Esophageal cancer is the eighth most common cancer worldwide, with 456,000 new cases and 400,000 related deaths in 2012 [1]. The majority of esophageal cancers are classified into two main histological subtypes: esophageal adenocarcinoma (EAC) and esophageal squamous cell carcinoma (ESCC). In recent decades, the incidence of EAC has increased among the white population of high-income countries. This increase is thought to be due to the rising prevalence of obesity. EAC typically arises from the metaplastic columnar epithelium, called Barrett’s esophagus (BE), in the lower third of the esophagus. EAC and BE develop as a result of long-standing gastroesophageal reflux disease (GERD). Since bile acids are contained in gastric juices due to duodenogastric reflux, the esophagus of patients with GERD is exposed to a mixture of acid and bile acids. Therefore, the cytotoxic effect of bile acids can play a role in the development of BE and EAC. In fact, Menges et al. conducted simultaneous 24-h esophageal pH and bile reflux testing and found that acid and bile exposure were more extensive in 23 patients with BE than in 20 patients with esophagitis [2]. Nehra et al. investigated the spectrum of bile acids by using 15-h continuous esophageal aspiration with simultaneous pH monitoring [3]. The predominant bile acids detected were cholic acid (CA), taurocholic acid, and glycocholic acid, but there was a significantly greater proportion of secondary bile acids, deoxycholic and taurodeoxycholic acids (DCA and TDCA, respectively), in patients with BE.

In contrast, the association between ESCC and bile acid reflux is controversial. ESCC arises in squamous epithelial cells and usually occurs in the upper and middle third of the esophagus. Smoking and heavy alcohol consumption are the main factors that increase the risk of ESCC. The incidence of ESCC is declining in developed countries probably due to a decline in tobacco smoking; however, this incidence remains common in Africa and eastern Asia. These characteristics of ESCC imply the difference in underlying etiologies of ESCC and EAC. However, from another perspective, tobacco smoking and heavy alcohol consumption increase the risk of erosive esophagitis and BE, especially in Asians [4]. More interestingly, in several rat duodenal and gastroduodenal contents reflux models, both ESCC and EAC can develop without stimulus from any known carcinogens [5]. Therefore, we can not deny the possibility that bile acids might play a role in the development of ESCC.

In the following sections, we provide an overview of recent studies related to EAC and bile acids.

Since esophageal reflux of gastric and duodenal contents is facilitated after esophagectomy or gastrectomy, several researchers assessed the relationship between the resection of the esophagus or stomach and development of BE or EAC. Avidan et al. reported that Billroth-1 gastrectomy, Billroth-2 gastrectomy, vagotomy, and pyloroplasty were not associated with BE [6]. However, O’Riordan et al. showed that BE occurs frequently after lower esophagectomy [7]. Among 48 patients with a median follow-up period of 26 months (range = 12–67 months) postesophagectomy, 24 (50%) developed columnar metaplasia, and of these, 13 had specialized intestinal metaplasia. The prevalence of specialized intestinal metaplasia increased over time possibly due to chronic acid and bile exposure. In addition, Tsiouris et al. reported that concomitant fundoplication with resection of the gastroesophageal junction had some protective effect against the development of BE [8]. Thus, the association is still controversial, but careful long-term observation should be recommended for patients after the resection of the lower esophageal sphincter.

There are two distinct hypotheses on the mechanisms of metaplastic conversion of esophageal squamous epithelium to BE. One possibility is the conversion of differentiated squamous cells or stem cells of the squamous epithelium in the basal cell layer. Alternatively, cells at the gastroesophageal junction or transitional zone may colonize the distal esophagus in response to noxious luminal contents.

There are extensive in vitro evidences that suggest the transdifferentiation of squamous cell lineage to BE. CDX2, a member of the caudal-related homeobox gene family, may play a major role in the development of BE. Since CDX2 regulates intestinal cell differentiation, stomach-specific transgenic overexpression of CDX2 induced intestinal metaplasia in mice stomach [9]. CDX2 is not expressed in squamous epithelial cells in normal human esophagus; however, it is aberrantly expressed in BE [10]. Therefore, several researchers investigated CDX2 expression in esophageal squamous cell cultures during exposure to acid and bile and found a high expression of CDX2 in response to bile acids via the activation of nuclear factor-κB (NF-κB) [11–13]. Kong et al. generated transgenic mice that expressed the CDX2 transgene in esophageal squamous tissues [14]. They found that ectopic CDX2 transgene expression in esophageal squamous cells reduced basal epithelial cell proliferation and barrier function and altered cell morphology in vivo.

On the other hand, recently, two studies on animals models strongly suggested that BE arises from a gastric cardia lineage of BE-like metaplasia. Wang et al. showed that p63-deficient mice rapidly developed intestine-like metaplasia with gene expression profiles similar to BE [15]. Using this model, they reported that Krt7-positive epithelial cells at the squamocolumnar junction are the origin of BE. They concluded that BE does not develop because of genetic alterations but because of competitive interactions between cell lineages driven by opportunity. Quante et al. observed the novel BE-model mice (L2-IL-1β mice) in which human IL-1β was overexpressed in the esophagus [16]. They also showed that the migration of gastric cardia progenitor cells, including leucine-rich repeat-containing G protein-coupled receptor 5-positive cells, may cause BE development. In addition, oral administration of DCA accelerated intestinal metaplasia and dysplasia in this model. However, as mentioned in the previous section, BE and EAC developed even after the resection of esophagogastric junction or total gastrectomy. Thus, if progenitors of BE exist only in the esophagogastric junction or gastric cardia, the development of BE after the surgery cannot be explained.

The other possibilities include multipotent stem cells in the submucosal glands of the esophagus or migrated cells from the bone marrow. However, we need further investigations to conclude the origins of BE.

Various in vitro investigations showed that bile acids enhance cell proliferation and confer resistance to apoptosis in BE and EAC cells [17–20]. Tselepis et al. reported that c-myc expression was upregulated in BE and EAC cells during exposure to chenodeoxycholic acid (CDCA) or DCA [21]. The production of reactive oxygen species (ROS) can explain some of these phenotypic changes during the exposure to bile acids. ROS also cause oxidative DNA damages, leading to carcinogenesis [22, 23]. In addition, ROS are known to activate NF-κB and enhance downstream signals, such as epidermal growth factor receptor (EGFR), IL-1β, IL-8, and cyclooxygenase (COX)-2 [22, 24–26]. COX-2 was also reported to be regulated by cAMP response-binding protein (CREB) and activation protein-1 (AP-1) through ROS-mediated activation of PI3K/AKT (phosphatidylinositol 3-kinase/protein kinase B) signaling pathway and extracellular signal-related kinases (ERK) 1/2 during exposure to CDCA or DCA [27].

Since ROS production in esophageal epithelial cells was inhibited by diphenyleneiodonium chloride (an NADPH oxidase [NOX] inhibitor) or N(G)-monomethyl-l-arginine (a nitric oxide synthase [NOS] inhibitor), NOX and NOS play roles in ROS production during exposure to bile acids [28]. McAdam et al. reported that DCA induces inducible NOS (iNOS) expression and produces nitric oxide (NO) in esophageal epithelial cells [29]. Hong et al. reported that TDCA increased NOX5-S expression and hydrogen peroxide (H₂O₂) production in EAC cells [17]. They also revealed that TDCA-induced increase in NOX5-S expression might depend on sequential activation of phosphoinositide phospholipase Cγ2 (PI-PLCγ2) and mitogen-activated protein kinases (MAPK)/ERK signaling cascade.

Two types of bile acid receptors have been identified in BE: the cell membrane receptor Takeda G protein-coupled receptor 5 (TGR5) and the nuclear receptor farnesoid X receptor (FXR). Hong et al. observed that TDCA activates TGR5, leading to the upregulation of NOX5-S expression in BE and EAC cells [19]. On the other hand, we observed that oncogenic microRNA-221/222 was upregulated through the activation of FXR during exposure to CA or CDCA in BE and EAC cells. The target of these microRNAs, p27Kip1, was downregulated, and proteasomal degradation of CDX2 was enhanced by the activation of FXR [30]. Thus, both FXR and TGR5 could play roles in the progression from BE to EAC. In addition, the expressions of bile acid transporters, such as the apical sodium-dependent bile acid transporter (ASBT), ileal bile acid-binding protein (IBABP), and multidrug-resistant protein 3 (MRP3), are increased in BE [31], although these functions are not distinguished enough.

There is a remarkable association between abdominal obesity and GERD, including BE and EAC. Classically, abdominal obesity was thought to increase intragastric pressure and gastroesophageal pressure gradient. However, Anggiansah et al. observed that esophageal mechanical function was not associated with increased reflux in obese individuals [32]. Alternatively, recent observations have suggested that other mechanisms, including the release of humoral mediators from visceral adipose tissue, may provide a better explanation for the association between obesity and EAC [33].

One of the promising possibilities is the alteration of the bile acid composition related to lifestyle or obesity. Chen et al. investigated the bile acid composition of the bile juice in rats [34] and noted that high dietary animal fat increased the concentration of taurine conjugates in the bile juice. In addition, they carried out esophagojejunostomy for reflux of the duodenal contents and compared sequential morphological changes between rats fed with low soybean-oil diet and those with high cow-fat diet for up to 30 weeks after surgery. The animals with reflux in the high cow-fat group had a significantly higher incidence of BE and Barrett’s dysplasia than those in the low soybean-oil group, and the incidence of EAC in the high cow-fat group was also slightly higher than that in the low soybean-oil group. Since bile acid composition is quite different in humans and rodents, more human studies are warranted.

Peng et al. investigated whether ursodeoxycholic acid (UDCA) protected against DCA-induced injury in patients and in vitro [23]. They took biopsies of BE from 21 patients before and after esophageal perfusion with DCA at baseline and after 8 weeks of oral UDCA treatment. Baseline esophageal perfusion with DCA significantly increased the levels of phospho-H2AX and phospho-p65 in Barrett’s metaplasia, whereas oral UDCA increased the levels of glutathione peroxides 1 (GPX1) and catalase in Barrett’s metaplasia and prevented DCA perfusion from inducing DNA damage and NF-κB activation. At the cellular level, DCA-induced DNA damage and NF-κB activation were prevented by 24-h pretreatment with UDCA, but not by a combination of UDCA with DCA. UDCA activated nuclear factor erythroid 2-related factor 2 signaling that increased GPX1 and catalase expression, and protective effects of UDCA pretreatment were blocked by siRNA knockdown of these antioxidants. Rizvi et al. investigated the efficacy of the combination of UDCA and aspirin [35]. They showed that UDCA-aspirin combination reduced the risk of adenocarcinoma in animals with reflux, decreased the proliferation of esophageal adenocarcinoma cells, and downregulated a key cell cycle regulator, cyclin-dependent kinase 2 (CDK2). In addition, they noted that GLI1, a hedgehog-regulated transcription factor, was upregulated during esophageal carcinogenesis, and GLI1 could bind to the CDK2 promoter and activate its expression. The UDCA-aspirin combination could downregulate GLI1. Thus, the chemopreventive effect of UDCA is worthy of being verified by an observational study or a clinical trial.