscher-Firzlaff et al., 1999; Wierstra and Alves, 2006d, 2008). In particular, E7 enhances the transactivation of the human c-myc promoter by FOXM1c (Wierstra and Alves, 2006d, 2008). This positive effect of E7 on the FOXM1c-mediated transactivation of the c-myc promoter seems to require the binding of E7 to RB (Wierstra and Alves, 2006d, 2008) because the E7 mutant E7ΔRB, which retained the FOXM1c interaction domain (Lüscher-Firzlaff et al., 1999), but did not bind to RB (Münger et al., 1989), failed to increase the transactivation of the c-myc promoter by FOXM1c (Wierstra and Alves, 2006d, 2008).

In addition to enhancing the transcriptional activity of FOXM1c (Lüscher-Firzlaff et al., 1999; Wierstra and Alves, 2006d, 2008), E7 appears to upregulate the foxm1 mRNA expression (Fig. 6.1) because overexpression of E7 increased the foxm1 mRNA level in senescent HMFs (human mammary fibroblasts), which had first been immortalized with hTERT (human telomerase reverse transcriptase) and SV40 large T antigen and were then induced to senesce by inactivation of SV40 large T (Rovillain et al., 2011). Accordingly, the FOXM1 expression level correlated with the E7 expression level in HPV16-infected invasive cervical carcinoma (Rosty et al., 2005).

The oncoprotein E7 is essential for HPV16-induced cervical carcinogenesis (see above), suggesting that its target FOXM1c could be implicated in the pathology of cervix cancer.

Indeed, FOXM1 is overexpressed in cervical SCC (Table 6.1), namely, in early-stage tumors and in particular in late-stage tumors (Chan et al., 2008; He et al., 2012; Pilarsky et al., 2004; Rosty et al., 2005; Santin et al., 2005). FOXM1 is also highly expressed in cervical cancer cell lines (HeLa, CaSki, SiHa, HCC94, C33A) (Chan et al., 2008; He et al., 2012). With respect to clinicopathological features of cervical SCC, the FOXM1 expression level correlated with tumor stage (Chan et al., 2008; He et al., 2012), recurrence, and poor prognosis (He et al., 2012).

Furthermore, FOXM1 is part of the CCPC (cervical cancer proliferation cluster), which was defined in HPV16/HPV18-positive invasive cervical carcinoma (Rosty et al., 2005). The expression level of the CCPC genes was positively correlated with the E7 mRNA level and to a lower extent with the HPV DNA load (Rosty et al., 2005). The average expression level of all CCPC genes was higher in tumors with an unfavorable outcome (i.e., early relapse) than in tumors with a favorable course so that it may be indicative of poor disease prognosis (Rosty et al., 2005).

In accordance with a possible implication of FOXM1 in cervix cancer, FOXM1 enhanced the anchorage-independent growth of cervical cancer cells (C33A, SiHa) on soft agar (Chan et al., 2008), the tumorigenicity of cervical cancer cells (HCC94, SiHa) in nude mice (measured as volume and weight of xenograft tumors) (He et al., 2012), and the migration and invasion abilities of cervical cancer cells (HCC94, SiHA) (He et al., 2012), which in each case were significantly increased by FOXM1 overexpression (C33A, HCC94), but significantly decreased by shRNA-mediated knockdown of FOXM1 (SiHa) (Chan et al., 2008; He et al., 2012).

E Direct Approach via a Possible Tumorigenic Viral Protein

Wound tumor virus apparently does not induce neoplasia in its insect vectors or transformation in cultured vector cells. Therefore, there is a possibility that a viral tumorigenic protein might act as a repressor and combine with host plant DNA at a specific site but not with the DNA of the insect vector.

It has been demonstrated that the repressor of the lac gene of Escherichia coli is a protein estimated to be 150,000 to 200,000 daltons (Gilbert and Müller-Hill, 1966). The repressor responsible for the maintenance of lysogeny of λ phage and for the suppression of replication and function by one or more genes of superinfecting related phages is a protein having subunits of ˜27,000 daltons (Ptashne, 1967a; Maniatis and Ptashne, 1976). Both the λ repressor protein (Ptashne, 1967b) and the lac repressor protein (Gilbert and Müller-Hill, 1967) were shown to be highly specific in their binding to the operator DNA sites. The operator for the repressor of λ phage has 17 base pairs included in 2 turns of the DNA helix. The binding is tight though reversible, occurs with native DNA but not with denatured DNA, and occurs only under certain determined conditions in solution. Both repressors exert a negative control, that is, they regulate by interfering with the binding of RNA polymerases and so hinder or prevent transcription of RNA from genes adjacent to or between binding sites.

The twelve proteins encoded by the WTV genome are estimated to have different molecular weights between 160,000 and 19,000 (Reddy and Black, 1977; Nuss and Peterson, 1980). All but one of them are large enough to operate like the two repressors identified above. They can be labeled with radioisotopes in vivo and in vitro (Nuss and Peterson, 1980), and there is a possibility that the putative tumorigenic protein, and consequently its tumorigenic genome segment, can be identified by specific binding of the protein to appropriate DNA followed by its release after making appropriate changes in the solution in which the repressor-operator complex is suspended. Any specificity of differential complexing to plant and vector DNA's could be investigated further by using DNA from at least two host insect vectors and a number of the more than fifty susceptible plant species. Also DNA from only the tumorous cells of the overgrowths on plants infected with WTV might already be combined with a viral tumorigenic polypeptide and therefore might not combine with additional viral protein. This possible difference could also be examined between the DNA of the nontumorous cortical cells of many stem tumors and the DNA of the central neoplastic mass of cells enveloped by the cortex. Other phytoreoviruses might be expected to have a similar viral protein that combines with DNA from their grass hosts, but the nontumorigenic RDV might lack a protein that combines similarly with DNA of the rice plant.

In studies such as the above, highly purified WTV can be obtained in which ˜25–33% of the virions are still active (Reddy and Black, 1973a). This means that the seven structural proteins of WTV can be obtained for experimental tests virtually free of other proteins [small quantities of purified WTV can be efficiently collected from solution on suitable filters (Kimura and Black, 1976)]. The nonstructural proteins of WTV are not so readily obtained free of other proteins (Nuss and Peterson, 1980), but theoretically the binding and release of a repressor protein from its operator DNA might provide a means of both identifying and purifying such a protein.

Study of wound tumor disease along these lines might therefore provide an opportunity to learn how a foreign protein coded by viral dsRNA regulates eucaryotic cells to develop neoplasia.