A PRE-CLINICAL MODEL OF GLIOBLASTOMA RECURRENCE TO IDENTIFY PERSONALIZED THERAPEUTIC TARGETS

Abstract

Glioblastoma (GBM) is the most common and lethal primary tumour affecting the central nervous system in adults. Despite aggressive, multi-modal treatment consisting of surgical resectioning of the tumour followed by radiotherapy and chemotherapy, GBM remains incurable. Almost all patients experience relapse 7-9 months post-diagnosis and median survival has not extended beyond 15 months over the past decade. Extensive research in the molecular and cellular biology of GBM has revealed extensive inter- and intra-tumoural heterogeneity caused by dysregulation at genomic, epigenomic, transcriptomic and proteomic levels. Although this has led to the identification of molecular targets for therapeutic development, large body of GBM research has focused on the study of primary GBM, with little exploration of the biological landscape of recurrent GBM. Recent genomic studies suggest that recurrent GBM evolves significantly during the course of therapy and represents a distinct biological entity and therefore therapies developed based on primary GBM biology will not present efficacy against recurrent GBM. Thus, I postulate that models that capture the evolution of GBM biology in response to standard-of-care (SoC) chemoradiotherapy will allow for the identification of therapeutic targets specific to recurrent GBM and can be used for personalized medicine.

Here I show the development of an in vitro and in vivo model of GBM recurrence that can be used as a surrogate to identify personalized therapeutic targets for recurrent GBM. We use established cancer stem cell models combined with patient-derived glioblastoma stem cells (GSC) to profile and characterize the evolution of GBM through in vitro and in vivo adapted SoC. Through our in vitro model, I identified that combined chemoradiotherapy leads to increased sphere formation capacity of GBM and the global gene expression profiling of treatment-refractory GBM populations identified a poor-prognostic subtype of GBM. Next, I used patient-derived recurrent GBM to identify tyrosine kinases EphA2 and EphA3 as therapeutic targets in recurrent GBM and developed a bispecific antibody to co-target these receptors for therapeutic benefit. Lastly, I show the establishment of a novel patient-derived xenograft SoC model to profile the clonal evolution of GSCs through therapy. I show that this model can be coupled with multiple technologies, such as single cell RNA-sequencing and cellular DNA barcoding, to characterize the minimal residual cellular populations driving recurrence and identify personalized therapeutic targets for the treatment of GBM recurrence.

Altogether my thesis highlights the importance of developing clinically relevant models of GBM recurrence and using poly-targeting approaches for the treatment of recurrent GBM.