Preclinical models to study the impact of the blood-brain barrier in brain tumor chemotherapy

High-grade gliomas, in particular Glioblastoma Multiforme (GBM), are the most common primary brain tumors in adults and among the deadliest of human cancers. Their location and the extensively infiltrative character of tumor cells into surrounding normal brain structures is an impediment for all therapeutic interventions. The blood-brain barrier (BBB) is of pivotal importance to maintain homeostasis of the central nervous system (CNS) as it closely regulates the composition of the interstitial fluid in the brain, thereby protecting this delicate organ against the influence of harmful toxic substances. Unfortunately, malignancies that grow within the CNS may evade chemotherapeutic drugs using the same barrier, making this disease refractory to most chemotherapy regimens. Inadequate drug entry into the brain is partly mediated by drug efflux transporters localized in the BBB, which transport virtually all drugs back into the bloodstream. Besides the well described drug efflux transporter P-glycoprotein (P-gp, MDR1, ABCB1), this thesis demonstrates the important role of the more recently identified drug efflux transporter Breast Cancer Resistance Protein (BCRP, ABCG2), as another gate-keeper next to P-gp in the BBB by using P-gp-/-Bcrp-/- knockout mice. Absence or inhibition of both P-gp and BCRP results in increased brain penetration of cytotoxic (topotecan and temozolomide) and molecular-targeted (erlotinib) anticancer agents. Importantly, the improved brain penetration of temozolomide translated into a significantly better antitumor response in an experimental intracranial tumor model. The majority of preclinical studies are performed in xenografts, in which established mouse and human glioma derived cell lines are implanted into the flank or brain of immunodeficient mice. Although these models are standardized and generate reproducible tumors, they do not appropriate recapitulate many of the features that characterize high-grade gliomas and have therefore been less predictive of human response to drugs. More recently, genetically engineered mouse (GEM) models of glioma have been generated. Somatic GEM models are based on defined genetic alterations and result in spontaneous tumor formation in a more or less normal microenvironment in immunocompetent mice, and resemble many features of the human disease. Although these models are very valuable, they have not yet been exploited for preclinical studies, partly because of the lack of imaging modalities to diagnose tumors in a timely fashion. We therefore generated a new high-grade glioma models using a replication defective, self-deleting lentiviral vector that drives Cre-expression under the glial fibrillary acidic protein (GFAP) promoter. Deletion or activation of genes was induced by stereotactic intracranial injections of the lentivirus into compound Cre-LoxP conditional Ink4a/Arf;K-rasV12, Pten;Ink4a/Arf;K-rasV12 and p53;Ink4a/Arf;K-rasV12 adult mice and resulted in formation of high-grade gliomas WHO grade III anaplastic astrocytoma and grade IV GBM). Because these mice also carry a conditional Luciferase Reporter gene, tumor inducement and response can be visualized non-invasively using bioluminescence imaging, thereby making this model more suitable for intervention studies