Effects of long-term aspirin use on molecular alterations in precancerous gastric mucosa in patients with and without gastric cancer
The risk of gastric cancer (GC) remains even after H. pylori eradication; thus, other combination treatments, such as chemopreventive drugs, are needed. We evaluated the effects of aspirin on genetic/epigenetic alterations in precancerous conditions, i.e., atrophic mucosa (AM) and intestinal metaplasia (IM), in patients with chronic gastritis who had taken aspirin for more than 3 years. A total of 221 biopsy specimens from 74 patients, including atrophic gastritis (AG) cases without aspirin use (control), AG cases with aspirin use (AG group), and GC cases with aspirin use (GC group), were analyzed. Aspirin use was associated with a significant reduction of CDH1 methylation in AM (OR: 0.15, 95% CI: 0.06–0.41, p = 0.0002), but was less effective in reversing the methylation that occurred in IM. Frequent hypermethylation including that of CDH1 in AM increased in the GC group compared to the AG group, and CDH1 methylation was an independent predictive marker of GC (OR: 8.50, 95% CI: 2.64–25.33, p = 0.0003). In patients with long-term aspirin use, the changes of molecular events in AM but not IM may be an important factor in the reduction of cancer incidence. In addition, methylation of the CDH1 gene in AM may be a surrogate of GC.
Helicobacter pylori (H. pylori) infection causes non-atrophic gastritis, which progresses to atrophic gastritis, intestinal metaplasia (IM), dysplasia, and finally, gastric cancer (GC)1. Thus, the International Agency for Research on Cancer has concluded that H. pylori is a class I human carcinogen2. To date, some meta-analyses have shown that H. pylori eradication reduced the risk of GC in patients with chronic gastritis who underwent endoscopic resection (ER) for early GC3,4,5,6,7. However, a recent study from Japan showed that even after H. pylori infection was cured and gastric inflammation was eliminated, there was still a risk of GC in the long-term8. Additionally, metachronous GC occurred to some degree in patients who had H. pylori infection eradicated following ER for early GC9,10,11,12,13. Thus, it remains controversial if H. pylori eradication suppresses the development of GC. To reduce the risk of GC after H. pylori eradication, other combination treatments such as anti-inflammatory agents and dietary or nutritional intervention are needed.
Some studies including meta-analyses have reported that aspirin and other nonsteroidal anti-inflammatory drugs (NSAIDs) are associated with a reduced risk of both colorectal cancer and GC14,15,16,17. Their anti-carcinogenetic effects have been attributed to inhibition of the cyclooxygenase pathway and their anti-inflammatory abilities18,19. The roles of a number of genetic and epigenetic alterations, including microsatellite instability (MSI) and promoter hypermethylation of multiple tumor-related genes, are reportedly involved in GC and precancerous conditions of the stomach20,21,22,23,24,25,26,27,28,29,30,31,32,33. The CpG island methylator phenotype (CIMP), characterized by extensive hypermethylation of multiple CpG islands within the genome, is currently recognized as one of the major mechanisms in GC30. In addition, IM with a colonic phenotype, as detected by the Das-1 monoclonal antibody (mAb), was shown to be strongly associated with GC32,33,34,35. However, to date, no study has compared changes in molecular phenotype in patients with chronic gastritis with or without GC who have taken aspirin on a long-term basis.
It was recently reported that the risk of GC was reduced in patients who took aspirin on a regular basis for more than 3 years16. In this study, we examined the effects of aspirin on genetic and epigenetic alterations, as well as mAb Das-1 reactivity in precancerous conditions, i.e., atrophic mucosa (AM) and IM, in patients with chronic gastritis who regularly took aspirin for more than 3 years. We also determined the molecular markers linked to carcinogenesis risk in those patients. Finally, we compared the differences in molecular abnormalities between patients with AM and IM.
The characteristics of the patients are shown in Table 1. The mean duration of aspirin use was 6.3 ± 2.7 years (range 3–15 years) in the atrophic gastritis (AG) group and 6.5 ± 3.4 years (3–15 years) in the GC group; thus there was no significant difference between groups. There was also no significant difference in mean age among the three groups, although there were more males in the GC group than in the controls (p = 0.002). H. pylori infection rate was not significantly different among the three groups; 24 individuals (75.0%) in the control group, 21 patients (87.5%) in the AG group, and 11 patients (61.1%) in the GC group were negative for H. pylori. Of these patients, 21 cases in the control group, 6 in the AG group and 3 in the GC group had undergone H. pylori treatment. The remaining 26 patients had not been treated for H. pylori infection; thus, their infection was considered naturally eradicated. Of these patients, severe mucosal atrophy (open type according to the endoscopic classification by Kimura and Takemoto)36 was identified in 66.7% (2 of 3) of cases in the control group, 60.0% (9 of 15) of cases in the AG group and in 75.0% (6 of 8) of cases in the GC group.
The incidence of MSI and hypermethylation of seven genes are shown in Tables 2, 3, 4 and 5. Most of the molecular alterations, with the exception of E-cadherin (CDH1) methylation, were more frequently found in IM compared to AM in each group (i.e., control, AG, and GC groups).
Molecular changes in AM by aspirin use
The methylation of CDH1 and methylated-in-tumor-31 (MINT31) in AM significantly decreased in the AG group compared to the control group (p < 0.0001 and p = 0.03, respectively). Multivariate analysis showed that aspirin use was associated with a significant reduction of CDH1 gene methylation (odds ratio [OR]: 0.15, 95% confidence interval [CI]: 0.06–0.41, p = 0.0002) (Table 2). In contrast, the frequency of methylation of CDH1 gene, and MINT1 and MINT31 loci significantly increased in the GC group (p < 0.0001, p = 0.02, and p = 0.004, respectively). Multivariate logistic regression analysis showed that CDH1 methylation was an independent risk factor of significant gastric dysplasia (OR: 8.50, 95% CI: 2.64–25.33, p = 0.0003) (Table 3). Also, when adjusting for gender in multivariate analysis, a similar result was obtained (OR: 7.71, 95% CI: 2.34–25.42, p = 0.0008). The sensitivity, specificity, positive predictive value, and negative predictive value of CDH1 methylation for the development of gastric dysplasia were 59%, 86%, 68%, and 80%, respectively.
Molecular changes in IM by aspirin use
The frequency of CIMP in IM significantly decreased in the AG group compared to the control group after aspirin use (p = 0.02) (Table 4). On the other hand, although CIMP rate tended to be higher in the GC group than in the AG group (p = 0.08), no significant differences in other molecular events between the two groups were found (Table 5).
Comparison of molecular events in AM and IM in different parts of the stomach in the AG and GC groups
Methylation of CDH1, MINT1 and MINT31 in AM was observed in the stomach in both the AG and GC groups. Similarly, in IM, CpG island hypermethylation of most of the genes analyzed was identified in different portions of the stomach in both groups (Fig. 1A and B). The frequency of CDH1 methylation in AM was significantly higher in biopsy specimens taken from the greater curvature of the corpus in the GC group than in those from the AG group (p = 0.0002).
mAb Das-1 reactivity
There was no significant difference in reactivity to IM among the three groups (Fig. 2). However, mAb Das-1 reactivity against IM was highest in the angulus compared to the other portions of the stomach.
To the best of our knowledge, this is the first study to show the effects of aspirin on molecular alterations related to carcinogenesis in patients with chronic gastritis with and without gastric dysplasia. The patients who regularly took aspirin were followed up to 15 years. Additionally, we evaluated the molecular differences between the two precancerous conditions of AM and IM, and identified biomarkers linked with gastric dysplasia risk. We found that long-term aspirin use was associated with a significant reduction of CDH1 methylation in AM and CIMP in IM, and that CDH1 methylation in AM was a significant biomarker for gastric dysplasia in patients taking aspirin.
H. pylori infection causes aberrant DNA hypermethylation of specific genes and induces CIMP, which is an important epigenetic mechanism of tumorigenesis37,38. Previous studies have reported that CDH1 methylation is strongly associated with H. pylori infection20,23,24,