Tumor Biol. (2013) 34:713–722 DOI 10.1007/s13277-012-0599-5

RESEARCH ARTICLE

Beclin 1 activation enhances chemosensitivity and predicts a favorable outcome for primary duodenal adenocarcinoma Xiang-Yuan Wu & Jie Chen & Qing-Hua Cao & Min Dong & Qu Lin & Xin-Juan Fan & Qing Xia & Zhan-Hong Chen & Quentin Liu & Xiang-Bo Wan

Received: 19 September 2012 / Accepted: 16 November 2012 / Published online: 7 December 2012 # International Society of Oncology and BioMarkers (ISOBM) 2012

Abstract We and others had proven that hypoxia-induced autophagy was essential to regulate cancer cell destiny under anticancer therapeutic stress. Here, we addressed the clinicopathologic effect of HIF-1α and autophagic Beclin 1 in primary duodenal adenocarcinoma (PDA). HIF-1α and Beclin 1 expression level were semi-quantitatively evaluated using tissue microarrays and immunohistochemistry (IHC) staining in 141 PDA patients. Among these patients, 77 acted as training set to select HIF-1α and Beclin 1 IHC cutoff score for patient outcome, and 64 cases were used as testing set to evaluate their prognostic effect. We found that Beclin 1 was cytoplasmic overexpressed, defined by training set fixed cutoff point, in 49.6 % PDA tissue, compared to 46.8 % patients had HIF-1α high expression. In testing

set, Beclin 1 overexpression predicted a superior 5-year overall survival (OS) in both univariate (P00.010) and multivariate (P00.017) analyses. However, we did not detect any correlation between HIF-1α level and patient prognosis (P00.989). Significantly, among Beclin 1 overexpressed patients, radical surgery plus adjuvant chemotherapy had a 23.1-month OS improvement than given radical surgery alone (59.2 vs 36.1 months; P00.01). For Beclin 1 lowly expressed patients, radical surgery plus adjuvant chemotherapy and given radical surgery alone had the similar OS (P00.283). Contrary to previous studies, we failed to detect any correlation between Beclin 1 and HIF-1α levels in PDA (correlation coefficient 0.217, P00.099). In conclusions, our results confirmed that Beclin 1 was a favorable prognostic biomarker for PDA, and might be used to identify particular patients for more selective therapy.

X.-Y. Wu, J. Chen, and Q.-H. Cao contributed equally to this paper. X.-Y. Wu (*) : J. Chen : M. Dong : Q. Lin : Z.-H. Chen : X.-B. Wan (*) Department of Medical Oncology, the Third Affiliated Hospital, Sun Yat-sen University, 600 Tianhe Road, Guangzhou 510630, China e-mail: [email protected] e-mail: [email protected] Q.-H. Cao Department of Pathology, the First Affiliated Hospital, Sun Yat-sen University, 58 Zhongshan II Road, Guangzhou 510080, China X.-J. Fan Department of Pathology, Gastrointestinal institute, the Sixth Affiliated Hospital, Sun Yat-sen University, 26 Yuancun Erheng Road, Guangzhou 510655, China Q. Xia : Q. Liu State Key Laboratory of Oncology in South China, Sun Yat-sen University Cancer Center, 651 Dongfeng Road East, Guangzhou 510060, China

Keyword Autophagy . Beclin 1 . HIF-1α . Prognostic biomarker . PDA Abbreviations HIF-1α Hypoxia-inducible factor-1α HER2 Human epidermal growth factor receptor 2 EGFR Epidermal growth factor receptor HMGB1 High mobility group box-1 protein Bcl-2 B-cell lymphoma 2 Bcl-xl B-cell lymphoma extra-large UVRAG Ultraviolet radiation resistance-associated gene CRC Colorectal cancer ESCC Esophageal squamous cell carcinoma BNIP3 BCL2/adenovirus E1B 19 kd-interacting protein 3 BNIP3L Bcl-2/adenovirus E1B 19 kDa proteininteracting protein 3-like LC 3 Light chain 3 UPR Unfolded protein response

714

Introduction Though accounts for 63.2 % in small bowel adenocarcinoma, primary duodenal adenocarcinoma (PDA) is still a rare (0.35 %) cancer in the gastrointestinal tract [1]. Currently, radical surgery and adjuvant chemotherapy are the main therapeutic regimens for PDA. Although radical resection produces favorable outcome for early-stage patients, local recurrence and distant metastasis remains the major cause of cancer mortality in advanced PDA patients. However, the low incidence rate had precluded the potential to develop molecular targeting as well as prognostic biomarker for PDA. In the past decades, biomarker-guided target therapy and prognosis prediction had greatly improved the personalized medicine for a variety of cancers. HER2 overexpression, for example, was significantly correlated with lymph nodes status as well as tumor stage, and predicted an inferior overall survival for breast cancer and gastric cancers [2, 3]. HER2 fully humanized monoclonal antibody trastuzumab plus chemotherapy obtained a 2.7-month overall survival improvement than given chemotherapy alone for advanced gastric or gastroesophageal junction cancer [4]. For colorectal cancer, EGFR monoclonal antibody Cetuximab plus FOLFOX4 chemotherapy prolonged 2.6 months progression-free survival for metastatic patients [5]. However, due to the rare morbidity, the novel molecular markers, that not only can predict the prognosis individually but also direct the therapeutic regimen selection, had not yet been identified for PDA. Autophagy, an ubiquitous phenomenon in eukaryotic cells, transports materials and organelles from the cytoplasm to lysosomes for degradation by double or multi-membrane compartments [6–9]. As a stress responser, autophagy emerges to protect cells from fatal attacks at inhospitable microenvironment, such as ischemia, hypoxia, nutrient deprivation as well as other forms of stress reactions [8–13]. Beclin 1, a novel yeast autophagy gene apg6/vps30-alike Bcl-2interacting protein, might be functioned in a hypoxia (HIF1α)-mediated autophagy (Beclin 1) pathway to regulate tumorigenesis and disease progression [14, 15]. In head and neck cancer, Beclin 1 was closely correlated with lymph node metastasis as well as clinical stage [16], and was synergized by HIF-1α in promoting tumor distant metastasis [17]. Besides its essential role in regulating tumorigenesis, Beclin 1 was also an independent prognostic biomarker and a potential therapeutic targeting for a variety of tumors [12, 18–22]. For colorectal adenocarcinoma, allelic loss of Beclin 1 expression was correlated with poor prognosis, whereas overexpression of Beclin 1 linked with tumor HIF-1α level and aggressive clinical behavior [23]. In acute lymphoblastic leukemia, mTOR inhibitor RAD001 increased Beclin 1 expression and converted LC3I to autophagic LC3II, leading to massive reduction of leukemic mass and a strong survival advantage [24]. These results suggested that autophagic Beclin 1 might

Tumor Biol. (2013) 34:713–722

own the dual function as prognostic biomarker as well as therapeutic target for tumor individualized treatment. In this study, we examined HIF-1α and Beclin 1 expression using immunohistochemistry (IHC) staining, and further addressed their prognostic effect in 141 PDA patients. We found that overexpression of Beclin 1, defined by training set receiver operating characteristic curve (ROC) analysis generated cutoff point, predicted a superior OS in both testing set and overall patients. Further multivariate analysis proved that Beclin 1 was indeed an independent prognostic biomarker. Importantly, in Beclin 1 overexpressed patients, the subgroup received adjuvant chemotherapy had an evident OS improvement than that of given radical surgery alone. However, the correlation between Beclin 1 and HIF-1α expression level, and the prognostic effect of HIF1α were not detected for PDA.

Materials and methods Cohorts A total of 150 cases of nonmetastatic and histologically proved PDA in the Third Affiliated Hospital, the First Affiliated Hospital, and Cancer Center of Sun Yat-sen University (Guangzhou, China) from March 1999 to January 2008 were recruited in the present study. Due to nine cases had insufficient or detachment tissues in tissue microarray construction and immunohistochemistry staining process, 141 cases of PDA were brought into this study. All patients received radical surgery, and part of cases was given postoperative adjuvant chemotherapy. The postoperative adjuvant chemotherapy regimens consist of FOLFOX and FOLRIRI. Patients were included with the following inclusion criteria: pathologically confirmed as PDA; without ontological surgery, chemotherapy, and radiotherapy history; and all patients had the completed follow-up information and paraffin-embedded specimens. Moreover, patient would be excluded for any of the following reason: previously received any anticancer therapy; prior malignancy; pregnancy; the cancer of the periampullary region, including pancreatic cancer, bile duct carcinoma, and the lesion of unknown histological origin. The patient stage was redefined according to 2010 AJCC staging system for small intestinal cancer [25]. This study was approved by the Human Ethics Committee of the Third Affiliated Hospital, Sun Yat-sen University. Tissue microarray construction We firstly re-reviewed hematoxylin and eosin-stained slides, and selected the tumor zone in the paraffin-embedded specimens for tissue microarray (TMA) construction. Tissue

Tumor Biol. (2013) 34:713–722

microarrays were constructed in accordance with a previously described method [26]. For each case, two cores taken from the selected tumor area and additional one core from normal adjacent mucosa were used to construct the TMA. Briefly, a hollow needle was utilized to punch and remove bipartite cylinders tissue cores (1.0 mm in diameter) from selected donor tissues regions. Further, the punched tissue cores were inserted into a recipient paraffin block with a precisely spaced, array pattern, using an automatic tissue arraying instrument (Beecher Instruments, Silver Spring, Maryland, USA). Semi-quantitative assessment of immunohistochemical staining The TMA blocks were cut into 4-μm sections, and immunohistochemistry staining was performed as we previously described [27]. TMA slides were incubated at 4°C overnight with primary antibodies, including rabbit anti-Beclin 1 antibody (Santa Cruz, SC-11427) and mouse anti-HIF-1α antibody (Millipore, MAB5382) that were diluted at 1:200 and 1:100, respectively. In the meantime, negative controls were also utilized by replacing the primary antibodies with nonimmune serum immunoglobulin at 1:200 or 1:100 dilutions. The brown granules in cytoplasm or nuclei were regarded as positive staining [28]. Beclin 1 and HIF-1α expression level was evaluated by integrating the percentage of positive tumor cells and the intensity of positive staining. The intensity of staining was scored as follows: negative (score 0), bordering (score 1), weak (score 2), moderate (score 3), and strong (score 4). We scored the staining extent according to the percentage of positive stained cells in the field: negative (score 0), 0–25 % (score 1), 26–50 % (score 2), 51–75 % (score 3), and 76–100 % (score 4). The sum of the intensity and extent score was considered as the overall IHC score. Immunohistochemical staining level was assessed and scored by two independent pathologists (Fan XJ and Cao QH), who were blind to the clinicopathological and followup information. Western blot analysis Liquid nitrogen-conserved PDA tissues were ground and lysed with the RIPA buffer on ice. The protein concentration was determined by the Bradford method with bovine serum albumin as the control. Equal amounts of tissue lysates (50 μg) were run by SDS-PAGE, and electro-transferred on a polyvinylidene difluoride membrane. The membrane was then blocked and incubated with mouse anti-β-actin antibody (1:1,000, Santa Cruz, SC-81178) and rabbit anti-Beclin 1 antibody (1:1,000, Santa Cruz, SC-11427).

715

Selection of cutoff scores for Beclin 1 and HIF-1α “positive” expression The ROC curve analysis was applied to select the rational Beclin 1 and HIF-1α IHC cutoff score [29]. Briefly, the sensitivity and specificity for the patient outcome being studied at each score was plotted to generate a ROC curve. The IHC score localized closest to the point at both maximum sensitivity and specificity, the point (0.0, 1.0) on the curve, was selected as the cutoff score that leading to the greatest potential to determine having or not having the outcome. To facilitate ROC curve analysis, the survival Table 1 Beclin 1 expression status in relation to patient characteristics Variable

Age ≥56 years <56 years Gender Male Female Tumor stage T1+T2

Training set (n077)

Testing set (n064)

High Low P staining staining value

High Low P staining staining value

25 16

15 21

0.093 13 16

20 15

0.334

21 20

25 11

0.106 20 9

17 18

0.103

10

6

0.411

5

4

0.513

24

31

0.215a 29 0

32 3

0.245a

0.739 24 5

23 12

0.128

0.739 8 21

11 24

0.743

0.560 22 7

27 8

0.906

0.321a 11 18

6 29

0.062

0.311

T3+T4 31 30 Node stage N0+N1 41 34 N2+N3 0 2 Tumor–node-metastasis stage I+II 31 26 III+IV 10 10 WHO tumor differentiation Poor 10 10 Moderate+high 31 26 CEA level Normal 32 26 High 9 10 Bilirubin level Normal 7 3 High 34 33 Tumor size >2.5 cm 14 15 ≤2.5 cm Oxygen supply Hypoxia Normoxia

0.503 12

19

27

21

17

16

21 20

15 21

0.409 15 20

15 14

0.671

a

Denotes Fisher’s exact test. Patients with elevated HIF-1α expression were classified as hypoxia and those with low expression were regarded as normoxia WHO World health Organization, CEA carcinoembryonic antigen

716

Tumor Biol. (2013) 34:713–722

features were dichotomized: death vs others (censored, alive, or death from other causes). Clinical outcome assessment All patients were followed-up until the date of death or when censored at the latest date (November 6th 2011). Overall survival was defined as the time from PDA diagnosis to death or when censored at the latest date if patients were still alive.

difference between patient subsets was estimated by the log-rank test. The multivariate Cox proportional hazards regression model was used to assess the potential independent prognostic factors and 95 % confidence intervals (CI) of hazard ratios (HR). A two-tailed P<0.05 was considered statistically significant.

Statistical analysis All statistical analyses were conducted using SPSS software package v. 17.0. Clinicopathological variables, such as age, gender, tumor stage, node stage, TNM stage, pathohistological type, CEA level, and bilirubin level were included in this study. The relationships between Beclin 1 expression and other variables were evaluated by the chi-square test. Kaplan–Meier analysis was performed to determine the relationships between the Beclin 1, HIF-1α expression, and OS. In univariate analysis, the OS probabilities Table 2 Chemotherapy status in relation to patient characteristics in overall patients Variable

Adjuvant chemotherapy Without (n073)

Gender Male 39 Female 34 Age ≤56 years 34 >56 years 39 Tumor–node-metastasis stage I+II 57 III+IV 16 WHO differentiation Poor 25 Moderate+high 48 Beclin-1 Low expression 37 High expression 36 Oxygen supply Hypoxia 32 Normoxia 41 Subgroups Testing set Validation set

43 30

P value

With (n068)

44 24

0.174

34 34

0.684

47 21

0.227

14 54

0.070

34 34

0.935

39 29

0.109

34 34

0.292

Patients with high HIF-1α expression were defined as hypoxia and those with low expression were regarded as normoxia WHO World health Organization

Fig. 1 Beclin 1 and HIF-1α expression in human PDA and normal adjacent epithelium. a Beclin 1 was low expressed in tumor zone and b high expressed in normal duodenal epithelium (original magnification, ×50). The right panel displayed representative Beclin 1 expression in epithelial area and tumor zone with magnified view (original magnification, ×400). c HIF-1α was overexpressed in tumor tissue, compared to d low expression in normal adjacent epithelium (original magnification, ×50). The right panels displayed representative HIF-1α expression in tumor zone and matched epithelial area with magnified view (original magnification, ×400). e Western blot analysis of Beclin 1 expression in representative PDA tissues (T) and paired normal adjacent epithelia (N). Equal loading of protein was determined by β-actin

Tumor Biol. (2013) 34:713–722

Results Patient characteristics The patient characteristics of these 141 PDA cases, consist of the training set (n077) and another independent cohort of testing set (n064), were shown in Tables 1 and 2. For the overall patients, the 1-, 3-, and 5-year cumulative survival rates were 71.0, 43.0, and 37.0 %, respectively. The median follow-up time was 38.8 months (range, 0.63–113.4 months) for training set, and was 38.7 months for the testing set (range, 1.83–132.5 months). Moreover, we observed a similar clinicopathological features, including gender, age, T classification, N classification, TNM stage, WHO pathological type, preoperative serum CEA level, preoperative serum bilirubin level, tumor size, HIF-1α, and Beclin 1 expression level, between training set and testing set (Table 1). A total of 68 patients, including 34 patients in training set and 34 cases in testing set, received radical PDA surgery plus adjuvant chemotherapy, and 73 PDA patients obtained the radical PDA surgery alone (Table 2). Immunohistochemistry staining and ROC analysis Given the pivotal role of hypoxia in regulating autophagic cancer cell survival [30], we next tested Beclin 1 and HIF1α expression level in PDA. As shown in Fig. 1, Beclin 1 was cytoplasmic lowly expressed in PDA tumor nest zone (Fig. 1a), whereas was overexpressed in the normal adjacent Fig. 2 Beclin 1 and HIF-1α levels were compared using t test in the subgroups with dichotomized 5-year survival features and T stage. For Beclin 1, the features were dichotomized: a survival (alive vs dead), b T stage (T2–3 vs T4). For HIF-1α, the features were dichotomized: c survival (alive vs dead), d T stage (T2–3 vs T4)

717

duodenal glandular epithelium (Fig. 1b). Conversely, we found that HIF-1α was cytoplasmic or nuclear overexpressed in the PDA tumor area (Fig. 1c), compared to the negligible level in the normal adjacent epithelium (Fig. 1d). Western blot analysis showed the similar finding regarding to Beclin 1 expression in PDA and normal adjacent epithelial tissue (Fig. 1e). When stratified the overall patients according to their survival condition after 5-year follow-up, we detected a significant Beclin 1 expression difference between alive (7.4±4.5) and dead (6.1±4.3) patients (P00.025, Fig. 2a). However, we failed to detect a close correlation between Beclin 1 level and T stage (T4 vs T2–3, 7.0±4.5 vs 6.2±4.3, P00.261, Fig. 2b). For HIF-1α, the IHC scores were comparable between alive and dead patients (4.4±4.3 vs 4.3± 4.6, P00.691, Fig. 2c). Moreover, HIF-1α expression levels were not correlated with T stage (T4 vs T2–3, 4.4±4.5 vs 4.2±4.5, P00.352, Fig. 2d). Contrary to previous studies [16, 17], we did not detect any correlation between Beclin 1 and HIF-1α expression levels in PDA (correlation coefficient 0.217, P00.099). For Beclin 1, the ROC analysis generated IHC cutoff score for overall survival was 5.5 in the training set. In the testing set, we defined Beclin 1 IHC score ≤5.5 as low expression (71/141, 50.4 %), and Beclin 1 IHC score >5.5 were classified as high expression. For HIF-1α, we found that cutoff point of 4.5 was rational to distinguish the training set into high (66/141, 46.8 %) and low expressed subgroups.

718

Beclin 1, HIF-1α expression, and OS As shown in Fig. 3, Beclin 1 overexpression strongly predicted a superior 5-year OS rate both in testing set (54.2 vs 20.9 %, P0 0.010, Fig. 3a), training set (47.8 vs 26.9 %, P00.013, Fig. 3b), Fig. 3 Kaplan–Meier estimated of overall survival based on Beclin 1 expression in testing set, training set, and overall patients. Beclin 1 overexpression predicted a favorable overall survival in both testing set (a), training set (b), and overall patients (c). d For the Beclin 1 overexpressed patients, postoperative adjuvant chemotherapy prolonged overall survival than given radical surgery alone. e For the Beclin 1 lowly expressed patients, radical surgery plus adjuvant chemotherapy or given radical surgery alone had the similar overall survival

Tumor Biol. (2013) 34:713–722

and overall patients (38.0 vs 24.0 %, P00.001, Fig. 3c). In Beclin 1 overexpressed patients, radical surgery plus adjuvant chemotherapy had a 23.1 months OS improvement than given radical surgery alone (59.2 vs 36.1 months; P00.010, Fig. 3d). For Beclin 1 lowly expressed patients, radical surgery plus

Tumor Biol. (2013) 34:713–722

adjuvant chemotherapy and given radical surgery alone had the similar OS (with vs without chemotherapy, 38.5 vs 31.4 months, P00.283; Fig. 3e). We did not detect any prognostic effect of HIF-1α in testing set (high vs low expression, 34.5 vs 40.9 %, P0 Fig 4 Kaplan–Meier estimated of overall survival according to HIF-1α expression in testing set, training set, and overall patients. HIF-1α low and high expressed patients had the similar overall survival in both testing set (a), training patients (b), and overall patients (c). d For the HIF-1α overexpressed patients, postoperative adjuvant chemotherapy was tending to prolong the overall survival than given radical surgery alone. e For the HIF-1α lowly expressed patients, radical surgery plus adjuvant chemotherapy or given radical surgery alone had the similar overall survival

719

0.989, Fig.4a), training set (42.3 vs 34.5 %, P 00.206, Fig. 4b), and overall patients (32.2 vs 36.4 %, P00.380, Fig. 4c). In HIF-1α overexpressed patients, cases with postoperative adjuvant chemotherapy were prone to have an OS advantage than that of given surgery alone (50.1 vs 32.5 %,

720

Tumor Biol. (2013) 34:713–722

P00.058, Fig. 4d). For HIF-1α lowly expressed patients, cases with and without postoperative adjuvant chemotherapy had the similar 5-year OS probability (45.6 vs 23.6 %, P00.094, Fig. 4e). Multivariate analysis As shown in Table 3, Beclin 1 (testing set, P00.017; training set, P00.042) and TNM stage (testing set, P00.008; training set, P 00.005) were the independent prognostic factors for OS both in testing set and overall patients. Moreover, we also noticed that tumor size (P00.038; HR, 0.443) and histological type (P00.056; HR, 0.451) in the testing set, and tumor stage (P00.005; HR, 2.770) in training were the independent predictive factor for OS. Importantly, we found that Beclin 1 and TNM stage had the comparable hazard ratio (Beclin 1, testing set, 1/0.4240 2.358; training set, 1/0.51301.949; TNM stage, testing set, 2.971; overall patients, 2.770) in refining the risk to death for PDA patients.

Discussion Recently, molecular biomarkers had been proven to be critical in guiding targeting therapy and predicting prognosis in a variety of tumors. In stage IV colon adenocarcinoma, EGFR immunohistochemical reactivity of 2+ and 3+ were found to have the strong correlation with decreased overall survival [31]. Moreover, EGFR monoclonal antibody Cetuximab showed a gratified survival improvement when added with FOLFOX [32], and significantly reversed the FOLFIRI chemoresistance for advanced colorectal cancer [33]. Those results suggested that EGFR might function not

only as a prognostic biomarker but also as a molecular targeting for colorectal cancer. Due to the rare incidence, the EGFR-alike molecular biomarkers, that might guide prognosis prediction and direct target therapy, had not yet been characterized in PDA. In this study, we addressed the prognostic effect of autophagic Beclin 1 in PDA. To obtain a solid result, we assigned the PDA patients from one institute as a training set (77 cases), and further validated in an independent cohort (testing set) that came from another two institutes (64 cases). As shown in Table 1, these two cohorts had the comparable patient characteristics, and thus were ideal for further survival analysis. Our finding proved that Beclin 1 was indeed an independent prognostic biomarker for PDA (Table 3). Importantly, the stratified survival analysis demonstrated that, for Beclin 1 overexpressed patients, postoperative adjuvant chemotherapy had a 23.1-month OS improvement than given radical surgery alone (59.2 vs 36.1 months; P00.010, Fig. 3d). However, this survival difference was not observed in the Beclin 1 lowly expressed patients. This interesting finding raised a potential clinical application in determining the therapeutic regimen for PDA: (1) for Beclin 1 overexpressed patients, aggressive postoperative adjuvant chemotherapy might be of survival benefit; (2). for Beclin 1 lowly expressed patients, postoperative adjuvant chemotherapy might mean overtreatment to these patients, and therefore should be avoided. Autophagic digestion occurs as a survival mechanism to maintain cellular homeostasis or to respond to environmental and therapeutic stresses [34, 35]. As the central player for autophagy, Beclin 1 overexpression augmented cis-diamminedichloroplatinum-induced apoptosis in MKN28 human gastric cancer cells [36]. Similarly, our present study confirmed that Beclin 1 high expression benefited PDA patients

Table 3 Multivariate Cox proportional-hazards analysis Variable

Testing set

Training set

HR

95 % CI

Tumor stage (T3+T4 vs T1+T2) Node stage (N0+N1 vs N2+N3) TNM stage (III+IV vs I+II) CEA level (high vs normal) Bilirubin level (normal vs high) Tumor size (≤2.5 cm vs >2.5 cm)

1.440 0.688 2.971 1.532 0.870 0.443

0.805–2.575 0.253–1.871 1.326–6.653 0.673–3.488 0.387–1.955 0.206–0.954

Differentiation (moderate+high vs poor) Beclin 1 level (high vs. low) Hypoxia (vs. normoxia)

0.451 0.424 0.755

0.199–1.021 0.210–0.857 0.373–1.530

P value

P value

HR

95 % CI

0.219 0.464 0.008 0.310 0.735 0.038

3.840 5.292 2.770 1.164 1.535 0.917

1.320–11.170 0.814–34.428 1.370–5.599 0.535–2.533 0.555–4.247 0.483–1.739

0.014 0.081 0.005 0.701 0.409 0.790

0.056 0.017 0.436

0.529 0.513 0.621

0.261–1.070 0.269–0.978 0.327–1.178

0.077 0.042 0.145

The Wald test was used to calculate P values. Patients with high HIF-1α expression were defined as hypoxia and those with low expression were regarded as normoxia HR hazard ratio, CI confidence interval

Tumor Biol. (2013) 34:713–722

from postoperative adjuvant chemotherapy (Fig. 3). For the underlying mechanism, we supposed that Beclin 1 activation might trigger the autophagy process, and facilitate to inducing the autophagic death for the tumor cell under chemotherapeutic stress. Conversely, Beclin 1 inactivation would compromise autophagy-induced apoptosis, and survive the tumor cell from chemotherapy. Indeed, Beclin 1 high expression had been reported to be a favorable and independent prognostic factor in a variety of tumors [17, 37–41]. In hepatocellular carcinoma, deficiency of Beclin 1 synergized with Bcl-xL might facilitate tumor malignant differentiation, and predicted a poor OS as well as PFS [41]. Here, our finding also confirmed that Beclin 1 was an independent prognostic biomarker for PDA (Fig. 3 and Table 3). In addition, Cox regression analysis demonstrated that Beclin 1 and TNM stage had the comparable hazard ratio in refining the risk to death for PDA (HR, 2.358 vs 2.971, Table 3). Therefore, supplemented Beclin 1 with TNM staging might lead to more accurate prognosis prediction, and potentially segregate particular patients for more selective therapy in PDA. Hypoxia acts as an effective driving force for angiogenesis and invasion, and represents a compensable mechanism against tissue ischemia in tumor progression [42]. In glioma and breast cancer cells, HIF-1α-induced autophagic cell death was mediated by upregulating BNIP3 [43]. In esophageal squamous cell carcinoma (ESCC), Beclin-1-positive and HIF-1α low expression was correlated with a favorable prognosis [16]. Our previous study also confirmed that co-expression of HIF-1α and Beclin 1 predicted an even poorer prognosis for nasopharyngeal carcinoma [17]. However, the present study was failed to detect any prognostic value for HIF-1α. Moreover, the correlation analysis showed that HIF-1α and Beclin 1 were not correlated with each other. Survival analysis also confirmed that HIF-1α did not have any synergistic effect to Beclin 1 in improving risk definition for OS (data not shown). Indeed, in the extreme hypoxia conditions, Beclin 1/BNIP might be regulated by AMPK/mTOR and UPR signaling in driving autophagic cancer cell survival [44]. Those results confirmed that autophagic Beclin-1-induced clinicopathological effect was not driven by hypoxia in PDA.

Conclusions Our study confirmed that Beclin 1 was a favorable prognostic biomarker for PDA. Beclin1 overexpression might benefit particular patient from postoperative adjuvant chemotherapy, suggesting that Beclin 1 might be a promising targeting to identify subset patients for more selective therapy in PDA. Acknowledgments This work was supported by National Natural Science Foundation of China (No.81000934 to X.B. Wan), the Fundamental Research Funds for the Central Universities (to X.B. Wan),

721 Science and Technology Foundation of Guangdong Province (2009B060700024 to X.Y. Wu and 2011B03180076 to Q. Lin).

Reference 1. Zenklusen HR, Landmann J, Feess A, Durig M, Kasper M, Oberholzer M. Primary duodenal carcinoma arising in a nonvaterian tubulo-villous adenoma. A case report with immunocytochemical analysis and review of the literature. Virchows Arch A Pathol Anat Histopathol. 1989;414(6):529–33. 2. Gravalos C, Jimeno A. HER2 in gastric cancer: a new prognostic factor and a novel therapeutic target. Ann Oncol. 2008;19 (9):1523–9. doi:10.1093/annonc/mdn169. 3. Begnami MD, Fukuda E, Fregnani JH, Nonogaki S, Montagnini AL, da Costa Jr WL, et al. Prognostic implications of altered human epidermal growth factor receptors (HERs) in gastric carcinomas: HER2 and HER3 are predictors of poor outcome. J Clin Oncol. 2011;29(22):3030–6. doi:10.1200/JCO.2010.33.6313. 4. Bang YJ, Van Cutsem E, Feyereislova A, Chung HC, Shen L, Sawaki A, et al. Trastuzumab in combination with chemotherapy versus chemotherapy alone for treatment of HER2-positive advanced gastric or gastro-oesophageal junction cancer (ToGA): a phase 3, open-label, randomised controlled trial. Lancet. 2010;376 (9742):687–97. doi:10.1016/S0140-6736(10)61121-X. 5. Giantonio BJ, Catalano PJ, Meropol NJ, O’Dwyer PJ, Mitchell EP, Alberts SR, et al. Bevacizumab in combination with oxaliplatin, fluorouracil, and leucovorin (FOLFOX4) for previously treated metastatic colorectal cancer: results from the Eastern Cooperative Oncology Group Study E3200. J Clin Oncol. 2007;25(12):1539– 44. doi:10.1200/JCO.2006.09.6305. 6. Lamparska-Przybysz M, Gajkowska B, Motyl T. BID-deficient breast cancer MCF-7 cells as a model for the study of autophagy in cancer therapy. Autophagy. 2006;2(1):47–8. doi:2149. 7. Yang Z, Klionsky DJ. Mammalian autophagy: core molecular machinery and signaling regulation. Curr Opin Cell Biol. 2010;22(2):124–31. doi:10.1016/j.ceb.2009.11.014. 8. Shintani T, Klionsky DJ. Autophagy in health and disease: a double-edged sword. Science. 2004;306(5698):990–5. doi:10.1126/science.1099993. 9. Levine B, Klionsky DJ. Development by self-digestion: molecular mechanisms and biological functions of autophagy. Dev Cell. 2004;6(4):463–77. 10. Levine B, Kroemer G. Autophagy in the pathogenesis of disease. Cell. 2008;132(1):27–42. doi:10.1016/j.cell.2007.12.018. 11. Mizushima N, Levine B, Cuervo AM, Klionsky DJ. Autophagy fights disease through cellular self-digestion. Nature. 2008;451 (7182):1069–75. doi:10.1038/nature06639. 12. Li JL, Han SL, Fan X. Modulating autophagy: a strategy for cancer therapy. Chin J Cancer. 2011;30(10):655–68. doi:10.5732/ cjc.011.10185. 13. Zhuang WZ, Long LM, Ji WJ, Liang ZQ. Rapamycin induces differentiation of glioma stem/progenitor cells by activating autophagy. Chin J Cancer. 2011;30(10):712–20. doi:10.5732/ cjc.011.10234. 14. Liang XH, Jackson S, Seaman M, Brown K, Kempkes B, Hibshoosh H, et al. Induction of autophagy and inhibition of tumorigenesis by beclin 1. Nature. 1999;402(6762):672–6. doi:10.1038/45257. 15. Wang K, Liu R, Li J, Mao J, Lei Y, Wu J, et al. Quercetin induces protective autophagy in gastric cancer cells: involvement of AktmTOR- and hypoxia-induced factor 1alpha-mediated signaling. Autophagy. 2011;7(9):966–78. 16. Chen Y, Lu Y, Lu C, Zhang L. Beclin-1 expression is a predictor of clinical outcome in patients with esophageal squamous cell

722

17.

18.

19.

20.

21.

22.

23.

24.

25. 26.

27.

28.

29.

30.

Tumor Biol. (2013) 34:713–722 carcinoma and correlated to hypoxia-inducible factor (HIF)-1alpha expression. Pathol Oncol Res. 2009;15(3):487–93. doi:10.1007/ s12253-008-9143-8. Wan XB, Fan XJ, Chen MY, Xiang J, Huang PY, Guo L, et al. Elevated beclin 1 expression is correlated with HIF-1alpha in predicting poor prognosis of nasopharyngeal carcinoma. Autophagy. 2010;6(3):395–404. Ahn CH, Jeong EG, Lee JW, Kim MS, Kim SH, Kim SS, et al. Expression of beclin-1, an autophagy-related protein, in gastric and colorectal cancers. APMIS. 2007;115(12):1344–9. doi:10.1111/ j.1600-0463.2007.00858.x. Li BX, Li CY, Peng RQ, Wu XJ, Wang HY, Wan DS, et al. The expression of beclin 1 is associated with favorable prognosis in stage IIIB colon cancers. Autophagy. 2009;5(3):303–6. Domart MC, Esposti DD, Sebagh M, Olaya N, Harper F, Pierron G, et al. Concurrent induction of necrosis, apoptosis, and autophagy in ischemic preconditioned human livers formerly treated by chemotherapy. J Hepatol. 2009;51(5):881–9. doi:10.1016/j.jhep. 2009.06.028. Abedin MJ, Wang D, McDonnell MA, Lehmann U, Kelekar A. Autophagy delays apoptotic death in breast cancer cells following DNA damage. Cell Death Differ. 2007;14(3):500–10. doi:10.1038/ sj.cdd.4402039. Liu B, Bao JK, Yang JM, Cheng Y. Targeting autophagic pathways for cancer drug discovery. Chin J Cancer. 2012. doi: 10.5732/ cjc.012.10010. Koukourakis MI, Giatromanolaki A, Sivridis E, Pitiakoudis M, Gatter KC, Harris AL. Beclin 1 over- and underexpression in colorectal cancer: distinct patterns relate to prognosis and tumour hypoxia. Br J Cancer. 2010;103(8):1209–14. doi:10.1038/sj.bjc.6605904. Crazzolara R, Bradstock KF, Bendall LJ. RAD001 (Everolimus) induces autophagy in acute lymphoblastic leukemia. Autophagy. 2009;5(5):727–8. Edge SB, Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti A et al. AJCC cancer staging manual (7th ed). New York: Springer; 2010. Kononen J, Bubendorf L, Kallioniemi A, Barlund M, Schraml P, Leighton S, et al. Tissue microarrays for high-throughput molecular profiling of tumor specimens. Nat Med. 1998;4(7):844–7. Wan XB, Long ZJ, Yan M, Xu J, Xia LP, Liu L, et al. Inhibition of Aurora-A suppresses epithelial-mesenchymal transition and invasion by downregulating MAPK in nasopharyngeal carcinoma cells. Carcinogenesis. 2008;29(10):1930–7. doi:10.1093/carcin/bgn176. Mylonis I, Chachami G, Paraskeva E, Simos G. Atypical CRM1dependent nuclear export signal mediates regulation of hypoxiainducible factor-1alpha by MAPK. J Biol Chem. 2008;283 (41):27620–7. doi:10.1074/jbc.M803081200. Xu J, Wan XB, Huang XF, Chan KC, Hong MH, Wang LH, et al. Serologic antienzyme rate of Epstein-Barr virus DNase-specific neutralizing antibody segregates TNM classification in nasopharyngeal carcinoma. J Clin Oncol. 2010; 28(35):5202–9. Amaravadi RK, Thompson CB. The roles of therapy-induced autophagy and necrosis in cancer treatment. Clin Cancer Res. 2007;13(24):7271–9. doi:10.1158/1078-0432.CCR-07-1595.

31. Goldstein NS, Armin M. Epidermal growth factor receptor immunohistochemical reactivity in patients with American Joint Committee on Cancer Stage IV colon adenocarcinoma: implications for a standardized scoring system. Cancer. 2001;92(5):1331– 46. doi:10.1002/1097-0142(20010901)92:5<1331::AIDCNCR1455>3.0.CO;2-M. 32. Bokemeyer C, Bondarenko I, Makhson A, Hartmann JT, Aparicio J, de Braud F, et al. Fluorouracil, leucovorin, and oxaliplatin with and without cetuximab in the first-line treatment of metastatic colorectal cancer. J Clin Oncol. 2009;27(5):663–71. doi:10.1200/ JCO.2008.20.8397. 33. Cunningham D, Humblet Y, Siena S, Khayat D, Bleiberg H, Santoro A, et al. Cetuximab monotherapy and cetuximab plus irinotecan in irinotecan-refractory metastatic colorectal cancer. N Engl J Med. 2004;351(4):337–45. doi:10.1056/NEJMoa033025. 34. Mizushima N, Levine B. Autophagy in mammalian development and differentiation. Nat Cell Biol. 2010;12(9):823–30. doi:10.1038/ ncb0910-823. 35. Levine B, Mizushima N, Virgin HW. Autophagy in immunity and inflammation. Nature. 2011;469(7330):323–35. doi:10.1038/ nature09782. 36. Furuya D, Tsuji N, Yagihashi A, Watanabe N. Beclin 1 augmented cis-diamminedichloroplatinum induced apoptosis via enhancing caspase-9 activity. Exp Cell Res. 2005;307(1):26–40. doi:10.1016/ j.yexcr.2005.02.023. 37. Pirtoli L, Cevenini G, Tini P, Vannini M, Oliveri G, Marsili S, et al. The prognostic role of Beclin 1 protein expression in high-grade gliomas. Autophagy. 2009;5(7):930–6. 38. Shi YH, Ding ZB, Zhou J, Qiu SJ, Fan J. Prognostic significance of Beclin 1-dependent apoptotic activity in hepatocellular carcinoma. Autophagy. 2009;5(3):380–2. 39. Zhou WH, Tang F, Xu J, Wu X, Yang SB, Feng ZY et al. Low expression of Beclin 1, associated with high Bcl-xL, predicts a malignant phenotype and poor prognosis of gastric cancer. Autophagy. 2012;8(3):389–400. doi:10.4161/auto.18641. 40. Dong LW, Hou YJ, Tan YX, Tang L, Pan YF, Wang M, et al. Prognostic significance of Beclin 1 in intrahepatic cholangiocellular carcinoma. Autophagy. 2011;7(10):1222–9. doi:10.4161/ auto.7.10.16610. 41. Ding ZB, Shi YH, Zhou J, Qiu SJ, Xu Y, Dai Z, et al. Association of autophagy defect with a malignant phenotype and poor prognosis of hepatocellular carcinoma. Cancer Res. 2008;68(22):9167– 75. doi:10.1158/0008-5472.CAN-08-1573. 42. Boutin AT, Weidemann A, Fu Z, Mesropian L, Gradin K, Jamora C, et al. Epidermal sensing of oxygen is essential for systemic hypoxic response. Cell. 2008;133(2):223–34. doi:10.1016/ j.cell.2008.02.038. 43. Azad MB, Chen Y, Henson ES, Cizeau J, McMillan-Ward E, Israels SJ, et al. Hypoxia induces autophagic cell death in apoptosis-competent cells through a mechanism involving BNIP3. Autophagy. 2008;4(2):195–204. 44. Mazure NM, Pouyssegur J. Hypoxia-induced autophagy: cell death or cell survival? Curr Opin Cell Biol. 2010;22(2):177–80. doi:10.1016/j.ceb.2009.11.015.