**3.1 HPV and head and neck cancer**

Approximately 15% of malignant diseases are caused by infectious agents. HPV can be frequently found in oral carcinomas, especially tonsillar cancer. A group of HPV-infected tumours shows clear signs for a virally induced transformation process: high-risk HPVs can be detected in all tumour cells, the viral oncogenes E6 and E7 are constantly expressed and lead to upregulation of cellular p16(INK4a), a cyclin-dependent kinase inhibitor. The patients frequently lack typical risk factors associated with head and neck cancers such as drinking and smoking. An association of herpes viruses with head and neck cancer has been for long time suspected and there is good evidence there, but only for a relationship of EBV nasopharyngeal carcinoma. HPV aetiology is now accepted for up to 20% of head and neck cancer. However, the relationship between carcinogenesis and HPV infection is not as clear as with cervical cancer. HPV is not detectable in many head and neck cancers, and it is frequently detected in normal oral mucosa, which is why HPV infections were in head and neck cancer is often regarded only as an accompanying infection. Initial investigations into a causal role for HPV in the aetiology of head and neck lesions relied upon electron microscopy (EM) or immunohistochemical staining. Human papillomavirus virions could be identified by EM from specimens of papillomas (Frithiof & Wersall 1967), fibromas (Gross et al. 1982), verruca, condyloma acuminatav (Shaffer, Reimann & Gysland 1980; Syrjanen, K. J. & Surjanen 1981), focal epithelial hyperplasia, and oral nodular leukoplakias (Jenson et al. 1982). Immunohistochemical staining has revealed the presence of HPV capsid antigens in HPV-infected cells (Loning et al. 1984). Capsid antigen, however, has rarely been

Role of the Epstein-Barr Virus ZEBRA Protein

specimens (Tung et al. 1999).

NPCs in areas endemic for NPC.

and HPV in the Carcinogenesis of Nasopharyngeal Carcinoma 51

coinfection with both viruses was observed in 34% of patients in Morocco. Tung et al. showed that among 88 fresh NPC specimens from Chinese population, coexistence of EBV

These results are in agreement with other studies reporting the same prevalence of HPV DNA in NPC cases. In fact, using the same consensus primers, HPV DNA was detected in 31 of 103 NPC samples (30%). Moreover, Krishna et al. have shown that HPV DNA was detected in 38.8% of 36 southern Indian NPC cases (Krishna et al. 2004). Tung et al. in Eighty-eight fresh tissue samples of NPC showed that HPV DNA was detected in 51% of the

With regard to HPV genotypes, HPV31 was the most frequent genotype in Moroccan NPC patients (20.8 %). The same genotype was also frequently found in tonsils and nasopharyngeal cells in western Mexico NPC cases (Lopez-Lizarraga et al. 2000). The second prevalent HPV type detected in Moroccan NPC biopsies is HPV59 (16,7%). Of interest, HPV-16 and -18, which are the most virulent genotypes associated with CC in Moroccan woman (35% to 45 %) (Khair et al. 2009), were detected in very few NPC cases (8.3 %), and similar data were reported in an Iranian study (Mirzamani et al. 2006). However, a recent study suggests that WHO-I NPC may be associated with oncogenic HPV. Oncogenic HPV was detected by in situ hybridization in half of the WHO-I NPCs but only 5% of the WHO-II/III NPCs. In addition in a HPV genotyping cohort study, oncogenic HPVs were detected equally in WHO-II/III NPCs (31%, 13/42) and nasopharyngeal controls (35%, 14/40). Tumour high-risk HPV status did not correlate with the prognosis of patients with NPC. In the high-risk HPV in situ hybridisation cohort, 14 (88%) of the 16 oncogenic HPV-positive WHO-II/III NPCs showed a unique cytoplasmic/perinuclear staining pattern, which is distinct from the typical dot/punctate nuclear staining pattern indicating HPV genome integration. In addition, oncogenic HPVs were not always retained in NPC cells during the process of metastasis (Lo et al. 2010). Therefore, considering the fact that oncogenic HPV has not been consistently detected in NPC specimens from different endemic regions, it is likely that HPV infection may not be essential in the carcinogenesis of EBV-associated WHO-II/III

and HPV DNA was observed in 42% of samples (Tung et al. 1999).

**3.3 Insights into the molecular mechanisms of HPV carcinogenesis** 

These inconsistent results are likely to reflect a difference in life cycles of the different HPV subtypes in different mucosal locations, with an associated difference in mucosal immune responses. The high risk subtypes of HPV involved in cervical carcinogenesis have been defined (Munoz et al. 2003). We speculate the HPV subtypes associated with NPC are broadly similar (but not identical) (figure 2) with those seen in cervical carcinoma. Briefly, through wounds or abrasions, the papillomaviruses infect basal epithelial cells, which are the only actively dividing cells in the epithelial layer. The viral DNA is maintained in the nuclei of infected basal epithelial cells as a low-copy-number plasmid (Stubenrauch & Laimins 1999). Squamous epithelial cells normally undergo differentiation as they move from the basement membrane towards the surface epithelium, and HPV-DNA replicates to a high copy number only in terminally differentiated cells near the epithelial surface (Stubenrauch & Laimins 1999). Similarly, the late viral genes, which encode the L1 and L2 proteins that constitute the virus particle, are expressed only in the highly differentiated cells, where infectious progeny virus is produced and released. Three critical steps can be

detected in high-grade neoplasias or invasive cancer, probably because such tissue contains limited numbers of highly differentiated squamous cell epithelial cells. Consequently, the majority of head and neck lesions that contained HPV structural antigens were either benign or precancerous. Inconsistencies in antigen detection also resulted from sampling error, variable expression or lack of HPV capsid antigens, destruction of antigens during cellular processing or long term storage, or lack of sensitivity to a particular assay (Koutsky, Galloway & Holmes 1988; Syrjanen, S. M. 1990). From published studies, the overall antigen positivity in noncancerous head and neck lesions was about 34.4% (Adler-Storthz et al. 1986; Syrjanen, K., Syrjanen & Pyrhonen 1982). Whereas light microscopy, EM, and immunohistochemistry have resulted in inconsistent or irreproducible findings, the use of DNA hybridization has revolutionized the detection of HPV DNA types in benign and malignant lesions.

New data from case–control studies suggest that HPV is an independent risk factor for oral and oropharyngeal squamous-cell carcinomas (Rosenquist et al. 2005; Schwartz et al. 1998). Moreover, a systematic review showed an overall prevalence of HPV infection of 25.9% in specimens obtained from 5046 patients with head and neck squamous-cell carcinoma that had been analyzed in 60 separate studies (Kreimer et al. 2005). Using PCR detection from 26 countries which included 5046 cases of squamous cell cancers; 2642 oral cancers, 969 oropharyngeal cancers and 1435 laryngeal cancers. HPV prevalence was 35.6% in oropharyngeal cancers, 23.5% in oral cancers and 24.0% in laryngeal cancers. Overall prevalence of HPV in HNSCC was estimated at 26%. HPV 16 was by far the commonest subtype in all types of HPV+ cancers; 86.7% of oropharyngeal, 68.2% of oral and 69.2% of laryngeal cancers. HPV 18 was next most common but found in only 1% of oropharyngeal, 8.0% of oral and 3.9% of laryngeal cancers. A more recent meta-analysis by Termine et al 2008 (Termine et al. 2008), estimated that from studies utilising only FFPE samples, the pooled prevalence of HPV detected in these HNSCC (defined as SCCs originating in the oral, pharyngeal and laryngeal cavities only) was 34.5% (Goon et al. 2009).

#### **3.2 HPV and NPC**

NPC is one of the most striking examples of human malignancies that have been found strongly associated with the EBV, and interest in HPV as a cofactor in NPC occurrences has emerged over the last few years (Lin 2009). EBV has been detected in a large proportion of patients with WHO-II/III NPC, but a significant subset of patients with WHO-I NPC are EBV negative (Hording et al. 1994; Rassekh et al. 1998; Tsai et al. 1998). High-risk HPV may contribute to the development of NPC, given HPV's acknowledged role in the pathogenesis of oropharyngeal carcinomas.

Furthermore, it has been suggested that normal human oral epithelial cells, especially nasopharyngeal cells, could be very susceptible to persistent HPV and EBV co-infections and that EBV and high-risk HPV co-infections may play an important role in the initiation of a neoplastic transformation of human oral epithelial cells (Al Moustafa et al. 2009). To date, different degrees of associations between HPV and NPC have been described, yet no conclusive data have been obtained.

Coinfection by HPV and EBV has not been well documented and the significance of the presence of both viruses in nasopharyngeal cells has not been determined. In a recent study,

detected in high-grade neoplasias or invasive cancer, probably because such tissue contains limited numbers of highly differentiated squamous cell epithelial cells. Consequently, the majority of head and neck lesions that contained HPV structural antigens were either benign or precancerous. Inconsistencies in antigen detection also resulted from sampling error, variable expression or lack of HPV capsid antigens, destruction of antigens during cellular processing or long term storage, or lack of sensitivity to a particular assay (Koutsky, Galloway & Holmes 1988; Syrjanen, S. M. 1990). From published studies, the overall antigen positivity in noncancerous head and neck lesions was about 34.4% (Adler-Storthz et al. 1986; Syrjanen, K., Syrjanen & Pyrhonen 1982). Whereas light microscopy, EM, and immunohistochemistry have resulted in inconsistent or irreproducible findings, the use of DNA hybridization has revolutionized the detection of HPV DNA types in benign and

New data from case–control studies suggest that HPV is an independent risk factor for oral and oropharyngeal squamous-cell carcinomas (Rosenquist et al. 2005; Schwartz et al. 1998). Moreover, a systematic review showed an overall prevalence of HPV infection of 25.9% in specimens obtained from 5046 patients with head and neck squamous-cell carcinoma that had been analyzed in 60 separate studies (Kreimer et al. 2005). Using PCR detection from 26 countries which included 5046 cases of squamous cell cancers; 2642 oral cancers, 969 oropharyngeal cancers and 1435 laryngeal cancers. HPV prevalence was 35.6% in oropharyngeal cancers, 23.5% in oral cancers and 24.0% in laryngeal cancers. Overall prevalence of HPV in HNSCC was estimated at 26%. HPV 16 was by far the commonest subtype in all types of HPV+ cancers; 86.7% of oropharyngeal, 68.2% of oral and 69.2% of laryngeal cancers. HPV 18 was next most common but found in only 1% of oropharyngeal, 8.0% of oral and 3.9% of laryngeal cancers. A more recent meta-analysis by Termine et al 2008 (Termine et al. 2008), estimated that from studies utilising only FFPE samples, the pooled prevalence of HPV detected in these HNSCC (defined as SCCs originating in the

NPC is one of the most striking examples of human malignancies that have been found strongly associated with the EBV, and interest in HPV as a cofactor in NPC occurrences has emerged over the last few years (Lin 2009). EBV has been detected in a large proportion of patients with WHO-II/III NPC, but a significant subset of patients with WHO-I NPC are EBV negative (Hording et al. 1994; Rassekh et al. 1998; Tsai et al. 1998). High-risk HPV may contribute to the development of NPC, given HPV's acknowledged role in the pathogenesis

Furthermore, it has been suggested that normal human oral epithelial cells, especially nasopharyngeal cells, could be very susceptible to persistent HPV and EBV co-infections and that EBV and high-risk HPV co-infections may play an important role in the initiation of a neoplastic transformation of human oral epithelial cells (Al Moustafa et al. 2009). To date, different degrees of associations between HPV and NPC have been described, yet no

Coinfection by HPV and EBV has not been well documented and the significance of the presence of both viruses in nasopharyngeal cells has not been determined. In a recent study,

oral, pharyngeal and laryngeal cavities only) was 34.5% (Goon et al. 2009).

malignant lesions.

**3.2 HPV and NPC** 

of oropharyngeal carcinomas.

conclusive data have been obtained.

coinfection with both viruses was observed in 34% of patients in Morocco. Tung et al. showed that among 88 fresh NPC specimens from Chinese population, coexistence of EBV and HPV DNA was observed in 42% of samples (Tung et al. 1999).

These results are in agreement with other studies reporting the same prevalence of HPV DNA in NPC cases. In fact, using the same consensus primers, HPV DNA was detected in 31 of 103 NPC samples (30%). Moreover, Krishna et al. have shown that HPV DNA was detected in 38.8% of 36 southern Indian NPC cases (Krishna et al. 2004). Tung et al. in Eighty-eight fresh tissue samples of NPC showed that HPV DNA was detected in 51% of the specimens (Tung et al. 1999).

With regard to HPV genotypes, HPV31 was the most frequent genotype in Moroccan NPC patients (20.8 %). The same genotype was also frequently found in tonsils and nasopharyngeal cells in western Mexico NPC cases (Lopez-Lizarraga et al. 2000). The second prevalent HPV type detected in Moroccan NPC biopsies is HPV59 (16,7%). Of interest, HPV-16 and -18, which are the most virulent genotypes associated with CC in Moroccan woman (35% to 45 %) (Khair et al. 2009), were detected in very few NPC cases (8.3 %), and similar data were reported in an Iranian study (Mirzamani et al. 2006). However, a recent study suggests that WHO-I NPC may be associated with oncogenic HPV. Oncogenic HPV was detected by in situ hybridization in half of the WHO-I NPCs but only 5% of the WHO-II/III NPCs. In addition in a HPV genotyping cohort study, oncogenic HPVs were detected equally in WHO-II/III NPCs (31%, 13/42) and nasopharyngeal controls (35%, 14/40). Tumour high-risk HPV status did not correlate with the prognosis of patients with NPC. In the high-risk HPV in situ hybridisation cohort, 14 (88%) of the 16 oncogenic HPV-positive WHO-II/III NPCs showed a unique cytoplasmic/perinuclear staining pattern, which is distinct from the typical dot/punctate nuclear staining pattern indicating HPV genome integration. In addition, oncogenic HPVs were not always retained in NPC cells during the process of metastasis (Lo et al. 2010). Therefore, considering the fact that oncogenic HPV has not been consistently detected in NPC specimens from different endemic regions, it is likely that HPV infection may not be essential in the carcinogenesis of EBV-associated WHO-II/III NPCs in areas endemic for NPC.

#### **3.3 Insights into the molecular mechanisms of HPV carcinogenesis**

These inconsistent results are likely to reflect a difference in life cycles of the different HPV subtypes in different mucosal locations, with an associated difference in mucosal immune responses. The high risk subtypes of HPV involved in cervical carcinogenesis have been defined (Munoz et al. 2003). We speculate the HPV subtypes associated with NPC are broadly similar (but not identical) (figure 2) with those seen in cervical carcinoma. Briefly, through wounds or abrasions, the papillomaviruses infect basal epithelial cells, which are the only actively dividing cells in the epithelial layer. The viral DNA is maintained in the nuclei of infected basal epithelial cells as a low-copy-number plasmid (Stubenrauch & Laimins 1999). Squamous epithelial cells normally undergo differentiation as they move from the basement membrane towards the surface epithelium, and HPV-DNA replicates to a high copy number only in terminally differentiated cells near the epithelial surface (Stubenrauch & Laimins 1999). Similarly, the late viral genes, which encode the L1 and L2 proteins that constitute the virus particle, are expressed only in the highly differentiated cells, where infectious progeny virus is produced and released. Three critical steps can be

Role of the Epstein-Barr Virus ZEBRA Protein

EBV carcinogenic pathways.

dysplasia and invasive carcinoma.

**4. Conclusion** 

and HPV in the Carcinogenesis of Nasopharyngeal Carcinoma 53

1990). Targeting p53 may be a common requirement for the replication of many types of DNA viruses (Prayitno 2006). In addition, B cells transfected with EBV latent membrane protein lost the regulatory effects of the retinoblastoma (RB) protein, and the HPV E7 transcript has been shown to immunoprecipitate the RB protein (Giovannelli et al. 2002). Thus, the functional loss of the RB protein might be one event common to both the HPV and

Fig. 2. A hypothetical model of HPV-associated NPC development. Most information has been deciphered from oral carcinogenesis; there are fewer data on the other subsites of NPC. Light cells: Normal epithelium. Cells with a white ring: Koilocytes, as a result of an active viral replication. Dark cells: Dysplastic epithelium with increased chromosomal instability reflected by increasing aneuploidy of the cells and leading to progression to high grade

EBV is associated with the development of both B-cell and epithelial cell malignancies. The capacity of EBV to transform B lymphocytes has been well documented. EBV latent proteins are known to contribute to cellular transformation. Several lines of evidence demonstrated that reactivation of the latent viral genome in EBV associated cancers can cause cancer cell death. However, the underlying molecular mechanisms are unclear. Although ZEBRA plays an important role in immunomodulation, its capacity to reprogram the hosT-cell cycle control machinery is also notified in some tumour cell lines. Therefore, gene delivery techniques might be a novel therapeutic strategy for treating EBV positive malignancies

HPV-associated nasopharyngeal carcinoma represents a distinct clinical and biological entity with many unresolved issues that will be studied in future translational, clinical

especially NPC, via the induction of lytic viral transcription in certain tumour cells.

discriminated in this model (figure 2): the conversion of a single mutated stem cell in a patch into a group of stem cells without proper growth control (field); the eventual transforming event, which turns a field into an overt carcinoma showing invasive growth and metastasis; and the development of metastasis. Both aneuploidy and the accumulation of cancerassociated genetic changes in fields are linked to the risk of malignant progression.

It has been shown that these subtypes (particularly 16) are able to transform and immortalise cells in vitro. These effects are predominantly due to the E6 and E7 oncogenes, which bind and enhance degradation of p53 and Rb tumour suppressor genes respectively. There is evidence that immortalisation of oral keratinocytes and epithelial cells occur quite readily (Park et al. 1991).

HPV integration usually leads to disruption and/or deletion of HPV E1 or E2 open reading frame (ORF), which are important for viral replication and transcription. E2 functions also as a repressor of E6 and E7 and disruption of E2 activity allows increased E6 and E7 expression, thus maintaining the immortalised phenotype (zur Hausen 2009). Integration of HPV 16 DNA also correlates with a selective growth advantage and may allow the cancerous cell to outgrow its rivals; this may be an important step in the pathway of oncogenesis (Jeon, Allen-Hoffmann & Lambert 1995). However, despite the dominance of the integrated HPV genome in terms of cervical carcinogenesis, 15-30% of cervical cancer contains HPV only in the episomal form (Watts et al. 2002). In some of these cases, investigators have found deletions in the YY1-binding sites of the LCR (long control region) of HPV 16 episomal DNA which may allow elevated activity of the E6/ E7 promoter (Dong et al. 1994). It is clear that the actual molecular pathway to cervical carcinogenesis is far from homogeneous. The situation in head and neck cancers is less than clear but heterogeneity and the existence of multiple pathways to carcinogenesis is highly likely. Koskinen et al (2003) reported that in their series of head and neck cancers, 61% were HPV DNA positive. HPV 16 was the dominant subtype, and found in 84% of HPV+ cancers. Tonsillar carcinomas have been reported to have the highest prevalence rate of HPV DNA contained within cancerous cells (51%) of all the forms of head and neck cancers (Syrjanen, S. 2004). Mellin et al 2002 reported that all 11 cases of HPV+ tonsillar carcinomas in their series contained HPV DNA in episomal form (Mellin et al. 2002). Another study in 1992 reported two HPV 16+ tonsillar carcinomas which contained episomal HPV DNA, and two HPV 33+ tonsillar carcinomas in which one was integrated and the other had mixed forms (Snijders et al. 1992). It is unclear why tonsillar carcinomas appear to have a higher predominance of episomal HPV DNA than other types of head and neck cancer. It is likely though, that these various observations suggest a high heterogeneity and variation in the oncogenic pathways among these tumours.

#### **3.4 Similarities between HPV and ZEBRA**

HPV and EBV co-infections have not been well documented and the significance of the presence of both viruses in nasopharyngeal cells has not been determined. It has been shown that ZEBRA, an EBV immediate early protein expressed during lytic replication that activates early EBV genes, binds to p53 (Quinlivan et al. 1993). The physical interaction of the ZEBRA and p53 protein prevents p53 from activating p53-responsive promoters (MauserSaito, et al. 2002). Similarly, HPV has been found to interact with p53, suggesting that this interaction promotes cell growth and thereby enhance viral replication (Levine

discriminated in this model (figure 2): the conversion of a single mutated stem cell in a patch into a group of stem cells without proper growth control (field); the eventual transforming event, which turns a field into an overt carcinoma showing invasive growth and metastasis; and the development of metastasis. Both aneuploidy and the accumulation of cancer-

It has been shown that these subtypes (particularly 16) are able to transform and immortalise cells in vitro. These effects are predominantly due to the E6 and E7 oncogenes, which bind and enhance degradation of p53 and Rb tumour suppressor genes respectively. There is evidence that immortalisation of oral keratinocytes and epithelial cells occur quite

HPV integration usually leads to disruption and/or deletion of HPV E1 or E2 open reading frame (ORF), which are important for viral replication and transcription. E2 functions also as a repressor of E6 and E7 and disruption of E2 activity allows increased E6 and E7 expression, thus maintaining the immortalised phenotype (zur Hausen 2009). Integration of HPV 16 DNA also correlates with a selective growth advantage and may allow the cancerous cell to outgrow its rivals; this may be an important step in the pathway of oncogenesis (Jeon, Allen-Hoffmann & Lambert 1995). However, despite the dominance of the integrated HPV genome in terms of cervical carcinogenesis, 15-30% of cervical cancer contains HPV only in the episomal form (Watts et al. 2002). In some of these cases, investigators have found deletions in the YY1-binding sites of the LCR (long control region) of HPV 16 episomal DNA which may allow elevated activity of the E6/ E7 promoter (Dong et al. 1994). It is clear that the actual molecular pathway to cervical carcinogenesis is far from homogeneous. The situation in head and neck cancers is less than clear but heterogeneity and the existence of multiple pathways to carcinogenesis is highly likely. Koskinen et al (2003) reported that in their series of head and neck cancers, 61% were HPV DNA positive. HPV 16 was the dominant subtype, and found in 84% of HPV+ cancers. Tonsillar carcinomas have been reported to have the highest prevalence rate of HPV DNA contained within cancerous cells (51%) of all the forms of head and neck cancers (Syrjanen, S. 2004). Mellin et al 2002 reported that all 11 cases of HPV+ tonsillar carcinomas in their series contained HPV DNA in episomal form (Mellin et al. 2002). Another study in 1992 reported two HPV 16+ tonsillar carcinomas which contained episomal HPV DNA, and two HPV 33+ tonsillar carcinomas in which one was integrated and the other had mixed forms (Snijders et al. 1992). It is unclear why tonsillar carcinomas appear to have a higher predominance of episomal HPV DNA than other types of head and neck cancer. It is likely though, that these various observations suggest a high heterogeneity and variation in the oncogenic pathways

HPV and EBV co-infections have not been well documented and the significance of the presence of both viruses in nasopharyngeal cells has not been determined. It has been shown that ZEBRA, an EBV immediate early protein expressed during lytic replication that activates early EBV genes, binds to p53 (Quinlivan et al. 1993). The physical interaction of the ZEBRA and p53 protein prevents p53 from activating p53-responsive promoters (MauserSaito, et al. 2002). Similarly, HPV has been found to interact with p53, suggesting that this interaction promotes cell growth and thereby enhance viral replication (Levine

associated genetic changes in fields are linked to the risk of malignant progression.

readily (Park et al. 1991).

among these tumours.

**3.4 Similarities between HPV and ZEBRA** 

1990). Targeting p53 may be a common requirement for the replication of many types of DNA viruses (Prayitno 2006). In addition, B cells transfected with EBV latent membrane protein lost the regulatory effects of the retinoblastoma (RB) protein, and the HPV E7 transcript has been shown to immunoprecipitate the RB protein (Giovannelli et al. 2002). Thus, the functional loss of the RB protein might be one event common to both the HPV and EBV carcinogenic pathways.

Fig. 2. A hypothetical model of HPV-associated NPC development. Most information has been deciphered from oral carcinogenesis; there are fewer data on the other subsites of NPC. Light cells: Normal epithelium. Cells with a white ring: Koilocytes, as a result of an active viral replication. Dark cells: Dysplastic epithelium with increased chromosomal instability reflected by increasing aneuploidy of the cells and leading to progression to high grade dysplasia and invasive carcinoma.
