Therapeutic Target Discovery and Development of a Novel Patient-Derived Model in High-Risk Medulloblastoma

Abstract

Medulloblastoma (MB) is the most common malignant pediatric brain tumor. MB arises from the cerebellum during embryogenesis and is subdivided into four groups; WNT, SHH, Group 3 (G3) and Group 4 (G4). These groups have distinct molecular features and are further subdivided into multiple subtypes which present distinct clinical outcomes. Treatment options are limited to standard of care (SoC), consisting of tumor resection, chemotherapy and cranio-spinal irradiation. Advancements in SoC have greatly favored patient outcomes, although 30% of patients still succumb to the disease. These patients are termed high-risk medulloblastoma and mostly consist of G3-MB and G4-MB, display metastasis upon diagnosis and have the highest percentage of relapse. Improved treatment options are urgently warranted for high-risk MB patients. G3-MB has established pre-clinical (in vitro and in vivo) models, whereas G4-MB pre-clinical models are currently lacking. In order to discover novel therapeutic targets in G3-MB, I firstly employ an integrated target discovery platform, combining comparative liquid-chromatography mass-spectrometry and a genome-wide CRISPR KO screen to uncover lipidomic vulnerabilities. I discovered a prominent role of the de novo lipogenesis (DNL) pathway in G3-MB, with Stearoyl-CoA Desaturase 1 (SCD1) as a prominent therapeutic target and prognostic marker. Next, I utilize CUT&RUN-seq and RNA-seq to study the role of the epigenetic master-regulator BMI1 within the context of G3-MB. By overlapping these datasets, I uncover differentiation programs that are suppressed by BMI1 and may be exploited for co-treatment together with BMI1 inhibition. Lasty, I generated a novel G4-MB model. This model functions both as an in vitro cell line and as a patient-derived xenograft model (PDX). I demonstrated, with this model, that extrapolating G3-MB specific therapies can be a rational approach for G4-MB target discovery, indicating that this model is able to facilitate preclinical G4-MB research.