Clinical correlation of p16 expression with lymphatic invasion and epithelial-mesenchymal transition (EMT) in oropharyngeal carcinomas

Introduction

Head and neck cancers (HNC) comprise a group of neoplasms involving the oral cavity, oropharynx, nasopharynx, hypopharynx, and larynx. It is typically associated with risk factors, such as excessive alcohol intake, tobacco exposure, or both. However, oropharyngeal tumors have been found to be strongly associated with human papillomavirus (HPV) infection, particularly HPV-16 and HPV-18 strains (Johnson et al., 2020; Saxena et al., 2019). The main treatment approach for any pharyngeal malignancy is chemoradiation, predominantly owing to the higher functional morbidity observed following surgery in such patients. In general, HPV-associated HNC has been found to a better prognosis after treatment than non-HPV-associated HNC. Hence, multiple ongoing studies are exploring the scope of therapeutic dose de-escalation in the management of HPV-associated diseases (Johnson et al., 2020).

The most prevalent histological subtype in HNC is the Squamous Cell Carcinoma variant known as head and neck squamous cell carcinoma (HNSCC) (Saxena et al., 2019). It is the second-most prevalent form of cancer in India and comprises more than 30% of all the malignancies in India (Sung et al., 2021). This increased incidence rate is because the majority of the Indian population has a lower socioeconomic status, increased trend in habits such as tobacco chewing and alcohol consumption, insufficient exposure to new diagnostic measures, and subsequent delayed reporting. Despite recent advances in surgery and radiotherapy techniques over the last few decades, there has not been a considerable increase in the survival rates of these patients (Huber et al., 2011). Therefore, an extensive understanding of the prognostic factors in the interplay is of great importance for improving possible outcomes.

Although HPV-associated tumors of Oropharyngeal Squamous Cell carcinoma (OPSCC) often have advanced nodal disease at presentation, they respond better to treatment and, thereby, have a better prognosis. However, the exact molecular mechanisms underlying this phenomenon remain unknown (Wakisaka et al., 2015). Tumor invasion generally occurs in two unique forms: single-cell invasion and collective cell invasion. Epithelial-Mesenchymal Transition (EMT) is a process in which tumor progression occurs via single-cell invasion, where epithelial tumor cells acquire mesenchymal cell properties, leading to dissemination and metastasis. The characteristic features of EMT include changes in epithelial characteristics, such as decreased expression of E-cadherin and β-catenin, and an increase in mesenchymal characteristics, such as increased expression of fibronectin, vimentin, and proteolytic enzymes, resulting in intercellular reorganization. EMT has been reported to be involved in tumor invasion and its distant spread (Kalluri & Weinberg, 2010; Thiery, 2002). EMT has been observed in HNSCC and is generally associated with a poor prognosis (Cappellesso et al., 2015; Jensen et al., 2015; Schrader et al., 2015; Frederick et al., 2007). The association of these EMT markers with the HPV status of patients and their prognosis in OPSCC has not been investigated in detail, despite limited studies including a smaller number of patients (HATAkEYAMA et al., 2014; Wakisaka et al., 2015).

Podoplanin is a type-1 transmembrane glycoprotein, which in humans, are seen to be expressed in renal podocytes, placenta, skeletal muscle, heart, lungs, myofibroblasts of the salivary glands, breast, mesothelial cells, and osteoblasts (Breiteneder-Geleff et al., 1999; Martín-Villar et al., 2005; Ordóñez, 2006). It is generally used as a specific surrogate marker for lymphangiogenesis, as it is exclusively expressed in the lymphatic vessels (Breiteneder-Geleff et al., 1999).

Hence this study was conducted with an aim to establish a clinical correlation between p16 expression, lymphatic invasion and Epithelial Mesenchymal Transition (EMT) markers in OPSCC with the T-N status at presentation, response to treatment, survival and metastatic propensity.

Methods

Immunohistochemistry

The respective biopsy slides were reviewed with hematoxylin and eosin, and 3-micron thin sections were cut from the tumor blocks with the maximum tumor content and area of deepest invasion. Antibodies to p16, EMT markers (E-cadherin and Vimentin), and podoplanin were used for immunohistochemistry (IHC), and their expression patterns were assessed. Positive controls for all the above-mentioned markers were selected and used for the verification of each marker.

A strong, diffuse, and homogenous pattern of expression of p16 in both the cytoplasm and nuclei was considered positive p16 expression. E-cadherin, Vimentin and Podoplanin expression was assessed in the area of deepest invasion using 20× magnification in two separate fields with at least 200 tumor cells. E-cadherin and vimentin expression was considered as positive when >20% of the cells showed immunoreactivity. Patients with loss of E-cadherin expression and positive expression of vimentin were taken as EMT positivity. If >10% of the cells showed immunoreactivity to podoplanin antibody, it was considered as positive. The staining pattern of podoplanin was divided into six different types: diffuse and strong, diffuse and moderate, diffuse and weak, focal and strong, focal and moderate, and focal and weak. Patients with diffuse, strong, diffuse, moderate, focal, and strong patterns were considered to have high podoplanin staining. Patients with the remaining patterns were considered to have low or no podoplanin staining. Lymphatic vessel density was assessed in sections stained with anti-PDPN. Using low power, the area with the largest density was chosen, and counting was done using 100× magnification. Lymphatic Vessel Count (LVC) >3 was considered high lymphangiogenesis in the tumor (Wakisaka et al., 2015).

The IHC-stained slides were assessed separately by two pathologists. To avoid bias during assessment, they had no access to the patients’ clinical data. During the occurrence of inter-observer variation in assessment, specific slides were made to be viewed together, and their consensus opinion was taken into consideration ( Figures 1-4).

Figure 1. Expression of p16.

Figure 2. Expression of E-CADHERIN.

Figure 3. Expression of VIMENTIN.

Figure 4. Expression of PODOPLANIN.

Results

Survival analysis and correlations

Of the 60 patients included in this study, 11 received palliative treatment. These patients were excluded, and the remaining 49 were included in the survival analysis. Of these 49 patients, only 23 showed a complete response after treatment. Hence, only these patients were considered in the calculation of the LRRFS.

Median OS for p16 positives patients was 14 months (IQR: 3-30 months) and the median OS among p16 negative subjects was 10 months (IQR: 5.75 – 12.25 months). 1-year OS rate was found to be significantly higher in p16-positive patients when compared to the of p16-negative patients (80% vs. 44.1%, p=0.045).

Median PFS for p16 positive patients was 6 months (IQR: 3-30 months) and the median PFS among p16 negative patients was 9.5 months (IQR: 3.75 – 12.25 months). 1-year PFS rate in p16-positive patients in our study was found to be 60% whereas, in p16-negative patients, it was found to be 31.8%.

The median LRRFS for p16 positive patients was 29 months (IQR: 17-29 months), and the median LRRFS among p16 negative patients was 12 months (IQR: 9.5– 18.5 months). 1-year LRRFS rate of p16-positive patients was found to be 75%, whereas it was 42% in p16-negative patients ( Table 6, Figures 5-7).

Table 6. Correlation of HPV status with Survival parameters.

Survival parameter	HPV-Positive	HPV-Negative	Total	p
OS at 1 year
Not Survived	1 (20%)	29 (65.9%)	49	0.045
Survived	4 (80%)	15 (34.1%)
PFS at 1 year
Not Survived	2 (40%)	30 (68.2%)	49	0.209
Survived	3 (60%)	14 (31.8%)
LRRF at 1 year
Not Survived	1 (25%)	11 (57.8%)	23	0.714
Survived	3 (75%)	8 (42.2%)

Figure 5. Kaplan Meier Curve 1 - Correlation of OS at 12 months with p16 status.

Figure 6. Kaplan Meyer Curve 2 - Comparison of PFS at 12 months with p16 status.

Figure 7. Kaplan Meier Curve 3 - Correlation of LRRFS at 12 months with p16 status.

Discussion

According to previous studies, patients with HPV-associated OPSCC have been found to have aggressively growing tumors with advanced nodal disease at presentation, despite having a relatively better prognosis after primary treatment. Our study is in agreement with this finding.

A study conducted by Brittany J. Cline et al including 31702 patients revealed that the mean age at diagnosis was 60.3 years of which 77.1% were males and 22.9% were females (Cline et al., 2020). Similarly, in the present study, the mean age at diagnosis was found to be 58.2. Ninety% of patients in our study were males, and the remaining 10% were females. According to a study by Tristan tham et al including 23297 OPSCC patients, the majority (96.4%) of the patients belonged to the subsite of tonsil. In the same study, it was found that the most common histological grade was poorly differentiated SCC (38.9%), followed by moderately differentiated SCC (31.1%) (Tham et al., 2020). In our study, the most common site was found to the tonsil (48.3%). The most common histological grade in our study was moderately differentiated SCC (60%) followed by well-differentiated SCC (30%).

Compared to our Indian community, the prevalence of HPV association in OPSCC in the Western population was substantially higher. the current HPV prevalence in OPSCC in the US population is 70%, according to the CDC. The prevalence pattern of HPV infection related to OPSCC was as follows in a study by Catherine de Martel et al. to quantify the worldwide burden of cancer attributable to infections done in the year 2008 encompassing 12.7 million cases: North America, 56%; Northern and Western Europe-39%, Eastern Europe-38%, Southern Europe-17%, Australia-45%, Japan-52%, and rest of the globe-13% (de Martel et al., 2012). Research conducted in India by Bahl et al. revealed 105 individuals with oropharyngeal cancer had a 22.8% HPV positive rate (Bahl et al., 2014). Similar incidence rates of 20% and 15% were observed for Indian patients with oropharyngeal cancer by Murthy et al and Sannigrahi et al, respectively (Murthy et al., 2016; Sannigrahi et al., 2016). In our investigation, it was shown that 11.7% of OPC had HPV infection. The slight variations in prevalence rates may be attributable to the chosen detection technique, which may include HPV DNA in situ hybridization, HPV DNA polymerase chain reaction, or immunohistochemistry (IHC) for the interpretation of p16 expression.

Wakisaka et al. conducted a study that included 53 OPC patients, in which 58.4% had strong E-cadherin expression, 64.1% had strong vimentin expression, 33.9% had EMT, and 62.2% had higher expression of podoplanin. In the present study, the median LVC was 9. Based on this value, 51% of the patients had low LVC count (Wakisaka et al., 2015). In our study, the prevalence of strong E-cadherin expression was found to be 70%, and that of vimentin was found to be 5.7%. Only one patient (1.9%) had EMT positivity. 37.7% patients had High staining for podoplanin. In the present study, the median LVC was 3. According to this value, 56.6% of the patients had low LVC.

The influence of these parameters on the HPV status has been studied in multiple literature worldwide. However, few studies pertain to the Indian subpopulation. In a study by Wakisaka et al., 40.9% of all the HPV positive patients were found to have strong expression of E-cadherin, 90.9% were found to have vimentin expression, 60% were found to have EMT positivity, 63.6% were found to have low staining for podoplanin, and 63.6% were found to have low LVC. The same values for the HPV negative patients in their study were 70.9%, 45.2%, 19.4%, 9.7%, and 61.3%, respectively. In their study, 35.8% of the patients were found to have higher T stage (T3-4), 60.3% had higher N stage (N1-N3) and 75.4% had advanced stage disease (Stage III-IV).31.8% of HPV positive patients had T3-T4 stage, 77.2% had N1-N3 disease and 81.8% had Stage III-IV disease (Wakisaka et al., 2015). In our study, out of all HPV positive patients, all the patients had E-cadherin expression (100%), none of the patients had vimentin expression, 100% had EMT positivity (only one patient had EMT positivity in the total sample), 71.4% were found to have low staining for podoplanin, and 85.7% had low LVC (Median LVC=9). The same values for HPV negative patients in our study were 66%, 100%, 0%, 62.3%, and 86.8%, respectively. Among the HPV positive patients present in our study, 85.8% had T3-T4 stage, 100% of the patients had N2-N3 disease and 95% had Stage IV disease. The low prevalence of vimentin is the primary reason why EMT was observed in only one patient. It is also evident that, although statistically insignificant, there is a definite correlation between positive E-cadherin expression and lower podoplanin staining with the association of HPV.

In a study by Hesham Mohammed et al. including 202 patients, out of all the E-cadherin positive patients, 53.1% had moderate well-differentiated disease, 42.3% had T3-T4 stage, 79% had N1-N3 stage and 82.1% had Stage III-IV disease. Out Of all the Vimentin positive patients in their study, 49.5% had moderate well-differentiated disease, 49.5% had T3-T4 stage, 79.1% had N1-N3 stage and 82.4% had stage III-IV disease (Mohamed et al., 2020). In our study, among all the E-cadherin positive patients, 90.5% had moderate well-differentiated disease, 83.4% had T3-T4 stage, 78.6% had N2-N3 stage and 92.9% had stage III-IV disease. All vimentin-positive patients in our study had T3-T4 stage, N2-N3 stage and Stage IV disease. It is clear from this that Vimentin expression can be considered as a prognostic indicator for an advanced disease.

Various studies have shown that podoplanin expression corresponds to nodal progression, and thereby, to advanced disease (Wakisaka et al., 2015). In our study, among the patients with higher Podoplanin expression, 81.8% had moderate well-differentiated disease, 86.4% had T3-T4 stage, 91% had N2-N3 disease and 86.3% had Stage IV disease. The same values for patients with low podoplanin expression were 94.7%, 91.6%, 71.1%, and 71%, respectively. These data are in good agreement with literature data showing that podoplanin can be considered a surrogate marker for lymphatic progression.

Posner et al studied (TAX 324) the association between HPV status and survival parameters in 264 patients with OPC. The respective tumor biopsies were subjected to PCR for HPV type 16 and assessed for correlation with demographic characteristics, OS, and PFS. Out Of the 264 patients, 42% had biopsy specimens which were evaluable of which 50% were found to be HPV-associated. 1 year-OS and PFS rates in HPV-related patients were significantly better than those in HPV-negative patients [93% vs. 69% (OS), 85% vs. 52% () respectively]. HPV-positive patients also had significantly better locoregional control than the HPV-negative patients (locoregional failure: 13% vs. 42%) (Posner et al., 2011). Gypsyamber D’Souza et al. conducted a study that included 1356 patients, of which 517 were OPSCC and which 184 patients (35.6%) were found to be HPV-positive. HPV-positive patients had a significantly lower risk of mortality than HPV-negative patients. 3-year OS was significantly better in in HPV-positive patients (82%) when compared to HPV-negative patients (45%) (D’Souza et al., 2016). In 2015, Arya et al. compared the HPV status of 3952 OPSCC patients with their respective OS rates. In this study, 62% patients were found to be HPV-positive. 2-year Overall Survival rate of HPV-positive patients was significantly higher compared to that of HPV-negative patients (93.1% vs. 77.8%) (Amini et al., 2016). In our study, it was found that p16-positive patients had significantly higher 1-year OS rates when compared to that of p16-negative patients (80% vs. 44.1%, p = 0.045). Although not statistically significant, the 1-year PFS and 1-year LRRF survival rates were also found to be evidently improved among HPV-positive patients [60% vs 31.8% (PFS), 75% vs 42.2% (LRRFS)].

Based on the results of our study, we concluded that HPV-positive status in OPSCC patients is significantly correlated with advanced nodal disease at presentation and improved clinical outcomes. Although not statistically significant, loss of podoplanin expression was found to have a clear correlation with nodal progression in HPV-positive status, similar to previous studies. In contrast to other studies, loss of E-cadherin, Vimentin expression and EMT positivity were not found to have any significant correlation with p16 expression. The improved survival pattern observed among HPV-positive patients was in agreement with previous studies. Hence, more intensive research is required to determine the exact molecular mechanisms involved in HPV-associated OPSCC. De-intensification of definitive treatment schedules can be considered for HPV-positive OPSCC after the results of prospective trials are available. HPV prevalence can be evaluated using better diagnostic methods, such as HPV PCR or HPV ISH, to obtain more accurate results. Moreover, further studies need to focus on using E-cadherin, Vimentin and Podoplanin as prognostic markers for HPV-negative patients.

Ethics and consent

approval from the institutional ethics committee (Kasturba Medical College, Mangalore with IEC KMC MLR 12-2020/419 dated 24/12/2020). As the study involved retrospective analysis of the previously available specimens from treated cancer patients, the consent of the participants was waived off from the institutional ethics committee.