Exploring the Roles of Long Non-Coding RNAs in Glioblastoma Tumor Recurrence and Therapy Resistance

Abstract

Glioblastoma (GBM) is the most common and devastating primary CNS brain tumor with a median survival time of around 14 months. Most patients succumb to re-current disease which is often more malignant than the primary tumor and is frequently therapy resistant. There have not been significant advances in the treatment of GBM despite decades of research. This is partly due to the lack of accurate preclinical models and of the focus on primary rather than recurrent tumors. We created a 350 gene custom GBM-specific panel which contains 16 molecular signatures including molecular sub-typing signatures. We have demonstrated concordance of drug responses across multi-ple patient-derived models using this gene panel. We also developed a novel preclinical model of GBM tumor recurrence and therapy resistance utilizing patient-derived xeno-lines (PDXs) which have undergone serial in vivo selection to radiation therapy (RTS). The non-coding genome/transcriptome has in recent years become a focus for biological and oncology researchers. Non-coding RNAs (ncRNAs) have historically been relegated as junk transcripts, but now the biological importance of ncRNAs has become apparent. Long non-coding RNAs (lncRNAs) are a class of molecules that have become an area of research interest in normal as well as aberrant biology in the past decade. LncRNAs can affect transcriptional programs directly or through epigenetic regulation of chromatin and DNA remodeling. These transcripts can act as signals/guides, molecular sponges, or can be structural as part of ribonucleoprotein complexes. We have molecularly charac-terized our PDX GBM RTS models using paired transcriptomic and kinomic analysis. We developed an in silico analysis pipeline aimed at exploring the roles of lncRNAs in our models. The primary associations of lncRNA expression with transcriptional regula-tion were made by evaluating direct DNA-binding potential of the lncRNAs and looking at the cis-regulatory potential of genes proximal to binding sites. Our data show asso-ciations and correlations of expression of more than 50 lncRNA transcripts with tran-scriptional signatures related to proliferation, stemness, chromatin remodeling, and DNA damage response. We also identified altered kinase enrichment in our models re-lated to lncRNA regulatory networks which may represent druggable targets for recur-rent GBM tumors.