The Co-regulations of Transcription Factors and MicroRNAs in Lung Adenocarcinoma

• Main Authors: Yun-Ju Sun, 孫韻如
• Other Authors: Chung-Ming Chen
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/87599702704802250711

碩士 === 國立臺灣大學 === 醫學工程學研究所 === 99 === The purpose of this study is to uncover the relations of certain transcription factors (TFs) and microRNAs (miRNAs) regulating their target genes seperately using genomics data, and to further predict the co-regulation of TF and miRNA on their target genes, which may be involved in the lung adenocarcinoma (ADC) in human. In terms of gene regulation, TFs are known to play a major role on regulating their targets (TF-targets), while miRNAs play a fine-tuning role at the targets'' post-transcriptional levels (miRNA-targets). MicroRNAs are a group of short and 22nt ~ 28nt non-coding RNAs, which bind to their targets'' 3''UTR sites to carry out post-transcriptional regulation. Previous studies have shown that many miRNAs may be involved in cancers. By analyzing the differential expressions of miRNAs, different miRNAs were uncovered to be involved in different kinds of cancer. MicroRNA could be classified into tumor suppressor miRNAs or oncomirs by their functions of inhibiting oncogenes or tumor suppressor genes. The death rate of lung cancer has long been ranked the top one among all cancers in Taiwan, of which 80% of patients are non-small cell lung cancer (NSCLC), in which the adenocarcinoma subtype is the most common. There many of studies on TF-gene regulations, but how miRNAs are involved in the regulation of lung ADC is not well understood. This study proposes a model to infer which TF and which miRNA co-regulate their target genes in lung ADC. A two-stage fuzzy method was constructed to model the processes of TF and miRNA regulating targets sequentially using genomics data, and a fuzzy inference system (FIS) was used to infer the co-regulation of TFs and miRNAs. The first stage modeled the process of TFs'' binding to their targets. We first extracted DNA sequence data and inputed them into AdaFuzzy which learned and infered the TF-target regulation. In the second stage, AdaFuzzy and FIS modeled whithin a target gene in a predicted TF-target pair was further regulated by a miRNA, thus this triple had a co-regulation relation. In addition to sequence information, microarray gene expression data (MGED) containing lung ADC tissues and normal lung tissues was also used in the FIS. Integrating MGED, the TF-target binding strength predicted in the first stage and the miRNA-target relation strength as inputs, the FIS could infer the co-ragulations resonably. The miRNA-target were predicted by AdaFuzzy which learned binding strength scores from miRanda and TargetScan. The results show that the predicted co-regulations include a validated triple E2F─Cyclin E─miR-15 which has been confirmed by lung ADC cell lines. Because the upstream genes of cyclin E were not significantly expressed in the cell cycle pathway, and both E2F and miR-15 differentially expressed, we infered that this triple may be involved in the tumorigenesis of lung ADC. Candidate co-regulation of TF-miRNAs, which consisted of cell cycle-related E2F─Cyclin E─miR-15ab/195 /185/107 and p53 signaling pathway-related p53─TIMP3─miR-21 were further predicted from the triples having similar or higher score of the valified triple. These results are attempts to depict the roles of miRNAs play and their co-regulations in lung ADC. On the other hand, a potential therapy may be designed via the relationship of TF-miRNA co-regulation to kill cancer cells efficiently.