Brain metastases (BMs) are the most common brain tumor in adults, developing in about 10% of adult cancer patients. It is not the incidence of BM that is alarming, but the poor patient prognosis. Even with aggressive treatments, median patient survival is only months. Despite the high rate of BM-associated mortality, very little research is conducted in this area. Lack of research and staggeringly low patient survival is indicative that a novel approach to BMs and their treatment is needed. The ability of a small subset of primary tumor cells to produce macrometastases is reminiscent of brain tumor-initiating cells (BTICs) or cancer stem cells (CSCs) hypothesized to form primary brain tumors. BTICs are considered stem cell-like due to their self-renewal and differentiation properties. Similar to the subset of cells forming metastases, BTICs are most often a rare subpopulation. Based on the functional definition of a TIC, cells capable of forming a BM could be considered to be brain metastasis-initiating cells (BMICs). These putative BMICs would not only have the ability to initiate tumor growth in a secondary niche, but also the machinery to escape the primary tumor, migrate through the circulation, and invade the neural niche.

Brain tumours represent the leading cause of childhood cancer mortality, of which medulloblastoma (MB) is the most frequent malignant pediatric brain tumour. Current molecular Nsubgroups of MB recognize distinct disease entities of which activated Wnt signaling (monosomy 6, exon 3 mutations in CTNNB1, and Wnt gene signature) is associated with a distinct subgroup and the best overall outcome. In contrast, only non-Wnt MBs are characterized by metastatic disease, increased rate of recurrence, and poor overall survivorship. Given the excellent clinical outcome in patients with Wnt-driven MB, we aimed to convert treatment-resistant MB subgroups into an ostensibly benign tumour through selective targeting by small molecule Wnt agonists (Wnt3A), GSK3 inhibitors (CHIR99021), and transgenic lines containing a stabilized beta-catenin mutant. Activated Wnt signaling resulted in decreased in vitro self-renewal and promoted differentiation within primary human MB stem cells. The clinical relevance of these findings were demonstrated with an in vivo survival advantage in mice containing orthotopic injections of cells containing a stabilized beta-catenin mutant representative of constitutively active Wnt signaling. Xenografts generated from Wnt-activated tumours were much smaller in size, maintained a much lower rate of proliferation, and reduction in key MB stem cell self-renewal genes (Bmi1, Sox2, Msi1, FoxG1). Our work establishes activated Wnt signaling as a novel treatment paradigm in childhood MB, while providing evidence for the context-specific tumour suppressive function of the canonical Wnt pathway.

Despite aggressive multimodal therapy, glioblastoma (GBM) remains incurable and inevitably relapses. Recent data have implicated intratumoral heterogeneity as the driver of therapy resistance and tumour relapse in GBM. Models that capture the evolution of GBM biology in response to standard-of-care (SoC) chemoradiotherapy will allow for the identification of cellular mechanisms that govern GBM therapy failure. In this study, we coupled cellular DNA barcoding technology with our novel patient-derived xenograft SoC model (combined temozolomide and radiation treatment) to profile the clonal evolution of GBM stem cells (GSCs) through therapy. We report the successful barcoding of patient-derived primary, treatment-naive GSCs at a single cell resolution that were expanded into clonal populations, intracranially engrafted in immune-deficient mice, and treated with SoC therapy. We performed MRI imaging to identify spatial recurrence patterns of GSCs through the in vivo chemoradiotherapy model. We then interrogated the temporal fate of clonal barcoded GSC populations through SoC therapy model to identify differential barcode selection in response to treatment. Through this, we determined dynamics patterns of a pre-existing or a therapy-driven GSC subpopulation(s) seeding GBM tumour relapse. Profiling the dynamic nature of heterogeneous GBM subpopulations through disease progression and SoC treatment may lead to the identification of the modes of therapy resistance utilized by GBM to drive disease relapse.

Glioblastoma (GBM) is a uniformly fatal primary brain tumor, characterized by extensive cellular heterogeneity. Numerous studies have implicated CD133+ brain tumor initiating cells (BTICs) as drivers of chemo- and radio-resistance in GBM. We have recently demonstrated that a CD133-driven gene signature is predictive of poor overall survival and targeting CD133+ treatment-refractory cells may be an effective strategy to block GBM recurrence. Chimeric antigen receptors (CARs) and bispecific T-Cell engaging antibodies (BiTEs) present promising immunotherapeutic approaches that have not yet been validated for recurrent GBM. Using CellectSeq, a novel methodology that combines use of phage-displayed synthetic antibody libraries and DNA sequencing, we developed the CD133-specific monoclonal antibody ‘RW03’. We constructed CD133-specific BiTEs that consist of two arms; one recognizes the tumor antigen (CD133) and the second is specific to CD3 antigen that bind to human GBMs and PBMC-derived T cells, respectively. We observed BiTEs redirecting T cells to kill GBMs, with greater efficiency observed in CD133^high GBMs, validating BiTE target specificity. Incubating T-cells with BiTEs and the CD133^high GBMs resulted in increased expression of T cell activation markers. In parallel, we derived the single chain variable fragment (scFv) from RW03, added a myc-tag and generated a second-generation CAR with ΔNGFR as a selection marker. CD133-specific CAR-T cells were cytotoxic to CD133+ GBMs. Co-culturing CD133 CAR-T cells with GBMs triggered T cell activation and proliferation. Treatment of GBM tumor-bearing mice with CD133-specific CAR-T cells yielded extended survival in mice and significant reductions in brain tumor burden. Furthermore, we uniquely adapted the existing chemoradiotherapy protocol for GBM patients for treatment of immunocompromised mice engrafted with human GBMs. Within this model, we have initiated treatment of recurrent GBM directed against CD133+ BTICs, to allow for a direct prospective comparison of toxicity and efficacy of BiTEs and CAR T cell strategies.