The severity of duodeno-esophageal reflux influences the development of different histological types of esophageal cancer in a rat model


  • Disclosure Statement: The authors declare that no financial or other conflicts of interest exist in relation to the content of the article.


The mechanism through which each histological type of carcinoma arises from the esophageal mucosa remains unknown. This study was designed to investigate whether there is an association between the severity of duodeno-esophageal reflux and the histological type of esophageal cancer. A series of 120 male Fischer rats, weighing ∼180 g, were randomized to receive one of the following procedures: duodeno-forestomach reflux (DFR) with reduced exposure to duodenal contents, duodeno-esophageal reflux (DER) with increased exposure to duodenal contents and three control operations (DFR, DER control and sham). The reflux of bile was estimated with 99mTc-PMT scintigraphy. All animals were fed a standard diet without carcinogen. The esophageal mucosa was assessed 50 weeks after surgery for carcinoma. The median scanned fraction rate of duodeno-esophageal reflux was significantly lower for the rodents in the DFR group than those in the DER group. Five of 28 rodents in the DFR group and 17 of the 22 rodents in the DER group developed esophageal carcinoma. None of the controls developed carcinoma. The five rodents in the DFR group developed SCC. Of 22 esophageal carcinomas for the DER group, nine were SCC, 12 ADC and one was adenosquamous carcinoma. The fraction of esophageal SCC for the DFR group was significantly higher than that for the DER group, while the fraction of esophageal ADC for the DFR group was significantly lower than that for the DER group. These observations suggest that the severity of duodeno-esophageal reflux in rodents is related to the development of different histological types of esophageal carcinoma.

Esophageal squamous cell carcinoma (ESCC) and esophageal adenocarcinoma (EADC) are the two main histological types of esophageal cancer.1 Population based studies in the Western world have shown steady and rapid rise in the incidence of EADC.1, 2 In contrast, the incidence of ESCC, the most common histological type of cancer in the Eastern world, remains stable.3 The cause of the increased incidence of esophageal cancer remains obscure.

A rise in EADC may be related to an increase in the incidence of gastroesophageal reflux disease (GERD) over the past few decades. Population studies have demonstrated a strong and causal relation between GERD and risk of developing EADC.4 Data from a large multicenter US study identified four major risk factors for EADC: GERD, obesity [as measured by body mass index (BMI)], cigarette smoking and a diet low in fruits and vegetables.4, 5 The progression of GERD to EADC involves the initial development of inflammation induced hyperplasia and metaplasia at the gastro-esophageal junction. This is followed by multifocal dysplasia and ADC.6

The relationship between ESCC and GERD has not been investigated or reported. Prior studies support the hypothesis that GERD plays a role in laryngopharyngeal carcinogenesis.7, 8 A population-based cohort study of patients who underwent gastrectomy for peptic ulcer disease showed a long-term increased risk of laryngeal and pharyngeal cancer.9 Hypopharyngeal multichannel intraluminal impedance-pH was used to characterize laryngopharyngeal reflux (LPR).10 The LPR, especially bile acid reflux, has a role in the tumorigenesis of the upper aerodigestive tract.11, 12 More than 95% of head and neck malignancies are SCC; therefore, ESCC may also be related to GERD.

Cigarette smoking, alcohol consumption and a low fruit and vegetable diet are associated with ESCC.5 Cigarette smoking and alcohol consumption are considered major causative factors in the development of ESCC.1, 13, 14 Alcohol15 and smoking16 reduce lower esophageal sphincter pressure and increase the risk of strain-induced reflux. Smoking may also contribute to the trend through an early stage carcinogenic effect.17–19 Alcohol users, tobacco smokers and those with the genetic polymorphisms of aldehyde dehydrogenase type (ALDH) 2-2 allele and multiple lugol-voiding lesions of the esophageal mucosa, have an increased risk of superficial SCC within the head and neck region.20 Obesity is associated with an increase in intra abdominal pressure, which also increases the likelihood of GERD. Most studies have shown an inverse relationship between BMI and ESCC risk. General and abdominal obesity was strongly related to a higher risk of EADC4, 21, 22 and ESCC.23, 24

The role of duodenal juice in promoting carcinogenesis of the digestive tract has been demonstrated both clinically and experimentally.25–28 Results of several studies have shown that gastric acid and duodenal content reflux play an important role in the pathogenesis of esophageal mucosal injury and subsequent columnar re-epithelization of the distal esophagus.29–31 Individuals with a history of gastrectomy often suffer from severe alkaline reflux esophagitis.32 A history of gastrectomy is associated with an increased rate of lower third ESCC.33, 34 The long term complications include a greater incidence of Barrett's metaplasia and esophageal carcinoma.32–35

We have established a rodent duodeno-esophageal reflux model that produces three histologically different esophageal cancer types: EADC, ESCC and adenosquamous carcinoma (EASC).27, 28 It remains unclear how each type of carcinoma arises from the esophageal mucosa. Our previous study showed that EADC arose from dysplastic changes to the columnar-lined epithelium at the anastomosis in the lower esophagus. However, ESCC accompanied squamous proliferative hyperplasia or dysplasia and developed proximal to the columnar-lined epithelium.28

This study was designed to investigate whether the severity of duodeno-esophageal reflux is associated with a specific histological type of esophageal cancer.

Materials and Methods

The Institutional Animal Care and Use Committee of the Graduate School of Medical Science, Kanazawa University approved this study.

Experimental animals

One hundred twenty Fisher male rats, weighing ∼180 g, were used in the study. The animals were housed three per cage and maintained at a constant room temperature of 22 ± 3°C and 55 ± 5% humidity with a 12-hr light–dark cycle. They were fed standard solid chow CRF-1 (Charles River, Japan) and tap water free of carcinogens.

Surgical procedures

After a 24-hr fast, an upper abdominal incision was made under diethyl-ether inhalation anesthesia. The surgical procedures illustrated in Figure 1 were performed on each rat.

Figure 1.

Diagram of Surgical Procedure. The surgical procedures completed to establish varying degrees of duodeno-esophageal reflux and matched controls. DER induced increased amounts of duodenal reflux, while DFR was protective. D, Duodenum; E, Esophagus; FS, Forestomach; GS, Glandular stomach; JE, Jejunum.

Duodeno-forestomach reflux (n = 38)

The glandular stomach was removed and the duodenal stump was closed with sutures. The forestomach was then anastomosed end-to-side with jejunum at ∼4 cm distal to Treitz's ligament. This procedure allowed the duodenal contents to flow back directly into the forestomach and esophagus through the anastomosis.

Nonduodeno-forestomach reflux (n = 22)

Glandular stomach was removed and the duodenal stump was closed with sutures. Jejunum was cut and suture-closed at 4 cm from the Treitz's ligament. Forestomach was then anastomosed end-to-side at the suture-closed end of the jejunum. In addition, the proximal jejunum was anastomosed side-to-side to the jejunum at ∼10 cm distal to the forestomach-jejunal anastomosis. This procedure prevented duodenal reflux into the forestomach and esophagus.

Duodeno-esophageal reflux (n = 30)

Both the glandular and the forestomach were removed (total gastrectomy) and the duodenal stump was closed with sutures. The esophageal end was anastomosed end-to-side to the jejunum at ∼4 cm distal to Treitz's ligament. This procedure allowed the duodenal contents to flow back directly into the esophagus through the stoma.

Nonduodeno-esophageal reflux (n = 16)

Total gastrectomy was performed and the duodenal stump was ligated. Jejunum was cut ∼4 cm distal to Treitz's ligament and the distal end was ligated. The esophageal stump was anastomosed end-to-side with the distal jejunum near its distal ligated end. The proximal jejunal end was anastomosed side-to-side to the jejunum at ∼10 cm distal from the esophago-jejunal anastomosis. This procedure prevented duodenal reflux into the esophagus.

Sham operation (n = 14)

Animals underwent a laparotomy with blunt manipulation of the abdominal viscera.

In all the surgical procedures, intestinal anastomosis was carried out with interrupted sutures through all layers of the gut using a 7-0 atraumatic silk-braided suture. After surgery, the animals were allowed to drink water immediately while fasting was continued for 24 hr following the procedure. The rats were weighed every 5 weeks throughout the experiment. All the animals were killed 50 weeks after surgery.

Specimen extraction

The animals were sacrificed by diethyl-ether inhalation, after which the abdomen was opened. A ligature was placed around the afferent and efferent jejunal loop near the forestomach-jejunal and esophago-jejunal anastomosis. The esophagus was ligated at the level of the thyroid cartilage through a thoracotomy. The esophagus and the anastomosed jejunum were then removed.

Pathological assessment

Excised organs were washed with 10% formalin, spread and pinned on a cork plate with the mucosal side up. After fixing the organs with 10% formalin solution for at least 24 hr, the esophagus was cut along the longitudinal axis into slices at 3 mm intervals and embedded in paraffin. Five-micrometer thick sections of each embedded paraffin block were prepared for histological analysis with hematoxylin and eosin (H&E) stain.

Definition of pathological findings

Histological findings in the esophagus were classified into the following four categories:

  • 1Proliferative squamous hyperplasia: Proliferative hyperplasia is a condition characterized by a thickened epithelium to twice that of a normal epithelium with acanthosis, elongation of the papillae and parakeratosis. Other features include the thickening of the basal layer of the squamous epithelium and the preservation of a stratified appearance.
  • 2Squamous dysplasia: Here, the epithelium is composed of dysplastic squamous cells with large and polymorphic nuclei with deeply stained chromatin and increased number of mitotic figures. Squamous dysplasia may involve lamina propria of the epithelium but dose not invade the submucosal layer.
  • 3Barrett's metaplasia: Replacement of esophageal squamous epithelium with columnar-lined epithelium, comprised of gastric and/or intestinal cells.
  • 4Carcinoma: Carcinoma is defined as cellular and structural atypism with epithelial invasion into the submucosal layer. ADC consists of dysplastic glandular cell growth with both atypia and invasiveness and has two types of histology: tubular or papillary ADC and mucinous ADC. SCC is a type of squamous cell dysplasia with marked cellular and structural atypism. This carcinoma may be divided into, well-differentiated and poorly differentiated types according to the presence or absence of cancer pearls, respectively. ASC has components of both ADC and SCC.

Hepatobiliary scintigraphy

Five animals in each group received an intravenous injection of 37MBq of [99mTc]N-pyridoxyl-5-methyltryptophan (99mTc-PMT: Japan Mediphics, Japan) under ether anesthesia, for serial hepatobiliary scanning in the supine position using a gamma camera. After radioisotope injection, pictures were taken every 2.5 min, for 120 min.

Statistical analysis

Differences were analyzed for significance using Student's t test, Mann-Whitney Rank Sum test, Fisher Exact test or Log-Rank test as appropriate. Data management and statistical analysis were performed using Stat View software (SAS, Berkeley, CA). Data values were considered significant when the p value was <0.05.


General observations

Ninety-one of 120 animals survived after surgery and were used in our present study. The effective number of animals was 28 in the duodeno-forestomach reflux (DFR) group, 16 in the non-duodeno-esophageal reflux (non-DFR) group, 22 in the duodeno-esophageal reflux (DER) group, 12 in the non-duodeno-esophageal reflux (non-DER) group and 13 in the Sham operation (SO) group. The mortality rate was 24% (29 of 120). The causes of death were malnutrition (n = 11), ileus (n = 10), peritonitis (n = 4) and unknown causes (n = 4). At the end of the experimental period, 91 animals were evaluated. The median body weights at 50 weeks in the SO group was significantly greater than that in the DFR, non-DFR, DER and non-DER groups. In contrast, the median body weight did not differ among the survivors between DFR, non-DFR, DER and non-DER groups.

Hepatobiliary Scintigraphy

Among the five animals tested in the DFR group, 99mTc-PMT was detectable in the forestomach after 10 min and into the lower esophagus after 15 min. Reflux of 99mTc-PMT in DFR group was observed only in the lower esophagus but not in middle or upper esophagus. Among the five animals in the DER group, 99mTc-PMT began to flow into the lower esophagus after 10 min and reached the upper esophagus at 20 min. In both the DER and DFR groups, the maximum reflux occurred after 20 min and went down to undetectable levels after 90 min.

A region of interest (ROI) drawing over the entire esophagus was used to quantify the severity of reflux. The DER group achieved a greater maximum reflux (Median: 16.2%) when compared with the DFR group (Median: 6.2%; Fig. 2). No esophageal reflux of the 99mTc-PMT was observed at 120 min in the non-DFR group, non-DER group and SO group.

Figure 2.

Hepatobiliary Scintigraphy was used to quantify the severity of reflux with each surgical model. The severity of reflux was quantified by drawing the ROI over the entire esophagus. The maximum reflux in the DER group (Median: 16.2%) was significantly greater as compared to the maximum reflux in the DFR group (Median: 6.2%).

Macroscopic findings

The esophagus and forestomach in the DFR group was contracted with thickening of the wall. The lower esophagus was dilated. The epithelium of the lower esophagus and forestomach revealed tortuous longitudinal folds and sporadic erosions. Obvious tumor was visible near the anastomosis (Fig. 3a). The esophagus in the DER group rodents were wide and thickened and the epithelial surface contained longitudinal folds extending along the lower two third of the esophagi (Fig. 3b). These changes were also seen in the forestomachs of the DFR group rodents. These findings are consistent with severe esophagitis. In contrast, the esophagi of animals in the non-DFR (Fig. 3c) and non-DER (Fig. 3d) group did not display any of these findings. The surface was whitish, smooth and glistening. The esophagi had a sharp horizontal demarcation line between the esophageal or forestomach and jejunal epithelia.

Figure 3.

Representative macroscopic findings from each surgical group and Schematic diagram. (a) The esophagus and forestomach in the DFR group was contracted, with thickening of the wall and the lower esophagus was dilated. The epithelium of lower esophagus and forestomach showed tortuous longitudinal folds and sporadic erosions. Obvious tumor is visible near the anastomosis. (b) The esophagus in the DER group was wide and thickened and the epithelial surface contained longitudinal folds extending along the lower two third of the esophagus as forestomach in the DFR group. The single arrow is obvious tumor and double arrow is visible near the anastomosis. (c) (non-DFR); (d) (non-DER): The forestomach and esophagus of a rat from the control group exhibited a slightly uneven surface, but display no tumor. (e,f) Schematic diagram indicating the relationship between each type of carcinoma and background epithelial patterns in rats with duodeno-esophageal reflux. Red triangle is the esophagus.

Microscopic findings

Five of 28 animals for the DFR group and 17 of the 22 animals for the DER had developed esophageal carcinoma. No carcinoma was found in any of the controls. All of five rats for the DFR group displayed SCC (Fig. 4a). Of 22 esophageal carcinomas for the DER group, 9 were SCC, 12 ADC (Fig. 4b) and 1 ASC. The overall incidence of esophageal carcinoma was significantly greater for DER group than the DFR group (p < 0.05; Table 1).

Figure 4.

A photomicrograph of the esophagus subjected to surgically induced reflux showing: (a) ESCC with squamous pearls; (b) EADC; (c): Proliferative squamous hyperplasia (PSH); (d) Barrett's metaplasia (BM).

Table 1. Incidence of carcinoma of esophagus and histological findings
inline image

The fraction of ESCC in the DFR group was significantly higher than the DER group (p < 0.05), while the incidence of EADC was significantly lower for the DFR group than the DER group (p < 0.05; Table 1). However, in DFR group, of 24 the forestomach carcinomas, 8 were SCC, 15 ADC and 1 was ASC. There was no significant difference in the proportions of forestomach SCC, ADC and ASC for the DFR group when compared with esophageal SCC, ADC and ASC for the DER group (Table 2). Proliferative squamous hyperplasia (Fig. 4c), squamous dysplasia and Barrett's metaplasia (Fig. 4d) were more frequently found in the DER group rodents than DFR group rodents (p < 0.05). As expected, ADC arose from the columnar-lined epithelium near the anastomosis, while SCC arose from the squamous esophagitis (Figs. 3e and 3f).

Table 2. Incidence of carcinoma of forestomach in the DFR and esophagus in the DER
inline image


This study suggests that the severity of duodenal reflux in rodents is related to the development of different histological types of esophageal carcinoma. In brief, high exposure to duodenal contents promotes the development of EADC and low exposure induces ESCC.

The importance of duodenal reflux in the pathogenesis of esophageal mucosal injury is established.25–28 The histological spectrum of esophageal cancer is divided into SCC, ADC and ASC.27, 28 In our previuos study, ESCC was shown to arise from dysplastic changes to squamous cells naturally found in the esophagus. EADC arose in the areas of columnar-lined epithelium adjacent to the surgical anastomosis.28 Pera et al. investigated the histogenesis of EASC and proposed that “chronic duodenal reflux induced the development of metaplastic cells with glandular differentiation from the stem cells of squamous epithelium and that glandular metaplastic foci are the morphological element from which tumors with a dual differentiation arise.”30 The mechanism through which each histological type of carcinoma arises from the esophageal mucosa remains unknown. We proposed to use a surgical rat model to induce varying amounts of duodenal reflux to investigate whether the severity of duodenal reflux is associated with a specific histological type of esophageal cancer.

The rats were divided into five different groups and underwent operations that were intended to increase the reflux of duodeno-esophageal contents (DER and DFR), to produce no significant reflux at all with the Roux-Y bypass or sham surgery. Hepatobiliary scintigraphy was used to confirm increased duodeno-esophageal reflux in the DER rat model when compared with the DFR rat model. The study found that esophageal reflux high in duodenal contents (DER) resulted in the development of EADC, while esophageal reflux with lower duodenal contents (DFR) resulted in the development of ESCC. This suggests that the amount of duodeno-esophageal reflux has an important effect on the histogenesis of EADC and ESCC.

ESCC develops in the upper 2/3 of the esophagus and is associated with tobacco and alcohol use.36 The cancer follows chronological progression from proliferative squamous hyperplasia to squamous dysplasia and ultimately ESCC. This study illustrates the development of ESCC with reflux alone. The DER and DFR groups both displayed an increased incidence of proliferative squamous hyperplasia. Each group developed a significant proportion of ESCC, while none of the controls progressed to carcinoma. These results suggest that duodeno-esophageal reflux may promote the development of ESCC. This association has already been established in laryngopharyngeal cancer.7, 8

Barrett's metaplasia is the known pre-cursor to the development of EADC and commonly develops at the esophagogastric junction. The DER group revealed a 100% incidence of Barrett's metaplasia, while the DFR revealed no evidence of Barrett's metaplasia. Thus, increased exposure to duodeno-esophageal reflux is responsible for the development of Barrett's metaplasia and subsequent progression to EADC. The lack of a sufficient amount of reflux in the DFR and control groups prevented the development of Barrett's metaplasia and subsequent progression to EADC. The effects of increased duodenal reflux were also observed in the retained forestomach of DFR rats. The anatomical location of the forestomach in the DFR model predisposes the forestomach to increased duodenal reflux similar to that seen in the DER group. The forestomach tissue developed cancer in similar histological proportions as those seen in the lower esophagus of the DER group. These results suggest that the degree of duodenal reflux exposure plays an important role in carcinogenesis. The specific amount of reflux required to cause these changes requires further investigation.

The most significant question that remains is why do varying amounts of duodenal reflux promote different histological types of carcinogenesis? Nishioka et al found that the proliferation of ESCC showed a biphasic reaction to bile acid exposure. The growth of esophageal cancer cell lines was stimulated at a low concentration (maximally at 20–30 μM), but was suppressed at a higher one. Only a low dose of bile acid induced the expression of cyclin D1 and CDC25A and showed high Rb phosphorylation and high CDK2 kinase activity. In brief, bile acid at a low dose stimulates the proliferation of ESCC by inducing G1-regulating molecules such as cyclin D1, CDC25A and Rb. This biphasic reaction may reflect the clinical aspects of the bile acid effect on the squamous epithelium of the esophagus, that is, carcinogenesis (low concentration) and inflammation (high concentration).37

McQuaid et al also reported that elevated concentrations of bile acids could stimulate oesophageal squamous epithelial cells and Barrett's epithelial cells to produce cytokines that might promote esophageal inflammation (e.g. IL-8 and COX-2). Additionally, bile acids may cause oxidative stress and DNA damage to those same cells. Bile acids induce esophageal squamous cells to change their gene expression patterns to resemble those of intestinal cells and can cause Barrett's epithelial cells to increase their expression of intestinal-type genes.38 The sodium choleate (CA), sodium deoxycholate (DCA) or a 1:1 mixture (MIX) were the most toxic to esophageal squamous mucosal cells.39

Bonde et al. used a rat model to confirm that gastroduodenal reflux leads to increased DNA damage and down regulation of the DNA mismatch repair pathway in ESCC.40 Although exposure to risk factors is important for esophageal carcinogenesis, genetic polymorphisms also play a role. Genetic variability and patterns of mutation seem to be more diverse in ESCC than in EADC and this variability can be regional in its distribution.41 The molecular mutation most commonly associated with cancer occurs in the p53 tumor suppressor gene TP53, but this mutation is only one of many possible aberrations that can lead to esophageal carcinogenesis. For example, the prevalence of a TP53 mutation in patients with ESCC is high in Europe and Asia.42

Souza et al. summarized the molecular pathways behind the development of Barrett's metaplasia. Acid and bile exposure are thought to activate the Cdx promoter gene. This results in a transformation of the esophagus to Barrett's metaplasia.43

A reasonable extension of this hypothesis is that a minimum amount of acid and bile reflux is required to induce the expression of Cdx gene. Thus, those rodents with increased duodeno-esophaeal reflux developed ADC and increased rates of carcinoma compared with rodents without a significant amount of duodeno-esophageal reflux. It is also possible that a similar dose phenomenon could effect proteolytic enzymes such as trypsin and others from the pancreas and it could be those enzymes that are interacting with the esophageal mucosa accounting for the differences between the two groups.44 The presence of ESCC in all noncontrol groups may be explained by localized inflammatory changes through an alternative mechanism. These findings will require further investigation.

A number of events occur in patients who have EADC or ESCC, at both a mechanical and a cellular level, as the process of carcinogenesis unfolds.45, 46 Although ESCC and EADC differ in their histology and epidemiologic distribution, many of their risk factors and mechanisms of carcinogenesis are the same.47

In this study, we sacrificed the experimental rats at 50 weeks after surgery. Our previous sequential study suggested that the incidence of EADC is 45 and 44% in the 40 and 50 week after surgery using the DER model. In contrast, the incidence of ESCC is 9 and 13% in the 40 and 50 week after surgery.28 There were no differences in the time to development of adeno and squamous carcinogenesis.

We used Fischer 344 male rats, since they have a lower incidence of spontaneous solid malignant tumor than Wistar or Sprague Dawley (SD) male rats. Tatematsu et al reported that male SD (Crj:CD), WKY (WKY/NCrj), Lewis (LEW/Crj), Wistar (Crj:Wistar) and Fischer 344 (F344/DuCrj) rats (40 per strain), were given drinking water containing 100 μg/ml MNNG for 30 weeks and then normal tap water. All rats were sacrificed at weeks 10, 30 and 50 of the experiment. ADCs of the glandular stomach were found in nine of 15 SD rats (60%), 8 of 12 WKY rats (67%), 8 of 15 Lewis rats (53%), 3 of 13 Wistar rats (23%) and 1 of 18 F344 rats (6%) at week 50. This incidence of carcinomas in SD, WKY and Lewis rodents were significantly higher (p < 0.01) than that of F344 rats.48 In this study, the incidence of esophageal carcinoma is similar when compared with our prior studies using Wistar or SD rats with duodeno-esophageal reflux.28, 49 Therefore, different rodents reproduce similar results with our surgical reflux models.

This study establishes the role of duodenal contents alone without other carcinogens in the development of esophageal cancer. The results suggest that the continued rise of EADC in the Western world may be attributable to an increasing amount and concentration of duodenal reflux. This study also establishes an association between low levels of duodeno-esophageal reflux and ESCC. Obesity is known to raise intra-abdominal pressure and promote gastro-duodenal reflux in humans.21, 22 The increased prevalence of obesity is the most likely driving force behind of increased rates of ESCC and EADC. Further human studies will be required to establish a causal relationship between duodeno-esophageal reflux and the development of different histological type of esophageal cancer.