Detection of biomarker mirnas and their functions by co-expression network analysis in gastric cancer (gc)
,1 Pouya salehipour
,2 Shadi mahdipour
,3 Sina majidian
1. Department of Medical Genetics, School of Medicine, Tehran University of Medical Sciences
2. Department of Medical Genetics, School of Medicine, Tehran University of Medical Sciences
3. Department of Medical Genetics, School of Medicine, Tehran University of Medical Sciences
4. School of Electrical Engineering, Iran University of Sciences and Technology
Gastric cancer (gc) is the fourth most common cancer and the second leading cause of death among all cancers around the world. in 2008, a total of 989600 new cases of gastric cancer were discovered and as many as 738,000 died from gastric cancer. therefore, studies that explore the mechanisms of cellular and molecular gastric cancer development and the validation of novel biomarkers are critically required. many studies have shown that mirnas are associated with development and progression of gastric cancer and can act as diagnostic, prognostic and therapeutic biomarkers. thus, identification of differentially expressed mirnas (dems) may contribute to early diagnosis and prediction of survival in gc. by using several approaches including whole genome studies and co-expression analysis among mirnas, lncrnas, and mrnas we can recognize corresponding mirnas in gc. this finding of mirna and their target genes help us to identify the signaling pathways in which they are involved and provide a promising therapeutic outcome. this study was designed to achieve this goal by analyzing 351 samples of gc using tcga-stad data along with co-expression analysis of mirnas, lncrnas, and mrnas.
Transcriptome profiling data of 351 gc primary tumors and 32 primary normal samples, as well as transcriptome profiling data of 21 metastatic primary tumors and 351 non-metastatic primary tumors, were extracted from the tcga-stad project. differential expression analysis and co-expression network analysis among mrnas, mirnas, and lncrnas were performed to create network pathways that reveal active mirnas. in addition, experimental targets of these mirnas were determined by mirtarbase database. to find the functions of these mirnas in gc, nested network analysis of identified mrnas was performed using kegg and string databases to identify altered signaling pathways in gc.
In this study, extracted data from the project related to samples of primary tumors and primary normal samples identified hsa-mir-15b-5p as a candidate biomarker in gc. this mirna was up-regulated in gc with a log2fc of -0.813 and a p-value of 1.156e-6. however, in another project of tcga-stad, including samples of metastatic primary tumors and non-metastatic primary tumors, there was no difference in the expression of hsa-mir-15b-5p between two groups. network analysis determined that mir-15b-5p was associated with pi3k-akt signaling pathway (adjusted p-value: 5.60e-06) and with ras signaling pathway (adjusted p-value:1.20e-03) by altering the expression of insr, bcl2, ccnd1, ccne1, igf1r, fgf2, fgfr1, vegfa, and kdr genes.
Patients with gc have a poor prognosis and in spite of advances in gc therapy, the five-year survival rate is 5%-20%. therefore, identification of prognostic biomarkers is essential to improve clinical treatment and management of gc patients. over the past years, many types of research indicated the role of mirnas as diagnostic, prognostic and therapeutic biomarkers. it is believed that mirna alterations occur early in the cascade of the pre-neoplastic events. in this study, we aimed at finding a mirna as a biomarker in gc by network analysis we found that hsa-mir-15b-5p is a potential biomarker in this type of cancer. as for mir-15b-5p, a great number of target genes are associated with signaling pathways; insr, bcl2, ccnd1, ccne1, igf1r, fgf2, fgfr1, vegfa genes in ras signaling and pi3k-akt pathway whose expression is deregulated in gc. since these pathways are tightly regulated and controlled, any alteration in the genes that are involved in these pathways would further affect cell signaling and lead to the initiation or progression of cancer. mir-15b-5p is up-regulated in primary tumors compared to primary normal samples. this finding confirmed the result of a previous study on mir-15b in gc. however, the expression of mir-15b-5p in metastatic primary tumors compared to non-metastatic primary tumors demonstrated no difference. therefore, it is suggested that mir-15b-5p can affect tumorigenesis, but it has no effect on metastasis. identification of a single mirna in blood is not ideal for diagnosis of gc because of tumor heterogeneity. thus, analysis of a combination of plasma mirnas may help improve their diagnostic performance for gc.
Mirna, gastric cancer , network analysis, biomarker