Shanghai Biochip Company provided the human EAC high density tissue microarray (HGEj-Ade130Sur-01), which includes primary tumor specimens and matched adjacent tissues from 65 EAC patients. Between September 2006 and December 2009, all patients had EAC resection. Preoperative chemotherapy or radiotherapy was not given to any of these patients. All cases were pathologically confirmed. The tumor-node-metastasis (TNM) stage was determined using the 7th version of the American Joint Committee on Cancer Criteria (AJCC).

The tissues were fixed in 10% formalin at 4°C for 24 h before being embedded in paraffin. Tissue samples were sectioned into 4 µm-thick slices, and paraffin-embedded slices were dewaxed in dimethylbenzene and rehydrated in a graded ethanol solution. Antigen retrieval was carried out in citrate solution (10 mmol/L, pH 6.0) at 100°C for 30 min. After 15 min in 0.3% hydrogen peroxide solution to block endogenous peroxidase activity, tissues were incubated overnight at 4°C with the polyclonal rabbit anti-IL-33 (1:2,000; MBL Co., Ltd., Nagoya, Japan for human tissue; 1:200; Proteintech Group, Inc., Chicago, IL for rat tissue). The tissues were then incubated for 30 min at 37°C with the horseradish peroxidase (HRP) labeled goat anti-mouse/rabbit IgG polymer (1:5,000; Fuzhou Maixin Biotechnology Co., Ltd.). Color development was carried out using the enhanced DAB chromogenic kit (Fuzhou Maixin Biotech Co., Ltd.). After counterstaining at room temperature for 5 min with hematoxylin, the samples were dehydrated in a gradient of ethanol and fixed with neutral resin. The slides were then observed and photographed using a light microscope. The IL-33 immunostaining score was calculated by multiplying the intensity score by the positive rate score. The intensity of staining was measured as follows: 1) 0, no staining; 2) 1, weak staining; 3) 2, moderate staining; and 4) 3, strong staining. The positive staining cell rate was measured as follows: 1) 0, 0%–5%; 2) 1, 5%–25%; 3) 2, 26%–50%; 4) 3, 51%–75%; and 5) 4, >75%.

Wistar rats aged 5 weeks, weighing 220–300 g, were selected to establish the esophageal adenocarcinoma model of gastroduodenal mixed reflux by anastomosing the jejunum with the esophagus, following the established method (18). A 2 cm incision was performed in the upper abdomen to separate the connective tissue between the liver and stomach, retain the vagus nerves on both sides of the esophagus, separate the abdominal esophagus form the cardia, and cut off and suture the distal end. The lower esophagus was intermittently sutured with the distal ileum of Treitz ligament end to side with 7/0 silk thread. The rats in the control group were also anesthetized, opened, and separated, but not anastomosed. 50 mg/kg iron dextran was administered intramuscularly every 4 weeks postoperatively. The rats in the control group were killed 4 weeks after surgery, and the rats in the surgery group were killed 4, 10, and 16 weeks after surgery. After cardiac blood collection, the esophagus and the proximal ileal segment from the anastomosis (0.5 cm) were completely removed in each group. Excluding the dead rats, there were 10 rats in the control group, 8 rats in the 4w group, 9 rats in the 10w and 16w groups. Low-grade dysplasia occurred in all rats in the 4w group, 1 rat in the 10w group, and 3 rats in the 16w group. High-grade dysplasia occurred in 7 rats in the 10w group and 2 rats in the 16w group; EAC occurred in 1 rat in the 10w group and 4 rats in the 16w group.

Total RNA was extracted from tissue using TRIzol^® reagent (Invitrogen; Life Technologies, Inc.). The obtained cDNA was subjected to RT-qPCR using TB Green®Premix Ex Taq™ (Takara Bio, Inc.) on the LightCycler/LightCycler 480 System (Roche, Inc.) after the contaminated DNA was removed and reverse transcribed using Primescript™ RT Reagent Kit with gDNA Eraser (Takara Bio, Inc.). GAPDH was used as internal reference. Each sample was performed in triplicate, and relative RNA levels were calculated using the 2^−ΔΔCq method. The sequences were as follows: IL-1β: Forward: 5′-AAT​CTC​ACA​GCA​GCA​TCT​CGA​CAA​G-3′ and reverse: 5′- TCC​ACG​GGC​AAG​ACA​TAG​GTA​GC-3′; IL-4: Forward: 5′- CAA​GGA​ACA​CCA​CGG​AGA​ACG​AG-3′ and reverse: 5′- TTC​TTC​AAG​CAC​GGA​GGT​ACA​TCA​C-3′; IL-6: Forward: 5′- CCG​CAA​GAG​ACT​TCC​AGC​CAG​TTG-3′ and reverse: 5′- CGG​AAC​TCC​AGA​AGA​CCA​GAG​CAG​A-3′; IL-10: Forward: 5′- GGC​AGT​GGA​GCA​GGT​GAA​GAA​TG-3′ and reverse: 5′- TGT​CAC​GTA​GGC​TTC​TAT​GCA​GTT​G-3′; IL-33: Forward: 5′-GAA​CCC​GCC​AAA​AGA​TAT​TCA​C-3′ and reverse: 5′-AAG​TTC​CTT​GGA​TAC​TCA​GTG​TG-3′; GAPDH: Forward: 5′-CCA​TCA​ACG​ACC​CCT​TCA​TT-3′ and reverse: 5′-GAC​CAG​CTT​CCC​ATT​CTC​AG-3′; ST2: Forward: 5′-CGT​TAC​CTT​CCT​GTG​CCA​TT-3′ and reverse: 5′-CTC​CAT​TTG​CCA​ATC​ATG​TG-3′.

ATCC supplied the human esophageal adenocarcinoma cell lines (OE19 and OE33), as well as human normal esophageal epithelial cells (HEECs). OE19 and OE33 were cultured in RPMI-1640 medium (Gibco, Grand Island, NY, United States), while HEECs were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Hyclone; Cytiva), supplemented with 10% fetal bovine serum (Gibco, Grand Island, NY, United States), L-glutamine, 100 U/ml penicillin, and 0.1 mg/ml streptomycin (SigmaAldrich; Merck KGaA) at 37°C, 5% CO₂ air atmosphere.

The effect of IL-33 on the proliferation of OE19 and OE33 cells was assessed using the CCK8 assay. The cells were seeded at a density of 2,000 cells per well in a 96-well plate. After 24 h, the cells were incubated with IL-33 (Proteintech Group, Inc., Chicago, IL, United States) in increasing concentrations (0, 20 and 50 ng/ml) for 24 h, 48 h and 72 h, respectively. We used sST2 to block the IL-33/ST2 pathway (200 ng/ml) (Immune Technology Corp, NY, United States), which was administered 24 h before adding IL-33. The cells were then washed with PBS and incubated for 3 h with 100 µl of CCK8 solution. The multifunctional fluorescence microplate reader (Polarstar Otima, BMG, Australia) was then used to measure the optical density (OD) at 450 nm.

The capacity for cell migration was evaluated using the wound-healing experiment. In 12-well flat-bottomed plates, cells were planted to attain 80%–90% monolayer confluence. A straight line was scratched in the middle of the well with the tip of a sterile pipette. The wound margins of different treatment groups were inspected and photographed after washing the cell debris and administering 50 ng/ml IL-33 after 24 and 36 h. The migration rate was reported as the proportion of the average migration distance to the average initial (0 h) wound distance.

The invasive experiment was conducted utilizing a Matrigel-coated 24-well Transwell (BD Biosciences, San Jose, CA, United States) (8-m pore size; Corning, NY, United States). The upper chamber has 2 * 10⁵ cells suspended in 200 µl of serum-free media with PBS, 50 ng/ml IL-33, 200 ng/ml sST2 and 50 ng/ml IL-33 + 200 ng/ml sST2; The lower chamber contains 500 µl of medium supplemented with 10% FBS. After 24 h, cotton swabs were used to remove the cells on the upper surface of the membrane. The lower surface cells were subsequently fixed with paraformaldehyde and stained with crystal violet. Finally, invasive cells were enumerated and collected in five microscopic fields.

The whole cell proteins were extracted by RIPA lysis buffer. After being boiled for 5 min, protein samples were separated by 10% SDS-PAGE, transferred to PVDF membranes, and incubated overnight at 4°C with primary antibodies against IL-33 (1:1,000, Proteintech Group, Inc., Chicago, IL, United States), E-cadherin (1:25,000, Proteintech Group, Inc., Chicago, IL, United States) and N-cadherin (1:1,000, Abcam, Cambridge, United Kingdom). The next day, the membranes were washed and incubated with the HRP-conjugated secondary antibody (1:5,000, Proteintech Group, Inc., Chicago, IL, United States) for 2 h at room temperature. GAPDH (1:2,000, Proteintech Group, Inc., Chicago, IL, United States) was used as internal reference.

Enzyme-linked immunosorbent assay (ELISA) for IL-33 and sST2 in serum was performed using the IL-33 Quantikine ELISA Kit (R&D Systems, Inc.) and the sST2 Quantikine ELISA Kit (R&D Systems, Inc.). The supernatant of the cells was collected and centrifuged at 4°C at 1,000 g for 10 min. IL-4 Quantikine ELISA kit (R&D Systems, Inc.) and IL-6 Quantikine ELISA kit (R&D Systems, Inc.) were used to measure the protein secretion of IL-4 and IL-6 in esophageal adenocarcinoma cells following the manufacturer’s instructions.

All data analysis was performed as the mean ± SD using SPSS 22.0 (SPSS, Inc.). All experiments were repeated at least three times. Data was analyzed using unpaired t-tests for two groups and one-way ANOVA followed by Scheffe’s F test for multiple comparisons. All tests were two-sided with a significance level of p < 0.05.

To investigate the expression of IL-33 in EAC, a high-density tissue microarray containing 65 pairs of EAC and surrounding tissue was utilized (Figure 1A). IHC showed that IL-33 was mainly expressed in the cytoplasm in EAC cells and in the nucleus of adjacent normal epithelial cells (Figure 1B). The staining scores revealed that EAC cells had considerably greater cytoplasmic expression of IL-33 than adjacent normal epithelial cells (Figure 1C), but significantly lower nucleus expression of IL-33 than adjacent normal epithelial cells (Figure 1D).

In the esophageal adenocarcinoma rat model, the mRNA expression of IL-33 was significantly higher in the 10w and 16w groups than that in other groups (Figure 2A), and increased with the upgrade of pathological stage (Figure 2B). IHC showed that IL-33 was mainly expressed in the cytoplasm in esophageal adenocarcinoma cells (Figures 2C,D). The content of IL-33 in serum also increased with the upgrade of pathological stage (Figure 2E). However, the content of sST2 is too low to be detected.

Two esophageal adenocarcinoma cell lines (OE33: poorly differentiated adenocarcinoma; OE19: moderately differentiated adenocarcinoma) and human normal esophageal cells (HEECs) were used to detect IL-33. OE19 and OE33 cells have substantially greater IL-33 mRNA and protein levels than HEECs. (Figures 3A,B).

The CCK8 assay demonstrated that IL-33 increased the proliferation of OE19 (Figure 3C) and OE33 (Figure 3D) in a time- and dose-dependent manner. Blocking the IL-33/ST2 pathway with sST2 could attenuate the effect of IL-33 on the proliferation of OE19 (Figure 3E) and OE33 (Figure 3F).

In the wound healing assay, IL-33 greatly increased the migration of OE19 (Figure 4A) and OE33 (Figure 4B), which can be attenuated by pretreatment with sST2 (Figures 4C,D). As OE19 cannot pass through the Matrigel transwell, we only detected the effect of IL-33 in OE33 for the invasion assay. We found that IL-33 considerably enhanced the invasion capacity of OE33, and this effect could be weakened by sST2 (Figures 4E,F).

We investigated further the function of IL-33 in EMT, which plays a crucial role in tumor migration and invasion. OE19 and OE33 were stimulated with IL-33 (50 ng/ml) for 12 h to investigate the effect of IL-33 on the expression of EMT markers. In OE19 (Figures 5A–C) and OE33 (Figures 5D–F), IL-33 dramatically decreased E-cadherin expression while increasing N-cadherin expression, which could be inhibited by pretreatment with sST2.

We detected the tissue mRNA expression of IL-1β, IL-4, IL-6, and IL-10 in the esophageal adenocarcinoma rat model (Figure 6A). The mRNA expression of IL-6 was significantly higher in the adenocarcinoma group than that in the normal group (Figure 6A). However, we did not detect the alteration of IL-6 content in rat model serum (data not shown). To investigate whether IL-33 can promote the expression of IL-1β, IL-4, IL-6, and IL-10 in esophageal adenocarcinoma cells, the mRNA levels of the cytokines were detected with or without IL-33 stimulation. The IL-33 stimulation group in OE19 had considerably higher levels of IL-4 and IL-6 mRNA than the control group (Figure 6B). And the level of IL-6 mRNA in the IL-33 stimulation group was considerably higher than in the control group in OE33 (Figure 6C). We further confirmed the increased secretion of IL-6 in the supernatant of OE19 (Figure 6D) and OE33 (Figure 6E) after IL-33 stimulation. However, the content of IL-4 did not change significantly (Figures 6F,G).