Formalin-fixed paraffin-embedded blocks were obtained, containing tissue samples from 40 patients with histologically proven PNETs and who underwent surgical resection at Kagoshima University Hospital (Kagoshima, Japan) between January 1995 and December 2015. All resected tissue specimens were histologically examined using hematoxylin and eosin staining according to the tumor-node-metastasis classification and WHO classification 2010 systems (4,15). The present study was approved by the Institutional Ethics Review Board of Kagoshima University Hospital, and written informed consent was obtained from all patients.

The rabbit anti-human polyclonal antibody directed against Snail (catalogue no. ab85936; Abcam, Cambridge, UK) was diluted at 1:500. The mouse anti-human monoclonal antibody against E-cadherin (catalogue no. M361229-2; Dako; Agilent Technologies, Inc., Santa Clara, CA, USA) was diluted at 1:500.

All specimens were fixed in 10% formalin at room temperature and processed routinely. Sections (3 µm-thick) were cut from the paraffin blocks of primary tumors. Following deparaffinization in xylene and rehydration in graded solutions of ethanol, endogenous peroxidase activity was blocked by immersing the slides in absolute methanol solution containing 3% hydrogen peroxide for 30 min at room temperature. The sections were then treated with 1% goat serum albumin for Snail (catalogue no. S-1000) or 1% horse serum albumin for E-cadherin (catalogue no. S-2000) (both Vector Laboratories, Inc., Burlingame, CA, USA) for 30 min at room temperature to block nonspecific reactions. Heat-induced antigen retrieval via autoclave pretreatment (120°C for 5 min) in citrate buffer solution (pH 6.0) was performed. The sections were rinsed with PBS and incubated with the Snail antibody (dilution, 1:500) for 180 min and the E-cadherin antibody (dilution, 1:500) for 60 min at room temperature. Following incubation, the specimens were visualized using an Vectastain Elite ABC IgG Rabbit (catalogue no. PK-6101) or Mouse (catalogue no. PK-6102) kits (both Vector Laboratories, Inc.) and the Liquid DAB+ Substrate Chromogen system (catalogue no. K3468; Dako; Agilent Technologies, Inc.), according to the manufacturer's protocol. The slides were counterstained with hematoxylin prior to mounting. All reactions were performed with positive controls (mouse heart tissue for Snail and normal pancreatic tissue for E-cadherin). For the negative control, the primary antibody was replaced with PBS. No significant staining was observed in the negative-control sections.

All tissue sections were simultaneously assessed by two investigators who were blinded to the patient clinicopathological data. Each slice was observed under an Olympus CX31 optical microscope (Olympus Corporation, Tokyo, Japan) (magnification, ×400). Stained cells were assessed and quantified in five randomly selected fields. The immunohistochemical staining of Snail was classified as high expression if ≥10% of neoplastic cells exhibited nuclear staining, or as low expression if <10% of neoplastic cells were stained. The expression of E-cadherin was compared between tumor cells and normal islet cells located adjacent to the tumor. Tumor cells that exhibited equivocal staining to normal islet cells were considered to have preserved expression of E-cadherin, whereas those that exhibited less intense staining patterns compared with the normal islet cells, or those that did not stain at all, were considered to have reduced expression of E-cadherin.

Associations between different categorical variables were assessed using Fisher's exact test and the χ² test. The Kaplan-Meier method was used for survival analysis, and variations in survival rates were estimated using the log-rank test. All statistical analysis was performed using SPSS version 22.0 (IBM SPSS, Armonk, NY, USA). All tests were two-sided. P<0.05 was considered to indicate a statistically significant difference.

Table I lists the clinicopathological features of the patients enrolled in the present study. The study group comprised 17 men and 23 women with an age range between 18 and 79 years (mean, 56.1 years). Lymph node metastasis was observed in 7 patients (17.5%). Distant metastasis was observed in 9 patients (22.5%), and the site of metastasis was the liver. Postoperative recurrence or an increase in the residual tumor size was observed in 10 (25.0%) patients.

Fig. 1 depicts representative images of Snail and E-cadherin expression. Snail expression was identified in the nucleus of neoplastic cells. The proportion of immunoreactive cells varied, ranging from 0 to >50% of neoplastic cells. The expression of E-cadherin in neoplastic cells was heterogeneous, ranging from 0 to ~100% of neoplastic cells. High and low Snail expression levels were observed in 11 (27.5%) and 29 (72.5%) patients, respectively. Preserved and reduced E-cadherin expression was observed in 19 (47.5%) and 21 (52.5) patients, respectively.

Table II demonstrates the association between Snail/E-cadherin expression and various clinicopathological factors. High Snail expression levels were significantly associated with high incidences of lymphatic and vascular invasion and liver metastasis (P=0.0078, P<0.0001 and P=0.0067, respectively; Table II). Reduced E-cadherin expression levels were also significantly associated with high incidences of lymphatic and vascular invasion and liver metastasis (P=0.0015, P=0.0270 and P=0.0214, respectively; Table II). No significant difference was observed in Snail and E-cadherin expression according to gender, tumor size and tumor functionality. Furthermore, on the basis of the expression profiles of Snail and E-cadherin, all patients were divided into two groups as follows: Low Snail and preserved E-cadherin expression, and the ‘other group’ (patients with high Snail and reduced E-cadherin, high Snail and preserved E-cadherin, and low Snail and reduced E-cadherin expression) (Table III). Patients in the ‘other group’ experienced a markedly increased lymph node metastasis, liver metastasis, G2 and NEC in the WHO classification, vascular invasion and lymphatic invasion, as compared with in the low Snail and preserved E-cadherin expression group. Patients with low Snail and preserved E-cadherin expression were not observed to exhibit lymph node metastasis or liver metastasis during the present study. In total, 1/22 patients with a G1 classified tumor experience liver metastasis; these particular tumor tissues exhibited high Snail and reduced E-cadherin expression levels.

The 5-year and 10-year survival rates for the patient cohort were determined to be 91.7 and 59.5%, respectively. Fig. 2 presents the overall survival (OS) and progression-free survival (PFS) rates following surgery in the aforementioned two groups, based on the expression levels of Snail and E-cadherin. The ‘other group’ exhibited a poorer OS rate, as compared with the low Snail and preserved E-cadherin expression group, although these differences were not statistically significant (Fig. 2A). By contrast, a significant difference was observed between the two groups with regard to the rate of PFS (P=0.0099; Fig. 2B).