In earlier reports, we proposed that differences (heritable or acquired) among individuals in the signal transduction pathways activated in esophageal squamous epithelium by gastroesophageal reflux might determine the development of Barrett’s esophagus (21
). In those studies, we found differences between esophageal squamous cell lines from patients with and without Barrett’s esophagus in how acid activates the MEK-ERK pathways that regulate cellular proliferation and differentiation (22
). In the present study, we have shown that acid and bile salts induce CDX2 mRNA and protein expression in esophageal squamous cells from GERD patients with Barrett’s esophagus, but not from GERD patients without Barrett’s esophagus. This demonstration that esophageal squamous cells from patients with and without Barrett’s esophagus differ in acid- and bile salt-induced expression of CDX2, a key homeotic gene for intestinal development, suggests that such differences might underlie the development of Barrett’s metaplasia.
For these studies, we used telomerase-immortalized, non-neoplastic esophageal squamous cell lines established from GERD patients with and without Barrett’s esophagus. These cell lines maintain morphologic features and cytokeratin expression patterns typical of normal esophageal squamous cells and, unlike transformed cells, demonstrate cell contact inhibition and anchorage-dependent cell growth. For studies on how reflux contributes to the development of benign Barrett’s metaplasia, we feel that using such cell lines is more appropriate than using transformed cells, which have numerous uncharacterized genetic abnormalities that might alter their molecular responses to stimuli. Those responses also might be affected by infection with simian virus 40 large T antigen (SV40Tag), which can cause DNA promoter methylation abnormalities. SV40Tag has been used to immortalize the HET-1A esophageal squamous cell line. Although it is conceivable that the forced expression of telomerase also might alter cellular responses, our group and others have reported that telomerase allows for the immortalization of normal, non-tumorigenic, epithelial cells without the introduction of major genetic alterations, and that these cells at least maintain normal cell cycle checkpoint responses (22
None of our telomerase-immortalized esophageal squamous cell lines expressed CDX2 mRNA or protein at baseline. After exposure to acid, bile salts or both for 7 days, CDX2 mRNA and protein were expressed in the lines derived from GERD patients with Barrett’s esophagus (NES-B3T and NES-B10T), but not in those from GERD patients without Barrett’s esophagus (NES-G2T and NES-G4T). Moreover, in the NES-B3T and NES-B10T lines, we found that exposure to acid, bile salts, or both increased mRNA expression of CDX2 target genes (CK20 and MUC-2), suggesting that the induced CDX2 is functional. We also studied CDX2 mRNA expression in esophageal squamous mucosal biopsy specimens taken during endoscopic examinations in GERD patients. In support of our in vitro
findings, we found CDX2 mRNA expression in 7 of 10 squamous biopsy specimens from patients with Barrett’s esophagus, whereas only 1 of 10 such specimens from patients who had GERD without Barrett’s esophagus expressed CDX2 mRNA. Moons et al
. also found expression of CDX2 mRNA in 6 of 19 esophageal squamous mucosal biopsy specimens taken from patients with Barrett’s esophagus, but did not study CDX2 expression in GERD patients without Barrett’s esophagus (34
Others have reported that acid and bile salts increase CDX2 mRNA expression in HET-1A cells and in certain cancer cell lines by demethylation of the promoter (14
). In contrast, we did not find that acid and bile salts changed the methylation status of the CDX2 promoter in any of our telomerase-immortalized cell lines, even though those agents clearly increased CDX2 mRNA levels in NES-B3T and NES-B10T cells. To confirm that those findings were not due to a technical problem, we treated the cells with the demethylating agent DAC. Although DAC demethylated the CDX2 promoter in all 4 cell lines, there was no associated increase in mRNA expression, suggesting that promoter demethylation alone is not sufficient to induce CDX2. The reasons underlying the disparity between our findings and those of earlier investigators are not clear, but may be related to the presence of the SV40Tag in HET-1A cells, and to the genetic abnormalities of cancer cells.
We transfected our cells with a human CDX2 promoter-reporter construct attached to luciferase to explore the effects of acid and bile salts on promoter activity. We found that those agents increased CDX2 promoter activity in NES-B3T and NES-B10 T cells, but not in NES-G2T or NES-G4T cells. Since this human CDX2 promoter construct contains an NF-κB binding site, we investigated whether acid and bile salts increase CDX2 transcriptional activity via NF-κB activation. We inhibited NF-κB activation using three different approaches, and found that all three abolished the transcriptional activation of CDX2 by acid and bile salts in NES-B3T and NES-B10T cells. These data suggest that reflux-induced transcriptional regulation of CDX2 is NF-κB-dependent in esophageal squamous cells, as it is in human colorectal carcinoma cells and adenocarcinoma cells of the gastroesophageal junction (29
The NF-κB subunits p65 and p50 have been found to exert opposing effects on Cdx2 transcriptional activity, with p65 inhibiting and p50 stimulating that activity (29
). In adenocarcinoma cells of the gastroesophageal junction, deoxycholic acid has been found to increase activity of the Cdx2 promoter by inducing nuclear translocation and promoter binding of only the p50 subunit (30
). We found that acid and bile salts induced nuclear translocation of both the p65 and p50 subunits in NES-B3T and NES-B10T cells, but only the p50 subunit bound to the CDX2 promoter. These findings suggest that, in esophageal squamous cells from patients with Barrett’s esophagus, acid and bile salts increase CDX2 mRNA expression through transcriptional activation via the nuclear translocation and binding of the p50 subunit of NF-κB.
In summary, we have shown that acid and bile salts increase expression of the homeotic gene CDX2 in esophageal squamous cells from patients with Barrett’s esophagus, but not in those from GERD patients without Barrett’s esophagus. We have also documented in vivo that expression of CDX2 mRNA occurs more frequently in the esophageal squamous epithelium of patients with Barrett’s esophagus. The acid- and bile- induced increase in CDX2 transcription in squamous cells from Barrett’s patients is dependent on the activation of NF-κB. Although both the p50 and p65 subunits of NF-κB undergo nuclear translocation after treatment with acid and bile, only the stimulatory p50 subunit binds the CDX2 promoter. These findings support our hypothesis that differences among individuals in the molecular pathways activated when their esophageal squamous epithelium is exposed to gastroesophageal reflux determine whether reflux-induced damage heals through squamous regeneration or through the development of Barrett’s esophagus. It remains unclear whether those differences are heritable or acquired.