Infectious Agents and Cancer

HPV and Cervical Cancer

Cervical cancer is the third most common malignancy among women worldwide, with approximately 530,000 newly diagnosed cases each year and about 275,000 deaths annually. About 80% of cervical cancer occurs in developing countries, where it is frequently the most common cancer of women, accounting for as many as one quarter of female cancers. It occurs less frequently in developed countries. In the United States, there are about 12,000 newly diagnosed cases annually, and about one third of these women will die of their malignant disease. The incidence of cervical cancer in the United States varies considerably among ethnic and socioeconomic groups, with the rate among African American women being about twice that of White women.

Most cervical cancers develop in the transformation zone, the region of the cervix where the columnar cells of the endocervix adjoin with the stratified squamous epithelium of the exocervix. About 85% of cervical cancers are squamous cell cancers, the remainder being adenocarcinomas and small-cell neuroendocrine tumors. The progression of normal cervical epithelial cells to malignant squamous cell carcinomas typically occurs through a series of dysplastic changes over a time span of many years, a process that is the basis of the Pap smear screening program. The histologic classifications of cervical intraepithelial neoplasia (CIN) grades 1, 2, and 3 correspond, respectively, to mild dysplasia, moderate dysplasia, and severe dysplasia or carcinoma in situ. Because of the long interval for the progression of cervical dysplasia to invasive cancer, Pap smear screening programs can identify the vast majority of premalignant lesions for appropriate treatment, thereby preventing the development of most cases of cervical cancer in countries with screening programs. Most CIN lesions do not progress to cancer but resolve; the lesser the degree of dysplasia, the more likely the lesion is to resolve.

Cervical cancer had been recognized for decades as linked to a sexually transmitted agent, long before sexually transmitted HPV infection was implicated in its pathogenesis. Venereal transmission of a carcinogenic factor with a long latency had been suggested by the early epidemiologic studies. Sexual promiscuity, an early age of onset of sexual activity, and poor sexual hygiene conditions were identified by these studies as risk factors in women for cervical carcinoma. The counterpart to cervical cancer in the male is penile cancer, because there is a correlation between the incidence rates of these two cancers in different geographic areas. Compelling evidence linking an HPV infection with cervical carcinoma followed the observation that some of the morphologic changes characteristic of cervical dysplasia seen on Pap smears were due to a papillomavirus infection. The cell with its characteristic perinuclear clearing and abnormally shaped nucleus that is diagnostic for a cervical papillomavirus infection is the koilocyte. The presence of papillomavirus particles, papillomavirus-specific capsid antigens, and HPV DNA within the cervical preneoplastic lesions provides confirmation of the HPV etiology of cervical dysplasia.

Harald zur Hausen and his colleagues identified the first papillomavirus DNAs, HPV16 and HPV18, in cervical cancer tissues in the 1980s. Using HPV DNAs as probes under conditions of reduced hybridization stringency, most cervical carcinomas were shown to harbor these or related HPV DNAs. Subsequent studies led to the identification of approximately 40 different HPVs, mostly from the alpha genus, associated with genital tract lesions, a subset of which are associated with human cervical cancer. In addition to HPV16 and HPV18, which account for approximately 70% of human cervical cancers, HPV types 31, 33, 35, 45, 52 and 58 account for a total of 95% of HPV-positive cancers. DNAs from these same HPV types are found in other human genital carcinomas, including penile carcinomas, some vulvar carcinomas, and some perianal carcinomas, as well as in the precancerous intraepithelial lesions of each of theses sites (PIN, VIN, and PAIN).

The genital-tract–associated HPVs have been classified as either high risk or low risk based on whether the lesions with which they are associated are at significant risk for malignant progression. The low-risk viruses such as HPV6 and HPV11 are associated with venereal warts, lesions that only rarely progress to cancer. The high-risk viruses such as HPV16 and HPV18 are associated with CIN and cervical cancer. The other high-risk viruses include HPV31, HPV33, HPV35, HPV39, HPV45, HPV51, HPV52, HPV56, HPV58, HPV59, HPV68, and HPV82. Virtually all cases of CIN3 and cervical cancer contain a high-risk HPV DNA. HPV-positive cervical cancers and cell lines derived from HPV-positive cervical cancer tissues often, but not always, contain integrated viral DNA. In those cancers in which the viral DNA is integrated, the pattern of integration is clonal, indicating that the integration event preceded the clonal outgrowth of the tumor. Integration of the viral DNA does not occur at specific sites in the host chromosome, although in some cancers the HPV DNA has integrated in the vicinity of known oncogenes. For instance, in the HeLa cell line (which is an HPV18-positive cervical carcinoma cell line), integration of the HPV18 genome is within approximately 50 kilobases of the c-myc locus on human chromosome 8. It is possible that such an integration event might provide a selective growth advantage to the cell and thus might contribute to neoplastic progression.

The Role of HPV in Cervical Cancer

In cervical cancers, only a subset of the viral genes is expressed, and there is no production of virus by the cancer cells. The integration of the viral genome appears to play an important role leading to the deregulated expression of the viral E6 and E7 genes. The E6 and E7 genes are invariably expressed in HPV-positive cervical cancers. Integration of the HPV genome into the host chromosome in the cancers often results in the disruption of the viral E1 or E2 genes. Because HPV E2 is a viral regulatory factor that negatively regulates expression of the E6 and E7 genes, the disruption of E2 results in the derepression of the E6/E7 promoter, leading to deregulated expression of E6 and E7. Indeed, the introduction of E2 into cervical cancer cell lines results in the induction of cellular senescence by repressing E6 and E7 expression.

The E6 and E7 genes of the high-risk genital-tract–associated HPVs function as oncogenes. Expression of E6 and E7 together is sufficient for the efficient immortalization of primary human cells, most notably primary human keratinocytes, the normal host cell for the human papillomaviruses. In contrast to the immortalization properties of the HPV16 and HPV18 E6 and E7 proteins, the E6 and E7 proteins encoded by the low-risk viruses are either inactive or only weakly active in the same assays.

The major cellular targets for E6 and E7 are the tumor suppressor proteins p53 and pRB, respectively. E6 and E7 are, however, polyfunctional proteins and have many other biochemical activities and biologic properties that may be relevant to their activities in cervical carcinogenesis. A common theme among the small DNA tumor viruses (i.e., the polyomaviruses, the adenoviruses, and the cancer-associated HPVs) is that the immortalization and transformation properties of their encoded oncoproteins are in part due to their interactions with critical cellular regulatory proteins ( Figure 6-3 ). The HPV E7 proteins share some amino acid sequence similarity to adenovirus E1A and with portions of the SV40 large T antigen, in regions that are critical for the transformation activities of these oncoproteins. These regions of amino acid sequence similarity shared by these viral oncoproteins specify the binding to the product of the retinoblastoma tumor suppressor gene, pRB, and the related pocket proteins p107 and p130. Studies have established that a major component of the transformation activities of these viral oncoproteins is due to their respective abilities to complex and functionally inactivate pRB and the related pocket proteins. The binding of these viral oncoproteins to pRB, p107, and p130 leads to cellular proliferation rough the activation of genes under the control of the E2F family of transcription factors. The transcriptional activities of the E2F family of transcription factors are modulated by pRB and the other pocket proteins. When complexed with E2F proteins, they act as transcriptional repressors, and when dissociated from the pocket proteins by E7, E1A, or SV40 T-antigen, the E2F proteins function to activate transcription of their target genes. In the normal life cycle of the papillomaviruses, the binding of E7 to pRB is essential for the activation of the cell-cycle DNA replication machinery in differentiated keratinocytes that had otherwise exited the cell cycle. The small DNA tumor viruses, including HPV, depend on the host cell DNA replication machinery for the replication of their viral genomes. Because this machinery is only expressed in the S phase of the cell cycle, these viruses must stimulate cellular proliferation and drive the cell into the S phase in order to replicate the viral genomes. In the case of HPV, this occurs through E7 binding pRB, freeing up the E2F family of transcription factors.

A number of genetic studies indicate that E7 binding to pRB and its related pocket proteins is not sufficient to account for its immortalization and transforming functions, indicating that there are additional cellular targets and activities of E7 that are relevant to cellular transformation. Indeed, a large number of putative cellular targets for E7 have been identified using a variety of biochemical approaches and proteomic approaches. The physiologic relevance of many of these interactions is not yet clear, but some of these targets appear to be relevant to cancer. For instance, E7 can interact with cyclin-dependent kinase inhibitors. Like Ad E1A, HPV16 E7 interacts with and abrogates the inhibitory activity of p21(cip1) and p27(kip1) and thus has effects on cell cycle progression and keratinocyte differentiation. In addition, high-risk HPV E7 can cause genomic instability in normal human cells. HPV16 E7 induces G1/S and mitotic cell cycle checkpoint defects and uncouples synthesis of centrosomes from the cell division cycle. This causes formation of abnormal multipolar mitoses, leading to chromosome missegregation and aneuploidy. Moreover, there is an increased incidence of double-strand DNA breaks and anaphase bridges, suggesting that in addition to numerical abnormalities, high-risk E7 proteins also induce structural chromosome aberrations. Abnormal centrosome duplication rapidly results in genomic instability and aneuploidy, one of the hallmarks of a cancer cell. This activity is therefore likely to be functionally relevant to the contribution of high-risk HPVs to malignant progression.

The immortalization/transformation properties of the E6 protein were first revealed by studies using primary human genital squamous epithelial cells. Efficient immortalization of primary human cells by HPV16 or HPV18 requires both the E6 and E7 genes. Like SV40 large T antigen and the 55-kDa protein encoded by adenovirus E1B, the E6 proteins of the high-risk HPVs can complex with p53. The interaction of E6 with p53 is not direct but is mediated by a cellular protein, called the E6 - a ssociated p rotein (E6AP). E6AP is a ubiquitin protein ligase and, in the presence of E6, directly participates in the ubiquitylation of p53. Multiubiquitylated p53 is then recognized and degraded by the 26S proteasome. Consequently, the half-life and level of p53 are low in E6-immortalized cell lines and in HPV-positive cancers. Through its ubiquitylation of p53, HPV 16 E6 can abrogate the transcriptional activation and repression properties of p53 and disrupt the ability of wt p53 to mediate cell-cycle arrest in response to DNA damage. The p53 protein can sense DNA damage and prevent the replication of mutated DNA through its transcriptional activation of the p21 cyclin-dependent kinase inhibitor. Thus, the functional abrogation of p53 by high-risk HPV E6 results in decreased genomic stability and accumulation of DNA abnormalities in high-risk HPV E6–expressing cells. Hence, E6 can be directly implicated in the establishment and propagation of genomic instability, a hallmark in the pathology of malignant progression of cervical lesions.

The development of centrosome abnormalities and aneuploidy, two important related pathologic processes, appears to be initiated before viral DNA integration and may contribute to this process. High-risk HPV can induce abnormal centrosome duplication, which can result in genomic instability and aneuploidy. The deregulation of this mitotic event appears to depend on both E6 and E7, with the latter protein being most responsible for the effect. Indeed, the deregulated viral oncogene expression may result in chromosomal instability and aneuploidy, enhancing the likelihood of viral DNA integration.

A number of additional cellular targets have now been identified for the high-risk E6 proteins in an attempt to define additional p53-independent cellular targets. The reader is referred to the current edition of Fields Virology for a more comprehensive discussion of these additional activities, some of which may be relevant to the role of E6 in cervical carcinogenesis. Two activities are of particular importance, however, and are discussed here. The first is the binding to cellular PDZ domain–containing proteins. Interestingly, the high-risk E6 oncoproteins contain an X-(S/T)-X-(V/I/L)-COOH motif at the extreme C terminus that can mediate the binding to cellular PDZ domain–containing proteins. This motif is unique in the high-risk HPV E6 proteins and is not present in the E6 proteins of the low-risk HPV types. E6 serves as a molecular bridge between these PDZ domain proteins and E6AP, facilitating their ubiquitylation and mediating their proteolysis. Among the PDZ domain proteins implicated as E6 targets are hDlg, the human homologue of the Drosophila melanogaster Discs large tumor suppressor, and hScrib, the human homologue of the Drosophila Scribble tumor suppressor. Additional PDZ domain proteins have also been shown to be capable of binding to E6. Several PDZ-containing proteins have been shown to be involved in negatively regulating cellular proliferation. Therefore, some of the p53-independent transforming activities of the high-risk E6 oncoproteins may be linked to their ability to bind and degrade some of these PDZ motif–containing proteins.

A second important p53-independent activity of HPV16 E6 is its ability to activate telomerase in keratinocytes through the transcriptional upregulation of the rate-limiting catalytic subunit of human telomerase (hTERT). Maintenance of telomere length is an important step in cancer and can occur through the transcriptional activation of hTERT expression or through the activation of the ALT recombination pathway. Activation of hTERT is observed in most human cancers, including HPV-positive cervical cancers. The mechanism by which E6 activates the hTERT promoter has not been yet fully elucidated but could involve the direct activation of a cellular transcription factor by E6 or perhaps the E6AP-dependent degradation of a negative regulator of the hTERT promoter.

Infection by a high-risk HPV does not always cause cancer. Indeed, cancer is a rare outcome of an HPV infection, even for HPV16 and HPV18. Expression of the E6 and E7 oncogenes is not sufficient for malignant progression. The time period between infection by a high-risk HPV and the development of invasive cancer can be several decades. Thus, infection with a high-risk HPV constitutes only the initial step in cervical carcinogenesis; the genetic information carried by the virus per se is not sufficient to cause cancer. Epidemiologic studies have suggested that smoking is a risk factor for developing cervical carcinoma. The recognition that other factors are involved in the progression to cervical carcinomas suggests that papillomavirus infections may work synergistically with these other factors.

Tumor progression is, however, a complex process that involves multiple additional genetic loci. Specific chromosomal abnormalities have been detected in cervical cancer, including the loss of heterozygosity on the short arm of chromosome 3 (3p). This locus contains the FHIT (fragile histidine triad) gene, and its expression is inversely correlated with the severity of the lesion and prognosis. In addition, loss of 11q23 may involve the tumor suppressor of lung cancer gene (TSLC1), which is implicated in cell adhesion. An additional possibility is that cellular mutations or epigenetic changes could be involved in downregulating HLA antigen class I alleles and the ability of an HPV-positive cancer cell to be recognized by the host cellular immune response.

HPV and Other Cancers

The high-risk genital tract HPV types can infect other genital areas that contain stratified squamous epithelium and cause intraepithelial neoplasias and cancer. HPV DNA, usually HPV16, can be found in a subset of cancers of the vulva, vagina, and penis. Giant condyloma acuminata, also called the Buschke-Lowenstein tumor, is a low-grade, locally invasive squamous cell carcinoma that involves the external genitalia and is associated with low-risk HPV types, usually 6 or 11.

Anal cancer is also associated with high-risk HPV infection, and the rate of anal HPV infection appears to be similar to that of cervical infection, although anal HPV infection has been studied less systematically than cervical infection. As with cervical cancer, high-risk HPV can be found in most anal cancers, usually HPV16; most anal cancers arise in the transition zone between columnar and squamous epithelium. The risk of anal cancer in the general population appears to be much lower than for cervical cancer, and the incidence of anal cancer in women is less than one tenth that of cervical cancer. The risk of anal cancer among individuals who are human immunodeficiency virus (HIV)-positive is much greater than in the general population, with especially high rates for HIV-positive male homosexuals.

HPV is linked to some head and neck cancers, although not to the majority of the cancers in this region. HPV16 accounts for about 90% of the HPV-positive tumors. Most of these HPV-associated cancers are located in the oropharynx, which includes the tonsils, tonsillar fossa, base of the tongue, and soft palate. It is not understood why the HPV-positive tumors preferentially develop in the oropharynx. In the United Sates, the incidence of these oropharyngeal cancers, which usually develop at a younger age than the HPV-negative cancers, increased more than threefold between 1988 and 2004. Genital-oral sex may be a risk factor for these tumors, and the risk of HPV infection and cigarette smoking may be more than additive. The HPV-positive tumors tend to have a characteristic basaloid pathology and share many molecular features with those of HPV-positive anogenital tumors. The tumors usually have integrated HPV DNA expressing E6 and E7. Their p53 and pRB genes are wild type, and the vast majority of them expresses p16, in contrast to the HPV-negative tumors, which tend to have mutant p53 and to be p16-negative. There is thus far, however, no clearly identifiable premalignant oropharyngeal lesion for HPV-positive tumors. HPV-positive oropharyngeal cancers carry a better prognosis than the HPV-negative ones.

Esophageal carcinomas in humans have also been reported to have some association with HPVs; however, the data as yet are not as convincing as they are with the anogenital cancers and with oropharyngeal cancers. The esophagus is lined by a squamous epithelium, and squamous cell papillomas of the esophagus have been described in humans. Additional studies seem warranted to investigate a possible role of HPV in human esophageal cancers. There have also been sporadic reports associating occasional human tumors, including colon cancer, ovarian cancer, prostate cancer, and even melanomas, with the presence of HPV DNA in the literature. In general it seems prudent to be skeptical of such reports until systematic and well-carried-out studies are confirmed in multiple laboratories.