Role of Yap1 in adult neural stem cell activation

Most adult hippocampal neural stem cells (NSCs) remain quiescent with only a minor portion undergoing active proliferation and neurogenesis. The molecular mechanisms that trigger eventually the transition from quiescence to activation are still poorly understood. Here, we found the activity of the transcriptional activator Yap1 to be enriched in active NSCs. Genetic deletion of Yap1 led to a significant reduction in the relative proportion of active NSCs supporting a physiological role of Yap1 in regulating the transition from quiescence to activation. Overexpression of wild type Yap1 in adult NSCs did not induce NSC activation suggesting tight upstream control mechanisms, but overexpression of a gain-of-function mutant (Yap1-5SA) elicited cell cycle entry in NSCs and hilar astrocytes. Consistent with a role of Yap1 in NSC activation, single cell RNA sequencing revealed the partial induction of an activated NSC gene expression program. Yet, Yap1-5SA expression also induced Taz and other key components of the Yap/Taz regulon previously identified in glioblastoma stem cell-like cells. Consequently, dysregulated Yap1 activity led to repression of hippocampal neurogenesis, promoting aberrant differentiation instead.

MV1035 Overcomes Temozolomide Resistance in Patient-Derived Glioblastoma Stem Cell Lines

Glioblastoma (GBM, grade IV glioma) represents the most aggressive brain tumor and patients with GBM have a poor prognosis. Until now surgical resection followed by radiotherapy and temozolomide (TMZ) treatment represents the standard strategy for GBM. We showed that the imidazobenzoxazin-5-thione MV1035 is able to significantly reduce GBM U87-MG cells migration and invasiveness through inhibition of the RNA demethylase ALKBH5. In this work, we focus on the DNA repair protein ALKBH2, a further MV1035 target resulting from SPILLO-PBSS proteome-wide scale in silico analysis. Our data demonstrate that MV1035 inhibits the activity of ALKBH2, known to be involved in GBM TMZ resistance. MV1035 was used on both U87-MG and two patient-derived (PD) glioma stem cells (GSCs): in combination with TMZ, it has a significant synergistic effect in reducing cell viability and sphere formation. Moreover, MV1035 induces a reduction in MGMT expression in PD-GSCs cell lines most likely through a mechanism that acts on MGMT promoter methylation. Taken together our data show that MV1035 could act as an inhibitor potentially helpful to overcome TMZ resistance and able to reduce GBM migration and invasiveness.

SURG-05. Α5Β1 INTEGRIN TARGETING LIPOSOME DELIVERY OF MIR-603 SUPPRESSED GLIOBLASTOMA STEM CELL STATE AND ENHANCED RADIATION SENSITIVITY

INTRODUCTION Despite the potential for clinical efficacy, therapeutic delivery of microRNAs (miRNA) remains a major translational barrier. Here, we explore a surgery mediated polyethylenimine (PEI)/liposome-based strategy for the delivery of miR-603, a master regulatory miRNA that suppresses glioblastoma stem cell state by simultaneous down-regulation of insulin-like growth factor 1 (IGF1) and IGF1 receptor (IGF1R). METHODS miR-603 was complexed with PEI, a cationic polymer designed to facilitate miRNA from the endolysosomal compartment. The miR-603/PEI complex was encapsulated into liposomes decorated with polyethylene glycol (PEG) and PR_b, a fibronectin-mimetic peptide that specifically targets the α5β1 integrin that is overexpressed in glioblastoma. RESULTS Patient-derived glioblastoma cells internalized PR_b coated liposomes but not the non-coated liposomes. The internalization of PR_b liposomes encapsulating miR-603/PEI was associated with orders of magnitude increase in intra-cellular miR-603 levels and decreased IGF1 and IGF1R mRNA/protein levels. Moreover, treatment of glioblastoma cells with the PR_b liposomes encapsulating miR-603/PEI showed altered morphology and decreased expression of stem cell marker, suggesting the treated cells have exited the cancer stem cell state. Finally, treatment of the PR_b liposomes encapsulating miR-603/PEI sensitized glioblastoma cells to ionizing radiation (IR) in patient-derived glioblastoma cells. These results were not observed in liposomes missing the PR_b peptide, PEI, or miR-603. CONCLUSION These results suggest that intra-tumoral injection of PR_b functionalized PEGylated liposomes encapsulating miR-603/PEI complexes hold promise as a strategy for glioblastomas therapy. A first-in-human trial is currently underway to test this strategy.

STEM-09. DEFINING THE ROLE OF CD97 IN GLIOBLASTOMA STEM CELL SELF-RENEWAL

Glioblastoma (GBM) is the most common and aggressive primary brain malignancy. Despite multimodal therapy, resistant GBM stem-like cells (GSCs) inevitably mediate disease recurrence. To identify novel vulnerabilities of GSCs, we performed an arrayed CRISPR/Cas9 screen against select adhesion G protein-coupled receptors (aGPCRs), many of which we found to be de novo expressed in GBM. Knockout of CD97 (ADGRE5), previously implicated in GBM cell migration, produced the most striking proliferative disadvantage in patient-derived GBM cultures (PDGC) among aGPCRs tested. We found high CD97 surface expression in all our PDGCs, while levels remained nearly undetectable in non-neoplastic brain cells, confirming that CD97 is de novo expressed in GBM. Upon shRNA-mediated knockdown of CD97 in PDGCs from all three TCGA transcriptional subtypes, we observed significantly reduced proliferation, as measured by Ki67 and Hoechst cell cycle analysis, and significantly diminished surface expression of CD133, a GSC marker. Notably, CD97 knockdown also significantly reduced tumorsphere initiation capacity in six PDGCs, as measured by extreme limiting dilution assays. These findings suggest that CD97 regulates GSC self-renewal in vitro. RNA-sequencing and GSEA pathway analysis from PDGCs following CD97 knockdown indicate an enrichment of aerobic respiratory gene sets, suggesting one of the major regulatory roles of CD97 is metabolic regulation. Indeed, metabolic assays show that CD97 knockdown alters oxygen consumption and glycolysis rates in PDGCs. Lastly, we have developed human synthetic antibodies to target CD97 in order to investigate its therapeutic potential. We have observed internalization of some of these antibodies, thus identifying candidates for the development of antibody-drug conjugates. In addition, other clones reduced GBM cell proliferation and elicited expression of various differentiation markers. Overall; our studies identify novel roles of CD97 in regulating the cellular hierarchy in GBM and tumor cell metabolism, and provide a strong scientific rationale for developing biologics to target CD97 in GBM.

CSIG-13. A DYNAMIC CAUSAL MODEL OF GLIOBLASTOMA STEM CELL SIGNALING PREDICTS EFFECTS OF KINASE INHIBITORS

Glioblastomas and glioblastoma stem cells are heterogeneous with respect to mutations, gene expression, and response to drugs. To make predictive responses of individual GBM stem cell lines to kinase inhibitors, we have constructed a causal model of glioblastoma stem cell signaling. The core model was built starting from pathways identified from TCGA mutation data with the addition of the Jak/STAT, Hedgehog, and Notch pathways. Elements and relations between them were validated and extended using the PCNet interaction database and the INDRA database which includes machine read extractions from the biomedical literature. The result is a high confidence executable model consisting of 209 elements (proteins, genes, RNAs) and 370 regulatory logic rules between the elements. Stochastic simulations of the model provide dynamic (quantile) changes in time and responses to perturbations. The output simulates activity of individual nodes as well as cell cycle progression, apoptosis, and differentiation. To simulate the responses of individual cell lines to kinase inhibitors, the model was initialized using DNA sequencing data, RNA-seq, and reverse phase protein array (RPPA) data from each cell line. Comparing the results of the simulations to the drug responses of 11 different kinase targets in 3 cell lines, the model was 88% accurate in predicting effects on growth and survival. The model was further tested by comparing the effects of Mek inhibition of each of the cell lines in the model to the results observed in the RPPA data which overlap by 127 elements. In this case, there was less than 65% concordance between the model and the data for individual nodes. Discrepancies between the model predictions and the data are being investigated to determine whether the model logic or extent needs to be revised to improve the model. This modeling approach is a step toward developing algorithms for personalized therapeutics for GBM.

SLUG and Truncated TAL1 Reduce Glioblastoma Stem Cell Growth Downstream of Notch1 and Define Distinct Vascular Subpopulations in Glioblastoma Multiforme

Glioblastomas (GBM) are high-grade brain tumors, containing cells with distinct phenotypes and tumorigenic potentials, notably aggressive and treatment-resistant multipotent glioblastoma stem cells (GSC). The molecular mechanisms controlling GSC plasticity and growth have only partly been elucidated. Contact with endothelial cells and the Notch1 pathway control GSC proliferation and fate. We used three GSC cultures and glioma resections to examine the expression, regulation, and role of two transcription factors, SLUG (SNAI2) and TAL1 (SCL), involved in epithelial to mesenchymal transition (EMT), hematopoiesis, vascular identity, and treatment resistance in various cancers. In vitro, SLUG and a truncated isoform of TAL1 (TAL1-PP22) were strongly upregulated upon Notch1 activation in GSC, together with LMO2, a known cofactor of TAL1, which formed a complex with truncated TAL1. SLUG was also upregulated by TGF-β1 treatment and by co-culture with endothelial cells. In patient samples, the full-length isoform TAL1-PP42 was expressed in all glioma grades. In contrast, SLUG and truncated TAL1 were preferentially overexpressed in GBMs. SLUG and TAL1 are expressed in the tumor microenvironment by perivascular and endothelial cells, respectively, and to a minor extent, by a fraction of epidermal growth factor receptor (EGFR) -amplified GBM cells. Mechanistically, both SLUG and truncated TAL1 reduced GSC growth after their respective overexpression. Collectively, this study provides new evidence for the role of SLUG and TAL1 in regulating GSC plasticity and growth.

P13.16 Metastatic potential of systemic glioblastoma stem cell lines in vivo

BACKGROUND Despite aggressive tumor behavior, extracranial metastases rarely develop in glioblastoma (GBM) patients. Two potential explanations have been suggested: 1) The blood-brain-barrier functions as a physical barrier that prevents the dissemination of GBM cells out of the central nervous system (CNS) or 2) that extracranial metastasis do occur, but the patients die before extracranial metastases manifest themselves. The first theory has been questioned based on the fact that circulating tumor cells (CTC) were found in blood samples of GBM patients without systemic metastases. To date it has not been proven if CTCs are able to reenter the brain and to what extent they are able to form systemic extracranial metastatic lesions. Therefore, the current study aimed at analyzing the dissemination patterns and the underlying mechanisms associated with the ability of GBM CTCs to form extracranial metastases. MATERIAL AND METHODS Five highly characterized human GBM stem cell (GSC) lines (P3, BG5, BG7, GG6, GG16), displaying GBM CNV patterns, were intracranially implanted in a first cohort, then transduced with a lentiviral Firefly Luciferase-eGFP vector and injected into the left cardiac ventricle of NOD/SCID mice in a second cohort. Mice were observed closely and tumor burden was assessed using in vivo as well as ex vivo bioluminescence imaging, MRI and PET. Mice were euthanized when the objective endpoint criteria (tumor burden) was met, then organs were harvested and fixed for further analysis. RESULTS First, a detailed characterization of the GSC line invasion patterns were assessed when grown as orthotopic xenografts in vivo dividing them into three categories: 1) Highly invasive without apparent angiogenesis (BG5) 2) Invasive with perivascular infiltration and angiogenesis (P3, BG7 and GG16) and 3) Angiogenic and highly circumscribed (GG6). Following intracardial injection, (7 out of 8) P3 animals developed extracranial and intracranial tumors with a distinctive pattern. Brain, adrenal gland, ovary and liver were amongst the organs most susceptible for tumor growth in the P3 group. For the BG5 and BG7 cell lines, no metastases were observed whereas only 1 animal out of 10 developed metastases in both groups GG16 and GG6. CONCLUSION Only one out of 5 GSC lines exhibited a strong metastatic potential when injected into the left cardiac ventricle. Compared to other tumors which exhibit a strong metastatic potential from the circulation, GSC lines do only to a very limited extent show this potential reflecting observations made in the clinic.

TRIM28 Selective Nanobody Reduces Glioblastoma Stem Cell Invasion

Glioblastoma (GB), is the most common and aggressive malignant primary brain tumour in adults. Intra- and inter-tumour heterogeneity, infiltrative GB cell invasion and presence of therapy-resistant GB stem cells (GSCs) represent major obstacles to favourable prognosis and poor therapy response. Identifying the biomarkers of the most aggressive tumour cells and their more efficient targeting strategies are; therefore, crucial. Recently, transcription factor TRIM28 has been identified as a GB biomarker and, in this study, we have shown high expression of TRIM28 in GB and in low grade gliomas as well as higher expression in GSCs vs. differentiated GB cells, although in both cases not significant. We demonstrated significant in vitro inhibition of GB cells and GSCs invasiveness and spread in zebrafish brains in vivo by anti-TRIM28 selective nanobody NB237. TRIM28 was also enriched in GB (tumour) core and associated with the expression of stem cell genes, but was not prognostic for overall survival. However, based on the above results, we conclude that TRIM28 nanobody NB237 offers a new opportunity as a GB therapeutic tool.

NGMA-3. Use of multi-omics data to initialize and validate a causal model of glioblastoma stem cell signaling

Glioblastomas and glioblastoma stem cells are heterogeneous with respect to mutations, gene expression, and response to drugs. To make predictive responses of individual GBM stem cell lines to drugs, we have constructed a causal model of glioblastoma stem cell signaling. The core model was built starting from pathways identified from TCGA mutation data with the addition of the Jak/STAT, Hedgehog, and Notch pathways. Elements and relations between them were validated and extended using the PCNet interaction database and the INDRA database which includes machine read extractions from the biomedical literature. The result is a high confidence executable model consisting of 209 element and 370 rules of interaction between the elements. Stochastic simulations of the model provide dynamic (quantile) changes in time and responses to perturbations. The output provides activity of individual nodes as well as a cellular output state of cell cycle progression, apoptosis, or differentiation. To simulate the responses of individual cell lines to kinase inhibitors, the model was initialized using DNA sequencing data, RNA-seq, and reverse phase protein array (RPPA) data from each cell line. Comparing the results of the simulations to the drug responses of 11 different kinase targets, the model was 88% accurate in predicting effects on growth and survival. The model was further tested by comparing the effects of Mek inhibition of each of the cell lines in model to the results observed in the RPPA data which overlap by 127 elements. In this case, there was 62% concordance between the model and data when binned into quintiles. Discrepancies between the model predictions and the data are being investigated to determine whether the model logic or extent needs to be revised to improve the model. This modeling approach is a step toward developing algorithms for personalized therapeutics for GBM based on multi-omics data.

AmotL2, IQGAP1, and FKBP51 Scaffold Proteins in Glioblastoma Stem Cell Niches

Glioma stem cells (GSCs) live in a continuous process of stemness reprogramming to achieve specific cell commitment within the so-called GSC niches, specifically located in periarteriolar regions. In this review, we analyze the expression levels, cellular and subcellular location, and role of three scaffold proteins (IQGAP1, FKBP51, and AmotL2) in GSC niches. Scaffold proteins contribute to cell differentiation, migration, and angiogenesis in glioblastoma. It could be of diagnostic interest for establishing stages, for therapeutic targets, and for improving glioblastoma prognosis, which is still at the experimental level.