Seven-week old male Sprague-Dawley rats (Kist, Taejun, Korea) were used in this study. Throughout the experiment, all rats were housed in a controlled environment with a 12-h light/dark cycle at 22 ± 2℃. After a 1-week acclimation period, a total of 60 rats underwent operative procedures. Solid food was withdrawn for 24 h and water for 12 h before surgery. Anesthesia was induced and maintained with an isoflurane-air mixture. Esophagogastroduodenal anastomosis (EGDA) was performed.25 In short, a midline laparotomy was performed, two 1-cm incisions were made on both the gastroesophageal junction and on the anti-mesenteric border of the duodenum and were anastomosed together with accurate mucosal to mucosal opposition. Care was taken not to reach the glandular stomach when the gastroesophageal junction incision was made. Rats were allowed to drink water at 6 h post-op and were fed the following day. The Ethics committee supervising animal studies at the Dongguk University College of Medicine approved the experimental design.

Fifty-one of 60 animals operated on survived the duration of the experiment. These were divided into the control group (n = 27), the 500 ppm sulindac (Sigma, St Louis, MO, USA)-treated group (n = 15), and the 1000 ppm sulindac-treated group (n = 9). The sulindac-treated groups (n = 24) were given rat chow premixed with sulindac. An additional 12 rats underwent a sham operation and were given commercial rat chow to obtain normal esophageal tissue. Each rat was weighed every week throughout the experiment. Randomly selected rats were killed by diethyl ether inhalation on the 20th and 40th weeks after surgery. Rats were given a single intraperitoneal injection of 0.1 mg/g body weight of bromodeoxyuridine (BrdU) (Sigma) 1 h before autopsy.

The lower esophageal diameters were measured 0.5 cm away from the anastomosis site., Formalin-fixed tissues were embedded in paraffin, sectioned at 4 µm, and stained with H&E. Lower esophageal inflammation was graded as mild, moderate, or severe for histological evaluation. Barrett's esophagus was diagnosed when mucus-secreting columnar cells and goblet cells surrounded by squamous epithelium were present above the anastomosis site. Adenocarcinoma was diagnosed when glands with malignant cytological features were present both at the superficial and deep portions of the wall.

Serial 4-µm sections were made and spread on poly-L-lysine coated slides. Paraffin sections were immersed in three changes of xylene and hydrated using a graded alcohol series. Antigen retrieval was performed routinely by immersing the sections in 0.01 M citrate buffer (pH 6.0) and autoclaving for 15 min. Endogenous peroxidase activity was blocked with 3% hydrogen peroxide for 15 min and incubated with monoclonal mouse anti-BrdU antibody (DakoCytomation, Carpinteria, CA, USA) overnight in a humidified chamber at 4℃ to localize proliferating cells. Staining was achieved with a DAKO LSAB + kit (Dako Cytomation) and developed with diaminobenzidine tetrahydrochloride (DakoCytomation). Sections were counterstained for 5 min with Meyer's hematoxylin and mounted. As a negative control, mouse IgG isotype (DakoCytomation) was used instead of primary antibody.

Tissue from three rats in the same group was mixed in a tube and homogenized. Tissues were suspended in lysis buffer [10 mM Tris-HCl (pH 7.4), 1 mM EDTA, 0.25 M sucrose, 1% Triton X-100] supplemented with Complete mini protease inhibitor mixture tablets (Boehringer Mannheim, Mannheim, Germany) on ice for 1 h. After removing cell debris by centrifugation, the cell lysate protein concentration was determined by a BCA protein assay reagent (Pierce, Rockford, IL, USA) with a bovine serum albumin standard. Forty µg of protein was separated by 12% SDS-polyacrylamide gel electrophoresis and transferred to the nitrocellulose membrane. The membranes were blocked with 5% skim milk in Tris buffered saline (TBS) for 1 h at room temperature and probed with antibody overnight at 4℃. The primary antibodies were polyclonal anti-COX-2 (1 :1000, Cayman Chemical, Ann Arbor, MI, USA), anti-β-actin (1 : 200, Santa Cruz Biotechnology, Santa Cruz, CA, USA), mouse monoclonal anti-PCNA (1 : 200, Santa Cruz Biotechnology), and anti-cyclin D1 (1 : 100, Santa Cruz Biotechnology). After washing with TBS-0.05% Tween 20, the blots were treated with horseradish peroxidase-conjugated anti-rabbit IgG antibody (1 : 3000; Zymed Laboratories Inc. San Franciso, CA, USA) or anti-mouse IgG antibody (1 : 3000. Zymed Laboratories Inc.) for 1 h at room temperature. Enhanced chemiluminescence (Pierce) and autoradiography were used for detection.

We used the Kruskal-Wallis test to compare the lower esophageal diameters among the groups. When lower esophagus inflammation and Barrett's esophagus and adenocarcinoma incidences were compared, the Fisher's exact or Chi-square tests were used. Statistical significance was assumed if the p value was less than 0.05.

The respective number of rats examined in the control and sulindac-treated group were as follows: 8, 6, and 3 rats at week 20 and 19, 9, and 6 rats at week 40 were in the control, 500 ppm sulindac group, and 1000 ppm sulindac group, respectively. A total of 9 rats (4 control and 5 sulindac-treated) died. There was no significant difference in mortality between the two groups. Final body weights (g) were 378 ± 21 in the control, 358 ± 23 in the 500 ppm sulindac, and 392 ± 57 in the 1000 ppm sulindac at 20 weeks. They were 398 ± 40 in the control, 394 ± 48 in the 500 ppm sulindac, and 395 ± 45 in the 1000 ppm sulindac at 40 weeks. No significant difference in final body weights was found between the control and sulindac-treated groups at either time point.

The lower esophageal diameters were 10.7 ± 5.2 mm in rats undergoing EGDA and 3.8 ± 0.3 mm in sham surgery rats. Rat undergoing EGDA showed an abnormally dilated esophagus compared with sham surgery rats (p < 0.05). At 20 weeks, the control rats showed a wider lower esophagus than sulindac-treated rats, but the difference was not significant (Fig. 1A) (p > 0.05). At 40 weeks, rats treated with 1000 ppm sulindac had narrower esophageal diameters than those of the control rats (Fig. 1B) (p < 0.05). The lower esophageal mucosa displayed whitish nodular patches and superficial ulcers, which were prominent in rats with a larger lower esophageal diameter. The stomach and duodenum did not show any pathologic abnormalities.

At 20 weeks, lower esophageal inflammation was milder in the rats treated with sulindac than in the control group, but not significantly so (Fig. 2A) (p > 0.05). At 40 weeks, rats treated with 1000 ppm sulindac showed milder inflammation than control rats or those treated with 500 ppm sulindac (Fig. 2B) (p < 0.05).

Rats undergoing EGDA showed intestinal metaplasia including goblet cells above the esophagoduodenal junction (Fig. 3A) and adenocarcinoma characterized by abundant mucin secretion (Fig. 3B). At 20 weeks, Barrett's esophagus incidence was 50% in control rats and the 500 ppm sulindac-treated group and 0% in the 1000 ppm sulindac-treated group. Adenocarcinoma was present in 37.5% of the control rats and absent in sulindac-treated rats (Fig. 4A). At 40 weeks, the Barrett's esophagus incidence was similar between the control and sulindac-treated groups, but the incidence of adenocarcinoma was significantly lower in the 1000 ppm sulindac-treated group than the control or 500 ppm sulindac-treated groups (Fig. 4B) (p < 0.05). BrdU-labeled columnar cells were located in the upper and lower portion of Barrett's esophagus, while in duodenal mucosa, they were mainly restricted to the isthmic portion (Figs. 3C and D). Sulindac treatment did not change this localization (data not shown).

As shown in Fig. 5, COX-2 expression was significantly increased in rats exposed to gastroduodenal reflux contents for 40 weeks compared with other groups, including the sham surgery group. Moreover, its expression was localized in the lower esophagus, which also showed higher PCNA expression than the stomach and duodenum (Fig. 6). Cyclin D1 expression was increased in rats undergoing EGDA compared with those undergoing the sham surgery and with the sulindac-treatment groups (Fig. 5).