Genesis of Barrett Esophagus: Has a Histologic Transition From Gastroesophageal Reflux Disease-Damaged Epithelium to Columnar Metaplasia Ever Been Seen in Humans?, The

Archives of Pathology & Laboratory Medicine, Feb 2005 by Riddell, Robert H

Has a histologic transition from gastroesophageal reflux disease-damaged epithelium to columnar metaplasia ever been seen in humans? The answer to this question seems to be that it has but that we either do not readily recognize it or it is not readily recognizable with regular light microscopy. There are at least 3 possible mechanisms for the genesis of Barrett esophagus. The first is ulceration at the gastroesophageal junction with subsequent repair by an epithelium that differentiates into Barrett epithelium. The second is metaplasia through multilayered epithelium. The third is creeping columnar metaplasia at the Z-line proximally followed by intestinalization. These 3 hypotheses may not be mutually exclusive, and all may be operative, depending on the local circumstances, amount of inflammation, erosion, ulcers, healing, acid and alkaline reflux, and use of proton pump inhibitors. Any of the epithelial types involved could be stable and not progress. They might even be reversible, which may also in part explain the mosaic of epithelial types that typify Barrett esophagus, and may be modified by any of the molecular mechanisms that turn protein transcription on and off (eg, promoter methylation, mutations). These mechanisms ultimately may also be involved in the genesis of neoplastic transformation.

(Arch Pathol Lab Med. 2005;129:164-169)

Has a histologic transition from gastroesophageal reflux disease (GERD)-damaged epithelium to columnar metaplasia ever been seen in humans? This question in one sense asks whether one can ever observe a junction between squamous mucosa exhibiting features of GERD and glandular mucosa showing features of Barrett esophagus (BE), which is seen regularly. However, the real question is whether BE has ever been seen to "form before our very eyes" and if so what this formation looks like morphologically. In other words, has an actual morphological sequence of events been observed that starts as squamous mucosa (presumably at some point GERD damaged) and finishes as BE?

If a reductionist approach is taken, one can list all of the features that can be seen in GERD-damaged mucosa in one column and all of those known in columnar metaplasia in another column (Table 1). Assuming there are no other mucosal types about which we are unaware, the transition must be present among these features. Any of the features that can be seen in the squamous mucosa in GERD patients can obviously be teamed up with similar changes in columnar mucosa, including cardiac mucosa. Although we assume that this list includes goblet cells as part of either complete or incomplete intestinal metaplasia, this feature is not always found, even in patients with overt BE.

There are several epithelial types that probably play no role in any transition, including in squamous mucosa the pseudoepitheliomatous hyperplasia that can occur when ulcers and granulation tissue become re-epithelialized by squamous mucosa (Figure 1), balloon cells1 that are a sequel of reflux-associated damage, pancreatic metaplasia at the cardia and in BE,2,3 and dysplasia and carcinoma that are well-known late complications of BE in glandular mucosa. There has long been a suspected relationship between GERD and esophageal squamous cell carcinoma,4-10 which is hardly surprising given the increased epithelial turnover associated with GERD in the squamous mucosa11 that particularly extends into the larynx and pharynx12-14 but is irrelevant to this discussion.

In Table 1, some pathologic changes are listed in both columns, which suggests that these changes might be involved in one (there may be more than one) of the potential mechanisms of the metaplastic process. This process begins in squamous mucosa but results in glandular mucosa that ultimately includes goblet cells, that is, it has also become intestinalized. Potential mechanisms include (1) ulceration at the gastroesophageal junction (the squamocolumnar junction [SCJ] or Z-line), with subsequent restitution by an epithelium that differentiates into Barrett epithelium, (2) metaplasia through multilayered epithelium, (3) creeping columnar metaplasia at the Z-line proximally followed by intestinalization, and (4) re-epithelialization following therapy, resulting in overgrowth of squamous epithelium over glandular/ Barrett epithelium or regrowth of glandular mucosa from esophageal glands/ ducts or from the cardia.

ULCERATION FOLLOWED BY DEVELOPMENT OF BARRETT EPITHELIUM

GERD-associated ulceration at the SCJ is an indication of attempts by the squamous mucosa to increase proliferation so it can keep up with the damage caused to that epithelium. Increased epithelial turnover is reflected in the classical criteria (described by Ismail-Beigi and colleagues15,16) of basal cell hyperplasia and papillary elongation, which are the morphological counterparts of increased turnover.11 Following ulceration at the SCJ, re-epithelialization can occur from the adjacent squamous mucosa but should not result in BE (at least immediately). However, if re-epithelialization occurs over an erosion, it can result in quite marked pseudoepitheliomatous hyperplasia with pseudoinvasion of the underlying granulation tissue (Figure 1, C and D). Repair by glandular mucosa can come from either the gastric cardiac mucosa directly (Figure 2) or, if the ulceration is in the squamous-lined esophagus, from the esophageal glands or the esophageal gland ducts. These ducts open directly into the lumen of the esophagus and therefore are available to initiate a regenerative process of both squamous mucosa from duct orifices lined by squamous mucosa or of glandular mucosa from ducts lined by columnar epithelium. Repair could also come from Z-line mucosa with subsquamous glands of cardiac type, which occasionally are found beneath squamous mucosa at the Z-line and likely represent a dynamic interplay at the Z-line by which it may shift cranially or caudally (up or down). This type of repair is reminiscent of the changes seen in BE that has regressed except that it is limited to a few glands (up to approximately 8) and by definition is nonintestinalized. Otherwise, this repair would be indistinguishable from regressed short-segment BE or intestinal metaplasia that occurred at the Z-line in cardiac mucosa followed by a minimal regrowth/overgrowth of squamous mucosa, presumably resulting from a decrease in gastroesophageal reflux. However, if regeneration comes from glandular mucosa, it still requires goblet cell metaplasia within the columnar mucosa to result in typical Barrett-type epithelium.