Comparative transcriptome analysis between metastatic and non-metastatic gastric cancer reveals potential biomarkers.

Abstract

The transcriptome of metastatic gastric cancer (GC) was compared to that of non-metastatic GC to identify metastasis-related biomarkers. The gene expression dataset GSE21328, comprising 2 metastatic GC samples and 2 non-metastatic GC samples, was downloaded from the Gene Expression Omnibus database. Differential expression analysis was performed with the package limma of Bioconductor to identify differentially expressed genes (DEGs). Gene Ontology (GO) enrichment analysis was performed to identify significantly altered biological functions. In addition, the transcriptional regulatory and protein-protein interaction networks were constructed with information from the UCSC genome browser and STRING database, respectively, followed by functional enrichment analysis of all of the genes in these two networks. A total of 584 DEGs were identified, of which 175 were upregulated and 409 downregulated. Clustering analysis confirmed that these genes can distinguish metastatic from non-metastatic GC. Upregulated genes were enriched for the xenobiotic metabolic process, while downregulated genes were enriched for immune response and related pathways. Among the 584 DEGs, six genes (DAND5, EGR2, FOXD1, LMO2, PRRX2 and STAT1) were shown to encode transcription factors, which were used to establish the transcriptional regulatory network with 169 target genes, forming 175 nodes. The proteins of this network were significantly enriched for the process of negative regulation of cell differentiation. In conclusion, this study identified a range of DEGs in metastatic GC, which may enhance our current knowledge on this disease. Among these genes, STAT1 and EGR2 may constitute potential biomarkers of GC metastasis.