The Role of the p16 and p53 Tumor Suppressor Proteins and Viral HPV16 E6 and E7 Oncoproteins in the Assessment of Survival in Patients with Head and Neck Cancers Associated with Human Papillomavirus Infections

Simple Summary The role of human papillomavirus (HPV) in the survival of patients with head and neck squamous cell carcinoma (HNSCC) is an important topic. The recognition of additional markers could play a significant role in survival prognosis. Our study aimed to assess the roles of different molecular and immunohistochemical factors in the survival of patients with HNSCC. We analyzed 106 HNSCC samples and confirmed the roles of HPV DNA and p16, p53, and HPV16 E6 and E7 proteins in different subgroups of HNSCC. In addition to p16, the immunohistochemical overexpression of HPV16 E6 protein should be used for patient survival prognosis. Abstract The role of HPV in the survival prognosis of patients with head and neck squamous cell carcinoma, especially patients with laryngeal squamous cell carcinoma (LSCC) and hypopharyngeal squamous cell carcinoma (HPSCC), is still somewhat ambiguous. The present study aimed to explore the significance of tumor suppressor proteins and HPV16 E6 and E7 oncoproteins in the assessment of survival in patients with oropharyngeal squamous cell carcinoma (OPSCC), LSCC, and HPSCC associated with high-risk (HR-) and low-risk (LR-) HPV infections. By utilizing molecular and immunohistochemical investigations of HNSCC samples and patient data, univariate and multivariate survival analyses were conducted. The presence of HPV DNA (LR- and HR-HPV) was associated with a better 5-year OS and DSS for OPSCC and LSCC. The IHC overexpression of HPV16 E6 protein and p16 protein was associated with better survival in the univariate (for OPSCC) and multivariate (OPSCC and HPSCC) survival analyses. The overexpression of p53 was associated with better survival in OPSCC. HPV infection plays a significant role in the tumorigenesis of HNSCC, and the immunohistochemical assessment of HPV16 E6 protein expression should be interpreted as a useful prognostic marker for OPSCC and HPSCC.


Introduction
As one of the most common cancers globally, HNSCC accounts for more than 660,000 new cases and 325,000 deaths annually [1]. According to the GLOBOCAN data, 98,412 new cases of OPSCC, 98,412 new cases of LSCC, and 84,254 new HPSCC were registered in 2020 [2].
The DNA extraction from fresh frozen tissue material was performed with the standard phenol/chloroform extraction method.
FFPE cancer samples were processed using a blackPREP FFPE DNA Kit (Analytik Jena, Germany) following the manufacturer's protocol. To avoid cross-contamination, separate sterile blades were used for each specimen.
To assess the concentration and quality of the extracted DNA, a spectrophotometric analysis was performed (Nanodrop ND-1000 Spectrophotometer, Thermo Fisher Scientific, Waltham, MA, USA). Beta-(β-) globin was used as a quality control for the isolated DNA [27]. Only β-globin-positive samples were included in the further investigation of the gathered specimens.
Standard RNA extraction with TRIzol LS Reagent from Thermo Fisher Scientific was accomplished for fresh frozen tissue specimens according to the producer's manual.
A PureLink FFPE Total RNA Isolation Kit (Thermo Fisher Scientific, USA) was used for RNA extraction from FFPE cancer samples, following the manufacturer's protocol. Each sample was sectioned separately with a new sterile blade.
A spectrophotometric analysis was used to assess the concentration and quality of the extracted RNA.

HPV DNA Detection Using MY09/11 and GP5+/6+ Consensus Primers
A polymerase chain reaction (PCR) with the consensus primers MY9/MY11 and GP5+/6+ was used for the initial detection of the broad range of HPV types (HR-HPV and LR-HPV types) [28,29]. Electrophoresis in a 1.7% ethidium bromide gel was used to assess the PCR results. Amplification products of appropriate lengths for the primers that were used were considered HPV-positive. Each reaction included positive and negative controls.

HPV Genotyping
Consensus PCR-positive samples were further subjected to HPV genotyping. Primers for HPV 16 and 18 (L1) and the Anyplex II HPV28 multiplex real-time-PCR (RT-PCR) were used for HPV genotyping.
The results were visualized via electrophoresis in 1.7% agarose gel with an assessment of appropriate amplification products [29]. Positive and negative controls were used in each reaction.
Anyplex II HPV28 multiplex RT-PCR was used following the manufacturer's recommendations (Seegene, Seoul, Republic of Korea).

HPV16 E6/E7 mRNA Detection
The detection of E6/E7 mRNA was performed using real-time PCR with the PreTect HPV-Proofer kit. The PreTect HPV-Proofer assay qualitatively detected the presence of HPV E6/E7 oncogene mRNA from HPV types 16, 18, 31, 33, and 45. It had an intrinsic sample control to assess specimen quality. Specimens with positive intrinsic controls were considered valid. Only HR-HPV-positive samples were used for E6/E7 mRNA detection.
Briefly, after the standard preparation process, the sections were incubated overnight with the primary antibodies at 4 • C. We used a monoclonal mouse anti-CDKN2A/p16INK4a antibody (Abcam, Cambridge, UK, 1:300 dilution, ab201980); a monoclonal mouse anti-p53 antibody (Santa Cruz Biotechnology, Inc., Dallas, TX, USA, 1:50 dilution, sc-47698); a monoclonal mouse anti-HPV16 E6 + HPV18 E6 antibody (Abcam, Cambridge, UK, prediluted, ab51931) [32][33][34]; and a monoclonal mouse anti-HPV16 E7 antibody (Santa Cruz Biotechnology, Inc., 1:50 dilution, sc-6981). We used a HiDef Detection HRP Polymer system and a diaminobenzidine tetrahydrochloride substrate kit (Cell Marque, Rocklin, CA, USA) to visualize the products of IHC reactions. Counterstaining of cell nuclei within a tumor section with Mayer's hematoxylin was used. In the negative controls of reactions, primary antibodies were omitted. The reaction results were assessed by two independent experienced investigators without knowledge of the clinical and molecular virology data.
A cut-off at 50% positive tumor cells for the p16 immunostaining was used, as proposed by Hong et al. (2013) [35].
The assessment of immunostaining for p53 was performed semiquantitatively. We considered a sample to be p53-postive (p53+) when the criteria described by Halec et al. (2013) were met [36]. The p53 overexpression (upregulation) was considered when p53 positivity was confirmed in >50% of tumor cells with intensity = 2 or >25% of tumor cells with intensity = 3. All FFPE specimens that did not reach these criteria were considered p53-negative (p53-; downregulation).
As all HR-HPV-positive specimens contained HPV16 DNA, only those were used for the IHC detection of E6 and E7 proteins. The IHC reaction results for the E6 and E7 viral proteins were estimated semiquantitatively in 20 randomly selected visual fields of each sample including the tumor and the surface epithelium of the regions of interest. To achieve enough statistical power, we used the expression levels of E6 and E7 at <10% as negative and at ≥10% as positive.

Statistical Data Analysis
All statistical analyses were performed using GraphPad Prism 9 (GraphPad Software, La Jolla, CA, USA). A standard statistical analysis was performed to assess the data distribution. A nonparametric Spearman's correlation analysis was used to find any correlations between the groups [37]. A univariate survival analysis was performed using the Kaplan-Meier method; overall and disease-specific survivals (OS and DSS) were assessed. A multivariate survival analysis was performed using the Cox regression method. p values less than 0.05 (p < 0.05) were considered statistically significant.
For the oropharynx, the OS rates were 26.82% and 0% for patients who were HPVpositive and HPV-negative, respectively, although this difference failed to reach statistical significance (p = 0.077; Figure 2A). The DSS rates were 27.78% and 0% for these groups of patients (p < 0.05; Figure 2B), respectively.  For patients with LSCC, the OS rates were 64.59% and 44.44% in patients who were HPVpositive and HPV-negative (p < 0.05; Figure 2C), respectively. The DSS rates were 68.90% and 50% for patients who were HPV-positive and HPV-negative (p < 0.05; Figure 2D), respectively.
As all HPSCC samples were HPV DNA+, a Kaplan-Meier survival analysis could not be performed.
3.6.2. OS and DSS, Depending on Immunohistochemical Expression of HPV16 DNA, HPV16 E6/E7 mRNA, and p16, p53, E6, and E7 Proteins We performed a Kaplan-Meier survival analysis with a stratification of patients depending on the location of the primary tumor. The OS and DSS were calculated. For most variables, a univariate survival analysis with the Kaplan-Meier method failed to reach statistical significance.
There were borderline statistically significant differences (p = 0.057, Figure 3A,B) between p16+ and p16− OPSCC for OS and statistically significant differences for DSS.
A Kaplan-Meier survival analysis of p53+ and p53− HPSCC showed statistically significant differences in OS and DSS ( Figure 3C,D).
The immunohistochemical overexpression of HPV16 E6 protein was associated with significantly better OS and DSS in patients with OPSCC ( Figure 3E,F).

Multivariate Cox Regression Analysis
The age; sex; hazards; applied treatment; immunohistochemical expression of p16, p53, E6 protein, and E7 protein; and the presence of HPV16 DNA and E6/E7 mRNA were included in the Cox model. First, a multivariate survival analysis was performed for all patients with head and neck tumors. Second, each anatomical location of the head and neck cancers was analyzed separately: oropharynx, larynx, and hypopharynx.

All HNSCC
The results of the analysis of all patients with head and neck cancer are summarized in Table 2.
The Cox regression analysis suggested that the T1 stage was associated with a lower risk of early death. While the results for each T stage did not reach statistical significance, there was a trend towards a higher early death risk with a higher T stage; patients with HNSCC with a T4 tumor had a 2.68-fold higher probability of death. The analysis also showed that a higher N stage was associated with a higher risk of early death. The N1 stage (in reference to N0) was associated with a 4.98-fold greater risk of early death, and the risk notably increased in the N3 stage. A lower tumor differentiation grade (G) was associated with a higher risk of early death. Patients with G3 tumors (well-differentiated) had an 81% lower risk of early death than patients with G1 tumors (undifferentiated).
The effect on survival was also statistically significant for treatment. Patients who received a combined treatment (RT+ChT+/−OP) showed a lower risk of early death.

OPSCC
This group encompassed 34 patients with 26 events (death). Two patients were excluded from the analysis due to missing values.     The overexpression of p16, p53, and HPV16 E6 protein showed much lower hazard ratios and was associated with significantly improved survival. On the contrary, the overexpression of HPV16 E7 protein was associated with a high risk of early death. A graphical analysis showed that the overexpression of p16 (p16+) in a tumor was associated with better survival than that of patients with p16-negative tumors ( Figure 4A). However, the overexpression of HPV16 E7 protein was associated with decreased survival. Moreover, when combining the two markers (p16 and HPV16 E7 protein), E7 protein overexpression (E7+) decreased survival, even in patients with p16+ tumors ( Figure 4A). Figure 4B shows that the best survival was seen in patients with p53-positive (p53+)/HPV16 E6 protein positive (E6+) tumors and that the worst was seen in patients with p53−/E6− tumors. There was no difference in survival between patients with p53−/E6+ and p53+/E6− tumors. A larger tumor size (T grade) negatively affected survival. An analysis showed a lower risk of early death for tumors with lower N grades; however, the difference was not statistically significant.
The patients who underwent radiotherapy had a significantly lower risk of early death than patients with other treatment modalities.  A larger tumor size (T grade) negatively affected survival. An analysis showed a lower risk of early death for tumors with lower N grades; however, the difference was not statistically significant.
The patients who underwent radiotherapy had a significantly lower risk of early death than patients with other treatment modalities.

LSCC
This group encompassed 41 patients with 16 events (death). A Cox regression model with all included variables was statistically significant (p < 0.001). An analysis showed that no variable significantly affected survival.

HPSCC
This group encompassed 31 patients with 29 events (death). Table 4 depicts the Cox regression analysis for patients with HPSCC. The Cox regression model showed that the expression of p16 and HPV16 E6 protein; the presence of HPV16 DNA; the hazards; and the T, N, and M grades statistically significantly affected survival. The effects of the other variables were not statistically significant.
The overexpression of p16 and HPV16 E6 protein was associated with an extremely low risk of early death ( Figure 5A,C). By combining the p16 status and the HPV16 E7 protein status, we found that E7 protein expression did not affect survival ( Figure 5B, overlaying of the curves). However, combining the p53 and HPV16 E6 protein statuses showed that patients with E6+ tumors had better survival and that p53 overexpression seems to increase survival even more in these patients ( Figure 5D). The worst survival was in the group of patients with p53−/E6− tumors.
The Cox regression analysis revealed that larger primary tumors are associated with a higher risk of early death. Patients with T3 tumors had 87% less risk of early death than patients with T4. Moreover, a lower N grade was associated with lower hazard ratios. Lastly, the presence of distal metastases was associated with a 22-fold increase in the risk of death. Smoking patients had a 57-fold increase in the risk of early death in comparison to non-smokers/non-drinkers.

Discussion
The current study aimed to assess the roles of HPV infection and associated markers such as p16, p53, HPV16 E6/E7 oncoproteins, the presence of HPV DNA, and E6/E7 mRNA in survival.
An initial univariate survival analysis (Kaplan-Meier) shows the potential role of not only HR-HPV but also LR-HPV infection in the survival of patients with OPSCC and LSCC, as 1/3 of the patients have a probability of LR-HPV infection. The study results suggest that patients with HPV-DNA-positive OPSCC and LSCC have a better 5-year OS and DSS. These results agree with other studies where patients with HNSCC and patients with tonsil cancer also had better survival rates if the tumors were positive for HPV DNA [38,39]. This is probably due to better radiosensitivity of HPV+ tumors, which means patients could benefit from the "softer" treatment applied to HPV-positive tumors and increases positive outcomes for the patients [38]. On the other hand, HPV-infected cells could be more visible to the host's immune system, allowing for easier identification as well as the destruction of virus-related tumor tissues. In that case, a deeper investigation of the HPV activity in patients with HNSCC and the interaction with their immune systems would be required.
It is well documented that patients with HPV-positive OPSCC have higher 3-and 5-year survival rates than patients who are HPV-negative [40], but the consensus is made for HR-HPV (mostly HPV16 and 18). For LSCC, many studies have shown no significant survival increase for HPV-positive tumors [7,41,42]. However, in recent years, there have been studies with results similar to ours, with better survival in patients with HPV-positive LSCC [9,43].
On the other hand, in our study, the stratification of patients with HNSCC by the tumor location and the identification of specific HPV types showed that the presence of HPV16 DNA in hypopharyngeal squamous cell carcinoma cases substantially decreased the survival rates of patients. This indicates that HPV16 may play a significant role in HPSCC development. Additionally, the immunological aspects should be considered. The presence of viral antigens could promote anti-tumor immunity and lead to better survival of the patients [44][45][46].
Head and neck cancers encompass a multitude of subsites for cancer development. Sometimes studies analyzing the effects of HPV on the survival of head and neck cancers can be confusing in that they unify the survival analysis without stratifying the primary tumors by location, especially hypopharyngeal and laryngeal cancers, which are sometimes combined in non-oropharyngeal cancers [14,47]. In our view, this could lead to incorrect conclusions. The oropharynx, larynx, and hypopharynx are three distinct locations with different prognoses based on lymphatic drainage alone. In our study, an analysis of all HNSCC in a Cox regression did not show p16, p53, or other variables to be significant factors affecting the survival of the patients. This indicates that patients should preferably be stratified by the primary location of the tumor to obtain a more comprehensive view of the potential risk factors.
This study reaffirmed the predictive role of p16 overexpression in OPSCC (univariate survival analysis), confirming better survival in patients with p16+ tumors [48,49]. This trend continued in the Cox regression analysis, with statistical significance further confirming its role as a distinct predictive marker for OPSCC. However, for HPSCC and LSCC, this could not be confirmed in the univariate survival analysis. The Cox regression analysis showed better survival and a lower risk of death for patients with p16+ HPSCC, suggesting that there might be a reason to consider it as a predictive marker. Several studies have shown similar findings [50,51]. There is also a question of p16's association with HPV activity in non-oropharyngeal squamous cell carcinoma, whether it can be used as a surrogate marker for HPV infection, and whether it serves as a suitable prognostic factor of survival. Several studies have shown that p16 often does not correspond to the HPV status in non-oropharyngeal cancers; however, it has a prognostic value for survival [52][53][54].
The lack of significance for many analyzed variables in OPSCC (univariate survival analysis) of our study could be due to the relatively small patient number in this subgroup, which could affect the statistical power of analysis. Additionally, the high number of smokers and alcohol abusers could also affect the significance of the results. This is accounted for in the Cox regression model.
The univariate survival analysis of p53 immunohistochemical expression showed significantly better OS and DSS in p53+ HPSCC. The trend persisted in the Cox regression, although without statistical significance. Similar findings were present for OPSCC in the Cox regression analysis; p53 overexpression (p53+) was associated with a significantly lower risk of death. This could be due to the tumor-suppressing properties of p53. However, there was a considerable number of HPV16-positive samples and even more HPV16 E6/E7 mRNA-positive OPSCC samples. A logical picture would be that in HPV-driven cancer, p53 is suppressed, resulting in a p53-negative result that is confirmed using immunohistochemistry. Published data suggest that HPV-driven tumors show p53 downregulation [55][56][57]. On the contrary, Hasegawa et al. [58] reported that p53 overexpression correlates with a better response to chemotherapy and is thus associated with better survival. Similar results were demonstrated by Sun et al. [59]. In these studies, however, the HPV status was not studied. Initially, in HPV-driven cancers, there could be p53 overexpression due to the degradation of pRb by E7 oncoprotein and increased stabilization of p53 [60]. A meta-analysis of oral tongue squamous cell carcinoma showed that p53 could not be used as a prognostic biomarker for these tumors [61]. Similar conclusions were made by Halec et al. for LSCC [36]. Unfortunately, our study did not include an assessment of TP53 gene mutations, which could have clarified some questions about the previously mentioned points [62,63]. Additionally, there is a possibility that p53 overexpression is unrelated to HPV infection, especially considering the high number of smokers in our study. Additional studies are needed to study the prognostic role of p53 in HNSCC, especially in OPSCC and HPSCC.
To the best of our knowledge, very few studies have been focusing on HPV oncoprotein E6/E7 immunohistochemical expression and its role in survival or prognostic values. As E6 and E7 are considered to be the main driving forces of HPV-mediated carcinogenesis, we found it interesting to study the role of these proteins in survival using immunohistochemistry. In the cases of both OPSCC and HPSCC, the immunohistochemistry results of HPV16 E6 protein expression showed that patients with positive staining in their tumor samples had a better survival rate. However, a high expression of either p16 or p53 was simultaneously found with E6, which could be considered a positive outcome marker for the patient. Moreover, there is a possibility that at a certain stage of viral activity, this oncogene (E6) did not have time to disrupt the cell cycle. For example, E6 initiates proteasome-dependent p53 degradation by recruiting the ubiquitin ligase E6AP. Furthermore, only the combined complex of E6 and E6AP is reactive with p53. This means that the expression of a single HPV16 E6 protein cannot affect p53 degradation (detection could be less informative for a patient's outcome prognosis) [64]. Unfortunately, E6AP activity was not studied in this research. A prospective study (of the dynamics with several time points) might better reveal HPV oncogenes' roles in the progression of an HNSCC tumor, as a persistent HPV infection is a major factor for carcinogenesis [65]. With this study, it is difficult to distinguish persistent from non-persistent HPV infections (sampling was performed only a single time). However, in patients with HPSCC, E6 protein was detected using only immunostaining, while E6 mRNA was not detected, and HPV16 DNA was still detectable. This could indirectly indicate the presence of a persistent HPV16 infection, which could be one of the reasons why the presence of HPV16 DNA in the samples of patients with HPSCC showed worse outcomes.
E7 is recognized as the major transforming protein of high-risk HPVs due to mutational analyses in transformation assays [66]. In addition, it was shown that E7 precisely drives early tumorigenesis [67]. The present study shows that the IHC overexpression of HPV16 E7 protein in OPSCC is associated with a poorer prognosis (Cox regression). However, in HPV-associated tumors, the E7 protein should be the driving factor for p16 overexpression, which is associated with better survival. On the other hand, some studies report that the overexpression of p16 has consistently and repeatedly been shown to be associated with a better response to therapy and a favorable clinical outcome in OPSCC, and not all cases of p16 overexpression could be related to HPV's oncogenic activity [68,69]. This suggests the presence of additional mechanisms of E7-protein-associated carcinogenesis. Several studies have shown that E7 induced the upregulation of several types of matrix metalloproteinases [70,71]. This process has been linked to the promotion of the invasiveness the tumors [72]. Additionally, the protein function of HR-HPV E7 has been associated with a more stable mitotic function that is needed for viral genome maintenance and replication [73,74]. These processes could lead to an invasive and potentially metastatic phenotype of cancer, and this could explain the poorer prognosis in OPSCC with IHC HPV16 E7 protein overexpression [66]. Oton-Gonzalez et al. [75] showed that patients with OPSCC with detectable HPV16 E7 protein in their serum had poorer relapse-free survival and OS. The authors also showed a correlation between E7 protein in serum and E7 mRNA expression. Thus, they concluded that the source of the E7 protein must have been HPV16-positive cancer, more specifically circulating tumor cells, suggestive of the metastatic process. It is worth noting that not all tumors are HPV-related, and it was shown that virus-induced oncogenesis takes a long time to develop and that some patients with HNSCC can have a concomitant HPV infection [66].
One limitation of our study is the relatively small number of patients for each region (oropharynx, larynx, and hypopharynx), which could result in insufficient statistical power and limit the conclusions drawn for some markers, especially if they did not reach statistical significance. However, it is hard to deny the observed trends of the studied markers and their effects on survival. The other limitation is that almost all HPSCC samples were FFPE due to possible genetic material degradation, especially that of RNA. On the other hand, all samples were viable for analysis based on the intrinsic control of the kit that was used (mRNA detection) or β-globin detection (DNA quality).

Conclusions
HPV infection plays a significant role in the tumorigenesis of HNSCC, especially OPSCC. It should be noted that not only HR-HPV but also LR-HPV could affect survival prognosis. The immunohistochemical assessment of HPV16 E6 protein expression should be interpreted as a useful prognostic marker for OPSCC and HPSCC.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The results related to the immunohistochemical assessment of p16, viral proteins HPV16 E6 and E7, and HPV DNA testing in the samples obtained from patients with hypopharyngeal and laryngeal cancer were published in Viruses (https://doi.org/10.3390/v13061008). The aforementioned article aimed to estimate the prevalence of aforementioned markers in laryngeal and hypopharyngeal cancers.