Gene Expression Profiling Interactive Analysis (GEPIA, http://gepia.cancer-pku.cn/index.html) was used to show the expression of AGR2, which was the target gene, in esophageal cancer.

Human ESCC cell lines TE2, TE5, TE8, TE11 and TE15 and human osteosarcoma cell lines U2OS and SAOS2 were obtained from the RIKEN BioResource Center Cell Bank. Human ESCC cell lines KYSE70, KYSE150 and KYSE170 were obtained from the Japanese Collection of Research Bioresources Cell Bank. The immortalized esophageal epithelium cell line Het-1A and mesothelial cell line MeT-5A were obtained from the American Type Culture Collection. The ESCC cell lines Het-1A and MeT-5A were maintained in Roswell Park Memorial Institute medium (Nakalai Tisque) and U2OS and SaOS2 cells were cultured in HyClone McCoy's 5A medium (Cytiva). The two media were supplemented with 10% fetal bovine serum (System Biosciences), 100 U/ml penicillin and 100 µg/ml streptomycin. All cell lines were cultured in a humidified 37°C incubator with 5% carbon dioxide.

Total RNA was extracted from the cell lines using a miRNeasy Mini kit (Qiagen) as per the manufacturer's instructions. Reverse transcription reactions were performed using a High Capacity cDNA Reverse Transcription kit (Applied Biosystems; Thermo Fisher Scientific, Inc.), as per the manufacturer's given procedure, at 25°C for 10 min, followed by 37°C for 120 min and 85°C for 5 min. The level of mRNA expression was measured using TaqMan Gene Expression Assays (Hs_00356521_m1 for AGR2 and Hs_02758991_g1 for GAPDH; Applied Biosystems; Thermo Fisher Scientific, Inc.), as per the manufacturer's protocol. Reverse transcription-quantitative (RT-q) PCR was performed using a StepOnePlus PCR System (Applied Biosystems; Thermo Fisher Scientific, Inc.) and the cycle threshold (Ct) value was calculated using StepOne Software v2.3 (Applied Biosystems; Thermo Fisher Scientific, Inc.). The thermocycling conditions were: 95°C for 10 min, followed by 40 cycles at 95°C for 15 sec, 55°C for 30 sec and 72°C for 30 sec. The results were examined using the 2^−ΔΔCq method relative to the expression of GAPDH (24).

The cultured cells were lysed using the M-PER Mammalian Protein Extraction Reagent (Thermo Fisher Scientific, Inc.) and protein was extracted. The protein concentration was adjusted using the Protein Assay Rapid kit Wako II (FUJIFILM Wako Pure Chemical Corporation). The prepared lysate containing 20 µg of total protein was subjected to 12 or 15% sodium dodecyl sulfate-polyacrylamide gel electrophoresis, followed by transfer to a polyvinylidene difluoride membrane (Cytiva). After blocking by 5% skimmed milk powder or 5% bovine serum albumin (Sigma-Aldrich, Merck KGaA) at room temperature for 1 h, the membrane was subsequently probed with each candidate antibody at 4°C overnight. Next day, the secondary antibody was added into the membranes at room temperature for 1 h and the level of protein expression in each case was visualized via SuperSignal West Dura (Thermo Fisher Scientific, Inc.) and evaluated using an Amersham Imager 680, software version 2.0.0 (Cytiva). β-actin (ACTB) was used as a loading control.

The antibodies were purchased and diluted as follows: anti-AGR2 antibody (rabbit, 1:5,000; cat. no. Ab76473) from Abcam; anti-phospho-p53 antibody (serine 15) (mouse, 1:1,000; cat. no. 9286S) from Cell Signaling Technology, Inc.; anti-p53 antibody (DO1, Mouse, 1:1,000; cat. no. sc-126) from Santa Cruz Biotechnology, Inc. and anti-ACTB antibody (Mouse, 1:20,000; cat. no. A5441) from Sigma-Aldrich. The secondary antibodies were: Anti-mouse IgG, HRP-linked anti body (horse; cat. no. 7076) and anti-rabbit IgG, HRP-linked anti body (goat; cat. no. 7074) both from Cell Signaling Technology, Inc. The dilution rate of the secondary antibody varied according to the dilution rate of the primary antibody.

The expression of AGR2 was downregulated using small interfering RNA (siRNA) targeting AGR2 (Stealth RNAi™ siRNA, Thermo Fisher Scientific, Inc.; cat. no. HSS116220; UUUCUUUAAAGCUUGACUGUGUGGG). ESCC and osteosarcoma cells were transfected either with control siRNA (Stealth RNAi™ siRNA Negative Control, Thermo Fisher Scientific; cat. no. 12935112) or siRNA targeting AGR2 at 10 nmol/l in a six-well culture plate using Lipofectamine^® RNAiMAX (Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. After transfection for 72 h at 37°C, AGR2 mRNA and protein expression was confirmed using RT-qPCR and western blotting, respectively.

Cells were seeded on 96-well plates, incubated for 24 h and transfected with either control or AGR2-targeting siRNA. After transfection, cells were incubated for 24 h at 37°C, and the number of viable cells 0, 24, 48, 72 and 96 h after treatment was determined using absorbance data at a wavelength of 450 nm using Cell Count Reagent SF (Nakalai Tesque).

TE2, TE5 and TE15 cells transfected with either control or AGR2-targeting siRNA were seeded in six-well culture plates at a density of 1,000 cells per well. After 2 weeks, the cells were fixed with 4% paraformaldehyde at room temperature for 10 min and briefly stained with crystal violet (Nakalai Tesque) at room temperature until all cells were stained enough. The areas of all stained cells were measured using the NIH ImageJ System (v1.53) (http://rsb.info.nih.gov/ij/).

Cell cycles were analyzed 72 h after transfection using flow cytometry. The cells were detached from the plate using trypsin-EDTA, treated with 0.2% Triton X-100 and stained with PI RNase staining buffer (Becton-Dickinson and Company). The cells were analyzed using a BD Accuri C6 (BD Biosciences) and cell cycle distribution was recorded using the BD Accuri Software (BD Biosciences). At least 10,000 cells were assessed for each measurement.

The cells were stained and evaluated using an Annexin V-FITC kit (Beckman Coulter, Inc.) 72 h after transfection according to the manufacturer's instructions. The proportions of early and late apoptotic cells were measured by flow cytometry using the BD Accuri C6. At least 10,000 cells were assessed for each measurement.

A total of 81 patients with ESCC who had undergone esophagectomy between 1999 and 2013 at the Kyoto Prefectural University of Medicine were enrolled in the present study. Written informed consent for the research was obtained from all patients prior to their respective treatments. Patients who had undergone non-curative resection or preoperative chemo- or radiation-therapy were excluded from the present study. The median length of the follow-up period for the censored cases was 5.98 years (range, 0.38–16.7 years). A total of 35 patients (43.2%) had recurrence within 5 years of surgery and 26 patients (32.0%) died of primary ESCC. The clinicopathological features were evaluated using the 8th edition of the International Union Against Cancer (UICC)/tumor-node-metastasis (TNM) Classification of Malignant Tumors (25). The present study was approved by the Faculty of Science Ethics Committee of the Kyoto Prefectural University of Medicine (approval no. ERB-C-1414-1).

IHC staining was performed using the labeled streptavidin-biotin method using a Vectastain Universal Quick kit (Vector Laboratories, Inc.) as per the manufacturer's protocol. The method can be briefly described as follows: antigen retrieval was performed using heated Dako Target Retrieval Solution (Dako; Agilent Technologies, Inc.) at 95°C for 60 min and endogenous peroxidase activity was disrupted by incubation in 0.3% H₂O₂ at room temperature for 20 min; then, 2.5% normal horse serum (Vector Laboratories, Inc.) was used to block non-specific binding at room temperature for 10 min. The slides were incubated with either the anti-AGR2 antibody 1:500 (cat. no. Ab76473; Abcam) or the anti-p53 antibody 1:200 (mouse, NCL-L-p53-DO7; Novocastra Laboratories Ltd.) at 4°C overnight. They were then incubated with a biotinylated pan-specific universal antibody at room temperature for 10 min and a streptavidin/peroxidase complex (Vector Laboratories, Inc.) at room temperature for 5 min, followed by incubation with 3,3′-diaminobenzidine tetrachloride. Counterstaining was performed using hematoxylin. A formalin-fixed ESCC cell line overexpressing AGR2 (TE15 cells) was used as the positive control.

The AGR2 staining was evaluated based on a score depicting the intensity and proportion of stained cancer cells. The intensity was scored as 3 (strong), 2 (intermediate), 1 (weak), or 0 (no staining). The proportion was evaluated as the ratio of stained areas in the cancer tissue and was scored from 0 to 1. Cancer tissues with an intensity score of 3 and a proportion ≥0.3 or an intensity score of 2 and a proportion ≥0.5 were considered the high expression group and the others comprised the low expression group.

p53 staining was evaluated in a manner similar to that of AGR2. The intensity was scored as 3 (strong), 2 (intermediate), 1 (weak), or 0 (no staining) and the proportion was scored from 0 to 1. Cancer tissues with an intensity score of 3 or 2 and a proportion ≥0.3 were considered the high expression group and the others comprised the low expression group.

Statistical analysis was performed using JMP Version 14 (ASA Institute). The Wilcoxon signed-rank test or Student's t-test was used to compare differences between the paired and unpaired samples. Differences between survival curves were examined using the log-rank test. Multivariate analysis of survival was carried out using the Cox regression method. Relationships between clinicopathological factors were assessed using Fisher's exact test or the Chi-square test. P<0.05 was considered to indicate a statistically significant difference.

The mRNA expression of AGR2 was significantly higher in EC tissue compared with the normal counterpart(P<0.001; Fig. 1A), according to GEPIA (26). Among the nine ESCC cell lines and two non-cancerous cell lines, the mRNA and protein expression of AGR2 was higher in the TE2, TE5 and TE15 cells compared with the other cells (Fig. 1B). p53 protein expression was also confirmed, a result that was partly consistent with the previously reported mutation status (Fig. 1B) (27).

AGR2-expressing cell lines TE15 and TE5, which are reported to be TP53-wild-type and mutant cell lines, respectively (27,28), were selected for the subsequent experiments. The transient introduction of the siRNA targeting AGR2 efficiently downregulated AGR2 mRNA and protein expression levels (Fig. S1). As determined using the proliferation assay, the downregulation of AGR2 expression inhibited the growth of TE15 cells but not TE5 cells (Fig. 1C). The downregulation of AGR2 also reduced colony formation in TE15 cells but not TE5 cells (Fig. 1D).

In addition, cell cycle analysis demonstrated that AGR2 downregulation increased the proportion of TE15 cells, but not TE5 cells, in the G2/M phase (Fig. 2A). The apoptosis assay demonstrated that early and late apoptosis was increased only for the TE15 cells (Fig. 2B). The TE2 cell line, which is a TP53-wild-type cell line with AGR2 expression, demonstrated similar changes in cell function as TE15 cells owing to AGR2 downregulation (Fig. S2A-E).

Based on the aforementioned results, the associations between AGR2 expression and the phosphorylation or mutation status of TP53 were examined. The phosphorylation of p53 Ser15 was markedly enhanced by the downregulation of AGR2 in TP53-wild-type TE15 cells (Fig. 2C). By contrast, in TP53-mutant TE5 cells, the phosphorylation of p53 was already high and not enhanced by the downregulation of AGR2 (Fig. 2C).

For further verification of these associations, similar examinations were performed using the osteosarcoma cell lines U2OS, which is a TP53-wild-type cell line and SAOS2, which is a TP53-deficient cell line (29). AGR2 and p53 expression was confirmed in U2OS cells, whereas in SAOS2 cells, AGR2 expression was very low and p53 expression was not detected (Fig. 3A and B). The present study therefore only examined the effects of AGR2 downregulation in U2OS cells. The downregulation of AGR2 also suppressed the proliferation of and enhanced p53 Ser15 phosphorylation in U2OS cells (Fig. 3C-E).

AGR2 was expressed in the esophageal gland but not the non-cancerous esophageal epithelium (Fig. 4A and B). In ESCC tissue, AGR2 was detected in the cell membrane and cytoplasm and the level of AGR2 expression was categorized as either low or high (Fig. 4C and D). No significant differences were observed in the 5-year overall survival rate was between the high and low expression groups (60.5 vs. 71.2%; P=0.32; Fig. 4E). In addition, no significant differences were observed in the clinicopathological features between these groups (Table I).

The result of IHC for p53 was also examined and categorized into low and high expression groups (Fig. S3A and B). No significant differences were observed in the 5-year overall survival rate between the high and low p53 expression groups (73.7 vs. 61.9%, P=0.28; Fig. S3C). However, with respect to AGR2 and p53 expression in combination, patients with high-AGR2/low-p53 expression demonstrated significantly worse prognosis than other patients. The 5-year overall survival rate was 40.0% in the high-AGR2/low-p53 expression group, whereas it was 70.3% in the low-AGR2/low-p53 expression group (P=0.046; Fig. 4F). The survival analysis of patients with low p53 expression demonstrated that a large tumor of >40 mm in size (HR=3.82, P=0.03) and high AGR2 expression (HR=3.66; P=0.03) were independent prognostic factors (Table II).