Study On Androgen Transcriptional Regulation Network In Prostate Cancer

Prostate cancer is the tumor with great incidence rate and lethality among Western males. During its therapy and development course, prostate cancer has a remakable characteristic of androgen dependence transformation; therefore prostate cancer has been the ideal model for studying cancer progression. Nowadays, there have been hundreds of researches focused on prostate cancer, especially based on microarray data; however, the molecular mechanism of prostate cancer carcinogenesis and progression still remained to be unrevealed, especially in respect of a global and systematical understanding. A major cause contributing to this non-achievement is that the present strategy and methods for extracting information from microarray data failed in considering the real biological process when performing data analysis. This leads to data extraction being either primitive, insufficient, or deviating from the real biological course. Therefore, the present microarray analysis can not accurately approach the biological molecular mechanism.Aiming at solving these problems, in this research we performed two works to provide accurate and effective bioinformatics’predictions for the biological experiment study to analyze cancer carcinogenesis and progression mechanisms. One work is based on random walk model to simulate and rebuild the course of cancer progressing to higher stage, in order to identify the gene pairs with significant co-expression changes during the progression and accurately provide the changes in molecular mechanisms happened during the progression. The other work is based on the time-course matched miRNA-mRNA microarray data; and the microarray was performed with cancer cells stimulated by androgen. We developed strategies and algorithms to identify the target genes directly regulated by androgen receptor (AR) and the target mRNAs directly modulated by miRNAs, in order to establish a systematic AR regulatory cascade network mediated by miRNA. In these two works, both the essential biological processes of cancer progression and the cellular response to external stimulation have been taken into account in the algorithms we proposed. Therefore, the result of the first work can not only highly approach the biological course of cancer development and progression, and also afford sound explanation for the confusing functions of some genes in previous literatures. Moreover, the novelly identified regulatory impacts and functions in the second work have been well verified by our biological experiments.Specially, for the first work, based on the evolutionary mutation in cancer cells and the pressure of natural selection from the outside, we presented a novel analytical method named’Stochastic process model for Identifying differentially co-expressed Gene pair’(SIG), to identify gene pairs with significant co-expression changes between the lower stage and the higher stage of cancer progression by using a rander walk model. This method has been applied to two well-known prostate cancer gene microarray data sets:androgen dependent (AD) versus androgen independent (A1), and cancerous versus healthy prostate tissue. Also, the data sets were analyzed by other two statistical methods, which were developed to identify gene pairs with significantly differential co-expression. After comparison of the results from these three methods, we found that SIG method had evident superiority compared to the other two methods in several respects:analysis of cancer development course, of gene pairs and that of pathways. In correlation with previous researches on gene function, we emphasize in analyzing the AD versus Al data set with SIG method, and identified the potential molecular mechanism of PPARG gene in the functional inactivation of the prostate cancer. Moreover, from this work, three gene regulatory networks inferred could provide clues for the mechanism of prostate cancer transformation from AD to Al. These networks include arachidonic acid metabolism pathway, TNF-induced NF-k B and apoptosis pathways, and Androgen receptor pathway. However, these results were still needed to be verified by experiment. To summarize, the SIG method reliably identifies cancer progression correlated gene pairs, and performs well both in gene pair ontology analysis and in pathway enrichment analysis. This method provides an effective means of understandingthe molecular mechanism of carcinogenesis by appropriately tracking down the process of cancer progression.For the second work, in order to completely establish AR regulatory cascade network mediated by miRNA, and systematically analyze t impact of androgen on prostate cancer formation, we used androgen dependent prostate cancer cell line LNCaP as a study model, and detected the expression in ten time points of time-course microarray of miRNA and mRNA of LNCaP cells stimulated by androgen. Based on real response process of cell, we presented a strategy to group the response of gene toward androgen into early response and late response. Then we construct Response Score (RS) parameter and method to identify genes significantly and directly targeted by AR, including miRNA and mRNA. Afterwards, based on a matched miRNA-mRNA time-course expression profiling, in consideration of the possible delay of miRNA’s regulation, we introduced Modulation Score (MS) parameter and method, to identify target mRNAs significantly and directly modulated by miRNA. Based on this analysis and experiment validation biological relevance between miRNAs and target genes, a novel androgen-stimulated and miRNA-mediated network is provided, in which miR-19a, miR-27a, miR-133b, miR-421, BRF2, IRS2, and PHLDA1 are identified as prominent AR targets, and several mRNAs involved in cellular key courses and carcinogenesis, are found significantly down regulated by miR-19a,-27a,-133b and -421. Based on the identified results and published biological reports, we established an AR-regulated, miRNA-medicated androgen signaling cascade regulatory network. Furthermore, we found that miR-19a,-27a, and-133b can promote the growth and production of cells, and miR-421 can help repress the growth of cancer cells. Especially, miR-19a has an important role in mediating a negative feedback regulation of AR expression and activity. Additionally, we also discovered another miRNA-mediated AR signaling cascade that explains hormone involved immunological tolerance in prostate cancer. In conclusion, our study provides a novel approach to construct miRNA-mediated AR signaling cascade, and specifically provides insights on androgen relevant mechanisms in prostate cancer.Furthermost, since we have obtained valuable results by developing bioinformatics’ methods to study prostate cancer hormone dependence transformation and AR regulatory network, now these results work as foundation and research direction for our further research on prostate cancer mechanisms. At present, we perform the detailed analysis and experimental validation to study the miRNA-affected cellular molecular pathway functions.