Identification of the feature genes of pancreatic cancer-associated diabetes mellitus based on the pancreatic cancer cohort from TGCA / 中华胰腺病杂志

ABSTRACT

Objective:To investigate the genetic characteristics of pancreatic cancer-associated diabetes mellitus (PCDM) and screen out the possible molecular markers for PCDM.Methods:The clinical data of pancreatic cancer (PC) cohort from The Cancer Genome Atlas (TCGA) were selected and collected, and the patients were divided into PCDM( n=11) and PC groups( n=109) according to whether the patients were diagnosed as diabetes within 2 years of PC diagnosis. Then, the mRNA microarray data of genome expression were extracted from TCGA PC cohort, and the differentially expressed genes (DEGs) were screened out by the " limma" package of R software based on (|log2 fold change|>2 and P<0.05). The functions of DEGs were revealed with gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis. Finally, a protein-protein interaction (PPI) network was constructed with the STRING database, and the hub genes were identified by the molecular complex detection (MCODE) module of Cytoscape software. Results:The analysis showed that among 20 531 genes, 47 genes were significantly upregulated, and 60 genes were significantly downregulated in the PCDM group. GO analyses revealed that 107 DEGs were mainly involved in the positive regulation of secretory function in terms of biological function (gene number=9, P<0.01); in the regulation of receptor function of molecular function (gene number=10, P<0.01); and in the intracavitary components of cytoplasmic microtubules of cellular components (gene number=8, P<0.01). The results of KEGG pathway enrichments revealed that DEGs mainly affected PCDM via cytokine interactions (gene number=8, P<0.01). Finally, five hub genes, including GNG8, CNR2, GALR2, CXCL13, and NPY2R, were identified for PCDM in PPI network analysis. Conclusions:The feature genes of PCDM are mainly different from PC in terms of secretion function, receptor function, cytoplasmic microtubule composition, and cytokine interaction. Five genes including GNG8, CNR2, GALR2, CXCL13, and NPY2R may become potential molecular markers for PCDM.