Abnormal p16 Expression and Prognostic Signi cance in Chinese Esophageal Squamous Cell Carcinoma

Xue Zhang Zhongshan Hospital Fudan University Weijie Chen Zhongshan Hospital Fudan University Xiaolei Zhang Zhongshan Hospital Fudan University Lei Xu Zhongshan Hospital Fudan University Feng Gao Zhongshan Hospital Fudan University Yufeng Huang Zhongshan Hospital Fudan University Wenyi Gu Zhongshan Hospital Fudan University Dongxian Jiang (  jiangdongxian3@aliyun.com ) Zhongshan Hospital Fudan University Yingyong Hou Zhongshan Hospital Fudan University


Background
Esophageal Cancer (EC) is the seventh most frequent malignancy and the sixth most common cause of cancer-associated mortalities with the estimated 572,000 new cases and 509,000 deaths worldwide in 2018 (1). Despite improvements in multidisciplinary treatments, including surgery, radiotherapy, and systemic therapy, the prognosis for EC remains poor in the "Esophageal Cancer Belt" including China (2).
The most common histological type is esophageal squamous cell carcinoma (ESCC), which was regarded as an important public health problem in China. Recently, the major focus of cancer research is identifying the molecular changes that occur in tumorigenesis and progression (3,4). A detailed search into these alterations can discover novel biomarkers, which may help classify patients at the same stage into different subgroups in terms of their prognosis and further guide surgery or adjuvant treatment.
Therefore, reliable biomarkers are urgently required in ESCC.
P16 gene, also known as MTS-1 (major tumor suppressor 1), INK4a (inhibitor of CDK4a), or CDKN2A (cyclin-dependent kinase inhibitor 2A), is located on chromosome 9p21.3. It consists of 3 exons and 2 introns, encodes a tumor suppressor p16, which play an important role in regulating the cell cycle pathway. P16 inhibits cyclin D dependent protein kinases (CDK4 and CDK6) therefore maintaining hypophosphorylation of the retinoblastoma protein (Rb). It ultimately inhibits the release of the transcription factor E2F, preventing cell conversion from G1 phase to S phase, and eventually suppressing cell proliferation (5,6). The decreased expression or inactivation of p16 attenuates the ability of Rb to inhibit cell proliferation and allows unregulated cell-cycle progression. P16 aberration is frequently observed in a wide variety of tumors (7,8).
P16 protein expression is frequently used as a surrogate marker for human papillomavirus (HPV) infection in many kinds of cancer, including cervical cancer (9) and head and neck squamous cell carcinoma (HNSCC) (10). In HPV related tumors, E7 oncoprotein integrates into host genome and leads to inactivation of Rb, which has a negative feedback on intracellular p16 levels leading to p16 overexpression. In addition to HPV-related cancer, including lung, pancreas, colorectal, bladder, and breast tumors, p16 function is lost by gene deletions, mutations, or epigenetic silencing, which results in negative immunohistochemistry (IHC) ndings (11,12). That's to say, there are two abnormal p16 expression patterns: absent and overexpressed. Many studies to date have explored the clinicopathological and prognostic signi cance of p16 expression in tumors (9,13,14).
Some studies that have examined p16 IHC results in ESCC within the past few decades (15)(16)(17), however, the conclusions were doubtful. In their results, p16 expression was detected in 5.8-88.3% ESCC. There was no detail demonstrate and comparison among absent, expression and overexpression. Its in uence on the prognosis of ESCC patients remains unclear. Some researches demonstrate P16 expression was associated with favorable prognosis (18). No prognostic signi cance was shown in other studies (19,20).
The inconsistent conclusions are obtained due to several reasons such as clinical stage, sample size, IHC evaluation criteria, or ethnicity. Further studies are required to investigate the in uence of p16 expression in ESCC patients. Hence, we performed this study to explore the status of P16 expression using IHC methods, and analysis the association of p16 expression with clinicopathological characteristics, and prognosis of large-scale ESCC patients.

Patients and tissues
The present study included 525 patients with primary ESCC who underwent surgery at Zhongshan Hospital, Fudan University between 2007 and 2010. Eligible patients had histologically proven squamous cell carcinoma of esophagus. Exclusion criteria included patients with neoadjuvant therapy and incomplete follow-up information. Clinical data were collected from their medical records, including sex, age at diagnosis, smoking, histological grading, tumor size and location, vessel and nerve invasion.
Tumor stage was determined according to the 8 th edition of American Joint Committee on Cancer tumor, node and metastasis (TNM) classi cation system. Informed consent was submitted by all patients. The present study was approved by the Institutional Review Broad of Zhongshan Hospital. All procedures performed in studies involving human participants were in accordance with the 1964 Helsinki declaration (and its subsequent updates) of the World Medical Association or comparable ethical standards.

Tissue microarray and immunohistochemistry
Tissue microarrays were constructed as described previously (21). Brie y, all donated cylinders were extracted from the most representative areas of the tumor within the para n block and transplanted into recipient tissue microarray (TMA) block. Available TMA sections were stained with p16 antibody (clone MX007, Maixin Biotechnology Co. Ltd, Fuzhou, China, monoclonal, 1: 400 dilution) on an automated immunostainer (Leica Biosystem) according to the manufacturer's protocol with appropriate controls. The results were evaluated by 2 pathologists who was blinded to the clinical outcome and all other data on the patients.

Evaluation of P16 expression status
The p16 nuclear expression and cytoplasmic expression were scored using a histochemical or H-score like method in which the percentage of cell staining was recorded for each intensity level (-, no staining; +, weak staining intensity; ++, moderate staining intensity; or +++, strong intensity staining). Different de nitions for positive p16 expression have been used in the literature. The main de nitions were summarized as follows: 1) p16 was considered positive when >10% of the cells showed both nuclear and cytoplasmic brownish staining; 2) P16 overexpression was de ned as positive when nuclear or cytoplasmic expression has an intensity of 2+/3+ and distribution ≥70% of cancer cells (22). As there is little consensus in the literature about what constitutes positive staining for P16 antibodies, we analysis and compared our results with the abovementioned de nitions.

Statistical analysis
Statistical analyses were performed with SPSS for Windows version 24 (IBM, Armonk, NY).X 2 analysis and Fisher exact test were used as appropriate to assess the relationship between P16 expression and clinicopathological characteristics, including sex, age, smoking, histological grading, tumor size and location, vessel and nerve invasion. Disease-free survival (DFS) was calculated from the date of primary treatment to the date of recurrence, progression, or death from esophageal cancer. Overall survival (OS) was de ned from the date of primary treatment to the date of death from any cause or the date of the last follow-up. The survival rates were calculated using Kaplan-Meier method and compared using a longrank test. The univariate Cox proportional hazards model was used to analyze covariates. Factors with P value <0.05 were included in a multivariable Cox proportional hazards model using a forward stepwise procedure. All of the statistical tests performed were two-tailed, with P values <0.05 considered statistically signi cant.

Clinicopathologic Features
As reported in Table 1, which summarizes the clinical and pathological features of these 525 patients, the median age was 61 years with a wide age distribution ranging from 34 to 83 years. Four hundred twentynine patients (81.7%) were males. A history of tobacco smoking was observed in 203 (38.7%) patients. As the anatomic site, 5% cases were located in the upper esophagus, 44.8% in the middle and the other 45.5% in the lower with a mean of tumor size of 3.4 cm of the biggest axis. The histopathological diagnoses consisted of well-differentiated ESCC in 20 patients (3.8%), moderately-differentiated ESCC in  Table 2. When a cutoff value of > 10% was used, p16 positive was seen in 65 specimens (12.4%) and negative in 460 specimens (87.6%). When a cutoff value of > 70%++ was used, p16 overexpression was seen in 29 (5.5%) specimens. Among 65 ESCC with p16 IHC positive, 29 (5.5%) cases were p16 overexpression with diffuse and strong staining, and the other 36 (6.9%) cases showed focal and weak staining (Fig. 1). When these patients were divided into p16 positive group (n = 65) and negative group (n = 460) de ned by a cut-off value of 10%, the positive group demonstrated a better outcome compared with the negative group, however, this did not reach statistical signi cance (DFS P = 0.088 and OS P = 0.115) ( Fig. 2A and  2B). When these patients were divided into p16 overexpression group (n = 29) and non-overexpression group (n = 496) de ned by a cut-off value of 80% ++, there was no difference concerning DFS (P = 0.888) and OS (P = 0.933) of patients with p16 overexpression compared to those without overexpression ( Fig. 2C and 2D). When these patients were divided into p16 negative group (n = 460), focal expression group (n = 36) and overexpression group (n = 29) de ned by the above mentioned cut-off values, the survival of focal expression group tended to be better compared with negative group (DFS P = 0.040 and OS P = 0.052) and overexpression group (DFS P = 0.201 and OS P = 0.258), and there was no survival difference between negative group and overexpression group (DFS P = 0.780 and OS P = 0.837) ( Fig. 2E and 2F). The univariate analysis indicated a signi cant association between poor differentiation, vessel invasion, nerve invasion, higher clinical stage and poorer survival, and an association between p16 focal expression and favorite survival. Then multivariate analysis for OS and DFS was performed and included above mentioned factors, and only clinical stage was found to be a signi cantly independent prognostic factor ( Table 3) (Fig. 3A and 3B). When these patients were divided into p16 negative group (n = 256), focal expression group (n = 22) and overexpression group (n = 12) de ned by the above mentioned cut-off values, the survival of focal expression group was better compared with negative group (DFS P = 0.015 and OS P = 0.019), tended to be better compared with overexpression group (DFS P = 0.405 and OS P = 0.432). And p16 overexpression group tended to have better outcome compared with the negative group, however, this did not reach statistical signi cance (DFS P = 0.233 and OS P = 0.254) (Fig. 3C and 3D). The univariate analysis only indicated p16 focal expression was associated with favorite DFS and OS ( When these patients were divided into p16 negative group (n = 204), focal expression group (n = 14) and overexpression group (n = 17) de ned by the above mentioned cut-off values, there were no survival difference between focal expression group and negative group (DFS P = 0.899 and OS P = 0.848) or overexpression group (DFS P = 0.976 and OS P = 0.853) ( Fig. 3E and 3F). In univariate statistical analysis, no prognostic factor was found in these patients (Table 4).

Discussion
P16, as an important tumor suppressor protein, plays an important role in cell cycle regulation and prevents tumor development. Serrano et al. rst cloned the cDNA of its encoding gene (CDKN2A) in 1993 (23). Since then it has been widely studied in the eld of cancer research. Although some studies have explored the clinicopathological and prognostic signi cance of p16 aberration in ESCC, the results remain inconclusive because of the differences in sample sizes, methods, study populations, and evaluating criteria (16,(18)(19)(20). Therefore, we conduct this study in a larger sample size to explore and validate the association between p16 status and clinicopathological factors including survival in ESCC patients.
There are two abnormal p16 expression patterns: absent and overexpression in tumors. In HPV-driven tumors such as HNSCC and cervical cancer, viral E7 oncoprotein functionally inactivates RB protein and have a negative feedback on intracellular p16 leading to its protein accumulation. Therefore, a positive p16 expression is considered when there is a diffuse block staining with strong nuclear or nuclear plus cytoplasmic staining, and focal or patchy nuclear staining and exclusive cytoplasmic staining is interpreted as negative (9,22). In non-HPV-driven tumors such as lung, breast, pancreas and colon cancer, p16 function is lost as a result of various alterations including complete point mutation, promoter methylation, homozygous deletion and loss of heterozygosity (11,12). IHC negative for p16 protein expression is believed an accurate and relatively simple method for evaluating P16 gene inactivation, and observed at a higher frequency.
At present, there is little detailed discussion about the expression pattern in ESCC. Recent global data including Chinese study, indicates that HPV has no signi cant etiological role for ESCC (3,(24)(25)(26). P16gene inactivation through gene mutation and promoter methylation is reported in ESCC (4,17), and IHC expression is reported in 5.8-88.3% patients (15)(16)(17). In our study, we used cut-off values of 10% and 70%++, to evaluate p16 focal and diffuse expression. The former was commonly used in various tumors (27,28). The latter was similar to the new p16 criteria, which AJCC adopted in oropharyneal cancer and included for its TNM staging (22). p16 overexpression was found in 5.5% of our ESCC. p16 overexpression using this new criteria was also reported in other non-HPV-driven tumors (14). The mechanisms that lead to p16 overexpression in these tumors are not well understood. Our p16de ciency rate was 87.6%, consistent with some reports (83.1-94.2%) about ESCC (15,20,29), which demonstrate p16 inactivation might be an important molecular event for esophagus tumorigenesis. Our negative rate was higher than some older reports (30), which tended to have small sample size (< 200 cases), different cut-off values and environmental factors. Our study is one of the largest ESCC studies about p16 expression to date in china.
The prognostic effect of p16 abnormal expression on survival was explored in both HPV-driven tumors and non-HPV-driven tumors (9,27,28). P16 positive has been widely identi ed as a prognostic factor for a better outcome. The higher the p16 expression, the greater the effect it is on cell-cycle arrest and survival. At present, the prognostic studies in ESCC are still limited and variable, with some controversy. High p16 expression supposedly correlated with favorable prognosis in ByIsrid's study with 53 ESCC patients (18).
However, other retrospective studies did not nd p16 to function as a signi cant outcome predictor (19,20). In our study, 6.9% patients with p16 focal expression had better DFS than those without p16 expression, and 5.5% patients with p16 overexpression had no prognostic difference with those without p16 expression. What's more, 5.5% patients with p16 overexpression tended to be associated with a more unfavorable survival than 6.9% patients with p16 focal expression. That's to say, either overexpression or absence of p16 tend to be associated with unfavorable outcomes in ESCC. Some studies in ovarian cancer also showed that overexpression and silencing of p16 could predict worse outcome(31, 32).
In our multivariate survival analysis of 525 ESCC patients, only clinical stage was found to be a signi cantly independent prognostic factor. Then we analyzed and veri ed the prognostic value of p16 expression in earlier stage (I-II stage) and later stage (III-IVa stage), separately. No signi cant association was observed between p16 expression and clinical stage. Among 290 I-II stage patients, p16 focal expression was associated with both favorite DFS and OS, which was not found in 235 III-IVa stage patients. In other words, reduced risk of progression and mortality in the p16-focal expression patients compared with the p16 negative only in earlier stage (stage I and II) ESCC patients. Similar phenomenon is also observed for other molecules(33-35). Abnormal p16 expression may lead to malignant, abnormal cell proliferation and accelerated tumor development in ESCC, however, many other markers may interact with p16 and also contribute to these processes, especially in III-IVa stage tumors (3,4,36). These might reduce the prognostic signi cance of individual marker, and further studies are needed in more advanced ESCC.

Conclusion
In this investigation, we used immunohistochemical staining of TMA blocks to evaluate p16 expression patterns and its prognostic role in a large cohort of Chinese ESCC patients. To conclude, p16 focal expression was signi cantly associated with better DFS, especially in I-II stage ESCC, and p16 overexpression or negative tend to be associated with unfavorable outcomes. As present, this study is the large sample size providing excellent power to examine differences of protein expression patterns in Chinese ESCC. Our study will help to identify a subgroup ESCC with excellent prognosis after surgical therapy, who might not need any further adjuvant therapy after surgery with curative intention at all. This, however, is only a hypothesis, and it remains to be elucidated in a prospective trial in the future.  Table   Table 4 not available with this version.