Methylation of cyclin D2 is observed frequently in pancreatic cancer but is also an age-related phenomenon in gastrointestinal tissues.

PURPOSE Hypermethylation of CpG islands in the promoters of selected genes is a common feature of neoplasia. Aberrant methylation of cyclin D2 has been observed in several cancers. We investigated the methylation of cyclin D2 in aging and pancreatic neoplastic development, and the utility of cyclin D2 methylation as a marker of pancreatic adenocarcinoma. EXPERIMENTAL DESIGN Methylation-specific PCR was performed on DNA from 165 resected pancreatic exocrine neoplasms [109 adenocarcinomas, 46 intraductal papillary-mucinous neoplasms (IPMNs), and 10 mucinous cystic neoplasms], 14 pancreatic intraepithelial neoplasms, 13 microdissected-normal pancreatic ductal epithelia, 25 normal pancreatic parenchyma, 51 specimens of pancreatic juice obtained perioperatively, 15 pancreatic cancer xenografts, 22 pancreatic cancer cell lines, 59 specimens of normal duodenum, and 49 gallbladders affected by cholecystitis. Cyclin D2 RNA expression was determined in pancreatic cancer cell lines, before and after 5-AZA-2'-deoxycytidine treatment, by reverse transcription-PCR. RESULTS Methylation of cyclin D2 was identified in 65.1% (71 of 109) of primary pancreatic adenocarcinomas, in 50% (23 of 46) of IPMNs, and in 70% (7 of 10) of mucinous cystic neoplasms, but was detected infrequently in microdissected samples of normal pancreatic epithelia [7.7% (1 of 13)] and in pancreatic intraepithelial neoplasms [14.3% (2 of 14)]. Cyclin D2 methylation was also recognized in 10 of 15 (66.7%) pancreatic cancer xenografts and in 19 of 22 (86.4%) pancreatic cancer cell lines. All of 10 pancreatic cancer cell lines completely methylated at cyclin D2 showed no expression by reverse transcription-PCR. Four of these 10 cell lines were treated with 5-AZA-2'-deoxycytidine, and all 4 showed increased RNA expression of cyclin D2 after treatment. In pancreatic juice, cyclin D2 methylation was detected in 9 of 22 (40.9%) samples from patients with pancreatic cancer and in 6 of 9 (66.7%) patients with IPMNs, but in none of 20 non-neoplastic controls, respectively (P = 0.0013 and P < 0.0001, respectively). Methylation of cyclin D2 was also observed more in non-neoplastic tissues and with increasing age (P = 0.041 in the pancreas, P = 0.047 in the duodenum, and P = 0.0008 in the gallbladder). CONCLUSIONS The promoter region of cyclin D2 undergoes age-related methylation in multiple tissues, but aberrant methylation is more often detected in tissues and juice samples of pancreatic cancer than in normal tissues. The detection of cyclin D2 methylation in pancreatic juice may aid in the diagnosis of pancreatic adenocarcinoma.