| Background and Aim: The incidence of pancreatic cancer(PC)has increased in recent years,which is characterized by a late presentation.For all stages of this cancer,the five-year survival rate is less than 5%.CA-199 is currently the most widely used tumor marker for pancreatic cancer,and the sensitivity of CA-199 for the diagnosis of pancreatic cancer is 70-84.9%.Nevertheless,its sensitivity and specificity are still unsatisfactory.With the advent of the post-genome era as well as the rapid development of bioinformatics technology,a new biological research model in the field of disease diagnosis has been established based on the existing data: theoretical speculation followed by experimental verification.The purpose of this study is to use bioinformatics research methods to screen for key genes related to pancreatic cancer and to verify its expression in pancreatic cancer in order to find a new tumor marker to improve the early diagnosis.Methods: The validated genes related to pancreatic cancer were retrieved from the KEGG PATHWAY,which is a main database in KEGG database.65 validated genes were extracted from the pathway hsa05212.Each gene was represented by 12,750 features derived from the GO terms and KEGG pathways.A reliable and widely used feature selection method,named m RMR method,was adopted to analyze all investigated 12,750 features.All features are sorted based on their relevance to targets measured by their mutual information(MI)values.STRING(Search Tool for the Retrieval of Interacting Genes/Proteins)online software(http://string.embl.del/)was used to draw the interaction network diagram of related gene expressing protein,and Cytoscape software was used to select the key genes by calculating the network and each node topology characteristics.Immunohistochemistry was performed in human pancreatic cancer tissue microarray which envolve 30 paires of pancreatic cancer and parapancreatic cancer cases to detect expression of ENO1.The serum levels of ENO1 in humans were detected by AlphaLISA kit.73 patients with pancreatic cancer and 30 healthy volunteers were enrolled in this study.The expression differences of ENO1 in serum were detected and analyzed as diagnostic ability for a new pancreatic cancer marker.Result: A total of 17 important KEGG signaling pathways and five pancreatic cancer related GO terms were calculated by m RMR and analyzed by the STRING online software for the interaction of the proteins encoded by the pancreatic cancer-related genes.Results showed that the interactions between these proteins encoded by genes were mainly concentrated in 11 proteins.Cytoscape constructed protein-protein interactions found that 7 of the encoded proteins were closely related to pancreatic cancer,including RAC1,AKT1,TP53,CCND1,SRC,CDKN1 A and ENO1.The positive rates of ENO1 in pancreatic cancers and parapancreatic cancers were 91±5.6% and37±22.2% respectively.Significant difference existed between them.The concentration of serum ENO1 in PC group was 33.08±22.87 ng/m L,which was significantly higher than the level of 10.40±9.41ng/ml in normal group(P< 0.001).12.88 ng/m L of ENO1 was used as the cutoff for predicting pancreatic cancer,the sensitivity and specificity of diagnosis were 75.8% and88.2%.the sensitivity and specificity of combination diagnosis were improved to 94.5% and 82% respectively.There was an expression of ENO1 in 30 pairs of human pancreatic cancer with immunohistochemical method.The positive rates of ENO1 in pancreatic cancers and parapancreatic cancers were 91±5.6% and 37±22.2% respectively.Significant difference existed between them.The concentration of serum ENO1 in PC group was33.08±22.87 ng/m L,which was significantly higher than the level of10.40±9.41ng/ml in normal group(P < 0.001).12.88 ng/m L of ENO1 was used as the cutoff for predicting pancreatic cancer,the sensitivity and specificity of diagnosis were 75.8% and 88.2%.The sensitivity and specificity of combination diagnosis were improved to 94.5% and 82% respectively.Conclusion: Based on bioinformatics technology,this study obtained 17 important KEGG signaling pathways and 5 pancreatic cancer-related GO terms with m RMR analysis.Enolase-1 and 6 pancreatic cancer closely related key genes were selected by further analysis.The level of ENO1 in peripheral blood of patients with pancreatic cancer is positively correlated with PDAC disease progression and lymph node metastasis,which is closely related to the prognosis of PDAC,and can be used as a tumor marker for pancreatic cancer. |
PDF Full Text Request