The genes, ALP1, MUC1, CCND1, CDK2 and FHIT significantly contributed to the 1st clustering between the EUS-FNA samples and the pancreatic juice samples (p < 0.05). On the other hand, in the EUS-FNA samples, the gene, CDK2A, CD44, S100A4 and MUC1 were specifically related to the 2nd clustering between cancer and non-cancer (p < 0.05). Figure 3 Hierarchical cluster of the human 25 genes expression pattern in 12 pancreatic samples. FNA, EUS-FNA specimens

(n = 6); PJ, pancreatic juice samples (n = 6); PC, pancreatic cancer (n = 5); CP, chronic pancreatitis (n = 3); IPMC, Selleckchem BIRB 796 intraductal papillary mucinous adenocarcinoma (n = 1); IPMA, intraductal papillary mucinous adenoma (n = CUDC-907 2); PET, pancreatic endocrine tumor (n = 1). Each color scale represents the signal intensity of each gene. Some genes that significantly contributed to the dividing of clusters (p < 0.005) were noted at the bottom of the panel. Line A shows the boundary of the gene expression pattern between EUS-FNA and pancreatic juice. Line B shows the boundary of cancer or non-cancer in the EUS-FNA specimens. Gene mutation analysis (K-ras codon 12/13) PCR amplification and gene mutation analysis for K-ras (codon12/13) were successful in the case of the samples with good quality total RNA. We extracted the total RNA and DNA from the same specimens in this study. When

SGC-CBP30 ic50 one nucleic acid could not be successfully prepared and analyzed, the other nucleic acid also could not be used. The degradation of the nucleic acid seems to be depended on the condition of sample storage after

EUS-FNA or collecting pancreatic juices. All of the analyzable pancreatic cancer samples showed a specific mutation of K-ras codon12 with Pregnenolone a single base change from GGT (Gly) to GAT (Asp) (See Table S3, Additional file 3), which is the most frequent mutation, as previously reported [13]. Additionally, one of the analyzable chronic pancreatitis samples showed a mutation from GGT (Gly) to GTT (Val), which is also frequent in pancreatic cancer as previously reported. No mutation could be detected in the samples of autoimmune pancreatitis and pancreatic endocrine tumor. Disscussion DNA microarrays can analyze plural gene expression changes at the same time. It is useful for the early detection of pancreatic cancer, evaluation of malignant potential and drug efficacy. There are some articles about the identification of genes that show chemosensitivity to anti-cancer drugs, such as gemcitabine and 5FU in the pancreatic cancer cell line [14, 15]. Gene expression profiling would especially help to predict the effectiveness of chemotherapy. This time, we inspected whether gene expression analysis by 3D microarray was possible using small amount samples obtained endoscopically. In EUS-FNA specimens, the sample storage method using RNAlater® seemed to improve the quality of the total RNA when compared with the method using liquid nitrogen storage.