Influence of p16 Protein Immunostaining on Histopathological Features of Pleomorphic Adenoma and Carcinoma ex- Pleomorphic Adenoma

The objective of this study was to evaluate the role of p16 in histologic characteristics and transition of Pleomorphic Adenoma (PA) to Carcinoma ex-PA (CxPA). So, 60 PA and 4 CxPA were histologic reviewed based on microscopic characteristics proposed by Hellquist, Triantafyllou and Dulguerov (PA) and Morais, Antony and Toluie (CxPA). Immunostaining for p16 was associated in different parenchyma and stroma of both tumors and Fisher’s/chi-square tests and Mann-Whitney test were performed (SPSS v20.0, p<0.05). In PA the periductal cells were predominantly p16- and that ductal and myoepithelial cells showed a significant increase in p16+ cells (p<0.001). In CxPA, none of the cases showed p16+ in periductal cells, most parotid cases showed p16+ in ductal cells, and one case of parotid and the submandibular case showed mild immunostaining for myoepithelial cells. There was a small reduction in p16+ in CxPA compared to PA (p=0.537), but in both tumors there was less p16+ cells in solid stroma than other (p<0.001). The p16+ cases of PA had a higher capsular thickness (p=0.047). So, the loss of p16 immunostaining does not seem to be associated with the transition from PA to CxPA, but in both tumors the loss of p16+ cells are related to microscopic aggressiveness.


Introduction
Pleophprmif adenoma (PA) consists of a benign salivary gland neoplasm, which arises due to the architectural pleomorphism, morphological complexity, and diversity in the tumor (Choi et al., 2019).Among the benign salivary gland tumors, Warthin tumor, myoepithelioma, and oncocytoma, PA is the most common tumor, corresponding to 91.3% of cases.Moreover, it is considered the most common neoplasm of all benign and malignant salivary tumors, accounting for 59.6% of the

RESEARCH ARTICLE
Influence of p16 Protein Immunostaining on Histopathological Features of Pleomorphic Adenoma and Carcinoma ex-Pleomorphic Adenoma most commonly in the parotid gland region, without the involvement of the facial nerve (Sergi et al., 2008).They are usually discovered during a routine physical examination, measuring about 2 to 5 cm.The smaller diameter tumors have firm and mobile palpation.In contrast, those of larger diameter may have protrusions that attenuate the overlying skin or mucosa (Thompson, 2006).Although PA manifests mainly in parotid glands, another preferred location is the palate.In addition to the palate, other intraoral affected sites are the upper lip, buccal mucosa, tongue, and gingiva (Erdem et al., 2011).
Histologically, PAs have an epithelial component, a myoepithelial cell component, a mesenchymal component (Almeslet, 2020), and has connective tissue capsule that may be incomplete (Ahmedi et al., 2017).In the epithelial region, tubular or solid structures contain luminal epithelial cells in the inner layer, and the non-luminal myoepithelial cells in the outer layer (Lee et al., 2000) dispersed in the mesenchyme of the lesion (Batsakis et al., 1983).The main cells that make up PA are myoepithelial cells, which have a morphological diversity due to their considerable degree of cytological modification.The spectrum of cytological transformation consists of the appearance of fusiform, transparent, cuboidal or plasmacytoid cells.The mesenchymal region of the tumor is formed by a variable fibromyxoid chondroid area.According to Lee and co-workers, in the PA tissues examined in their study, the epithelial and mesenchymal elements arise from the same cell clone, which may be a myoepithelial or ductal cell (Batsakis et al, 1983;Dardick and van Nostrand, 1985).
It is important to stress that PAs should have diligent treatment because of the tendency to recur and malignant transformation, which consists of change into carcinoma ex-PA (CxPA) (Jain et al., 2015).CxAP is a carcinoma arising from primary (de novo) or recurrent benign PA12.According to Andreasen and colleagues, the frequency of malignant transformation in patients with recurrent PA was 3.3% as opposed to 0.25% in patients with primary PA alone.Therefore, among the risk factors, it is essential to consider in the clinical evaluation the following characteristics facing the possibility of malignant transformation: age, tumors located in minor salivary glands, being the main sites of involvement of carcinoma ex-PA, and enlargement of regional lymph nodes (Seok et al., 2019;Andreasen et al., 2016).
The p16 protein encoded by the INK4a gene located on chromosome 9p21 acts as a negative regulator of the cell cycle, one of the primary mechanisms for preventing tumor formation (Romagosa et al., 2011).The function of p16 consists in the inhibition of cyclin-dependent kinase 4 (CDK4), which prevents the formation of the CDK4/Cyclin D complex, resulting in the inhibition of the phosphorylation of the retinoblastoma susceptibility protein (pRB) (Serrano et al., 1993).According to the author Jour, the p16 protein was expressed in benign and malignant salivary gland tumors, including the PA and CxPA.Furthermore, in benign tumors that express positive staining of p16, this protein demonstrates the ability to protect the cell from malignant transformation due to the control of the proliferation of oncogenic stimuli (Romagosa et al., 2011;Serrano et al., 1993;Jour et al., 2013).
Moreover, through immunohistochemistry with p16 antibody, one can detect relevant associations between HPV infection and the development of head and neck tumors since several articles prove that HPV is involved in the pathogenesis of squamous cell carcinoma (SCC).The literature describes a possible involvement of HPV in the oncogenesis of salivary gland tumors (Vageli et al., 2007).
So, as a better understanding of p16 overexpression in PA and CxPA will contribute to understanding the biological behavior of the cellular elements of these two such heterogeneous stromal tumors, the objective of this study was to evaluate the role of p16 immunostaining in histologic characteristics and transition of PA to CxPA.

Study type and patient samples
This is a quantitative and cross-sectional study.The biological samples of PA and CxPA were collected through a survey of clinical-pathological data of patients who underwent surgery for resection of salivary gland cancers at the Hospital Haroldo Juaçaba/Instituto do Câncer do Ceará (HHJ/ICC), between 2010 and 2020.We excluded specimens with incomplete clinicopathological data that did not have sufficient amounts of biological material to perform immunohistochemistry, slides and blocks damaged or deficient for microscopy, and tumors from patients who underwent neoadjuvant treatment.
After selecting the collected specimens based on the inclusion and exclusion criteria, the study counted 60 samples of PAs and four samples of carcinomas ex-PA fixed in formalin and embedded in paraffin from tissues removed by excisional biopsy, in which they were submitted to histopathological review by a pathologist with more than ten years of experience to confirm the diagnosis of the tumors.
The tumors were divided into two study groups, a group consisting of 60 PAs and a group composed of 04 carcinomas ex PA; each group was subjected to histomorphological and immunohistochemical analysis for p16, and the expression of the protein in the stromal patterns of each group was compared.

Immunohistochemical processing
Paraffin blocks containing the tumor tissue samples were sectioned to a thickness of 4 μm and mounted on silanized glass slides.Then, the sections were deparaffinized with xylene and rehydrated in graded concentrations of alcohol.Immediately after, antigen retrieval was performed by incubating the slide in Tris-EDTA solution under the temperature of 98ºC for 20 minutes in the PTLink® system (Dako ® ).Then the slides were cooled at room temperature for 20 minutes and washed with phosphate-buffered saline (PBS) solution.
Subsequently, endogenous peroxidase was blocked with 3% hydrogen peroxide diluted in buffer solution for 30 minutes, followed by further washes with buffer solution and overnight incubation with the primary mouse monoclonal antibody, clone D25 (Merck ® ), directed against the p16 protein.
After the incubation period with the primary antibody, the slides were washed with PBS and incubated with the Envision system (Dako ® ) for 30 minutes.Then, the slides were washed with PBS for 10 minutes and incubated with a solution prepared with DAB chromogenic reagent (3,3' -diaminobenzidine-tetrachloride) for 5 minutes for development.Finally, counterstaining of the slides with Harris hematoxylin for 10 seconds, dehydration in ethanol, clarification in xylene, and mounting for analysis were performed.
A p16+ oropharyngeal carcinoma sample was used as a positive control for the immunohistochemistry reaction.The same immunohistochemistry technique described above was used as a negative control but omitted the primary antibody's incubation step.

Histomorphology analysis of PA
PA slides were evaluated according to the criteria previously published: The criteria by Hellquist et al., (2019)

Ethical Aspects
This research was submitted to the Research Ethics Committee of the HHJ/ICC, observing the norms that regulate research with human beings, of the National Health Council under resolution 466/12.The opinion number was 2.251.564.

Characterization of the sample: PA
Our sample consisted of 60 PAs and 04 carcinomas ex PAs.Of the 60 PAs, most cases (n=37, 61.7%) were female.The mean age of the patients was 45.5±16.6, with the youngest patient being 19 years old and the oldest being 87 years old.Half of the sample was up to 45 years old.The most frequent location was in the parotid gland, affecting 47 patients (81.0%) (Table 1).
Regarding the parameters proposed by Hellquist et al., (2019), multifocal tumors were seen only in 3 cases (5.3%), and most cases were focal tumors.The most frequently found metaplasia was squamous metaplasia present in 28 patients (46.6%), followed by oncocytic metaplasia seen in 22 patients (36.7%); bony and sebaceous metaplasias were not seen in any case.Among the 28 patients with squamous metaplasia, 23 had this metaplasia in up to 50% of the sample, and 5 had keratin poles.The capsular invasion occurred in 10 cases (16.7%) with a mean depth of 694.44±680.28μm (Table 1).
The criteria by Triantafyllou et al., (2015) included: -Squamous metaplasias: when present, they were categorized to the extent as up to 50% of the tumor, more than 50%, and with keratin pearl formation; -Predominant stromal pattern: myxoid, chondroid, hyaline, bony, or adipocytic -Separation between cells and stroma: categorized by demarcated or indistinct.
The criteria by Dulguerov et al., (2017) included: -Capsular invasion and average capsule thickness: when present, the most significant depth was assessed.
-Other features: presence of pseudopodia or satellite nodules, compromised surgical margin, and intentional or unintentional rupture of the capsule.

Histomorphology analysis of CxPA
CxPA slides were evaluated considering these criteria previously published: The criteria by Morais et al., (2019) included -Presence of extracapsular invasion; The criteria by Antony et al., (2012) included: -Dominant cell profile: luminal or non-luminal cell; -Cellular features: presence of nuclear pleomorphism, mitosis figures, atypical mitoses, hemorrhage, necrosis, or association with another salivary gland tumor.

Immunohistochemical Analysis
The p16 protein expression pattern was classified according to the percentage of immunopositive cells in the different stromal patterns of the samples by two blinded researchers and mean of positivity was the sample unity.The cells with brown pigmentation observed within the nucleus and/or cytoplasm were considered positive.The value of the intraclass correlation coefficient was 0.922.

Statistical Analysis
The data were tabulated in a standard Microsoft Excel spreadsheet and exported to SPSS v20.0 software, in which the analyses were performed adopting a 95% confidence level.Clinical data and pathological characteristics were expressed as absolute and percentage frequency and morphometry data and rate of p16+ cells as mean and standard deviation.The percentage of p16+ cells was categorized based on the median, and the categories were associated with other characteristics  was 311.00±271.73micrometers.The maximum tumor size was 2.70±1.19cm (Table 1).
The mean and median percentage of immunostaining for p16 in the PAs was 11.35±13.93%and 5%, respectively, ranging from 0-50%.The profile for p16 protein was very variable, sometimes marking ductal, periductal, and sometimes myoepithelial cells.In most cases, periductal cells were negative for p16, whereas ductal and myoepithelial cells showed significantly increased p16 immunostaining (p<0.001).Regarding the immunostaining profile in the stroma, we can observe that the myxoid pattern showed a higher frequency of immunostaining for p16, followed by the hyaline pattern, and the solid pattern showed the lowest expression for p16 (p<0.001)(Table 1, Figure 1).

Characterization of the sample: CxPA
Only four cases of carcinomas ex PA were raised, with ages ranging from 40-66 years, 3 in males and 1 in females, 3 in the parotid gland, and 1 in the submandibular gland.
Total resection was performed as treatment in all cases, and the smallest free margin was 0.01-0.2cm.In none of the cases was facial nerve removal reported, and only 1 of the cases performed cervical emptying, and this was the neoplasm-free margins.The predominant pattern of the tumors was also non-luminal cells; only one case had another associated gland tumor, an adenoid cystic carcinoma (Table 2).
No cases had a perineural, vascular-lymphatic, or bony invasion.Pleomorphism was absent in one case, mild in two, and severe in another.Evident nucleoli were observed in the case that had the most cell atypia.The nucleus-cytoplasm ratio was observed in all cases, with one case having a high ratio.Mitosis figures were absent in one case, typical in one, and atypical in two.The extracapsular invasion was observed in one case, nuclear pleomorphism in three of the four cases, atypical mitoses in two, hemorrhage and necrosis in one case, which coincided with the case with the highest cell atypia and located in the submandibular gland (Table 2).
Regarding the immunostaining profile for p16, none of the cases showed immunopositivity in periductal cells, the three parotid cases showed immunostaining in ductal cells, and one parotid case and the submandibular case showed mild immunostaining for myoepithelial cells.In addition, myxoid stroma was seen in 2 cases, and the immunostaining profile ranged from 1-20% of labeled cells showing a mean of 7.00±8.83%(Table 2, Figure 2).There was no significant difference between p16 immunostaining of PA and CxPA (p=0.537).

Influence of immunostaining for p16 on the clinicopathological profile of PAs
Regarding the number of cells marked for p16 with the clinicopathological patterns, no pattern showed a significant association; however, those who did not have an adipocytic stromal profile in the histological feature of Triantafyllou (19) showed lower p16 expression, suggesting that this stromal profile is inversely associated with p16 protein expression (p=0.033).Additionally, cases with higher p16 expression had significantly higher mean fibrous capsule thickness than those with low expression (p=0.042), while p16 positive cases had higher mean maximum capsular thickness than p16 negative cases (p=0.047)(Table 3).
Regarding the influence of cell profile on p16 immunostaining in PAs, periductal and ductal cells did not influence the amount of total labeled cells.However, myoepithelial cells were directly associated with increased p16 expression (p<0.001) (Table 4).

Discussion
The study of clinicopathological characteristics of tumors, even benign ones, is essential to predict their biological behavior and, consequently, prognosis.PA is a benign tumor, but it may sometimes have aggressive characteristics that need to be associated with its cellular and biological profile.The present sample showed a predilection for females, close to the proportion reported in other studies, and a wide age range similar to that previously described (Almeslet, 2020;Tarakji et al., 2013;Dardick et al., 1982;Sharma et al., 2018).The most common location of PA was in the parotid gland, DOI:10.31557/APJCP.2023.24.11.3815 p16   followed by the submandibular gland, in line with other studies, such as that of Handa et al., (2009) whose series showed 80% of cases in the parotid gland and 12% in the submandibular gland.Moreover, multifocal tumors seem equally rare in our sample (Bartkowiak et al., 2022).
Regarding microscopic features, we observed a higher frequency of squamous metaplasia than the approximately 25% described in the literature (Dardick et al., 1982).According to Sharma et al., (2018), squamous cell foci are an integral feature of PA.Still, extensive squamous metaplasia is uncommon and may be misinterpreted as squamous cell carcinoma of salivary glands, completely changing the therapeutic management.In the present study, squamous metaplasia was the most prevalent, but when present, it was not extensive, reaching less than half of the tumor in 38.3% of the cases.Only five cases presented keratin pearls (8.3%).
The main cells constituting the epithelial region of the PA were non-luminal cells, present in 85.0% of cases (n=51).The stromal areas in all cases analyzed by Lee et al., (2000) were composed predominantly of non-luminal cells.In 2009, Handa et al., (2009)    Regarding the expression of p16 protein in PAs, the immunolabeling profile was quite variable since they sometimes marked ductal, periductal, and sometimes myoepithelial cells.In the present study, we observed negative immunostaining for p16 in most periductal cells and positive immunostaining for p16 in myoepithelial cells, corroborating the study by Jour et al., (2013).Furthermore, they described that PA expressed p16INK4A, especially in epithelial and myoepithelial cells.
According to Patel et al., (2007) PA and CxAP expressed p16INK4A in the neoplastic epithelial and myoepithelial components.Furthermore, they evidenced a higher expression of p16 in the benign epithelial component than a malignant component of CxPA, hypothesizing a "protective" effect for p16 against the progression of PA to carcinoma ex-adenoma pleomorphic.
Corroborating with the present study's findings, the immunolabeling profile of p16 in CxPA was lower than in PA.
Regarding the clinicopathological characterization and p16 immunostaining of the four ex-PA carcinomas analyzed in this study, it was evidenced that most cases were male, with ages ranging from 40-66 years.In the study by Hu et al., (2011) from the 50 CxAP analyzed, most cases were also male (n=36), with the age range also slightly higher than that of PA, ranging from 34 to 78 years.The cases of CxPA were mainly located in the parotid gland, which corroborates with the study of Okano et al., (2020) and the treatment mainly based on wide surgical resection (Gupta et al., 2019), and may or may not be indicated cervical emptying if nodal metastasis is suspected (Antony et al., 2012).In the present study, in only one of the cases, it was necessary to perform cervical emptying, the margins free of neoplasia.
Regarding microscopic features, the predominant pattern of the tumors was non-luminal cells.Still, according to the study by Demasi et al., (2009) the epithelial (luminal) component formed 75% of all CxAP cases (n=16).CxPA also presents severe nuclear pleomorphism with prominent nucleoli, frequent and/ DOI:10.31557/APJCP.2023.24.11.3815 p16 in Pleomorphic Adenoma and Carcinoma ex-Pleomorphic Adenoma or atypical mitotic figures, and vascular, capsular, and perineural invasion and hemorrhagic and necrotic foci (Di Palma, 2013;Khana et al., 2019).In the present study, only one case (25%) demonstrated a high degree of cellular atypia, being precisely the case that presented low p16 expression since it has been suggested that p16 immunostaining in carcinomas ex PAs is often downregulated.
The role of the p16 protein is to inhibit cyclin-dependent kinase 4 (CDK4), preventing the phosphorylation of pRb and blocking cell cycle progression from G1 to S phase (Zhang et al., 1999).In head and neck squamous cell carcinoma, it is common to have a loss of p16 expression that may be related to a worse prognosis of this tumor (Namazie et al., 2002;Worsham et al., 2006).According to Stephen et al., (2013) protein immunostaining is an important prognostic indicator in cases of oropharyngeal squamous cell carcinoma since patients positive for p16 had an improved overall survival for all sites analyzed in the study.
Thus, this study demonstrated that despite the wide variation in microscopic features of PA, its biological behavior is quite predictable and that loss of p16 immunostaining seems to be related to more aggressive microscopic features and possibly to evolution to CxPA.Perhaps the major limitation of this study is the small sample size of carcinomas ex PA due to the tumor's rarity.Still, the extensive histological review directs to an essential role in the control of cellular aggressiveness of these salivary gland tumors by p16 immunostaining.
The microscopic features of the PA and CxPA of the present sample are similar to those evaluated in the literature.The loss of expression for p16 in parts of its components seems to be related to microscopic features of aggressiveness.Cohort studies are needed to evaluate whether this marker can be a prognostic marker for these two tumors

Author Contribution Statement
Fabricio Bitu Sousa and Paulo Goberlanio Barros Silva design the study, they read and approve the final version of article.Kecynara Costa Barbosa and Iana Aragao Magalhaes performed histologic analysis of PA, they read and approve the final version of article.Gabriella Alves Juliao Costa and Osias Vieira Oliveira Filho performed histologic analysis of CxPA, they read and approve the final version of article.Jose Erialdo da Silva Junior performed p16 immunohistochemical analysis of PA and CxPA, he read and approve the final version of article.Lucio Flavio Gonzaga Silva and Sergio Ferreira Juaçaba revised clinic data and rescue paraffin blocks and medical records, They read and approve the final version of article.

Figure 2 .
Figure 2. Histological Features and Immunostaining for p16 in CxPA.A: carcinoma ex pleomorphic adenoma hypercellular pattern showing solid stroma (a); B: carcinoma ex pleomorphic adenoma hypercellular pattern showing hyaline (a) and myxoid (b) stroma; C: classical pattern pleomorphic carcinoma ex-adenoma showing large amounts of ductal cells (a); D: carcinoma ex-pleomorphic adenoma showing extensive cellular and nuclear pleomorphism; E: carcinoma ex-pleomorphic adenoma showing keratin pearl (a); F: carcinoma ex-pleomorphic adenoma showing large areas of intratumoral hemorrhage (a); G: positive immunostaining for p16 in luminal ductal cells, and negative in non-luminal ductal cells; H: positive immunostaining for p16 in myxoid stroma (myoepithelial cells); I: negative immunostaining for p16 in solid stroma. included:
in Pleomorphic Adenoma and Carcinoma ex-Pleomorphic Adenoma
also evidenced that