建立一個新穎的生物資訊平台來預測與驗證做為早期診斷的microRNA 生物標記組合

Abstract

微小核醣核酸〔microRNA/miRNA〕為一段長約由22個核苷酸組成的非編碼核醣核酸分子，廣泛地研究 miRNA在基因表現調控的角色。近年來研究發現，微小核醣核酸在血清、血漿等體液中穩定存在，循 環血液中的miRNA表現圖譜被應用於癌症、心血管、免疫及各種疾病。miRNA生物標記〔miRNA biomarker〕為非侵入性的診斷，具有極大的潛能成為疾病及癌症診斷與預測的工具。因此，發展 miRNA生物標記分析及驗證平台極為重要。在此研究計晝中，我們感興趣於大腸直腸癌〔colon cancer〕、冠狀動脈疾病〔coronary artery disease〕、肝癌〔liver cancer〕及妊娠毒血症〔preeclampsia〕的早 期檢測與診斷。

本研究將發展一個早期檢測和診斷miRNA生物標記〔miRNA biomarker〕s的分析及驗證平台，結 合資料庫、高通量定序實驗〔Next-generation sequencing〕與序列分析技術、生物資訊〔Bioinformatics〕、 統計分析與多重miRNA定量技術。研究目標如下：

〔a〕建立miRNA生物標記分析平台，包含miRNA生物標記知識庫建立、miRNA的功能註解、分析 miRNA表現圖譜找到及驗證新的miRNA生物標記及挑選表現及功能獨立的miRNA〔第一年〜 第三年〕。

〔b〕發展多重miRNA定量技術，同時測定多個miRNA生物標記基因表現〔第二年〜第四年〕。

〔c〕結合臨床檢體來驗證miRNA生物標記組，第一〜二年為妊娠毒血症，第三年為肝細胞癌和冠狀 動脈疾病，第四年為肝細胞癌。

本研究特色在於，我們將整合計算與實驗方法來挑選早期檢測及診斷miRNA生物標記，建立多重 miRNA定量技術同時測定多個miRNA生物標記基因表現。為了挑選出最佳的miRNA生物標記組合 一個最好的生物標記模組，利用表現及功能獨立的miRNA生物標記來最小化生物標記及最大化疾病 診斷的分別。我們將產出妊娠毒血症、肝癌、冠狀動脈疾病及大腸直腸癌的早期診斷miRNA生物標 記組將被發展及驗證。每個疾病的miRNA生物標記組，將利用至少200以上的臨床檢體來驗證，我 們將運用Area under the curve〔AUC〕、敏感性〔Sensitivity〕及特異性〔Specificity〕等指標來衡量疾病預測 模型的準確度。本研究成果將可對於所提出疾病之早期偵測及診斷做出很好的貢獻。亦可針對所提出 的miRNA生物標記之生物功能探討，做深入的分析。

MicroRNAs are 22-nucleotide-long non-coding RNA molecules. The crucial role of different miRNAs in gene-expression regulation has been widely studied. Recently, miRNAs have been detected in serum, plasma, and other body fluids, and circulating miRNA profiles have now been associated with the pathogenesis of cancer, along with cardiac- and immune-related diseases. Circulating miRNAs have great potential for use as clinical biomarkers in various diseases, including cancer, because noninvasive detection in various medical conditions is possible. Therefore, it is necessary to develop a discovery and validation platform for the systematic identification of miRNA biomarkers. Besides, in this project we are interested in early detection and diagnostics of colon cancer, liver cancer, coronary artery disease (CAD) and preeclampsia.

In this project, we plan to construct an integrative miRNA biomarker discovery and validation platform for early detection and diagnostics. We propose a bioinformatics-driven approach combining database, high-throughput sequencing, computational pipelines and a multiplex miRNA quantification method. The major aims of this project are summarized as follows:

(a) Establishing miRNA biomarker discovery platform including the construction of miRNA biomarker repository, functional annotations of miRNAs, identification of novel miRNA biomarkers using circulating miRNA expression profiles, and the selection of expressional and functionally-independent miRNA biomarkers (1st year 〜3rd year).

(b) Developing a multiplex miRNA quantification method for simultaneously detecting the expression levels of miRNAs in a diagnostic panel (2nd year 〜4th year).

(c) Incorporating clinical samples to validate the constructed miRNA biomarker panels for Preeclampsia (1st and 2nd year), colon cancer (3rd year), coronary artery disease (CAD) (3rd year and 4th year), and hepatocellular carcinoma (HCC) (4th year)

We will develop an integrative approach for selecting circulating miRNA biomarkers for early detection. A multiplex circulating miRNA quantification method will be developed for determining the expression levels of several miRNA biomarkers. As to select the best circulating miRNA biomarker for constructing a panel, we will develop a method for minimizing number of miRNA biomarkers and maximizing disease discrimination based on expressional and functional independence.

Panels of circulating miRNA biomarkers for early diagnostics of preeclampsia, hepatocellular carcinoma, Coronary artery disease, and colorectal cancer will be developed and validated, respectively. These panels will be validated by at least 200 samples of independent cohorts for each disease. Predictive models will be devised and the Area under the curve (AUC), sensitivity, and specificity will be assessed. The major contribution of this work is to establish an integrated platform for circulating miRNA biomarkers discovery and validation. Accurate predictive models comprising mRNA biomarkers of the four diseases can provide an alternative solution for early detection and diagnostics.