Abstract 279: Long non-coding RNAs in glioblastoma tumor recurrence and therapy resistance

Abstract Glioblastoma multiforme (GBM) is a devastating disease which invariably recurs and is often resistant to standard therapies. GBM patient-derived xenoline (PDX) models of tumor recurrence and radiation resistance were created using serial in vivo selection against radiation therapy (6 × 2Gy fractions over 2 weeks for 6+ rounds). This produced 8 isogenic pairs of patient matched radiation-sensitive, primary GBM PDX to radiation-resistant, recurrent GBM PDX. Long non-coding RNAs (lncRNA) represent novel regulatory mechanisms for therapy resistance and tumor recurrence in GBM. An in silico informatics pipeline employing differential expression, differential gene correlation, machine learning, and semantic networking was devised to identify coding and non-coding transcripts related to radiation resistance from deep total RNA sequencing. This analysis revealed 269 lncRNA and 947 coding transcripts that are associated with adaptive radiation resistance. LncRNA:RNA and lncRNA:DNA interaction prediction software was employed to further uncover mechanisms of epigenetic regulation by lncRNAs. A subset of the lncRNA transcripts were predicted to interact directly with other coding transcripts while other lncRNAs have predicted interactions with DNA in regulatory regions of the human genome. We are further validating lncRNA:DNA interactions through in silico methods to determine if genes proximal to these interaction sites are related to tumorigenesis or therapy response. Semantic linkages have been predicted from references in the current literature for lncRNAs and key molecular processes such as DNA damage response, extracellular matrix, and chromatin remodeling. Modules of highly correlated genes, including lncRNAs, have been identified that are differentially regulated between radiation sensitive and resistant tumors. Cross-validation between our in silico approaches have confirmed high-confidence associations of a subset of lncRNAs with acquired radiation resistance. Subsequently, we will validate the expression and phenotypic relevance of high-confidence transcripts in our cohort of PDX models derived from recurrent GBM tumors. In conclusion, recurrence of therapy resistant GBM is responsible for patient mortality. Not all patients qualify for surgical resection or for chemotherapy, but radiation is almost a universally tolerated therapy. We have generated a novel model of tumor recurrence and radiation resistance that recapitulates recurrent tumor physiology. We have discovered 269 lncRNAs associated with radiation resistance that suggests potential regulatory mechanisms through nucleic acid binding. Evidence suggests that lncRNAs most likely contribute to acquired resistance through epigenetic regulation of transcription and chromatin state. These transcripts may provide novel, druggable targets for treating therapy resistant, recurrent GBM. Citation Format: Christian Tyler Stackhouse, James R. Rowland, Jelai Wang, Thanh Nguyen, Zongliang Yue, Jake Y. Chen, Lara Ianov, G. Yancey Gillespie, Christopher D. Willey. Long non-coding RNAs in glioblastoma tumor recurrence and therapy resistance [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 279.