The p16INK4a gene (CDKN2a/INK4a) functions as negative regulator of the cell cycle progression through its inhibition of cdk4/6 and subsequent blockage of the cyclin-dependent phosphorylation of the Retinoblastoma gene (Rb). CDKN2a/INK4a located on 9p21 is frequently inactivated in oral epithelial pre-cancer and cancer via the following events: LOH, hypermethylation, deletion, mutation. The loss of p16INK4A expression defines a subgroup of oropharyngeal cancer patients with worse clinical outcome  whereas p16 protein over-expression has been proposed as surrogate marker of HPV infection initially in cervical cancer by Amortegui et al., [25, 26], followed by squamous cell carcinomas from other sites than uterine cervix , and especially in Head and Neck Cancers [28, 29].
In order to further explain differences between p16 expression and HPV-DNA in oropharyngeal cancer we performed Methylation-Specific PCR (MSP) to evaluate CDKN2a/INK4a promoter inactivation. Finally, MSP showed that the methylation of CDKN2a/INK4a is a frequent epigenetic alteration in oropharyngeal cancer and revealed that p16 was inactivated in 75% of OPSCCs associated to alcohol and tobacco risk factors. Therefore, we can conclude that OPSCC should be ideally subdivided in two groups: a) alcohol-tobacco-associated/HPV-DNA negative/CDKN2a-MSP methylated/p16-IHC negative or faint; b) alcohol-tobacco-unassociated HPV-DNA/HPV-DNA-positive/CDKN2a-MSP unmethylated/ p16-IHC positive. However, the situation in clinical setting is more complex because different risk factors frequently overlapped i.e. in HN-SCCs of young population, alcohol/tobacco consumption and HPV infection may be also associated. HPV positive OPSCC mostly occurs in younger patients and may also arise in people without a history of tobacco use. Furthermore, there are some cases showing discrepancy between promoter methylation and protein expression; i.e. cancer cases in which promoter is methylated (as evaluated by qualitative analysis) and the protein is unexpectedly expressed. Therefore, in these cases quantitative tests should be performed in order to establish the proportion of methylated/unmethylated alleles in the cancer cells clearly distinguishing the latter from alleles of non-cancerous cells. Although the microdissection-based quantitative tests are important research tools they cannot be easily performed in current clinical practice. Furthermore, we observed different levels of p16-IHC accuracy in the different cancer subpopulation studied. In details, in a cohort of prevalently alcohol/tobacco associated cancers from the south-west of Italy (Napoli) p16-IHC test showed a lower level of specificity in detecting HPV positive cases. In this cohort, there was an unacceptable large group of p16-IHC positive cancers that were diagnosed as negative by the combined ISH and PCR methods. It is possible that up-regulated p16 expression we observed in cancers was due to other molecular events not related to HPV. On the other hand, in a cohorts with higher number of HPV positive cases, i.e. the cohort from middle-east Italy (Ancona) the p16-IHC test increases its specificity in detecting HPV cases. This observation is confirmed by our analysis of OPSCCs cases which showed higher level of HPV infection than that of OSCC, in parallel with higher p16-IHC specificity levels as method of detection of HPV positive cases. In agreement with this hypothesis, a recent literature report demonstrates different p16 accuracy according to different anatomical sub-sites of the Head and Neck region . In this complex scenario the p16-IHC test alone or in association to CDKN2a promoter methylation could be used only as a screening method and need to be associated with molecular tests in order to detect HPV-DNA and to assess its integration status.
Integration of HPV-DNA into the host DNA is a well known topic in cervical cancer but there are few investigations in Head and Neck Cancers. Integration of HPV 16 DNA correlates with dysfunction of HPV E1 or E2 open reading frames (ORF), which are active during HPV replication.. E2 loss of function allows up-regulation of E6 and E7 oncoproteins, because E2 is a repressor of E6 and E7. The great percentage of cervical cancers harbors HPV in the integrated form, however, recently has emerged that cervical cancers may contain HPV episomic DNAs as well. It has been also demonstrated that HR-HPV episomal DNAs up-regulate the activity of E6/E7 promoter, which in turn gives rise to elevated E6 and E7 protein expression in cancer cells. As regard to HN-SCC HPV-DNA integration there are numerous interesting points to discuss. On the one hand current literature describes the HPV-related HN-SCC as almost exclusively HPV type 16 restricted cancers  and HPV type 16 is considered to have the highest capability to integrate into the host DNA in cervical high grade squamous intraepithelial lesions (HSILs) and invasive carcinomas [32, 33]. On the other hand, HN-SCCs, in particular tonsillar cancers (TCs), have been described as tumors with elevated frequency of HPV-DNA type 16 in integrated or episomal form producing E6 and E7 oncogenic proteins, since the early observations of Sniders PJF at al. . These observations have been further confirmed by Mellin H. et al. concluding that in oropharyngeal cancers HPV is almost exclusively not integrated and its carcinogenic activity is due to E6/E7 oncoproteins expressed from episomal viral sequences . Recently, some bias in the interpretation of HPV prevalence in HN-SCCs have been emerged since oral cancers (OSCCs) have not been clearly distinguished from oropharyngeal ones (OP-SCCs). Aim of this study was to demonstrate the integrative versus episomic HPV status as a diagnostic tool separating OSCC from OP-SCC in order to furnish site by site information regarding percentage and distribution of viral integration in host cancer cells.
For this purpose, the ISH technique for HPV is able to reveal the morphological context of viral DNA location. Our ISH study showed heterogeneous status of HPV integration in cancers originated from different Head and Neck regions. Among the HPV positive OSCCs only the case infected by HPV type 16 DNA showed clear integration signals, although not uniformly distributed throughout the entire tumor but restricted to some cancer fields. Other fields of the same tumor specimen showed clusters signals demonstrating episomal status. Therefore, the HR-HPV type 16 OSCC harbor both integrated and episomic HPV-DNA into the host transformed cells but with an heterogeneous distribution. Interestingly, the OSCC case with double LR-HPV type 44 and HR-HPV type 31 infection showed no signals of integration; however, the viruses have been localized exclusively in the cytoplasms of the cancer cells. Remarkably, we found a case of OSCC infected by the HPV type 53, recently included in a phylogenetic group of HR-viruses unrelated to HR-HPV 16 and HR-HPV 18 . Unfortunately this case with type 53 HPV virus has been not determined by ISH because commercially available systems do not include this type 53 in the HR-HPV probe panel. This is because most detection systems for HPV detection are tailored for female genital system and not for other organ systems.
As regard to OPSCC we found the most HPV-DNA positive cases in tonsillar cancers (TCs). As described in literature HPV16 is the almost unique HPV-DNA found in this anatomical sub-site of the OP region. We demonstrated that TC cancers had intense integrative and clusters signals of HR-HPV diffuse throughout the whole cancer specimen and not focally distributed as we have observed in the OSCC where the integrative signals were focally restricted to some fields of the cancer specimen. This is a further demonstration that HPV 16 plays a significant role in the pathogenesis of a subgroup of OPSCC and that the HPV 16 virus integration into the host genome begins in the tonsillar cripts leading to indirect stimulation of CDKN2a locus and p16INK4a over-expression [37–39].
Schache AG et al.  assessed the sensitivity, specificity, and prognostic ability of eight possible assays and assay combinations for HPV16, including the gold standard of RNA qPCR, in 108 cases of OPSCC from the United Kingdom. The investigators found that HPV16-positive patients with OPSCC were younger and smoked less than HPV negative patients, and the proportion of HPV16-positive cases increased from 15% to 57% between 1988 and 2009. When compared to RNA qPCR, p16-IHC/DNA qPCR showed sensitivity and specificity of 97% and 94%, respectively, and proved to be the best discriminator of favorable outcome. The p16-IHC/HR-HPV ISH had a specificity of 90%, but reduced sensitivity of 88% affected the prognostic utility. Used in isolation, p16-IHC, HR-HPV ISH, or DNA qPCR were not specific enough to be recommended for use in clinical trials. In agreement with our findings, the authors recommended "caution [...] in applying HPV 16 diagnostic tests because of significant disparities in accuracy and prognostic value" .