Are Barrett's mucosa and the neoplastic colonic mucosa evidence of a diffuse gastrointestinal inflammatory reaction?
A. Keshavarzian, SJ. Sontag (Hines)
We previously reported in a series of 65 patients with well-documented Barrett's esophagus that benign colon adenomas were present in about 30% of the patients and malignant colonic tumors were present in 15% of the patients. After adjusting for several types of statistical biases and deleting patients from the analysis who had already had colonic surgery before the diagnosis of Barrett's was made, we found that 5.5% of the remaining 55 patients had malignant colonic tumors and 29% had benign adenomas [1]. We had no appropriate control group. Since the true prevalence rate of colon cancer and benign colonic adenomas in the general population is not known, we could conclude only that a 5.5% prevalence rate of colon cancer appeared to be very high for the general population.
Since that publication, four other studies in patients with Barrett's esophagus have been published [2-5]. One of the studies [3] reported a 9% prevalence rate of colon cancer (three of 32 patients) and a 25% prevalence rate of benign adenomas (eight of 32 patients). In the remaining three studies [2,4,5], no cancers were found in a total of 97 patients.
First assumption
Without knowing for certain whether the presence of Barrett's esophagus truly correlates with the presence of benign and malignant colonic neoplasms, let us assume that Barrett's metaplastic mucosa and the colonic neoplastic mucosa respond similarly to several stimuli.
The first question asked is, "why would Barrett's esophagus and colonic neoplasia occur together?" One possible explanation is that certain dietary factors that affect
bile composition may be associated in some way with both conditions. Elements in the bile have been implicated as colonic mucosal tumor promoters. The same elements that affect the colon may also be involved in mucosal changes in the esophagus, especially since bile is often present in the esophagus in individuals with gastroesophageal reflux (GER). In other words, a tumor promoter or carcinogen in the bile could potentially affect any segment of the gut in which it makes contact. Individuals with GER therefore would likely be exposed to biliary elements that are travelling up to the distal esophagus as well as down to the colon. One might argue that prolonged reflux of acid promotes the gastric junctional type of columnar epithelium in the esophagus, while reflux of bile promotes the specialized intestinal type. A reasonable sequence of events might begin with "acid" reflux causing gastric-type Barrett's and progress to "acid and bile" reflux, causing goblet cell metaplasia with specialized columnar-type Barrett's. The scenario could end with subsequent malignant transformation as a result of the tumor-promoting elements in the bile.
Possible mechanisms
It is well known that acid is the most important factor in the development of esophagitis. The exact mechanism by which acid causes the inflammatory reaction and subsequent mucosal damage is not known. Reactive oxygen metabolites (ROM), produced by epithelial cells, phagocytic cells and neutrophils, can cause tissue injury both in peptic disorders such as gastric and duodenal ulcer and inflammatory disorders such as ulcerative colitis and Crohn's disease [6-8]. Recent studies have documented the production of ROM by the inflamed esophageal mucosa as well as by Barrett's mucosa, with or without inflammation [9], indicating that the presence of inflammation is not necessary for ROM production. Indeed, the uninvolved rectal mucosa of patients with colonic neoplasia produces ROM [10].
In the last decade, ROM as a factor in human diseases has gained considerable attention [11-13]. In rats and in humans, studies show that ROM are an important factor in increased colonic cell proliferation [14-17]. Since colon cancer is the final phase of the increased colonic cell proliferation sequence (normal mucosa -. proliferative mucosa --> adenomatous polyp --> carcinoma), the cancer may result from ROM control of mutagenesis and carcinogenesis [18]. In this regard, the entire gastrointestinal tract, especially the esophagus and colon, may be under the powerful influence of the dreadful "secret and invisible" ROM.
Several studies have shown that the technique of chemiluminescence is an accurate method to estimate ROM production [19,20]. The technique measures light production as a cellular by-product of oxidative metabolism and has been utilized to demonstrate the involvement of ROM in various diseases [14]. We used chemiluminescence to estimate ROM levels in different areas of the gut during health and disease [9,10,21,22]. The analysis consisted of noninflamed and inflamed esophageal squamous mucosa, noninflamed and inflamed Barrett's epithelium, Barrett's epithelium with and without low grade/high grade dysplasia, Barrett's epithelium with adenocarcinoma, gastric mucosa, normal colonic mucosa, uninvolved and involved
Figure 1. .Production of reactive oxygen metabolites (ROM) measured by chemiluminescence.
rectal and colonic mucosa harboring benign and malignant neoplasms, postsurgical colonic mucosa and rectal mucosa of inflammatory bowel disease and Crohn's colitis.
Figure 1 demonstrates that ROM is present in both normal and diseased tissues. The first column has three bars: the first bar shows that the rectum of normal healthy controls with no polyps or cancer has a level of ROM of about 800 counts/min/mg protein; the second bar shows that the uninvolved rectal mucosa of patients who had had previous polypectomies for benign adenomas produces twice as many ROM; the third bar shows that the uninvolved rectal mucosa of patients with previous cancer produces almost 3 times as many ROM, despite the fact that the benign adenomas and the malignant cancers had been removed. These data suggest that the rectal mucosa of patients who have had previous neoplasia continues to produce significantly higher levels of ROM than normal healthy mucosa.
The second column with two bars shows the very high levels of ROM produced by both the pericancer tissue and the malignant tissue itself. As would be expected, the uninvolved mucosa surrounding the cancer tissue produces similar quantities of ROM as the uninvolved rectal mucosa of patients with previous cancer (column 1, bar 3).
The last column has four bars demonstrating the ROM production of different disease states of the esophageal mucosa. The mucosa of patients with esophagitis, with Barrett's and no esophagitis and with Barrett's with esophagitis produces up to 4 times as many ROM as the normal squamous mucosa of healthy control patients. The high production of ROM seen in the diseased esophageal mucosa is present with or without inflammation.
We have tried to show that the esophageal and colonic mucosa produce ROM
regardless of the presence of tumor and regardless of the presence of inflammatory activity. From these data it appears that the gut mucosa (at least the esophagus and colon) respond in similar fashion to certain stimuli in the production of inflammatory mediators such as ROM.
References
6. Babbs CFC. Free radicals and the etiology of colon cancer. Free Radic Biol Med 1990;8:191-200