Summary Glioma, the most common brain tumor in adults, develops as a result of aberrant growth and invasion of astrocytic tumor cells. Even with aggressive treatment, survival is very poor and is attributed to the presence of therapy-resistant tumor-initiating cells (TICs), which are highly migratory and invasive and thus render complete surgical tumor removal impossible. Engineering therapies that target glioma tumor cells and TICs may enable enhanced efficacy and as a result longer clinical survival times in patients afflicted with this disease. Accordingly, this proposal is focused on the development of gene therapy strategies for glioblastoma multiforme (GBM), an aggressive form of glioma, based on the targeting of GBM tumor cells and TICs. Adeno-associated virus (AAV) has emerged as a safe and promising vector for gene delivery applications. However, viral vectors in general, and AAV in particular, do not display strong intrinsic cell tropism for glioma cells in the central nervous system (CNS), and in addition they experience a number of delivery and transport barriers for systemic delivery to clinical GBM, including biodistribution to the CNS, the blood brain barrier, and intraparenchymal and intratumoral transport to the primary and diffuse secondary tumors. Thus, it is highly desirable to develop vectors that can be systemically delivered and that are capable of overcoming these delivery barriers. We propose to engineer the coat proteins of AAV to target delivery to glioma tumor cells and TICs to greatly enhance delivery efficiency and reduce any biological off-target effects. We hypothesize that AAV directed evolution, a strategy we originally developed and have successfully employed to enhance viral vector properties, can be implemented to engineer AAV vectors in vivo for enhanced and potentially selective tropism for GBM tumor cells and TICs. Specifically, we propose to harness (1) a mouse model based on the xenografting of primary cultured, patient-derived GBM TICs that accurately represents the hallmarks of GBM, (2) highly diverse AAV vector libraries, and (3) a sophisticated directed evolution strategy that includes a stringent in vivo selection selective for viral particles that can localize to the CNS and transduce GBM tumor cells and TICs. We have successfully recovered viral genomes from the first round of evolution, highlighting the potential of this strategy. We also propose to characterize the resulting engineered AAV vectors by studying their tropism and biodistribution, essential gene delivery properties for clinical implementation. Furthermore, we propose to evaluate the therapeutic potential of engineered AAVs by delivering two promising therapeutic genes that can hamper tumor progression and extend the survival of our animal models, or that offer promise in future exploration of cancer immunotherapies. This blend of molecular virology, protein engineering, and a translationally accurate animal model will therefore enable the engineering of enhanced genetic delivery systems for the treatment of glioblastoma multiforme and in the future potentially other cancers.

PROJECT NARRATIVE Adeno-associated viral (AAV) vector mediated gene therapy has been clinically successful in treating an increasing number of diseases. Furthermore, it has been studied in preclinical models of cancer and is entering into the clinic for these applications. However, there are in general major delivery barriers that hinder the extension of gene therapies to glioblastoma multiforme (GBM), a particularly aggressive form of glioma with very poor survival prognosis due in part to the persistence of therapy-resistant, highly invasive tumor-initiating cells (TICs) that preclude complete surgical tumor removal. To address these challenges, we propose to apply our novel directed evolution approach to engineer optimized AAV vectors capable of highly efficient delivery of both marker and therapeutic genes to human GBM tumor cells and TICs in a murine xenograft model, work with implications for the development of next generation cancer therapies.