Drug and single-cell gene expression integration identifies sensitive and resistant glioblastoma cell populations [RNA-Seq]

Glioblastoma (GBM) remains the most common and lethal adult malignant primary brain cancer with few treatment options. One of the most significant issues hindering GBM therapeutic development is intratumor heterogeneity. GBM tumors consist of multiple neoplastic cell types with distinct transcriptional profiles. Identifying compounds that target these different cell types requires a platform that predicts differential sensitivity and resistance of cells within GBM tumors. To address this, we developed a novel framework to generate tumor cell-type-specific disease signatures and their predicted drug sensitivity. We performed single-cell RNA sequencing of newly diagnosed and recurrent GBM tumors and identified compounds from the LINCS L1000 database with transcriptional response signatures selectively discordant with distinct single-cell transcriptional states within GBM tumors. We then validated in silico predictions with in vivo findings in mice as the aurora kinase inhibitor alisertib induced a shift from NPC-like to MES-like transcriptional states upon treatment. To allow other investigators to predict selective tumor cell type drug sensitivity in silico we developed a platform termed ISOSCELES (Inferred cell Sensitivity Operating on the integration of Single-Cell Expression and L1000 Expression Signatures). Our studies suggest that ISOSCELES is a novel platform to identify cell populations that are sensitive and resistant to targeted therapies in GBM, which can inform treatment decisions in future clinical trials.