Identification of Long Non-coding RNAs as Novel Biomarkers for Heterogeneous Glioblastomas

点击翻译按钮获取中文摘要

PROJECT SUMMARY Glioblastoma multiforme (GBM) is the most malignant form of brain tumors with more than 18,000 newly diagnosed patients and 13,000 deaths annually in the United States. The prognosis for GBM remains dismal with a median survival time of GBM patients of 14 to 16 months after diagnosis. A hallmark of malignant GBM is their high heterogeneity within the tumors. This unique characteristic manifests to molecular subtypes of GBM that display unique patterns of pathogenesis, biology, and prognosis. While specific molecular markers are of value in clinical care (e.g., IDH1/2 mutations, 1p/19q co-deletion), significant improvement in prognostic stratification and targeted therapeutics are urgently needed. With the ultimate goal of realizing the full potential of personalized and precision medicine, we propose to interrogate long non-coding RNAs (lncRNAs) in a cohort of clinical GBM specimens. LncRNAs are a class of non-coding RNAs that have emerged as critical modulators in various cellular processes through gene regulation. Previous studies including The Cancer Genome Atlas (TCGA), though limited by their profiling technique not designed for non-coding RNAs, have suggested that lncRNAs are abundant in human cancers and are highly cancer-type-specific. In particular, lncRNAs have been implicated in brain function and glioma development. Specifically, in this exploratory project, we will apply the Ribo-Zero-based transcriptomic sequencing (RNA-seq) to comprehensively characterize lncRNAs in 100 clinical GBM samples that have been collected at the Northwestern University Brain Tumor Tissue Bank (Aim 1a). In contrast to the oligo(dT)-based RNA-seq that was used by previous studies, including TCGA, the Ribo-Zero-based technique is optimized for non-coding RNA transcripts, thus offering a great advantage for profiling all potentially functional lncRNAs in GBM. Notably, a novel detection algorithm based on machine learning will be developed to provide a more flexible and universal framework of lncRNA detection using RNA-seq. Though restricted to those lncRNAs shared between our GBM data and the oligo(dT)-based TCGA, we will evaluate the tissue-specificity of lncRNAs detected in GBM using TCGA data on several solid tumors (Aim 1b). After characterizing the landscape of lncRNAs in GBM, we will evaluate whether lncRNAs are associated with the clinical outcomes of GBM patients, and evaluate the feasibility of integrating lncRNAs and gene-level transcripts into a prognostic tool (Aim 2a). This proposal will enable us to employ these novel biomarkers for the prognosis of GBM as well as future functional studies. In addition, we will utilize co-expression network analysis to assign functions to the detected lncRNAs in GBM. An integrated, internet-based catalog will be constructed to provide a resource of lncRNAs in GBM that will benefit the general research community in this new area (Aim 2b). Finally, the PIs have assembled an outstanding research team with significant achievements in relevant research areas and complimentary expertise, therefore ensuring the success of this multidisciplinary and highly innovative project.