Relationship between p16 expression and prognosis in different anatomic subsites of OSCC

2.1 Study population and design

Samples and the follow-up data of 147 eligible histopathologically confirmed OSCC patients were obtained from the head and neck tumor bio-bank of Nanjing Stomatological Hospital, Medical school of Nanjing University. All the enrolled patients underwent surgeries in the department of Oral and Maxillofacial Surgery, Nanjing Stomatological hospital between February 1, 2013 and November 13, 2014. Specifically, the OSCC primary tumors were removed with adequate surgical margins, which were sent for histopathology analysis. If margins were not deemed to be tumor free, further resection was performed. A total of 136/147 (92.5%) of patients received follow-up examinations every two months after surgery until death. Follow-up visits continued until February 9, 2018, and only 7.5% (11/147) of patients did not attend the follow-up appointment because of an incorrect phone number or address.

This study was conducted according to the ethical principles of the Helsinki Declaration and approved by Nanjing Stomatological Hospital ethics committee (date of approval, 26th February 2018; approval number 2018NL-022 (KS)).

2.2 p16 immunohistochemistry

Paraffin sections (4- μ m thick) of tumor tissues were deparaffinized and treated with 10 mM citrate buffer at 92 ∘ C for 30 minutes for antigen retrieval. After blocking in 10% bovine serum albumin, the slides were immunostained for a mouse monoclonal anti-p16 (MTM Laboratories, Heidelberg, Germany). p16 expression was rated by two pathologists without patient details according to the fraction of positive cells: negative staining = 0, 1–5% = 1, 6–25% = 2, 26–75% = 3, $>$ 75% = 4.

2.3 DNA extraction from FFPE tissues

Samples that met the criteria for inclusion were then re-cut into 10- μ m thickness and mounted on adherent slides. After scraping the tumor and adjacent non-tumor cells onto the slides, DNA was obtained by using the QIAamp DNA FFPE Tissue kit (Qiagen Inc., Valencia, CA) according to the manufacturers’ direction. Then, DNA concentration and quality were assessed by Nanodrop.

2.4 Real-time PCR

All samples were run in triplicate in a 10- μ L qPCR mix containing 25 to 50 ng DNA, 5 μ L PowerUp™ SYBR ® Green Master Mix (Applied Biosystems, Life Technologies, Austin, TX, USA) and 0.2 μ M of each primer. The samples were classified as HPV positive if two or three wells of the triplicate showed an amplifying curve and only one peak in the melt curve. CaSki, an HPV-containing cell line, was used as the positive control and Fadu, an HPV negative cell line, was used as the negative control. In order to exclude environmental contamination, a negative control without DNA templates was also evaluated. A standard curve of HPV16 E6/E7 and E7 was developed using a 5-fold diluted DNA extracted from the Caski cell line. All primers are shown in Table 1.

OPEN IN VIEWER

Correlation of p16 expression to the clinical parameters was evaluated with the chi-square test. OS was defined as the time from the date of surgery to the date of death or the date of censorship. RFS was defined as the time of surgery to the date of recurrence and relapse or the date of censorship. MFS was defined as the time of surgery to the date of metastasis or the date of censorship. The OS, RFS and MFS were estimated with Kaplan-Meier methods for all patients and subpopulations with different anatomic subsites. Log rank tests were performed to test the differences in survival between p16-positive and p16-negative patients with 25% and 75% cut off values, which were previously used to determine p16 positivity. Univariate and multivariate analyses of all patients using Cox proportional hazards regression were performed using the following variables: age, primary tumor location, TNM, differentiation, metastasis, recurrence and p16 positivity. In all statistical tests, a p value of 0.05 or below was considered as statistically significant.

OPEN IN VIEWER

3.1 Clinicopathological characteristics of patients with OSCC

147 OSCC patients, including primary sites in the tongue ( n = 82, 56%), gingiva ( n = 32, 22%) and buccal mucosa ( n = 33, 22%) were enrolled in this study. The baseline characteristics of the patients are shown in Table 2. p16 expression was analyzed by IHC. p16 mainly expressed on the nuclear and cytoplasm of tumor cells and its expression patterns were clustered or diffused (Fig. 1). Overall, 86% (126 of 147) patients had no p16 staining and 14.3% (21 of 147) patients had p16 staining with the positivity from 25% to more than 75% (Fig. 1). As the cut off values for p16 positivity were different in various researches, we selected two reported cut off points for p16, i.e., 25% and 75% (Table 2). Then we investigated the correlation of p16 expression (25% cut off or 75% cut off values) with the clinical parameters in OSCC. Regardless of cut off values for p16 positivity, p16 expression did not correlate with patient age, sex, smoking history, tumor differentiation, T stage or N stage ( $p>$ 0.05, Table 3).

3.2 Prevalence of HPV16 was rare in OSCC patients

Although p16 expression has been regarded as a surrogate marker for HR-HPV in OPSCC, its diagnostic value in OSCC is still confusing because of the diverse anatomical tumor subsites and discordant definition of p16 positivity. Therefore, real-time PCR was performed to detect HPV16 E6/E7 and E7 DNA using the two kinds of primers in p16 positive OSCC patients. Moreover, in order to confirm the copy number of HPV16 per cell, the standard curves of HPV16 E6/E7 and E7 using a 5-fold dilution of CaSki gDNA samples with known copy numbers were set up.

Only 1 out of 21 p16 positive patients (cut off value 25% and 75%) showed a standard amplifying curve and melt curve for HPV16 E7 (Fig. 2B). The result of real-time PCR was confirmed by agarose gel electrophoresis, demonstrating the presence of HPV16 in the sample (Fig. 2C). The copy number of HPV E7 DNA was 0.312 copies/genome. Moreover, the matched adjacent group without tumors did not show a standard amplifying and melt curve for HPV16 E6/E7, showing that normal epithelial cells were not infected by HPV16 (Fig. 2D).

The information of the patient was as follows, age: 69, sex: female, smoking history: no smoking, primary tumor site: Buccal mucosa, Grade: poor differentiation, T stage: 2, N stage: negative, live status: live.

3.3 The expression of p16 had no prognostic value for OSCC patients

In order to evaluate the prognostic values of p16 in OSCC, we investigated the associations between p16 expression and patients’ clinical outcomes with 25% and 75% cut off values. Whether the cut off value of p16 expression was set at 25% or 75%, no significant differences in OS, DFS and RFS between patients with p16-positivity and p16-negativity was observed (Fig. 3).

The Cox regression model was then used to determine whether the known prognostic values and p16 expression status were associated with OS. Multivariate analysis showed that TNM, metastasis and recurrence were associated with worse OS. p16 status was not statistically significant in univariate and multivariate analysis. These results showed that p16 expression had no prognostic values in OSCC patients.

OPEN IN VIEWER

OPEN IN VIEWER

As molecular differences exist between different anatomic subsites in OSCC, which highlight the importance of precise anatomical subsites classification in OSCC diagnosis, we then examined the prognostic values of p16 expression in the patients with tumors derived from tongue, gingiva and buccal mucosa. Univariable analysis showed that only the patients with p16-negative (cut off 25%) tumors located in the buccal mucosa had more RFS than patients with p16-positive tumors ( p = 0.03) (Fig. 4F). However, no significant differences in OS, DFS and RFS were found between patients with tumors located in tongue and gingiva with cut off values at 25% or 75% (Fig. 4).