Carcinogenesis vol.21 no.8 pp.1587–1591, 2000
Duodenal-content reflux esophagitis induces the development of glandular metaplasia and adenosquamous carcinoma in rats
Miguel Pera, Maria J.Brito1, Manuel Pera3, Richard Poulsom2, Emilio Riera, Luis Grande, Andrew Hanby2 and Nicholas A.Wright2 Service of Gastrointestinal Surgery, IMD, Hospital Clinic, IDIBAPS, University of Barcelona Medical School, Villarroel 170, 08036 Barcelona, Spain, 1Department of Histopathology, Hospital Garcia de Orta, Almada, Portugal and 2Histopathology Unit, Imperial Cancer Research Fund, London, UK 3To
whom correspondence should be addressed Email:
[email protected]
Recent studies have demonstrated that refluxed duodenal contents cause esophageal carcinoma in rats without exposure to carcinogens. The histopathological spectrum of these carcinomas includes squamous-cell carcinoma, adenocarcinoma and adenosquamous carcinoma. Pure adenocarcinomas are thought to arise in areas of columnar metaplasia adjacent to the anastomosis, similar to Barrett’s esophagus in humans. In contrast, the histogenesis of adenosquamous carcinomas is unclear. The purpose here was to investigate the pathogenesis of esophageal adenosquamous carcinomas in a time-course experiment of chronic duodenal-content reflux without carcinogen. Forty-two 8-week-old male Sprague–Dawley rats were divided into seven groups and exposed to duodenal-content esophageal reflux during 10, 15, 20, 25, 30, 35 and 40 weeks, respectively. All animals underwent an esophagojejunostomy with gastric preservation in order to produce chronic esophagitis. The rats received a standard diet without addition of carcinogens. An increasing incidence of glandular metaplasia and carcinoma was observed over the time course, starting at 20 weeks. After 40 weeks of reflux, multiple foci of glandular metaplasia and adenosquamous carcinoma were found in 83 and 50% of the animals, respectively. Most of the carcinomas occurred in the middle and proximal esophagus and had a dual pattern of differentiation, glandular and squamous. These findings confirm that duodenal content reflux alone has a carcinogenic effect. We propose that chronic duodenal reflux induces 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 pattern of differentiation arise. Introduction The incidence of adenocarcinoma of the esophagus and esophagogastric junction has been increasing during the past decade (1). Most of these adenocarcinomas arise in Barrett’s metaplasia (2), in which columnar epithelium replaces the normal squamous epithelial lining of the lower esophagus, after damage by gastroesophageal reflux. Results of several Abbreviation: 2,6-DMNM, 2,6-dimethylnitrosomorphine. © Oxford University Press
studies have shown that not only gastric acid, but also duodenalcontent reflux plays an important role in the pathogenesis of esophageal mucosal injury and subsequent columnar reepithelization of the distal esophagus (3,4). At this time, it is widely accepted that Barrett’s esophagus is a premalignant lesion and there is increasing evidence supporting a progression from metaplasia to low-grade dysplasia, high-grade dysplasia and, finally, adenocarcinoma (5). Experimental animal models provide good tools for understanding the developmental mechanisms underlying carcinogenesis. Previous studies have shown that duodenal-content reflux, produced by means of an esophagojejunal anastomosis, exerts a co-carcinogenic effect on 2,6-dimethylnitrosomorphine (2,6-DMNM)-induced rat esophageal tumorigenesis (6–8). In addition to the development of squamous cell carcinomas, components of this mixed reflux (duodenal juices) promote histopathological changes culminating in the development of carcinomas with glandular differentiation (adenocarcinomas and adenosquamous carcinomas). These animal models have provided some insights about the role of different components of the duodenal-content reflux on the development of Barrett’s esophagus and esophageal adenocarcinoma in humans. Recent studies have demonstrated that refluxed duodenal contents per se cause esophageal carcinoma in rats without exposure to carcinogens (9,10). The histopathological spectrum of these tumors includes squamous-cell carcinoma, adenocarcinoma and adenosquamous carcinoma. Although adenocarcinomas usually occur in the distal esophagus, adenosquamous tumors are found in the middle and proximal esophagus. Pure adenocarcinomas are thought to arise in areas of columnar metaplasia adjacent to the anastomosis, similar to Barrett’s esophagus in humans (6,9,10). In contrast, the histogenesis of adenosquamous carcinomas is unclear. The progression of Barrett’s esophagus to adenocarcinoma has been the subject of many investigations. However, no experimental studies have focused on the pathogenesis of adenosquamous carcinomas. The purpose of the present study is to investigate the pathogenesis of esophageal adenosquamous carcinomas in a time-course experiment of chronic duodenal-content reflux without carcinogen. Materials and methods Animals Forty-two 8-week-old male Sprague–Dawley rats weighing 200–250 g were obtained from Harlan Interfauna, Iberica S.A. (Barcelona, Spain). The rats were housed two to a cage, kept under standard laboratory conditions [room temperature (22 ⫾ 2°C), 55 ⫾ 5% humidity and a 12 h light–dark cycle] and received a standard pellet diet (Panlab S.L., Barcelona, Spain) and water ad libitum, without addition of carcinogens of any kind. Diet composition included cereals, vegetables, fish, minerals and vitamins A, D and E. Solid food was withdrawn for 24 h before surgery, but rats were allowed free access to tap water. The study was approved by the Institutional Animal Care and Use Committee of the Hospital Clinic Research Foundation, University of Barcelona. All rats received humane care in accordance with the ‘Guide for the Care and Use of Laboratory Animals’ (NIH publication 85–93 revised 1985).
1587
M.Pera et al. Experimental design The animals were randomly divided into seven groups in a time-course design, and exposed to chronic duodenal-content esophageal reflux for 10, 15, 20, 25, 30, 35 and 40 weeks, respectively. All animals underwent an esophagojejunostomy with gastric preservation in order to produce chronic esophagitis. At the end of the appropriate time period for each group, the rats were killed with ether. Surgical model of duodenal-content reflux esophagitis Anesthesia was induced and maintained with an ether–air mixture. According to the Levrat model (11), the gastroesophageal junction was divided, the gastric end oversewn, and the esophagus reimplanted 2 cm beyond the ligament of Treitz in an end-to-side fashion into a loop of jejunum. Anastomosis was performed with 5-0 silk. All animals had free access to both water and food from 12 and 24 h, respectively, after the operation. Histology Immediately after death, the entire esophagus, contiguous anastomotic site and 5 mm of jejunal mucosa were removed and the lumen was longitudinally opened by sectioning through the dorsal aspect of the esophageal wall. With the mucosal surface upward, the margins of the specimen were fixed to a cork plate with pins for macroscopic examination, photographed and fixed in 10% neutral buffered formalin. After 24 h of fixation, the esophagus was divided into three full-thickness segments: proximal, middle and distal. The distal segment included the esophagojejunal anastomosis. These segments were embedded in paraffin, cut into 4 µm sections and stained with hematoxylin and eosin (H&E) and diastase Periodic-acid Schiff/alcian blue for microscopic examination. Histological findings of the squamous epithelium were classified into: (i) reactive changes, defined as the presence of basal cell hyperplasia, hyperkeratosis and papillomatosis; (ii) glandular metaplasia, characterized by mucus-secreting cells surrounded by squamous epithelium; (iii) squamous cell carcinoma; (iv) adenosquamous carcinoma and (v) adenocarcinoma.
Results Effective number of animals The mortality rate was 9.5% (four of 42). Two rats died within the first week after surgery for unknown reasons. Two rats died of respiratory distress before the time scheduled for termination. This was probably due to pulmonary aspiration secondary to reflux disease. At the end of the experiment period, 38 rats were evaluated. Macroscopic findings In all the rats, the middle and lower esophagus was abnormally dilated. The esophageal wall was thickened and the inner surface showed longitudinal folds extending along the entire length of the esophagus as well as whitish nodular patches of 1⫻2 mm, which covered the esophageal surface and gave it a cobblestone appearance (Figure 1). Together with these changes, there were superficial ulcers located mostly in the middle and lower thirds of the esophagus. All these morphological findings were present in all groups, although the changes were more intense in rats exposed to reflux for the longer periods. Macroscopically, no esophageal tumors were observed. Microscopic findings The results of this study are shown in Table I. Reactive changes were seen in all rats, almost always accompanied by extensive ulceration of the mucosa. These findings, characteristic of reflux esophagitis, mostly affected the middle and lower thirds of the esophagus. Columnar-lined epithelium extending ⬍4 mm above the esophagojejunostomy was observed in the distal esophagus in ⬎80% of the rats. Glandular metaplasia (Figure 2A, B and D), was observed from week 10 (in 17% of the cases) increasing over time to week 40 (in 83% of the rats). In some cases (Table I) there were multiple foci of glandular metaplasia. In six rats this glandular metaplasia was found in the middle and proximal thirds of the esophagus, far above the anastomosis. No pure squamous cell carcinomas or 1588
Fig. 1. Macroscopic picture of rat esophagus after 35 weeks of reflux. The arrow points out the site of the esophagojejunostomy.
adenocarcinomas were observed. Adenosquamous carcinomas, characterized by the presence of both glandular and squamous cell patterns of differentiation, were present from week 20 to week 40, with an increasing number of cases with time (Figure 2C, E and F). These carcinomas had no characteristic macroscopic features, being made up of irregular ulcerations in the surface of the mucosa. Most of the adenocarcinomas (eight of 10) were located in the middle and proximal segments of the esophagus. In three cases there were multiple foci of carcinoma in all three thirds of the esophagus. Discussion Our most remarkable finding is the high prevalence of glandular metaplasia and adenosquamous carcinoma increasing over the time. Since the rats did not receive any carcinogen, duodenal content reflux alone was considered to exert a carcinogenic effect and promote the development of these lesions. It has been shown previously that chronic duodenal-content reflux causes columnar metaplasia and adenocarcinoma of the distal esophagus in rats. In 1989, Pera et al. (6) reported that rats with chronic reflux esophagitis had an increased incidence of esophageal carcinomas induced by s.c. injection of 2,6DMNM. Interestingly, 50% of those tumors were adenocarcinomas and adenosquamous carcinomas, whereas control animals given only 2,6-DMNM showed only squamous cell carcinomas, suggesting that duodenal contents induced the glandular differentiation of the tumors. Most recently, it has been demonstrated that long-term duodenal or gastroduodenal reflux per se cause esophageal carcinomas in rats without the use of a carcinogen. Miwa et al. (9) reported 83 and 77% prevalence of carcinoma in rats with exposure to gastroduodenal or duodenal reflux, respectively, extending to 50 weeks. Columnar re-epithelization-associated adenocarcinoma was the most frequent histological type. In a similar study by Goldstein et al. (10), pure adenocarcinomas arising from columnarlined esophagus were observed in 73% of rats with an esophagoduodenal anastomosis after 31 weeks. In another
Induction of glandular metaplasia in rats
Table I. Histological findings Time (weeks)
Reactive changes Glandular metaplasia Adenosquamous carcinoma aMultiple bMultiple
10 (n ⫽ 6)
15 (n ⫽ 5)
20 (n ⫽ 6)
25 (n ⫽ 5)
30 (n ⫽ 5)
35 (n ⫽ 5)
40 (n ⫽ 6)
6 0 0
5 0 0
6 1 2
5 1 1b
5 2a 2
5 3a 2
6 5a 3b
foci of glandular metaplasia in 1, 2 and 3 cases (30, 35 and 40 weeks, respectively). foci of adenosquamous carcinoma in 1 and 2 cases (25 and 40 weeks, respectively).
Fig. 2. (A and B) Near-serial sections stained with H&E and diastase/Alcian blue/PAS, respectively, of a region of the mid esophagus exhibiting glandular metaplasia after 25 weeks of reflux. Part of (B) is shown at higher magnification in (D), in which a gland is seen to express both squamous (right arrow) and columnar mucous-secreting features (left arrow). (C) H&E-stained section of the full thickness of the mid esophagus above the muscularis mucosa, containing adenosquamous carcinoma after 25 weeks of reflux (arrow). (E and F) Near-serial sections stained with H&E and diastase/Alcian blue/PAS, respectively, of a region of the mid esophagus revealing another carcinoma with mixed squamous/mucinous phenotype after just 20 weeks of reflux. The arrow in (E) indicates the region that in (F) shows Alcian blue/PAS-positive cells (indicating mucous differentiation).
recent experiment (13), junctional adenocarcinomas were described in 48% of animals after only 16 weeks of exposure to reflux of gastroduodenal or duodenal juice without exogenous carcinogen administration. Therefore, esophageal carcinomas
can be induced experimentally after a relatively short period of duodenal-content reflux. We confirmed this in the present investigation, in which just 20 weeks of reflux were enough to induce carcinomas. Overall, the results from the quoted 1589
M.Pera et al.
animal models support the carcinogenic effect of duodenal reflux and the temporal progression from intestinal metaplasia to dysplasia and adenocarcinoma. Adenosquamous carcinomas have been described in several experimental studies of esophageal carcinogenesis (6,7,14) but only Miwa et al. (9) have previously reported squamous cell carcinomas with glandular differentiation without carcinogen. However, that investigation was not designed to study the pathogenesis of this type of tumor. In contrast to pure adenocarcinomas, which are known to arise in Barrett’s esophagus (2), the histogenesis of adenosquamous carcinomas is unclear. Based on experimental studies and clinical reports, several hypotheses have been proposed. It has been speculated that the glandular component of these neoplasms originates in esophageal glands or their ducts, in the light of the existence of subepithelial tumors covered by non-cancerous squamous epithelium (14,15). However, this idea does not explain the origin of the malignant squamous component. Others advocate a field effect wherein both the squamous epithelium and submucosal glands are similarly affected during carcinogenesis (16). The finding of adenosquamous carcinomas in rats with chronic reflux esophagitis does not support these two hypotheses, because there are no esophageal submucosal glands in rats. Another hypothesis suggests that carcinomas with glandular differentiation may arise from initiated stem cells of the basal layer of the squamous epithelium (10). There has been considerable discussion about whether stem cells in the esophageal squamous epithelium have the capacity to give rise to glandular epithelial cells. In the histogenesis of Barrett’s esophagus, Jankowski et al. (17) proposed that such squamous/glandular epithelial transdifferentiation is indeed possible, although Wright (18) was sceptical about this concept and proposed instead that stem cells in the ducts of esophageal glands were the origin of Barrett’s epithelium. Our present findings show unequivocally that the stem cells in the squamous epithelium of the esophagus are able to transdifferentiate to a glandular habitus. The destruction of the epithelial tissue architecture under the chronic effect of duodenal secretions may allow the proliferating stem cell population to become exposed to factors contained in the reflux material. We have shown previously that duodenalcontent reflux esophagitis in rats induces an expansion of the proliferative compartment in the squamous epithelium, with an increasing number of proliferating cells in the suprabasal layer (19). This provides a larger cellular target for subsequent mutations for progression from a benign proliferative lesion to full-blown malignancy. While of course a mere expansion of the suprabasal cell compartment will not provide additional targets per se, since many of these cells will be lost through shedding, it is possible that such expansion also initiates an increase in the stem cell population. A minority of initiated cells will eventually proliferate to form a focus with the potential for double differentiation, mucinous and squamous, from the basal layer of the squamous epithelium. These foci will undergo further genetic changes, which lead to the formation of carcinomas with areas of squamous cell carcinoma and adenosquamous carcinoma. The high incidence of glandular metaplastic foci found here suggests this idea. Glandular metaplastic foci are characterized by mucus-secreting structures surrounded by squamous epithelium occurring in the middle and proximal esophagus. In the experiment by Pera et al. (6), glandular foci were observed in the esophagi of 11 animals with chronic esophagitis lasting 1590
for 32 weeks, and the authors suggested that glandular metaplasia might represent the precursor lesion for adenocarcinoma. Miwa et al. (9) found several patches of ectopic columnar tissue in the inflamed squamous epithelium. Goldstein et al. (10) described islands of columnar metaplasia in the squamous epithelium far above the esophagoduodenal junction in one animal. In our study, glandular metaplasia was first identified at 20 weeks, simultaneously to adenosquamous neoplasms, suggesting a rapid temporal progression from metaplasia to this type of carcinoma. We propose that chronic duodenal reflux may induce 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. In conclusion, the present investigation is the first time-course experiment on chronic duodenal reflux esophagitis. It provides compelling evidence about the carcinogenic effect of duodenal reflux and the role of glandular metaplasia as a premalignant lesion of esophageal adenosquamous carcinoma. Acknowledgements This work was supported by Grant 97/1221 from the Fondo de Investigaciones de la Seguridad Social, Ministry of Health, Spain and the Imperial Cancer Research Fund, London, UK.
References 1. Pera,M., Cameron,A.J., Trastek,V.F., Carpenter,H.A. and Zinsmeister,A.R. (1993) Increasing incidence of adenocarcinoma of the esophagus and esophagogastric junction. Gastroenterology, 104, 510–513. 2. Cameron,A.J., Lomboy,C.T., Pera,M. and Carpenter,H.A. (1995) Adenocarcinoma of the esophagogastric junction and Barrett’s esophagus. Gastroenterology, 109, 1541–1546. 3. Champion,G., Richter,J., Vaezi,M., Singh,S. and Alexander,R. (1994) Duodenogastroesophageal reflux: relationship to pH and importance in Barrett’s esophagus. Gastroenterology, 107, 747–754. 4. Kauer,W., Peters,J., DeMeester,T., Ireland,A., Bremner,C. and Hagen,J. (1995) Mixed reflux of gastric and duodenal juices is more harmful to the esophagus than gastric juice alone. Ann. Surg., 222, 525–533. 5. MacArdle,J., Lewin,K., Randall,G. and Weinstein,W. (1992) Distribution of dysplasias and early invasive carcinoma in Barrett’s esophagus. Hum. Pathol., 23, 479–482. 6. Pera,M., Cardesa,A., Bombı´,J.A., Ernst,H., Pera,C. and Mohr,U. (1989) Influence of esophagojejunostomy on the induction of adenocarcinoma of the distal esophagus in Sprague–Dawley rats by subcutaneous injection of 2,6-dimethylnitrosomorpholine. Cancer Res., 49, 6803–6808. 7. Attwood,S.E.A., Smyrk,T.C., DeMeester,T.R., Mirvish,S.S., Stein,H.J. and Hinder,R.A. (1992) Duodenoesophageal reflux and the development of esophageal adenocarcinoma in rats. Surgery, 111, 503–510. 8. Pera,M., Trastek,V.F., Carpenter,H.A., Fernandez,P.L., Cardesa,A., Mohr,U. and Pairolero,P.C. (1993) Influence of pancreatic and biliary reflux on the development of esophageal carcinoma. Ann. Thorac. Surg., 55, 1386–1393. 9. Miwa,K., Sahara,H., Segawa,M., Kinami,S., Sato,T., Miyazaki,I. and Hattori,T. (1996) Reflux of duodenal or gastro-duodenal contents induces esophageal carcinoma in rats. Int. J. Cancer, 67, 269–274. 10. Goldstein,S.R., Yang,G., Curtis,S.K., Reuhl,K.L., Liu,B.C., Mirvish,S.S., Newmark,H.L. and Yang,C.S. (1997) Development of esophageal metaplasia and adenocarcinoma in a rat surgical model without the use of a carcinogen. Carcinogenesis, 18, 2265–2270. 11. Levrat,M., Lambert,R. and Kirshbaum,G. (1962) Oesophagitis produced by reflux of duodenal contents in rats. Dig. Dis. Sci., 7, 564–573. 12. Clark,G.W.B., Smyrk,T.C., Mirvish,S.S., Anselmino,M., Yamashita,Y., Hinder,R.A., DeMeester,T.R. and Birt,D.F. (1994) Effect of gastroduodenal juice and dietary fat on the development of Barret’s esophagus and esophageal neoplasia: an experimental rat model. Ann. Surg. Oncol., 1, 252–261. 13. Fein,M., Peters,J.H., Chandrasoma,P., Ireland,A.P., Oberg,S., Ritter,M.P., Bremner,C.G., Hagen,J.A. and DeMeester,T.R. (1998) Duodenoesophageal reflux induces esophageal adenocarcinoma without exogenous carcinogen. J. Gastrointest. Surg., 2, 260–268.
Induction of glandular metaplasia in rats 14. Cardesa,A., Bombı´,J.A., Pera,M., Ferna´ ndez,P.L., Campo,E., Pera,C. and Mohr,U. (1993) Spectrum of glandular differentiation in experimental carcinoma of the esophagus induced by 2,6-dimethylnitrosomorpholine under the influence of esophagojejunostomy. Exp. Toxicol. Pathol., 46, 41–49. 15. Bell-Thompson,J., Haggitt,R.C. and Ellis,F.H. (1980) Mucoepidermoid and adenoid cystic carcinomas of the esophagus. J. Thorac. Cardiovasc. Surg., 79, 438–446. 16. Kuwano,H., Ueo,H., Sugimachi,K., Inokuchi,K., Toyoshima,S. and Enjoji,M. (1984) Glandular or mucus secreting components in squamous cell carcinoma of the esophagus. Cancer, 56, 514–518. 17. Jankowski,J.A., Wright,N.A., Meltzer,S.J., Triadafilopoulus,G., Geboes,K., Casson,A.G., Kerr,D. and Young,L.S. (1999) Molecular evolution of the
metaplasia-dysplasia-adenocarcinoma sequence in the esophagus. Am. J. Pathol., 154, 965–973. 18. Wright,N.A. (1996) Migration of the ductular elements of gut-associated glands gives clues to the histogenesis of structures associated with responses to acid hypersecretory state: the origins of ‘gastric metaplasia’ in the duodenum of the specialized mucosa of Barrett’s esophagus and of pseudopyloric metaplasia. Yale J. Biol. Med., 69, 147–153. 19. Pera,M., Grande,L., Gelabert,M., Figueras,X., Pera,M., Palacı´n,A., Elena,M., Cardesa,A., Tiburcio,A.F. and Trastek,V.F. (1998) Epithelial cell hyperproliferation after biliopancreatic reflux into the esophagus of rats. Ann. Thorac. Surg., 65, 779–786. Received February 2, 2000; revised April 11, 2000; accepted April 13, 2000
1591
