Selective cyclooxygenase-2 blocker delays healing of esophageal ulcers in rats and inhibits ulceration-triggered c-Met/hepatocyte growth factor receptor induction and extracellular signal-regulated kinase 2 activation

Abstract

Nonsteroidal anti-inflammatory drugs, both nonselective and cyclooxygenase-2 (COX-2) selective, delay gastric ulcer healing. Whether they affect esophageal ulcer healing remains unexplored. We studied the effects of the COX-2 selective inhibitor, celecoxib, on esophageal ulcer healing as well as on the cellular and molecular events involved in the healing process. Esophageal ulcers were induced in rats by focal application of acetic acid. Rats with esophageal ulcers were treated intragastrically with either celecoxib (10 mg/kg, once daily) or vehicle for 2 or 4 days. Esophageal ulceration triggered increases in: esophageal epithelial cell proliferation; expression of COX-2 (but not COX-1); hepatocyte growth factor (HGF) and its receptor, c-Met; and activation of extracellular signal-regulated kinase 2 (ERK2). Treatment with celecoxib significantly delayed esophageal ulcer healing and suppressed ulceration-triggered increases in esophageal epithelial cell proliferation, c-Met mRNA and protein expression, and ERK2 activity. In an ex vivo organ-culture system, exogenous HGF significantly increased ERK2 phosphorylation levels in esophageal mucosa. A structural analog of celecoxib, SC-236, completely prevented this effect. These findings indicate that celecoxib delays esophageal ulcer healing by reducing ulceration-induced esophageal epithelial cell proliferation. These actions are associated with, and likely mediated by, down-regulation of the HGF/c-Met-ERK2 signaling pathway.

Figure 1.

COX-1 and COX-2 mRNA expression in normal esophageal tissue from sham-operated (SO) rats and ulcerated esophageal tissue (UL) 3 days after sham operation or ulcer induction. A: RT-PCR products obtained with use of specific primers for COX-1 (561 bp) and COX-2 (253 bp) and β-actin. B: Quantitative data for COX-1 and COX-2 mRNA expression. Data were obtained by computerized analysis of the optical intensity of amplified PCR products; each signal was normalized against corresponding β-actin signal and the results are expressed as COX-1/β-actin and COX-2/β-actin ratio. NS, not significant. Values are means ± SD. For each column, n = 6.

Figure 2.

COX-1 and COX-2 protein expression in normal esophageal tissue from sham-operated (SO) rats and ulcerated esophageal tissue (UL) 3 and 7 days after sham operation or ulcer induction. A: Immunoblotting with specific antibodies detected the specific 70-kd band for COX-1 and the specific 72-kd band for COX-2. B: Quantitative data for COX-1 and COX-2 protein expression. Data were obtained by a computerized video analysis of the Western blots. NS, not significant; *, P < 0.05 versus UL at 3 days. Values are expressed as optical intensity units and represent means ± SD. For each column, n = 6.

Figure 3.

Photomicrographs of the esophageal epithelium showing immunofluorescence staining for COX-2 (3 days after sham operation or ulcer induction). A: Epithelium of sham-operated rats, a faint COX-2 signal is detected in the cells of the basal zone. B: Epithelium of the ulcer margin, a strong COX-2 signal is present in the cytoplasm of almost all cells constituting ulcer margin. BZ, basal zone; SZ, stratified zone. Scale bars, 50 μm.

Figure 4.

Photomicrographs of esophageal epithelium showing immunofluorescence staining for c-Met (3 days after sham operation or ulcer induction). A: In the normal esophageal epithelium of sham-operated rats, c-Met staining is present mainly in membranes of cells in basal and stratified zones (green fluorescence signal). B: In the epithelium of the ulcer margin, cells in the stratified zone display strong membrane and cytoplasmic staining, whereas, basal cells show mainly diffuse cytoplasmic staining. BZ, basal zone; SZ, stratified zone. Scale bars, 50 μm.

Figure 5.

Effects of celecoxib on esophageal ulcer healing. Ulcer area was measured 3, 5, and 7 days after ulcer induction. Starting from 3 days after ulcer induction, rats were treated with either celecoxib (CEL) (10 mg/kg, once a day) or its vehicle (VHC) for 2 or 4 days. Ulcer area was measured by a computerized video analysis of the ulcers. NS, not significant; *, P < 0.05 versus 3 days. Values are means ± SD. For each column, n = 6.

Figure 6.

Photomicrographs of esophageal epithelium showing immunostaining for PCNA (7 days after ulcer induction). Starting from 3 days after ulcer induction, rats were treated intragastrically with either celecoxib (10 mg/kg, once a day) or its vehicle for 4 days. A: Epithelium distant from the ulcer in rats treated with vehicle. B: Epithelium distant from the ulcer in rats treated with celecoxib. C: Epithelium of the ulcer margin in rats treated with vehicle. D: Epithelium of the ulcer margin in rats treated with celecoxib. PCNA staining is present in the cell nuclei (dark color). E, epithelium; MM, muscularis mucosae; SM, submucosa. Scale bars, 100 μm.

Figure 7.

Quantitative data for PCNA expression in normal esophageal epithelium of sham-operated (SO) rats and rats with esophageal ulcers (UL). Starting from 3 days after ulcer induction, rats with esophageal ulcers were treated intragastrically with either celecoxib (CEL) (10 mg/kg, once a day) or its vehicle (VHC) for 2 or 4 days. The results are expressed as the percentage of increase in the number of PCNA-labeled cells in the epithelium of the ulcer margin (epithelium of the lower esophageal segment in SO rats) over that in the epithelium distant from the ulcer (epithelium of the upper esophageal segment in SO rats). *, P < 0.01 versus UL at 3 days. Values are means ± SD. For each column, n = 6.

Figure 8.

c-MET mRNA expression in normal esophageal tissue from sham-operated (SO) rats and ulcerated tissue (UL) 3, 5, and 7 days after sham operation or ulcer induction. Starting from 3 days after ulcer induction, rats were treated with either celecoxib (CEL) (10 mg/kg, once a day) or its vehicle (VHC) for 2 or 4 days. A: RT-PCR products obtained with specific primers for rat c-MET (320 bp) and rat β-actin. B: Quantitative data for c-MET mRNA expression. Data were obtained by computerized analysis of the optical intensity of amplified PCR products; each signal was normalized against corresponding β-actin signal and the results are expressed as c-MET/β-actin ratio. *, P < 0.05 versus UL at 3 days. Values are means ± SD. For each column, n = 6.

Figure 9.

HGF and c-Met protein expression in normal esophageal tissue from sham-operated (SO) rats and ulcerated tissue (UL) 3, 5, and 7 days after sham operation or ulcer induction. Starting from 3 days after ulcer induction, rats were treated with either celecoxib (CEL) (10 mg/kg, once a day) or its vehicle (VHC) for 2 or 4 days. A: Immunoblotting with anti-HGF-α antibody detected the specific 69-kd band for α chain of HGF protein and immunoblotting with anti-c-MET antibody detected a 140-kd band for β-subunit of c-MET protein. B: Quantitative data for HGF protein expression. C: Quantitative data for c-MET protein expression. Data were obtained by a computerized video analysis of the Western blots. NS, not significant; *, P < 0.05 versus UL at 3 days. Values are expressed as optical intensity units and represent means ± SD. For each column, n = 6.

Figure 10.

Phosphorylated ERK (pERK), total ERK2 protein expression, and ERK2 activity in normal esophageal tissue from sham-operated (SO) rats and ulcerated tissue (UL) 3, 5, and 7 days after sham operation or ulcer induction. Starting from 3 days after ulcer induction, rats were treated with either celecoxib (CEL) (10 mg/kg, once a day) or its vehicle (VHC) for 2 or 4 days. A: Immunoblotting with anti-phosphorylated Tyr 204 ERK antibody detected the specific 44-kd band for pERK1 protein and the specific 42-kd band for pERK2 protein. Stripping and reprobing with anti-ERK2 antibody showed total ERK2 protein levels. Bottom: Myelin basic protein (MBP) phosphorylated by ERK2 immunoprecipitated from esophageal tissue samples. B: Quantitative data for pERK2 levels. Data were obtained by a computerized video analysis of the Western blots. The pERK2 protein levels are expressed as a percentage of total ERK2 protein levels. C: ERK2 activity determined by measuring levels of radiolabeled [γ-³²P]ATP incorporated to MBP by ERK2 (pmol/mg). Values represent means ± SD. For each column, n = 6.

Figure 11.

Phosphorylated ERK (pERK) and ERK2 protein expression in esophageal mucosal explants pretreated with either 10 μmol/L of SC-236 (SC 10), 100 μmol/L of SC-236 (SC 100), or dimethyl sulfoxide (DMSO) for 6 hours and treated with either HGF (100 ng/ml) or its vehicle (PBS) for 30 minutes. A: Immunoblotting with anti-phosphorylated Tyr 204 ERK antibody detected the specific 44-kd band for pERK1 protein and the specific 42-kd band for pERK2 protein. Stripping and reprobing with anti-ERK2 antibody showed total ERK2 protein levels. B: Quantitative data for pERK2 levels. Data were obtained by a computerized video analysis of the Western blots. The pERK2 protein levels are expressed as a percentage of total ERK2 protein levels. NS, not significant; *, P < 0.05 versus DMSO + PBS group. Values represent means ± SD. For each column, n = 6.