We have shown that in esophageal adenocarcinoma cell lines (OE19/OE33), DCA decreases Notch receptors and Notch downstream proteins, thereby increasing the transcription factor Hath1 (–). The data suggest that DCA-induced Notch signaling inhibition is not via inhibition of γ-secretase, since Notch1C did not decrease in excess of the decrease in Notch1 (data not shown). Although Notch receptor proteins and mRNA were downregulated in both cell lines, there was modest variability in signaling responses between the two, likely due to their relative genetic backgrounds. Uncharacteristic changes are more likely to occur and to be more pronounced in the poorly differentiated OE33 than in the moderately differentiated OE19 cells.
The presence of a gene responsible for creating notches in the wings of fruit flies (Drosophila) was reported as early as 1919.17
Classic Mendelian mutational analysis showed that the gene was involved in cell fate decisions. The Notch gene encodes a 300-kD transmembrane receptor and consists of a relatively small family of cell surface signaling molecules. Mutations of the Notch receptor result in developmental abnormalities in all animals tested.18
Notch signaling has been shown to regulate proliferation and cell fate determination within the intestinal epithelium.19
Interestingly, conditional knockout mice in which the RBP-Jk transcriptional binding component is absent, and/or treatment with γ
-secretase inhibitor, results in goblet cell formation and suppression of enterocyte and epithelial proliferation.19,20
In contrast, constitutive Notch activation depletes the secretory cell lineage and increases proliferation in intestinal epithelium.11
Hath-1 (Atoh1 in Drosophila and Math1in the mouse) deficient mice exhibit the loss of several secretory cell lineages, suggesting that this Notch regulated transcription factor is important in early secretory cell fate decisions.15
Scoville et al nicely review the current knowledge of intestinal stem cells and signaling including Notch.8
Although several studies link inhibition of Notch signaling followed by induction of Hath1 in the intestine to production of goblet cells, to the best of our knowledge this link has not been studied in the esophagus.11,16,20
It is interesting that bile acids, well known to be present in the esophageal lumen of patients with Barrett’s esophagus, inhibit primary signaling pathways whose down regulation is involved in the determination of goblet cells differentiation. Notch receptors and ligands are highly expressed in basal layer of the esophagus and weakly in the more superficial epithelial layers. Acting in concert with other signaling cascades, they likely control the equilibrium between progenitors and mature esophageal cells.20
Disruption of this equilibrium by components of refluxed gastric juice including bile acids may lead to the development of an epithelium containing goblet cells characteristic of Barrett’s esophagus.
Notch inhibition either by knockdown of its signaling components or by γ
-secretase inhibition has been shown to induce regression of adenomas to post-mitotic goblet cells.22
In addition, increase in Hath1 following inhibition of Notch signaling reduces the prevalence of cancer in experimental colon cancer models.22,23
These data suggest the theoretical possibility that inhibition of Notch signaling and induction of Hath1 in patients with dysplastic Barrett’s esophagus may reverse the neoplastic process to a less aggressive goblet cell phenotype. The use of biologic agents affecting cell signaling is a well established clinical paradigm in 2009.
The transcription factor CDX2 is considered to be a key component responsible for the metaplastic intestinal phenotype of BE.26,27
It is absent from the normal squamous esophageal epithelium, but has been shown to be highly expressed in human Barrett’s esophagus, and is readily induced in experimental animal models of Barrett’s epithelium.28,29
As we and others have shown in other esophageal cell lines, DCA also induced an increase in CDX2 mRNA in OE19 cells (3-fold).24,25,30
As such, we investigated the relationship between inhibition of Notch signaling and the induction of CDX2. Inhibition of Notch signaling by a γ
-secretase inhibitor, induced Hath1 and CDX2 mRNA and protein, indicating that CDX2 induction might occur downstream of Notch signaling under these conditions ( and the scheme in , left
). Taken together, the observations that DCA also inhibits Notch signaling and the repeated observations that bile acid stimulation increases CDX2 expression, indirectly suggest that Notch signaling may be, in part or wholly, responsible for bile acid induced CDX2 changes. Further studies with greater mechanistic detail are necessary to prove causality.
Fig 8 A simplified scheme of the correlation between Notch signaling components and CDX2 amounts determined in the study with and without inhibition by DCA and DAPT. Left, γ-secretase inhibitor DAPT decreases the amounts of Notch1C, inhibits Hrt/Hes, (more ...)
The available knowledge of Notch signaling and its relationship to CDX2 expression in the intestine suggests that it is a highly complex process that is dependent on the cell and tissue context as well as on manipulations of the cells in health and disease. Several studies, where Notch signaling components have been knocked down and goblet cells induced, report no change in CDX2 expression, although most of these do not contain a stimulus such as bile acids.31,32
A negative feedback of CDX2 on Notch signaling has been shown. CDX2 over-expression can induce Notch signaling inhibition.33
A few words of caution are warranted: Experiments in which CDX2 is knocked down or over-expressed in OE19 cells may shed more light on the mechanistic details. Experiments should be conducted also in acidic pH and a mixture of bile acids to mimic the refluxate environment; and importantly, although neoplastic cell lines are commonly used as an in-vitro model to discern cellular mechanisms, the ultimate in-vitro model is “normal” primary esophageal cells.
In conclusion, bile acid exposure results in inhibition of Notch signaling in human esophageal adenocarcinoma cells and is accompanied by an increase in the Notch downstream transcription factor Hath1 and CDX2, factors known to be partially or wholly responsible for the intestinal phenotype (). These data suggest that Notch signaling and its alteration by esophageal luminal content may be one of the key processes contributing to the formation of the goblet enriched IM characterizing Barrett’s esophagus.