Comprehensive analysis of gene expression profiles provides insight into the pathogenesis of Crohn's disease

Abstract

Crohn's disease (CD) is a type of inflammatory bowel disease that cannot be fully cured by medication or surgery. In the present study, the aim was to understand the underlying mechanisms of CD. Two CD microarray datasets were downloaded from The Gene Expression Omnibus database: GSE36807 (13 CD and 7 normal samples) and GSE59071 (8 CD and 11 normal samples). A series of bioinformatics analyses were conducted, including weighted gene co-expression network analysis to identify stable modules, and analysis of differentially expressed genes (DEGs) between CD and normal samples. The common DEGs in the GSE36807 and GSE59071 datasets were screened. Subsequently, overlapping genes in the stable modules and the DEGs were selected to construct a protein-protein interaction (PPI) network using Cytoscape software. Enrichment analysis of genes in the network was performed to explore their biological functions. A total of 10 stable modules and 927 DEGs were identified, of which 234 genes were shared in the stable modules and the DEGs. After removal of 32 uncharacterized genes, 202 genes were selected to build the PPI network. Low density lipoprotein receptor (LDLR), toll-like receptor 2 (TLR2), lipoprotein lipase (LPL), forkhead box protein M1 (FOXM1) and neuropeptide Y (NPY) were revealed as key nodes with high degree. Pathway enrichment analysis demonstrated that LPL was enriched in the peroxisome proliferator-activated receptor (PPAR) signaling pathway. In conclusion, LDLR, TLR2, FOXM1 and NPY, as well as LPL in the PPAR signaling pathway may serve critical roles in the pathogenesis of CD.