Zoning of mucosal phenotype, dysplasia, and telomerase activity measured by telomerase repeat assay protocol in Barrett's esophagus

Abstract

Glandular dysplasia in Barrett's esophagus may regress spontaneously but can also progress to cancer. The human telomerase RNA template and the human telomerase reverse transcriptase enzyme which do not, of themselves, correlate strongly with telomerase activity, are too often overexpressed in Barrett's dysplasia to predict individual cancer risk. This study relates telomerase activity, mucosal phenotype, and dysplasia in Barrett's esophagus. Biopsies (n = 256) from squamous esophagus, columnar-lined esophagus every 2 cm, esophago-gastric junction, gastric body, and antrum from 32 patients with long-segment Barrett's esophagus were evaluated by telomerase repeat assay protocol (TRAP). Three biopsies for histology (n = 794) were simultaneously taken at each anatomical level. These and all prior and subsequent biopsies (n = 1917) were reviewed for mucosal phenotypes and dysplasia severity. Intestinal-type Barrett's mucosa was present at all levels in Barrett's esophagus. At least one Barrett's biopsy was TRAP(+) in 22 of 32 patients. TRAP positivity of intestinal-type Barrett's mucosa increased distally, possibly as a consequence of mucosal exposure to acid or bile reflux. Native gastric mucosa was rarely TRAP(+) (1/31 corpus, 2/32 antrum), whereas native squamous mucosa usually was TRAP(+) (31/32). Dysplasia almost always involved intestinal-type Barrett's mucosa (85/87; P <.00001), without evidence of proximal-distal zoning. TRAP could be positive without dysplasia and negative in extensive, even high-grade, dysplasia. TRAP activity merits evaluation as a candidate biomarker for increased risk of persistent dysplasia and cancer progression in Barrett's esophagus.

Figure 1

Anatomy of Barrett's esophagus and stomach, with biopsy sites. Proximal esophagus is lined by its native squamous epithelium. Distal to the “Z line,” this is replaced by characteristic glandular Barrett's mucosa. Z line is often irregular and there may be residual islands of squamous epithelium below it. Anatomical esophago-gastric (EG) junction is taken as the most proximal extension of the mucosal folds (rugae). Gastric biopsy sites are indicated in the body (proximally) and antrum (distally).

Figure 2

Example of TRAP assay. Lane 1 = molecular weight (MW) markers. In lanes 2 to 4, 5 to 7, and 8 to 10, the first and second lanes represent analysis of 2 or 0.2 µg of protein extract, whereas the third lane is a heat-inactivated control. BM = Barrett's mucosa. Lanes 2 to 4 = Barrett's mucosa at 34 cm. Lanes 5 to 7 = Barrett's mucosa at 35 cm. Lanes 8 to 10 = gastric body mucosa. Lane 11 is a positive control (C+; GLC4 cells) and lane 12 is a negative control (C-). ITAS is the internal PCR control.

Figure 3

Mean probability (expressed as a percentage) for all 32 patients of different mucosal types being identified in Barrett's esophagus as a function of anatomical level, on the vertical axis, from the Z line proximally (top) down to, but not including, the anatomical esophago-gastric junction distally (bottom). The grey shaded area in each box shows how often that particular component is found at that level. The five boxes represent (a) full-thickness squamous epithelium, (b) squamous islands over glandular mucosa, (c) “specialized” intestinal-type Barrett's mucosa, (d) cardiac-type mucosa, and (e) cardio-oxyntic or fundic mucosa.

Figure 4

The left-hand box (a) shows a stacked graph of the number of telomerase-positive biopsies and the total number of telomerase biopsies as a function of anatomical level, on the vertical axis, from the Z line proximally (top) down to, but not including, the anatomical esophago-gastric junction distally (bottom). Data for all 32 patients with long-segment Barrett's esophagus. Horizontal scale is the number of biopsies represented by each centile. The center graph (b) plots the mean percentage of telomerase biopsies, which were TRAP⁺ for all 32 patients with long-segment Barrett's esophagus, as a function of anatomical level on the vertical axis. There is pronounced fluctuation in the percentage of TRAP+ biopsies with anatomical level. The right-hand box (c) plots the average probability of dysplasia being present in “specialized” intestinal Barrett's mucosa as a function of anatomical level. Data are averaged for the 10 patients with definite dysplasia and seven patients indefinite for dysplasia. There is no evidence of any significant proximal-to-distal gradient in dysplasia frequency.