Identification of novel upregulated microRNAs in the pathogenesis of gastric cancer by the use of open access databases and bioinformatics tools.

Abstract

15 Background: Gastric cancer (GC) is the third leading cause of cancer death worldwide. Chronic infection by H.pylori and EBV is related to the multistep cascade of GC. This cascade hypothesizes that a specific sequence of lesions leads to GC. Epigenetic processes, such as DNA methylation or miRNAs, are new players in the pathogenesis of GC. We aim to identify novel upregulated microRNAs associated with the multistep cascade through the use of open access databases and bioinformatics tools. Methods: We selected 3 GEO DataSets with human miRNA expression levels in paired tumor and non-tumor samples from gastric cancer patients. The largest library, GSE 23739 (40 matched T/N samples) was used as a "discovery dataset”. Only miRNAs upregulated >1.2 fold T over N were considered. Upregulated miRNAs were further validated in two independent datasets: i) GSE28700 (22 fresh-frozen paired T and N samples) and ii) GSE54397 (16 formalin-fixed T and N samples) and confirmed with TCGA Stomach Adenocarcinoma (STAD) data (doi:10.1038/nature13480). Finally, upregulated miRNAs were analyzed by Diana Tools miRPath, for associated oncogenic pathways and target genes. Results: From the discovery dataset GSE 23739 we identified 231 upregulated miRNAs between 40 matched T/N samples. Only 12 of these candidates were found in both validation datasets GSE28700 and GSE54397. Only three of these miRNAs were undescribed in GC: miRNAs-505, -552 and -592. TCGA STAD data confirmed our results. Analysis in miRPath revealed association with three oncogenic pathways PI3K-AKT, TGFβ and Erbβ; as well as genes relevant in oncogenesis: MAP2K1, RPS6KB1, TGFA, SMAD3, IGF1, LAMC1, IFNAR1, ITGB6, PIK3CA, PPP2R2D, EIF4E, BRCA1, FOXO3 and CREB3L3. Conclusions: The use of open access databases and bioinformatics tools enabled us to identify three novel upregulated microRNAs with oncogenic potential in GC. The analysis of these miRNAs revealed associated oncogenic pathways and target genes. These findings need to be further inquired in an experimental manner in order to assess the potential role of these miRNAs in the multistep cascade of GC.

pETM: a penalized Exponential Tilt Model for analysis of correlated high-dimensional DNA methylation data

“…One possible explanation is the small sample size, where there are only 12 Her2 subtypes. We similarly investigated the top 10 genes out of the 94 genes uniquely selected by pETM from 109 uniquely selected sites and found that 9 genes were reported to be associated with cancers: CLTC (P€ arssinen et al, 2007) and NSD1 (Stephens et al, 2009) were reported to be associated with breast cancer; SLC25A2 (Motamedian et al, 2015) and ESRRA (Micci et al, 2014) were reported to be associated with ovarian cancer; FSD1 (Yamashita et al, 2006) and CREB3L3 (Wichmann et al, 2015) were reported to be associated with gastric cancer; CARD14 (Oudes et al, 2005) was reported to be associated with prostate cancer; LRRFIP1 (Ariake et al, 2012) was reported to be associated with colorectal cancer; and PARD3 was reported to be associated with both esophageal squamous cell carcinoma (Zen et al, 2009) and lung squamous cell carcinomas (Bonastre et al, 2015).…”