EXTH-51. DEVELOPMENT OF A NOVEL GENE THERAPY APPROACH TARGETING GLIOBLASTOMA FOLLOWING ARTIFICIAL INTELLIGENCE (AI)-DIRECTED IDENTIFICATION OF THE GBM STATE

Abstract

Listen

Translate article

Abstract BACKGROUND Profound heterogeneity has severely hampered therapeutic advancements in GBM. Remarkably, however, GBM exhibits broad clinical and histopathologic overlaps suggesting the presence of a common state. The GBM state embodies network restructuring forced by founding mutations and perpetuated in subclones of GBM stem-like cells (GSCs). Successful targeting of the GBM state promises to circumvent the heterogeneity. METHODS To decipher the GBM state, we applied NETZEN, an AI suite integrating a deep neural network with gene network-based ranking, to first generate the reference GBM gene network from TCGA’s entire GBM RNAseq collection, and then identify the altered master regulatory gene subnetwork in GBM using a dataset containing >30 diverse patient-derived GSC lines and their paired differentiated cells, 6 astrocyte and 3 neuronal precursor cell lines. To develop a gene therapy against the GBM state, we screened a rAAV capsid library through GBM patient-derived xenografts (PDX) to identify variants with specific tropism for GBM cells that can deliver targeting constructs intratumorally. RESULTS We discovered a putative GBM state anchored by developmentally restricted master regulators. To validate its critical role, we deconstructed it using shRNA in a panel of PDX and uniformly observed improved tumor control and survival compared to controls (p< 0.05 in all lines). More notably, when the core GBM state was forcibly reconstructed in astrocytes, transformation into GSC-like cells occurred, as measured by single-cell analysis, neurosphere formation, and most importantly, development of lethal infiltrating brain tumors in 15/15 mice. Finally, selected novel rAAV capsids with 10-40-fold higher specificity for GBM cells were utilized in a shRNA-based rAAV platform to target key master regulators of the validated GBM state. CONCLUSIONS The GBM state is established by a developmental master regulator subnetwork permitting the creation of a first-of-its-kind, heterogeneity-agnostic GBM therapy. This AI-directed R&D program can be expanded to target other tumors.