OL Oncology Letters 1792-1074 1792-1082 D.A. Spandidos 10.3892/ol.2017.7218 OL-0-0-7218 Articles Expression of Beclin 1 and Bcl-2 in pancreatic neoplasms and its effect on pancreatic ductal adenocarcinoma prognosis SongShanshan WangBaosheng GuShuailin LiXiaocheng SunShaolong Department of Pancreas and Endocrine Surgery, Shengjing Hospital of China Medical University, Shenyang, Liaoning 110004, P.R. China Correspondence to: Professor Shaolong Sun, Department of Pancreas and Endocrine Surgery, Shengjing Hospital of China Medical University, 36 Sanhao Street, Heping, Shenyang, Liaoning 110004, P.R. China, E-mail: sj_sunsl@126.com 12 2017 18 10 2017 14 6 7849 7861 12112015 13042017 Copyright: © Song et al. 2017 This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

Aberrant expression of Beclin 1 and B-cell lymphoma-2 (Bcl-2) has been identified in a variety of human tumors; however, little information is available for pancreatic neoplasms. The present study analyzed the expression of Beclin 1 and Bcl-2 in pancreatic ductal adenocarcinoma (PDAC) and solid pseudopapillary neoplasm (SPN) of the pancreas, and evaluated their prognostic significance for PDAC. The present study included 117 PDAC, 43 SPN and 32 chronic pancreatitis (CP) cases. Levels of Beclin 1 and Bcl-2 expression were evaluated semiquantitatively by immunohistochemistry, and their correlation with the survival of patients with PDAC was determined. Beclin 1 was upregulated in 74 (63.2%) PDAC, 26 (60.5%) SPN, and 14 (43.8%) CP cases. Bcl-2 was upregulated in 38 (32.5%) PDAC, 11 (25.6%) SPN and 24 (75.0%) CP cases. High Beclin 1 and low Bcl-2 expression was significantly correlated with poor differentiation and distant metastasis in PDAC, and associated with the presence of nuclear pleomorphism in SPN and with advanced Tumor-Node-Metastasis stage in PDAC. Beclin 1 and Bcl-2 levels were inversely correlated in PDAC, whereas they were positively correlated in SPN. Low Beclin 1 and high Bcl-2 expression was associated with improved disease-free survival and overall survival (OS). However, the association of Beclin 1 with survival was not significant in the Cox analysis, whereas Bcl-2 expression was significantly correlated with OS in the multivariate analysis. In conclusion, Beclin 1 upregulation exacerbated the progression and aggressiveness of pancreatic neoplasms, and Bcl-2 downregulated expression was an independently poor prognostic factor for PDAC.

 Beclin 1 B-cell lymphoma 2 pancreatic ductal adenocarcinoma solid pseudopapillary neoplasm immunohistochemistry Introduction

Pancreatic cancer is a highly aggressive cancer, with a 5-year overall survival (OS) rate of <1.0%, and is one of the most frequent causes of cancer-associated mortality worldwide (1). The poor outcome for patients with pancreatic cancer is associated with its late diagnosis due to the paucity of symptoms, rapid tumor progression and unresponsiveness to chemotherapy or radiotherapy. These parameters result in low resectability rates following diagnosis, early recurrence subsequent to resection and an overall poor survival rate (2). Therefore, pancreatic cancer is the object of extensive study, to obtain an improved understanding of the development and progression of the disease, and to improve the survival and quality of life of patients.

In total, >90% of pancreatic cancer cases arise from the ductal epithelium of the pancreas and are termed pancreatic ductal adenocarcinoma (PDAC). Extensive efforts have been made to identify potential biomarkers that may be used to develop anti-metastatic treatments and improve prognostic evaluation (3). Unlike PDACs, solid pseudopapillary neoplasms (SPNs) of the pancreas are distinctive tumors that exhibit a low malignancy potential and are associated with a favorable prognosis following resection (4).

Autophagy, which is known as type II programmed cell death, is an important pathway for the degradation and recycling of cellular components and the response to stress (5). At basal levels, autophagy controls cellular homeostasis by removing potential intrinsic toxic components and enables cells to perform structural remodeling, which serves an essential role in cellular differentiation, development and survival in response to starvation conditions (6,7). The aberrant regulation of these mechanisms contributes to the pathogenesis of a variety of diseases, including cancer (6,8).

The Beclin 1 gene was the first mammalian autophagy gene identified, and it is located on chromosome 17q21 (6). Beclin 1 has received special attention in studies investigating autophagy, as it is the first autophagy protein demonstrated to be a haploinsufficient tumor suppressor in cancer cell experiments (9,10). However, studies associating Beclin 1 with the pathogenesis of human cancer are limited, and the correlation is controversial. Decreased expression of Beclin 1 has been detected in glioblastoma (11), hepatocellular carcinoma (12), esophageal cancer (13), non-small cell lung cancer (14), bladder urothelial tumors (15) and breast cancer (16), whilst increased expression of Beclin 1 has been identified in ovarian cancer (17), colorectal cancer (18) and gastric cancer (19).

The BCL2, apoptosis regulator (BCL2) gene, located on chromosome 18q21, is a member of the Bcl-2 family, which serves an important role in apoptosis suppression (20,21). The anti-apoptotic function extends cell survival in normal and tumor cells by inhibiting different cell death mechanisms (22). Bcl-2 is involved in cellular morphogenesis and differentiation; it functions as an oncogenic and anti-death molecule by modulating various homeostatic, developmental and disease processes (23–25). Although overexpression of Bcl-2 confers a survival advantage to the cell that may lead to rapid and uncontrolled cellular proliferation, which characterizes the development of cancer, differences in the expression of Bcl-2 in numerous types of human cancer have been identified, similar to that found for Beclin 1. Bcl-2 downregulation has been revealed in breast cancer and pancreatic cancer (26,27), whereas increased expression of Bcl-2 upregulation has also been demonstrated (16,28). The significance of Bcl-2 expression for cancer prognosis remains unclear.

Beclin 1 was originally identified in a yeast two hybrid screen as a Bcl-2 family-interacting protein that serves a pivotal role in the process of autophagy (29). The binding between Beclin 1 and Bcl-2 is regulated by a variety of proteins and compounds that enhance or inhibit the Bcl-2/Beclin 1 interaction to repress or activate autophagy, and serve important roles in the crosstalk between autophagy and apoptosis (30,31). The cooperation between Beclin 1 and Bcl-2 may additionally affect tumorigenesis and progression.

Studies suggest that Beclin1 and Bcl-2 expression in tumor cells varies according to tumor and tissue type, which requires additional investigation (11–19,26–28). There is a study that assesses Beclin 1 and Bcl-2 expression levels in pancreatic tumors, their interactions, and the correlation between their expression and patient survival (32). The aim of the present study was to examine the expression of Beclin 1 and Bcl-2 by immunohistochemistry in pancreatic neoplasms, including PDACs, SPNs and chronic pancreatitis (CP) tissues as the controls for PDAC. The expression of Beclin 1 and Bcl-2 was evaluated in correlation with clinicopathological parameters and the survival of patients to assess their value as markers of tumor development and prognosis.

 Patients and methods <sec> <title>Patients and tissue samples

Formalin-fixed and paraffin-embedded tissue samples from 117 patients with PDAC, 43 patients with SPN and 32 patients with CP were selected from the archived materials of the Department of Pathology of Shengjing Hospital of China Medical University (Shenyang, China). The PDAC group consisted of 70 men and 47 women, the age of which ranged between 37 and 77 years (mean age ± standard deviation, 58.89±9.27). The SPN group consisted of 5 men and 38 women, the age of which ranged between 9 and 65 years (mean age ± standard deviation, 32.77±13.72). Patients diagnosed with PDAC, SPN or CP underwent pancreaticoduodenectomy or distal pancreatectomy between January 2011 and May 2014. Sufficient clinical and survival information of the PDAC and SPN patients was available.

Patients without complete survival information or those who succumbed to surgical complications were excluded from the study. No patients had received any neoadjuvant chemotherapy or chemoradiotherapy prior to the surgical resection. All patients were followed up by interview in the clinic or by phone call. Data on chemotherapy following surgery, disease-free survival (DFS) time and OS time were collected for each patient. Cases with distant metastasis or R1 resection were excluded from DFS evaluation. The total period of follow-up was 18–57 months.

All samples were initially stained with Mayer's hematoxylin for 5 min and 2% eosin for 1 min at room temperature; two independent investigators evaluated each section to confirm the foci and the differentiation of PDAC microscopically, and the most representative section from each case was used for immunohistochemical staining.

The present study was approved by the Clinical Ethics Committee of Shengjing Hospital of China Medical University.

 Immunohistochemistry

Immunohistochemical staining was performed using the biotin-free polymer detection system. The formalin-fixed, paraffin-embedded sections (4-µm thick) were deparaffinized and dehydrated via dimethylbenzene and a graded alcohol series, respectively.

Microwave antigen retrieval was achieved in sodium citrate buffer (pH 6.0) for Beclin 1, and in Tris-EDTA buffer (pH 9.0) for Bcl-2, each for 20 min at 100°C. Endogenous peroxidase activity was blocked by incubation with 0.3% hydrogen peroxidase at 37°C for 15 min. Sections were then incubated with primary mouse monoclonal antibodies against Beclin 1 (1:400; cat. no. ab114071; Abcam, Cambridge, UK) and rabbit monoclonal antibodies against Bcl-2 (1:350; cat. no. ab32124; Abcam) at 4°C overnight. A two-step immunohistochemical staining kit was used for expression analysis (cat. no. PV-9000; Zhongshan Jinqiao Biotechnology Co. Ltd., Beijing, China). Tissue sections were then incubated with a polymer auxiliary for 15 min at 37°C and a biotin-free polymeric horseradish peroxidase-linked antibody-conjugate agent from the kit for 25 min at 37°C. Following incubation, the reaction product was detected using diaminobenzidine (ZLI-9018; Zhongshan Jinqiao Biotechnology Co. Ltd., Beijing, China). Finally, the sections were counterstained with Mayer's hematoxylin as aforementioned, dehydrated, and mounted.

The normal ductal epithelium in breast tissue obtained from Pathology Department of Shengjing Hospital of China Medical University and the lymphocytes in the pancreatic stroma of the included pancreatic samples were used as positive controls for Beclin 1 and Bcl-2, respectively. The negative controls were obtained by replacing the primary antibodies with 0.01 mol/l phosphate-buffered saline.

Images were captured with a light microscope and NIS-Elements F3.0 software (Nikon Corporation, Tokyo, Japan).

 Evaluation of immunohistochemical staining

Immunohistochemical sections were independently evaluated by two investigators in a blinded manner. At least five fields at a magnification, ×400 were randomly chosen per section. Cytoplasmic or membranous staining for Beclin 1 and Bcl-2 was considered to represent a positive immunoreaction. Beclin 1 and Bcl-2 expression levels were estimated semiquantitatively based on the combination of staining intensity and the percentage of positively stained cells (proportion score). The staining intensity was scored as follows: 0, negative; 1, weak; 2, moderate; and 3, strong. According to the percentage of positively stained cells in whole foci, the staining extent was categorized into 5 grades as follows: 0, negative; 1, 0–25%; 2, 26–50%; 3, 51–75%; and 4, 76–100%. The intensity score and proportion score were multiplied together for a final score. Final scores were as follows for Beclin 1 and Bcl-2: <6, low expression; and ≥6, high expression.

 Clinical outcome assessment

DFS was calculated from the date of surgery to the date of detection of local recurrence/distant metastasis or to the date of mortality due to PDAC. DFS was only calculated in patients with R0 resection (free resection margins). Patients with R1 resection (resection margins invaded) were considered to have failure at time zero. OS was defined as the time from the date of surgery to the date of mortality or to the date of last follow-up if patients remained alive. The DFS of patients with PDAC was acquired in 97 cases with R0 resection; the mean DFS was 10.9 months (median, 6.0 months; range, 0.8–32.7 months). The OS of patients with PDAC was acquired in 117 cases; the mean OS was 13.6 months (median, 10.4 months; range, 1.37–42.0 months). None of the patients with SPN exhibited local recurrence/distant metastasis or had succumbed by the date of last follow-up.

 Statistical analysis

Statistical analysis was performed using SPSS v.18.0 software (SPSS, Inc., Chicago, IL, USA). Pearson's χ2 test and Fisher's exact test were used to evaluate the association between Beclin 1 and Bcl-2 expression and several clinicopathological variables, and the difference between the Beclin 1 and Bcl-2 expression in PDAC and CP cases. The correlation coefficients (r and P-values) between Beclin 1 and Bcl-2 among PDAC, SPN and CP tissues were obtained using the Spearman's test. Survival curves of DFS and OS were analyzed by the Kaplan-Meier method and compared with the log-rank test. Multivariate analysis of prognostic factors was tested with the Cox proportional hazards model. P<0.05 was considered to indicate a statistically significant difference.

 Results <sec> <title>Beclin 1 expression in pancreatic neoplasms

Assessment of Beclin 1 expression in pancreatic tissues demonstrated positive staining in cancerous foci, SPN foci and normal epithelium in CP tissues, and in the cytoplasm of intralobular ductal epithelial cells, acinar cells and pancreatic islet cells (Fig. 1). Pancreatic islet cells stained positive for Beclin 1, and were used as the positive control for Beclin 1 immunostaining in subsequent experiments. High Beclin 1 expression was observed in 74 (63.2%) PDAC cases, while 43 (36.8%) PDAC cases demonstrated low expression. PDAC tissues exhibited significantly higher Beclin 1 expression compared with the normal ductal epithelium in CP tissues (P=0.047) (Table I).

The association between Beclin 1 expression and the clinicopathological characteristics of patients with PDAC is described in Table II. High Beclin 1 expression was correlated with poor histological differentiation (P=0.001) and distant metastasis (P=0.021).

High Beclin 1 expression was detected in 26 (60.5%) patients with SPN and was correlated with pancreatic body or tail location (P=0.018) and the presence of nuclear pleomorphism (P=0.031). There was no significant correlation between Beclin 1 expression and other clinicopathological features (Table III).

 Bcl-2 expression in pancreatic neoplasms

Cytoplasmic or membranous Bcl-2 immunostaining was observed in the majority of the normal ductal (intralobular and interlobular) cells and certain acinar compartments, whilst no staining was observed in the endocrine component (Fig. 2).

Decreased Bcl-2 expression was detected in PDAC foci compared with the normal ductal epithelium (Table I). High Bcl-2 expression was detected in 38 (32.5%) PDAC cases, while 79 (67.5%) PDAC cases demonstrated low expression. The correlation between Bcl-2 and the clinicopathological characteristics of PDAC patients is summarized in Table II. Decreased Bcl-2 expression was correlated with cancerous pancreatic body or tail location (P=0.017), poor histological differentiation (P=0.002), distant metastasis (P=0.049) and advanced Tumor-Node-Metastasis (TNM) (33) stage (P=0.017).

High Bcl-2 expression was detected in 11 (25.6%) patients with SPN. There was no significant correlation between Bcl-2 expression and clinicopathological features in SPN (Table III).

 Association between Beclin 1 and Bcl-2 expression

As demonstrated in Table IV, Beclin 1 expression level was not associated with Bcl-2 expression in CP cases (P=0.629, r=0.073). However, there was a statistically significant inverse correlation (P<0.001, r=−0.342) between the level of Beclin 1 and Bcl-2 expression in PDAC cases. A statistically significant direct correlation (P=0.016, r=0.365) was detected between Beclin 1 and Bcl-2 expression in SPN cases.

 Association of clinicopathological characteristics, Beclin 1/Bcl-2 expression and patient survival in PDAC: A univariate survival analysis

Univariate survival analysis with the Kaplan-Meier method and log-rank test was used to calculate the effect of well-established clinical prognostic factors and the levels of Beclin 1 and Bcl-2 expression on the DFS and OS of patients with PDAC (Table V). The outcomes demonstrated that longer DFS was significantly correlated with head location (P=0.001), improved differentiation (P<0.001), limited tumor size and local invasion (T stage; P=0.003) and high expression of Bcl-2 (P=0.030). The median DFS of patients with high Beclin 1 expression was shorter compared with that of patients with low expression, although the difference did not reach statistical significance (Fig. 3A).

The effects of prognostic factors on OS was next analyzed. The results demonstrated that a longer OS was significantly associated with pancreatic head location (P=0.004), improved differentiation (P<0.001), limited tumor size and local invasion (P<0.001), negative lymph node metastasis (P=0.022), no distant metastasis (P<0.001) and lower TNM stage (P<0.001). Low expression of Beclin 1 and high expression of Bcl-2 were predictive factors of improved OS (P=0.038 and P=0.017, respectively; Fig. 3C and D).

 Independent prognostic factors of PDAC: A multivariate survival analysis

To avoid the bias of univariate analysis, the expression of Beclin 1, Bcl-2 and other characteristics were examined by multivariate Cox analysis (Table V). In the multivariate analysis, Beclin 1 and Bcl-2 were not independent prognostic parameters for DFS [Beclin 1: Hazard ratio (HR), 0.870; 95% confidence interval (CI), 0.532–1.423; P=0.580; Bcl-2: HR, 1.400; 95% CI, 0.810–2.421; P=0.228]. DFS was affected by improved differentiation (HR, 0.399; 95% CI, 0.190–0.839; P=0.015) and limited tumor size and local invasion (HR, 0.399; 95% CI, 0.190–0.839; P=0.014) independently.

Bcl-2 was an independent prognostic biomarker for OS (HR, 1.850; 95% CI, 1.053–3.249; P=0.032), whereas Beclin 1 was not an independent prognostic parameter for OS (HR, 0.927; 95% CI, 0.557–1.543; P=0.771). No smoking (HR, 0.546; 95% CI, 0.326–0.914; P=0.021), improved differentiation (HR, 0.394; 95% CI, 0.201–0.774; P=0.007), limited tumor size and local invasion (HR, 0.307; 95% CI, 0.175–0.539; P<0.001) and chemotherapy following surgery (HR, 2.415; 95% CI, 1.218–4.544; P=0.006) were also associated with improved OS.

 Discussion

The burden of pancreatic cancer is increasing worldwide, making it a serious health concern. Risk factors such as older age, smoking and long-term diabetes mellitus have been associated with pancreatic tumorigenesis (34). Surgical resection is considered the best potentially curative treatment; however, the majority of patients are diagnosed at an unresectable locally advanced or distal metastatic stage due to the aggressiveness of the disease (34). The effects of surgical resection on OS remains minimal, and the local failure rate is as high as 50–80% in patients who successfully undergo surgical resection, resulting in poor quality of life (35). Pancreatic cancer is difficult to prevent and diagnose early; it progresses rapidly and exhibits a poor prognosis. Therefore, extensive studies have been conducted to understand the mechanisms underlying pancreatic cancer development at the biochemical, genomic and proteomic levels to inhibit the progression of pancreatic cancer and to improve patient prognosis.

Autophagy and apoptosis have been studied extensively in association with tumor progression in previous years, particularly Beclin 1 and Bcl-2 as the primary factors in the 2 programmed cell death pathways. The correlation between autophagy and apoptosis is complex and varies with cell type and stress stages (31). Autophagy may initiate or inhibit apoptosis depending on the cellular context and stimulus, and the inhibition of autophagy may increase the sensitivity of cells to apoptotic signals (6). The present study focused on analyzing the correlation between Beclin 1 and Bcl-2 expression in pancreatic tissues.

In previous studies, Beclin 1 downregulation was detected in glioblastoma, hepatocellular carcinoma, esophageal cancer, non-small cell lung cancer, bladder urothelial tumors and breast cancer (11–16), whereas Beclin 1 upregulation was identified in colorectal, ovarian and gastric cancer (17–19). The correlation between Beclin 1 expression and clinicopathological characteristics differs among different cancer tissues. Decreased expression of Beclin 1 is correlated with poor differentiation or advanced TNM stage, inducing early tumor relapse and causing low OS in glioblastoma (11), hepatocellular carcinoma (12), non-small cell lung cancer (14), breast cancer (16), ovarian cancer (17) and gastric carcinoma (19). Han et al (18) performed a meta-analysis, including six studies, to evaluate the prognostic significance of Beclin-1 expression in colorectal cancer, and revealed that Beclin 1 overexpression is associated with tumor metastasis and a poor prognosis in affected patients. Therefore, the effects of Beclin 1 expression levels differ among malignant tissues.

In the present study, high Beclin 1 expression was observed in 74 (63.2%) PDAC cases, while 43 (36,8%) PDAC cases exhibited low expression. Compared with the normal ductal epithelia in CP tissues, patients with PDAC demonstrated high Beclin 1 expression, and the difference was more apparent in poorly differentiated tissues and patients with advanced TNM stage. In the analysis of the correlations between Beclin 1 expression and clinicopathological characteristics, high Beclin 1 expression was associated with poor histological differentiation and distant metastasis. The univariate analysis of survival indicated that, despite a longer median DFS and OS in patients with low Beclin 1 expression compared with that in those patients with high expression, the difference was not statistically significant. In the multivariate Cox analysis, although low Beclin 1 expression was associated with improved OS (P<0.05), it was not an independent prognostic factor for PDAC. The significant prognostic factors based on PDAC itself were the differentiation status, local advancement and distal metastasis. Taken together, these results suggest that Beclin 1 expression is upregulated in PDAC tissues, and increased Beclin 1 levels serve a significant role in PDAC progression. High Beclin 1 expression affects the survival of patients with PDAC by affecting tumor differentiation, local advancement and distal metastasis; however, it was not an independent indicator of prognosis in patients with PDAC in the present study.

To the best of our knowledge, there is only one previous study analyzing Beclin 1 expression in patients with PDAC. In that study, Kim et al (36) evaluated Beclin 1 expression levels in 63 PDAC cases, with positive expression in 14 (22.2%) cases and negative expression in 49 (77.8%) cases. The results of correlation and survival analyses were different from those of the present study, and the immunohistochemical pattern was also different. The opposite outcomes may be attributed to differences in the methods used. In the present study, the Power Vision two-step method was chosen to perform the immunohistochemical staining, whereas Kim et al (36) chose the streptavidin-peroxidase three-step method in their study. In our opinion, the streptavidin-peroxidase three-step method is not suited to pancreatic tissue.

Bcl-2 protein is considered a carcinogenic protein, as it contributes to apoptosis inhibition; its tumorigenic potential has been demonstrated in animal models (37) and is supported by its overexpression in a variety of tumors and in lymphomas, in which Bcl-2 acts as an oncogene (38,39). However, in certain solid tumors, Bcl-2 paradoxically appears to exert a tumor suppressor effect, and its expression is associated with favorable prognostic features. Callagy et al (40) performed a meta-analysis, including 17 studies and 5,892 breast cancer cases, to determine whether Bcl-2 is an independent prognostic marker in breast cancer. The meta-analysis clearly supported the prognostic role of Bcl-2 expression in breast cancer, as assessed by immunohistochemistry, and demonstrated that it is associated with DFS and OS. The mechanisms by which Bcl-2 exerts its protective effects remain unclear.

In the present study, it was demonstrated that Bcl-2 expression was lower in PDAC tissues compared with the normal ductal epithelium in CP tissues, and that it was associated with poor differentiation, positive distant metastasis, advanced TNM stage and a poor prognosis. The outcomes revealed that Bcl-2 is downregulated in PDAC tissues and that Bcl-2 downregulation promotes PDAC progression, suggesting that Bcl-2 is an independent prognostic indicator in PDAC. Consistent with the results of the present study, previous studies demonstrated decreased Bcl-2 expression in PDAC cancerous foci, and low Bcl-2 expression was associated with poor differentiation and high TNM stage cases (26,41,42). However, the prognostic significance of Bcl-2 expression in patients with PDAC remains controversial, which may be attributed to population differences.

Knowlton et al (43) analyzed cell cycle progression in breast cancer cells and demonstrated a decreased S phase fraction and an increased G1/G0 fraction in Bcl-2(−) cells, suggesting that Bcl-2 prolongs the cell cycle. Additionally, Bcl-2 also enhanced the G0 fraction and delayed G0 to S transition (44). Ke et al (45) revealed that Bcl-2 upregulation serves a critical role in the regulation of cell adhesion and migration via the formation of Bcl-2 and gelsolin complexes, leading to the inhibition of cell diffusion, given the established correlation of cell motility with cancer metastasis. Despite the antiapoptotic effects favoring tumor survival, these results may explain why the expression of Bcl-2 in certain tumor cell types reduces the potential for metastasis and improves patient prognosis.

Another notable reason for the high relapse and low 5-year OS of PDAC is the poor sensitivity to chemotherapy, radiotherapy or immunotherapy. Therefore, a number of patients with PDAC refuse adjuvant therapy subsequent to surgery. As an important adjuvant therapy following tumor resection, chemotherapy serves an essential role in preventing tumor relapse and prolonging OS. Chemotherapy is the adjuvant therapy method considered to be potentially effective in PDAC (2). The modulation of cell apoptosis is one of the mechanisms of chemotherapy. A meta-analysis performed by Yang et al (46) revealed that Bcl-2 expression is a predictive factor for chemotherapy sensitivity, and negative Bcl-2 expression is associated with a good chemotherapy response in patients with breast cancer. Therefore, the modulation of Bcl-2 expression may potentially benefit additional clinical treatment for PDAC.

In the present study, Beclin 1 and Bcl-2 levels were inversely correlated in PDAC tissues. The inverse correlation between Beclin 1 and Bcl-2 expression has been demonstrated in previous studies (16,27,28,47). Under normal conditions, Bcl-2 inhibits Beclin 1, whereas under conditions of stress, Beclin 1 dissociates from Bcl-2, allowing the activation of Vps34 and the subsequent stimulation of autophagy (48). Bcl-2 inhibits autophagy when bound to Beclin 1 and downregulates Beclin 1, which is associated with the outcomes of patients in the present study (48). Reduced Bcl-2 expression leads to decreased interaction with Beclin 1, resulting in increased Beclin 1 levels and the stimulation of autophagy, primarily in patients with PDAC with poor prognostic factors.

SPN was first described by Frantz in 1959 (4), and defined by the World Health Organization as a ‘solid pseudopapillary neoplasm’, a low grade malignant neoplasm of the exocrine pancreas in 2010 (49). SPNs primarily affect young women in their 20 sec and 30 sec, and generally form a well-demarcated mass with a mean diameter of 5–10 cm. Despite their large size, the majority of these tumors behave indolently; complete surgical resection of SPNs may lead to a favorable prognosis, with a 5-year survival rate >95% (50).

The present study detected Beclin 1 and Bcl-2 expression in SPN tissues, and detected 26 (60.5%) cases with high Beclin 1 expression and 11 (25.6%) cases with high Bcl-2 expression. High Beclin 1 expression was significantly correlated with the presence of nuclear pleomorphism. This result confirmed that increased Beclin 1 expression leads to poor tumor differentiation. Unlike PDAC, a direct correlation was detected between the levels of Beclin 1 and Bcl-2 expression in SPN cases. Additionally, the expression levels of Beclin 1 and Bcl-2 in normal ductal epithelium of CP tissues were not correlated. The differences in expression patterns indicated the tissue specificity of Beclin 1 and Bcl-2, suggesting that autophagy and apoptosis exhibit separate regulatory mechanisms in SPN. None of the patients exhibited tumor relapse, and all patients were alive by the date of last follow-up.

The present study has several limitations. Firstly, normalized chemotherapy for PDAC is generally performed based on the guidelines for pancreatic cancer treatment; however, few patients undergo chemotherapy due to the poor sensitivity of PDAC to adjuvant therapy and the ineffectiveness of primary chemotherapy strategies for PDAC. Therefore, the present study was unable to collect a large sample size of PDAC cases to analyze the association between the sensitivity to chemotherapy and Beclin 1 and Bcl-2 expression. Secondly, due to the unknown pathogenesis of SPNs, the present study was unable to detect changes in Beclin 1 and Bcl-2 expression in comparison with those in normal pancreatic tissues. Finally, due to the short follow-up time, none of the patients with SPN exhibited tumor relapse and all patients were alive by the date of last follow-up. Future studies with a larger sample size are necessary to overcome these limitations.

To conclude, the results of the present study suggest that Beclin 1 upregulation and Bcl-2 downregulation in PDAC promotes cancer progression, and that their expression is inversely correlated. High expression of Bcl-2 is a favorable prognostic indicator in PDAC. Unlike the expression pattern in PDAC, Beclin 1 and Bcl-2 are positively correlated in SPN. High Beclin 1 expression may lead to poor progression of neoplasms.

 Acknowledgements

The authors would like to thank the Science and Technology Department of Liaoning Province for supporting the present study (grant no. 2013225021).

 References XiaoAYTanMLWuLMAsraniVMWindsorJAYadavDPetrovMSGlobal incidence and mortality of pancreatic diseases: A systematic review, meta-analysis and meta-regression of population-based cohort studiesLancet Gastroenterol Hepatol14555201610.1016/S2468-1253(16)30004-828404111 ParkerSLTongTBoldenSWingoPACancer statistics, 1997CA Cancer J Clin47527199710.3322/canjclin.47.1.58996076 SinghDUpadhyayGSrivastavaRKShankarSRecent advances in pancreatic cancer: Biology, treatment and preventionBiochim Biophys Acta18561327201525977074 FrantzVKAtlas of tumor pathology, Section VII, Fascicles 27 and 28. Tumors of the Pancreas. Washington, D.C.: Armed forces institute of pathologyBrit J Surg473343341959 LevineBKlionskyDJDevelopment by self-digestion: Molecular mechanisms and biological functions of autophagyDev Cell6463477200410.1016/S1534-5807(04)00099-115068787 LiangXHJacksonSSeamanMBrownKKempkesBHibshooshHLevineBInduction of autophagy and inhibition of tumorigenesis by beclin 1Nature402672676199910.1038/4525710604474 WangCWKlionskyDJThe molecular mechanism of autophagyMol Med96576200312865942 ShintaniTKlionskyDJAutophagy in health and disease: A double-edged swordScience306990995200410.1126/science.109999315528435 QuXYuJBhagatGFuruyaNHibshooshHTroxelARosenJEskelinenELMizushimaNOhsumiYPromotion of tumorigenesis by heterozygous disruption of the beclin 1 autophagy geneJ Clin Invest11218091820200310.1172/JCI2003914638851 YueZJinSYangCLevineAJHeintzNBeclin 1, an autophagy gene essential for early embryonic development, is a haploinsufficient tumor suppressorProc Natl Acad Sci USA1001507715082200310.1073/pnas.243625510014657337 MiraccoCCosciEOliveriGLuziPPacentiLMonciattiIMannucciSDe NisiMCToscanoMMalagninoVProtein and mRNA expression of autophagy gene Beclin 1 in human brain tumoursInt J Oncol30429436200717203225 QiuDMWangGLChenLXuYYHeSCaoXLQinJZhouJMZhangYXEQThe expression of beclin-1, an autophagic gene, in hepatocellular carcinoma associated with clinical pathological and prognostic significanceBMC Cancer14327201410.1186/1471-2407-14-32724885292 ChenYLuYLuCZhangLBeclin-1 expression is a predictor of clinical outcome in patients with esophageal squamous cell carcinoma and correlated to hypoxia-inducible factor (HIF)-1alpha expressionPathol Oncol Res15487493200910.1007/s12253-008-9143-819130303 ZhouWYueCDengJHuRXuJFengLLanQZhangWJiDWuJAutophagic protein beclin 1 serves as an independent positive prognostic biomarker for non-small cell lung cancerPLoS One8e80338201310.1371/journal.pone.008033824260370 BaspinarSBircanSYavuzGKapucuogluNBeclin 1 and bcl-2 expressions in bladder urothelial tumors and their association with clinicopathological parametersPathol Res Pract209418423201310.1016/j.prp.2013.04.00623722017 DongMWanXBYuanZYWeiLFanXJWangTTLvYCLiXChenZHChenJLow expression of beclin 1 and elevated expression of HIF-1α refine distant metastasis risk and predict poor prognosis of ER-positive, HER2-negative breast cancerMed Oncol30355201310.1007/s12032-012-0355-023408367 CaiMHuZLiuJGaoJLiuCLiuDTanMZhangDLinBBeclin 1 expression in ovarian tissues and its effects on ovarian cancer prognosisInt J Mol Sci1552925303201410.3390/ijms1504529224675697 HanYXueXFShenHGGuoXBWangXYuanBGuoXPKuangYTZhiQMZhaoHPrognostic significance of beclin-1 expression in colorectal cancer: A meta-analysisAsian Pac J Cancer Prev1545834587201410.7314/APJCP.2014.15.14.555124969889 ChenYBHouJHFengXYChenSZhouZWZhangXSCaiMYDecreased expression of beclin 1 correlates with a metastatic phenotypic feature and adverse prognosis of gastric carcinomasJ Surg Oncol105542547201210.1002/jso.2215122095667 KorsmeyerSJBcl-2 initiates a new category of oncogenes: Regulators of cell deathBlood8087988619921498330 SetoMJaegerUHockettRDGraningerWBennettSGoldmanPKorsmeyerSJAlternative promoters and exons, somatic mutation and deregulation of the Bcl-2-Ig fusion gene in lymphomaEMBO J712313119882834197 ReedJCMiyashitaTKrajewskiSTakayamaSAime-SempeCKitadaSSatoTWangHGHarigaiMHanadaMBcl-2 family proteins and the regulation of programmed cell death in leukemia and lymphomaCancer Treat Res843172199610.1007/978-1-4613-1261-1_38724625 NaccaratoAGViacavaPVignatiSFanelliGBonadioAGMontruccoliGBevilacquaGBio-morphological events in the development of the human female mammary gland from fetal age to pubertyVirchows Arch436431438200010.1007/s00428005047010881736 ReedJCMiyashitaTTakayamaSWangHGSatoTKrajewskiSAimé-SempéCBodrugSKitadaSHanadaMBCL-2 family proteins: Regulators of cell death involved in the pathogenesis of cancer and resistance to therapyJ Cell Biochem602332199610.1002/(SICI)1097-4644(19960101)60:1<23::AID-JCB5>3.0.CO;2-58825412 TerakiYShioharaTApoptosis and the skinEur J Dermatol9413425199910417450 TomaszewskaRKarczDStachuraJAn immunohistochemical study of the expression of bcl-2 and p53 oncoproteins in pancreatic intraepithelial neoplasia and pancreatic cancerInt J Pancreatol26163171199910.1385/IJGC:26:3:16310732293 WonKYKimGYKimYWSongJYLimSJClinicopathologic correlation of beclin-1 and bcl-2 expression in human breast cancerHum Pathol41107112201010.1016/j.humpath.2009.07.00619762066 BaspinarSBircanSOrhanHKapucuogluNBozkurtKKThe relation of beclin 1 and bcl-2 expressions in high grade prostatic intraepithelial neoplasia and prostate adenocarcinoma: A tissue microarray studyPathol Res Pract210412418201410.1016/j.prp.2014.02.00824690321 LiangXHKleemanLKJiangHHGordonGGoldmanJEBerryGHermanBLevineBProtection against fatal sindbis virus encephalitis by beclin, a novel Bcl-2-interacting proteinJ Virol728586859619989765397 DecuypereJPParysJBBultynckGRegulation of the autophagic bcl-2/beclin 1 interactionCells1284312201210.3390/cells103028424710477 ZhouFYangYXingDBcl-2 and Bcl-xL play important roles in the crosstalk between autophagy and apoptosisFEBS J278403413201110.1111/j.1742-4658.2010.07965.x21182587 CaoSSSuAPDuXJLiAHuWMZhangZDTianBLLuHMKeNWHeZPThe expressions and correlation of bcl-2 and beclin-1 in pancreatic cancerSichuan Da Xue Xue Bao Yi Xue Ban431561602012(In Chinese)22650021 ChunYSPawlikTMVautheyJN8th Edition of the AJCC cancer staging manual: Pancreas and hepatobiliary cancersAnn Surg Oncol2017(Epub ahead of print) SongSWangBZhangXHaoLHuXLiZSunSLong-term diabetes mellitus is associated with an increased risk of pancreatic cancer: A meta-analysisPloS one10e0134321201510.1371/journal.pone.013432126222906 AsiyanbolaBGleisnerAHermanJMChotiMAWolfgangCLSwartzMEdilBHSchulickRDCameronJLPawlikTMDetermining pattern of recurrence following pancreaticoduodenectomy and adjuvant 5-flurouracil-based chemoradiation therapy: Effect of number of metastatic lymph nodes and lymph node ratioJ Gastrointest Surg13752759200910.1007/s11605-008-0762-x19089517 KimHSLeeSHDoSILimSJParkYKKimYWClinicopathologic correlation of beclin-1 expression in pancreatic ductal adenocarcinomaPathol Res Pract207247252201110.1016/j.prp.2011.02.00721420796 McDonnellTJKorsmeyerSJProgression from lymphoid hyperplasia to high-grade malignant lymphoma in mice transgenic for the t (14; 18)Nature349254256199110.1038/349254a01987477 McDonnellTJTroncosoPBrisbaySMLogothetisCChungLWHsiehJTTuSMCampbellMLExpression of the protooncogene bcl-2 in the prostate and its association with emergence of androgen-independent prostate cancerCancer Res526940694419921458483 PietenpolJAPapadopoulosNMarkowitzSWillsonJKKinzlerKWVogelsteinBParadoxical inhibition of solid tumor cell growth by bcl2Cancer Res543714371719948033089 CallagyGMWebberMJPharoahPDCaldasCMeta-analysis confirms BCL2 is an independent prognostic marker in breast cancerBMC Cancer8153200810.1186/1471-2407-8-15318510726 CampaniDEspositoIBoggiUCecchettiDMenicagliMDe NegriFColizziLDel ChiaroMMoscaFFornaciariGBevilacquaGBcl-2 expression in pancreas development and pancreatic cancer progressionJ Pathol194444450200110.1002/path.92511523052 CasneufVFFonteynePVan DammeNDemetterPPauwelsPde HemptinneBDe VosMVan de WieleCPeetersMExpression of SGLT1, Bcl-2 and p53 in primary pancreatic cancer related to survivalCancer Invest26852859200810.1080/0735790080195636318853313 KnowltonKManciniMCreasonSMoralesCHockenberyDAndersonBOBcl-2 slows in vitro breast cancer growth despite its antiapoptotic effectJ Surg Res762226199810.1006/jsre.1998.52779695733 ZinkelSGrossAYangEBCL2 family in DNA damage and cell cycle controlCell Death Differ1313511359200610.1038/sj.cdd.440198716763616 KeHParronVIReeceJZhangJYAkiyamaSKFrenchJEBCL2 inhibits cell adhesion, spreading, and motility by enhancing actin polymerizationCell Res20458469201010.1038/cr.2010.2120142842 YangDChenMBWangLQYangLLiuCYLuPHBcl-2 expression predicts sensitivity to chemotherapy in breast cancer: A systematic review and meta-analysisJ Exp Clin Cancer Res32105201310.1186/1756-9966-32-10524370277 JiangLCHuangSYZhangDSZhangSHLiWGZhengPHChenZWExpression of beclin 1 in primary salivary adenoid cystic carcinoma and its relation to Bcl-2 and p53 and prognosisBraz J Med Biol Res47252258201410.1590/1414-431X2013323124554038 PattingreSTassaAQuXGarutiRLiangXHMizushimaNPackerMSchneiderMDLevineBBcl-2 antiapoptotic proteins inhibit beclin 1-dependent autophagyCell122927939200510.1016/j.cell.2005.07.00216179260 KlöppelGHrubanRKlimstraDSolid-pseudopapillary Tumor of PancreasBosmanFTCarneiroFHrubanRHTheiseND3273332010 PotrcSKavalarRHorvatMGadzijevEMUrgent Whipple resection for solid pseudopapillary tumor of the pancreasJ Hepatobiliary Pancreat Surg10386389200310.1007/s00534-002-0836-114598141

Immunohistochemical staining for high Beclin 1 expression in PDAC tissues at (A) magnification, ×200 and (D) ×400, in SPN tissues at (B) magnification, ×200 and (E) ×400, and in normal ductal epithelium in CP tissues at (C) magnification, ×200 and (F) ×400. Low Beclin 1 expression in PDAC tissues at (G) magnification, ×200 and (J) ×400, in SPN tissues at (H) magnification, ×200 and (K) ×400, and in normal ductal epithelium in CP tissues at (I) magnification, ×200 and (L) ×400. Pancreatic islet cells were positively stained with Beclin 1 in each sample (arrow). PDAC, pancreatic ductal adenocarcinoma; SPN, solid-pseudopapillary neoplasm; CP, chronic pancreatitis.

Immunohistochemical staining for high Bcl-2 expression in PDAC tissues at (A) magnification, ×200 and (D) ×400 in SPN tissues at (B) magnification, ×200 and (E) ×400 and in normal ductal epithelium in CP tissues at (C) magnification, ×200 and (F) ×400. Low Bcl-2 expression in PDAC tissues at (G) magnification, ×200 and (J) ×400 in SPN tissues at (H) magnification, ×200 and (K) ×400 and in normal ductal epithelium in CP tissues at (I) magnification, ×200 and (L) ×400. Lymphocytes in the pancreatic stroma were used as the positive controls (arrow). PDAC, pancreatic ductal adenocarcinoma; SPN, solid-pseudopapillary neoplasm; CP, chronic pancreatitis; Bcl-2, B-cell lymphoma-2.

(A and B) Disease-free survival and (C and D) overall survival were analyzed by Kaplan-Meier analysis according to Beclin 1 and Bcl-2 expression status. Bcl-2, B-cell lymphoma-2.

Expression of Beclin 1 and Bcl-2 in PDAC and CP tissues.

Beclin 1 Bcl-2
Tissue type Cases, n Low, n (%) High, n (%) P-value Low, n (%) High, n (%) P-value
PDAC 117 43 (36.8) 74 (63.2) 0.047a 79 (67.5) 38 (32.5) <0.001a
CP   32 18 (56.3) 14 (43.8)   8 (25.0) 24 (75.0)

PDAC vs. CP, significantly differently by the χ2 test. Bcl-2, B-cell lymphoma-2; PDAC, pancreatic ductal adenocarcinoma; CP, chronic pancreatitis.

Association of Beclin 1 and Bcl-2 expression with clinicopathological characteristics in pancreatic ductal adenocarcinoma.

Beclin 1 Bcl-2
Characteristics Cases, n Low, n (%) High, n (%) P-value Low, n (%) High, n (%) P-value
Age, years (mean=58.98) 0.518 0.646
  <58.98   59 20 (33.9) 39 (66.1) 41 (69.5) 18 (30.5)
  ≥58.98   58 23 (39.7) 35 (60.3) 38 (65.5) 20 (34.5)
Sex 0.064 0.485
  Female   47 22 (46.8) 25 (53.2) 30 (63.8) 17 (36.2)
  Male   70 21 (30.0) 49 (70.0) 49 (70.0) 21 (30.0)
Smoker 0.170 0.332
  No   75 31 (41.3) 44 (58.7) 53 (70.7) 22 (29.3)
  Yes   42 12 (28.6) 30 (71.4) 26 (61.9) 16 (38.1)
Diabetes 0.631 0.629
  No 104 39 (37.5) 65 (62.5) 71 (68.3) 33 (31.7)
  Yes   13 4 (30.8) 9 (69.2) 8 (61.5) 5 (38.5)
CA19-9, U/mlb 0.585 0.871
  ≤37   27 9 (33.3) 18 (66.7) 18 (66.7) 9 (33.3)
  >37   79 31 (39.2) 48 (60.8) 54 (68.4) 25 (31.6)
CEA, ng/mlb 0.915 0.996
  ≤5   78 29 (37.2) 49 (62.8) 53 (67.9) 25 (32.1)
  >5   25 9 (36.0) 16 (64.0)
Location 0.235 0.017a
  Head   85 34 (40.0) 51 (60.0) 52 (61.2) 33 (38.8)
  Body/Tail   32 9 (28.1) 23 (71.9) 27 (84.4) 5 (15.6)
Differentiation 0.001a 0.002a
  Well   21 12 (57.1) 9 (42.9) 11 (52.4) 10 (47.6)
  Moderate   71 29 (40.8) 42 (59.2) 44 (62.0) 27 (38.0)
  Poor   25 2 (8.0) 23 (92.0) 24 (96.0) 1 (4.0)
T stage 0.786 0.577
  T1/T2   39 15 (38.5) 24 (61.5) 25 (64.1) 14 (35.9)
  T3/T4   78 28 (35.9) 50 (64.1) 54 (69.2) 24 (30.8)
N stage 0.390 0.351
  N0   73 29 (39.7) 44 (60.3) 47 (64.4) 26 (35.6)
  N1   44 14 (31.8) 30 (68.2) 32 (72.7) 12 (27.3)
M stage 0.021a 0.049a
  M0 100 41 (41.0) 59 (59.0) 64 (64.0) 36 (36.0)
  M1   17 2 (11.8) 15 (88.2) 15 (88.2) 2 (11.8)
TNM stage 0.235 0.017a
  I/II   85 34 (40.0) 51 (60.0) 52 (61.2) 33 (38.8)
  III/IV   32 9 (28.1) 23 (71.9) 27 (84.4) 5 (15.6)

Significantly differently by the χ2 test

preoperative serum CA19-9 was not measured in 11 patients and CEA in 14 patients. Bcl-2, B-cell lymphoma-2; CA19-9, carbohydrate antigen 19-9; CEA, carcinoembryonic antigen; TNM, Tumor-Node-Metastasis.

Association of Beclin 1 and Bcl-2 expression with clinicopathological characteristics in solid-pseudopapillary neoplasms.

Beclin 1 Bcl-2
Characteristics Cases, n Low, n (%) High, n (%) P-value Low, n (%) High, n (%) P-value
Age, years (mean=32.77) 0.207 0.399
  <32.77 28 13 (46.4) 15 (53.6) 22 (78.6) 6 (21.4)
  ≥32.77 15 4 (26.7) 11 (73.3) 10 (66.7) 5 (33.3)
Sex 0.633 0.096
  Male   5 1 (20.0) 4 (80.0) 2 (40.0) 3 (60.0)
  Female 38 16 (42.1) 22 (57.9) 30 (78.9) 8 (21.1)
Location 0.018a 0.116
  Head 16 10 (62.5) 6 (37.5) 14 (87.5) 2 (12.5)
  Body/Tail 27 7 (25.9) 20 (74.1) 18 (66.7) 9 (33.3)
Size, cm 0.859 0.802
  <6.75 26 10 (38.5) 16 (61.5) 19 (73.1) 7 (26.9)
  ≥6.75 17 7 (41.2) 10 (58.8) 13 (76.5) 4 (23.5)
Tumor feature 0.834 0.422
  Solid 16 6 (37.5) 10 (62.5) 13 (81.3) 3 (18.8)
  Cystic and solid 27 11 (40.7) 16 (59.3) 19 (70.4) 8 (29.6)
Nuclear pleomorphism, HPFs 0.031a 0.273
  <10 36 17 (47.2) 19 (52.8) 28 (77.8) 8 (22.2)
  ≥10   7 0 (0.0) 7 (100.0) 4 (57.1) 3 (42.9)
Peripancreatic invasion 0.446 0.967
  Yes   8 2 (25.0) 6 (75.0) 6 (75.0) 2 (25.0)
  No 35 15 (42.9) 20 (57.1) 26 (74.3) 9 (25.7)
Cytokeratin 0.749 1.000
  + 21 8 (38.1) 13 (61.9) 15 (71.4) 6 (28.6)
  − 16 7 9 11 (68.8) 5 (31.3)
Synaptophysin 1.000 1.000
  + 18 7 (38.9) 11 (61.1) 12 (66.7) 6 (33.3)
  − 13 5 (38.5) 8 (61.5) 8 (61.5) 5 (38.5)

Significantly different by the χ2 test. Bcl-2, B-cell lymphoma-2; HPF, high-power field.

Association between Beclin 1 and Bcl-2 expression.

Bcl-2 expression
Beclin 1 expression Low expression, n (%) High expression, n (%) r P-value
PDAC −0.342 <0.001a
  Low expression 20 (17.1) 23 (19.7)
  High expression 59 (50.4) 15 (12.8)
SPN 0.365 0.016a
  Low expression 16 (37.2) 1 (2.3)
  High expression 16 (37.2) 10 (23.3)
CP 0.073 0.692
  Low expression 5 (15.6) 13 (40.6)
  High expression 3 (9.4) 11 (34.4)

Significantly different by the χ2 test. Bcl-2, B-cell lymphoma-2; PDAC, pancreatic ductal adenocarcinoma; SPN, solid-pseudopapillary neoplasm; CP, chronic pancreatitis.

Univariate analysis and multivariate analysis of DFS and OS in pancreatic ductal adenocarcinoma.

DFS OS
Univariate analysis Multivariate analysis Univariate analysis Multivariate analysis
Variable Cases, n Median survival, months P-value HR (95% CI) P-value Cases, n Median survival, months P-value HR (95% CI) P-value
Age, years 0.856 0.926 (0.558–1.534) 0.764 0.567 1.071 (0.679–1.687) 0.769
  <58.98 46 6.0   59 9.9
  ≥58.98 51 6.5   58 11.1
Sex 0.292 0.913 (0.509–1.637) 0.760 0.639 1.469 (0.871–2.478) 0.149
  Female 37 7.5   47 12.1
  Male 60 5.9   70 10.0
Smoker 0.786 0.843 (0.482–1.475) 0.550 0.451 0.546 (0.326–0.914) 0.021a
  No 61 6.9   75 11.6
  Yes 36 5.9   42 8.7
Diabetes 0.421 1.722 (0.735–4.034) 0.211 0.499 1.336 (0.638–2.798) 0.443
  No 87 6 104 10.4
  Yes 10   6   13 10.0
Location 0.001a 0.576 (0.319–1.040) 0.067 0.004a 0.826 (0.495–1.379) 0.465
  Head 74 7.5   85 12.1
  Body/Tail 23 4.5   32 7.2
Differentiation <0.001a 0.399 (0.190–0.839) 0.015a <0.001a 0.394 (0.201–0.774) 0.007a
  Well/Moderate 86 7.0   92 12.9
  Poor 11 4.5   25 5.9
T stage 0.003a 0.500 (0.287–0.871) 0.014a <0.001a 0.307 (0.175–0.539) <0.001a
  T1/T2 38 12.0   39 20.8
  T3/T4 59 6.0   78 8.7
N stage 0.388 1.084 (0.628–1.870) 0.773 0.022a 1.154 (0.679–1.960) 0.596
  N0 71 6.9   73 12.8
  N1 26 5.0   44 7.9
M stage – – – <0.001a 0.380 (0.156–0.923) 0.033a
  M0 97 6.0 100 12.1
  M1   0 –   17 4.3
TNM stage 0.140 0.777 (0.349–1.731) 0.537 <0.001a 1.377 (0.647–2.930) 0.406
  I/II 85 6.5   85 12.9
  III/IV 12 3.6   32 6.2
Chemotherapy 0.871 0.949 (0.515–1.748) 0.866 0.235 2.415 (1.283–4.544) 0.006a
  No 80 6.0   98 9.3
  Yes 17 7.0   19 13.1
Beclin 1 0.238 0.870 (0.532–1.423) 0.580 0.038a 0.927 (0.557–1.543) 0.771
  Low expression 41 10.0   43 14.1
  High expression 56 6.0   74 9.1
Bcl-2 0.030a 1.400 (0.810–2.421) 0.228 0.017a 1.850 (1.053–3.249) 0.032a
  Low expression 61 6.0   79 9.1
  High expression 36 10.8   38 16.5

Significantly different by the Log-rank test. DFS, disease-free survival; OS, overall survival; TNM, Tumor-Node-Metastasis; Bcl-2, B-cell lymphoma 2.