Mutagen Sensitivity and Neoplastic Progression in Patients with Barrett's Esophagus: A Prospective Analysis

The incidence of esophageal adenocarcinoma has risen rapidly over the past 30 years in the United States and Western Europe (1-3). Most cases seem to arise in Barrett's esophagus, a metaplastic epithelium that develops in response to chronic gastroesophageal reflux disease (4-7). Although persons with Barrett's esophagus are at an elevated risk for progressing to esophageal adenocarcinoma, estimated at 0.5% to 1.0% per year (8-10), the vast majority of persons with Barrett's esophagus will not develop esophageal adenocarcinoma within their lifetimes. Thus, discrimination between persons at high risk of progression, who would benefit from more intensive prevention and surveillance programs, and those at relatively low risk, for whom lower-cost alternatives might be appropriate, is of critical importance. A variety of environmental and host factors are thought to play a role in the etiology of esophageal adenocarcinoma, including acid reflux (4, 5, 11-13), gender, race, obesity (14-18), Helicobacter pylori colonization, and cigarette smoking (14, 19, 20). The mechanisms of action of these factors are likely to directly or indirectly involve DNA damage. Increased levels of DNA damage have been detected in Barrett's mucosa (21) and may be associated with progression (22). Impaired ability to repair such damage may therefore play a role in progression to esophageal adenocarcinoma, as has been suggested in a study of polymorphisms in DNA repair genes in persons with Barrett's esophagus (23).

The bleomycin mutagen sensitivity assay is an indirect measure of an individual's constitutive ability to repair DNA damage (refs. 24, 25; reviewed in refs. 26, 27). The assay measures the number of unrepaired bleomycin-induced double-strand chromatid breaks in peripheral blood lymphocytes in vitro and is believed to reflect the equilibrium between mutation rate and DNA repair (26). Lymphocytes from individuals with cancer, and in one study, oral premalignant lesions, have been found to exhibit higher bleomycin sensitivity than healthy controls (28-33). Bleomycin sensitivity may capture both individual susceptibility (27, 34-36) and environmental exposures, such as those to tobacco smoke (37) or oral selenium (36), although most studies have found that sensitivity is not affected by such exposures (38-40). Prospective studies of mutagen sensitivity and cancer risk are more difficult to conduct because the assay requires viable cells. Previous prospective bleomycin sensitivity studies have analyzed cohorts of patients with cancer and used the recurrence of cancer (41-43) or mortality (44) as end points.

Progression in Barrett's esophagus is associated with the inactivation of tumor-suppressor genes, in particular CDKN2A/p16 [by mutation and chromosome 9p loss of heterozygosity (LOH)] and TP53 (by mutation and 17p LOH; refs. 45, 46). Inactivation of these genes could allow cells with DNA damage to progress through the cell cycle, possibly increasing cancer risk for bleomycin-sensitive individuals. The p16 tumor-suppressor gene is lost frequently and early during neoplastic progression in Barrett's esophagus. p16 loss is associated with loss of the wild-type late G₁ arrest (47) and entry into cell cycle. TP53 is a multifunction protein that mediates cell cycle arrest and apoptosis in response to a variety of conditions, including double-strand DNA breaks, hypoxia, and depleted nucleotide pools (48). Likewise, TP53 mutation and 17p LOH occur frequently in esophageal adenocarcinoma and have been shown to predict progression (49-51). Inactivation of the tumor-suppressor pathways contributes to the extensive genetic instability that characterizes the development of esophageal adenocarcinoma (46, 52, 53). This instability can become manifest as DNA content abnormalities, such as aneuploidy or tetraploidy, detectable by flow cytometry. Flow cytometric abnormalities have been shown to be predictors of esophageal adenocarcinoma risk in Barrett's esophagus (53-55).

We therefore conducted a prospective study of persons with Barrett's esophagus to determine whether mutagen sensitivity is associated with the subsequent development of cancer and the intermediate end point of aneuploidy. To determine the effects of bleomycin sensitivity in patients with inactivated tumor-suppressor genes, we compared the rates of progression in the subsets of patients with and without detectable alterations in p16 or TP53 coincident with or before the bleomycin sensitivity test. Our results indicate that bleomycin-sensitive Barrett's esophagus patients are at significantly increased risk for developing aneuploidy and may be at increased risk of developing esophageal adenocarcinoma.

Patients were enrolled in the Seattle Barrett's Esophagus Study, originally approved by the Human Subjects Division of the University of Washington in 1983 and renewed annually thereafter with reciprocity from the Fred Hutchinson Cancer Research Center Institutional Review Board from 1993 to 2001. Since 2001, the study has been approved by the Fred Hutchinson Cancer Research Center Institutional Review Board with reciprocity from the University of Washington Human Subjects Division. Endoscopic biopsies of Barrett's epithelium, acquired at 2-cm intervals in the esophagus according to a standard protocol (53, 55), were evaluated from 220 patients who had Barrett's esophagus without cancer at the time of the endoscopy associated with the blood draw for the bleomycin assay (the “baseline” endoscopy). Biopsies were evaluated by flow cytometry and sorted on the basis of proliferation/DNA content as described previously (56-58). All patients in this cohort received follow-up endoscopies (Table 1). Biopsies obtained at or before baseline from a subset of these patients were evaluated for 9p21 and 17p LOH using polymorphic microsatellite markers, as described previously (58, 59). One hundred eighty-three patients were evaluated for 9p LOH, 181 patients for 17p LOH, 178 for p16 mutation, and 180 for TP53 mutation.

A blood sample (10 mL) was obtained for all patients at the time of the baseline endoscopy. Blood was drawn into sodium heparinized tubes, packed on dry ice, and shipped overnight to the laboratory of Dr. Xifeng Wu in Houston, TX. One milliliter of whole blood was cultured in 9 mL RPMI I640 tissue culture medium (JRM Biosciences, Lenexa, KS) supplemented with 20% fetal bovine serum and 1.25 (volume for volume) phytohemagglutinin (Wellcome Research Laboratories, Research Triangle Park, NC). At 67 hours, bleomycin (Nippon Kayaku Co., Ltd., Tokyo, Japan) was added to each culture to a final concentration of 0.03 units/mL for 5 hours. During the last hour, cells were treated with 0.04 μg/mL Colcemid to induce mitotic arrest. Cells were treated with hypotonic 0.07 mol/L KCl solution for 12 minutes, fixed, washed with freshly prepared Carnoy's fixative (methanol and acetic acid 3:1), and air-dried on wet slides. Prepared slides were coded and stained with Giemsa solution. From stained preparations of each sample, 50 metaphases were examined under oil immersion and breaks were counted and expressed as the average number of breaks per cell. Gaps or attenuated regions were disregarded. Patients with an average of over 0.6 double-strand breaks per cell (the median number in our study) were deemed to be bleomycin sensitive. We also used 0.8 breaks per cell as a threshold for bleomycin sensitivity, as previously defined in ref. 25.

At the time of or before the baseline endoscopy, all 220 subjects underwent structured interviews carried out in person by trained staff as described in refs. (60, 61) to determine the use of tobacco, alcohol, and medications. Anthropometric measurements were taken at the time of this interview and follow-up visits using a standardized protocol.

Kaplan-Meier curves were used to plot the cumulative incidence of aneuploidy and cancer. A proportional hazards model was used to calculate the hazard ratios (HR), 95% confidence intervals (95% CI), and P values. The HRs were adjusted as necessary using patient age (as a continuous variable), gender, nonsteroidal anti-inflammatory drug use (current/former/never), tobacco (ever/never), and waist-to-hip ratio (above/below gender-specific median) using the subset of 219 patients (out of 220) for whom we had information on all of these factors. Statistical analyses were carried out using the R statistical computing language version 2.3.0 (62).

Patients were followed from 3 months to 10.1 years; patient follow-up is summarized in Table 1. Table 2 summarizes the characteristics and bleomycin sensitivity assay results for the 220 Barrett's esophagus patients in the cohort. The mean numbers of bleomycin-induced breaks were slightly but not significantly higher among males, persons over 70 years, those with lower waist-to-hip ratios, and current nonsteroidal anti-inflammatory drug users. The Kaplan-Meier curves describing cumulative incidence of cancer and aneuploidy stratified by bleomycin sensitivity are shown in Fig. 1.

Bleomycin-sensitive (having more than the median of 0.6 breaks per cell) patients had a statistically significantly greater risk of developing aneuploidy (adjusted HR, 3.71; 95% CI, 1.44-9.53, P = 0.006) and a nonsignificant 1.63-fold increased risk (95% CI, 0.71-3.75) of developing cancer (Table 3). The results were similar when the threshold for bleomycin sensitivity was raised from >0.6 to >0.8 or >1.0 [cutoffs commonly used in the literature (25); data not shown for the 1.0 cutoff] breaks per cell and when the interquartile (quartiles defined as <0.44 breaks per cell, 0.44 ≤ breaks < 0.6, 0.6 ≤ breaks < 0.87, and ≥0.87 breaks) HRs of increasing bleomycin sensitivity were compared (Table 3). Bleomycin sensitivity was not significantly associated with the future development of cancer when the number of bleomycin-induced breaks was modeled as a continuous variable (adjusted P = 0.23) or as an ordinal variable representing the bleomycin break quartiles (adjusted P = 0.24; see Table 3 for interquartile HRs).

We stratified the cohort based on the presence or absence of chromosome 9p (p16) LOH, 17p (TP53) LOH, p16 mutation, and TP53 mutation in the Barrett's segment either at or immediately before the endoscopy that coincided with the bleomycin sensitivity assay (Table 4). Among patients with detectable 9p or 17p LOH, bleomycin sensitivity (>0.6 breaks per cell) was associated with a significantly greater risk of developing aneuploidy. Among patients with 17p LOH, bleomycin sensitivity approached significance for cancer outcome (P = 0.053), and we found a significant trend modeling the number of bleomycin-induced breaks as a continuous variable. The HR from such a model comparing the third versus first quartiles (corresponding to 0.44 breaks) was 3.25 (95% CI, 1.62-6.53, P_trend < 0.001) for patients with 17p LOH and 1.20 (95% CI, 0.51-2.83, P_trend = 0.68) for patients without 17p LOH. The HR using aneuploidy as an end point is 35.5 (95% CI, 2.92-430, P_trend = 0.005) for patients with 17p LOH and 1.63 (95% CI, 0.85-3.11, P_trend = 0.14) for patients without 17p LOH. Bleomycin sensitivity was not a significant predictor of cancer in patients with (or without) 9p LOH, TP53 mutation, or p16 mutation.

Defects in DNA damage recognition and repair have been associated with a wide variety of cancers (63). Barrett's esophagus is characterized by chronic inflammation, cellular damage/repair, and increased proliferation (56, 64-66). Chronic inflammation is associated with oxidative damage and increased levels of double-strand DNA breaks. Thus, diminished DNA repair in Barrett's esophagus could lead to accelerated progression to esophageal adenocarcinoma. Here, we report that Barrett's esophagus patients whose peripheral blood lymphocytes were sensitive to bleomycin-induced double-strand DNA breaks were at significantly increased risk for subsequent development of aneuploidy, a validated intermediate marker of progression to esophageal adenocarcinoma (53-55), and, to a lesser extent, esophageal adenocarcinoma itself.

Progression in persons with Barrett's esophagus is associated with increasing chromosomal instability (46). LOH at the TP53 locus (17p LOH) generally precedes aneuploidy in persons with Barrett's esophagus (57, 67), and patients with 17p LOH are at increased risk for progression to esophageal adenocarcinoma (49). Although the bleomycin-sensitive patients in aggregate were not significantly more likely to progress to esophageal adenocarcinoma in our study, the risk among those with 17p LOH at or before baseline was significantly higher. We hypothesize that bleomycin-sensitive individuals have higher spontaneous chromosomal mutation rates and/or diminished DNA repair capacity, and, in that background, the loss of p53 function in the Barrett's epithelium allows cells to continue to cycle although chromosomal damage may not be fully repaired. G₁ arrest in cells with double-strand breaks is believed to be p53 dependent (68, 69), and there is strong evidence that alterations in a number of damage repair genes are associated with the development of cancer (70).

To our knowledge, ours is the first prospective study of bleomycin sensitivity (and perhaps mutagen sensitivity in general) and cancer risk before the onset of cancer. Because current technology does not allow us to measure the in vivo DNA repair capacity of individuals directly, we used the bleomycin assay as an indirect measure of the balance between DNA double-strand break formation and repair. A prospective study of mutagen sensitivity in cancer-free patients ensures that mutagen sensitivity is not influenced by the presence of cancer or its treatment. Detecting the association between bleomycin sensitivity and cancer risk in patients with 17p LOH was possible because of the frequent and long-term follow-up of the patients, the molecular characterization of our cohort using known biomarkers, and the relatively large number of cancer outcomes due to the increased risk of the cohort for developing esophageal adenocarcinoma. Our use of aneuploidy, a known risk factor for developing esophageal adenocarcinoma, as an intermediate end point strengthened the study. However, some of our analyses were hindered by the limited number of cancer outcomes, especially when stratifying the cohort on molecular criteria. There may have been insufficient cancer outcomes to conclusively determine whether bleomycin sensitivity predicts esophageal adenocarcinoma in our cohort in aggregate, although the quartile trend in Table 3 is consistent with bleomycin sensitivity being associated with the future development of esophageal adenocarcinoma.

This study supports the hypothesis that sensitivity to mutagens can increase the risk of developing cancer, particularly among those with inactivated tumor-suppressor genes. We also find an association between an in vitro mutagen sensitivity assay and the development of aneuploidy in Barrett's esophagus epithelium in vivo. Further studies would be required to determine whether mutagen sensitivity assays could help risk-stratify patients with Barrett's esophagus.

We thank David Cowan, Janine Kikuchi, and Terri Watson for database management; Christine Karlsen for coordination of patient care; Valerie Cerera for coordination of research biopsies and flow cytometry; Tricia Christopherson for management of patient interviews and data collection; and the study interviewers and participants in the Seattle Barrett's Esophagus Project for making this study possible.