No abstract

No abstract

The precise decoding of human genomes facilitated by the advancements in next-generation sequencing has led to a better understanding of genetic underpinnings of pediatric brain cancers. Indeed, it is now evident that tumours of the same type harbour distinct driving mutations and molecular aberrations that can result in different prognosis and treatment outcomes. The profounder insight into the the identity, amount and types of molecular aberrations has paved the way for the advent of targeted therapies in precision medicine. Nevertheless, less than 10% of pediatric cancer patients harbour actionable mutations. Strictly limited therapeutic options that are firstly available for brain cancers and secondly acceptable for children’s development further impede the breakthrough in the survival rate in pediatric brain cancers. This underscores a desperate need to delve beyond genomic sequencing to identify biomarker coupled therapies that not only featured with treatment efficacy in the central nervous system but also acceptable side effects for children. The Hudson-Monash Paediatric Precision Medicine (HMPPM) Program focuses on utilising genetic profiles of patients’ tumour models to identify new therapeutic targets and repurpose existing ones using high-throughput functional genomics screens (2220 drugs and CRISPR screen of 300 oncogenic genes). Using a large compendium of over sixty patient derived paediatric brain cancer models, we provide proof-of-concept data that shows an integrative pipeline for functional genomics with multi-omics datasets to perform genotype-phenotype correlations and, therefore, identify genetic dependencies. Herein, using several examples in ATRT, DIPG and HGG, we show how functional interrogations can better define molecular subclassification of tumours and identify unique vulnerabilities.

Cell lines represent the most versatile and widely used models of cancer and, as such, are critical for identifying and advancing new therapies. Strikingly, there is a significant gap in both the number of childhood brain cancer cell lines and their characterisation compared to their adult counterparts. To address this inequity, we established a childhood brain cancer cell line atlas (publicly available at vicpcc.org.au/dashboard) encompassing over 180 childhood CNS-derived cell lines, representing 20 tumour types and 11 molecular subtypes. Cell lines are characterized by whole genome, RNA-sequencing, phospho- and total proteomics, DNA methylation and ATAC-seq analyses. Multi-omic factor analysis revealed distinct lineage-specified classification of our cell line cohort. In parallel, high throughput drug and CRISPR/Cas9 screens were conducted to map the functional dependencies in over 70 childhood CNS cell lines, including 47 paediatric high grade glioma models. These screens identified both lineage and molecular-subtype specific genetic and drug dependencies, underscoring the utility of this wide-scale approach. Machine based learning approaches to predict genotype-phenotype correlations uncovered distinct paediatric-specific biomarkers of growth dependency, highlighting the unique genetic wiring underlying paediatric CNS tumours. Finally, by integrating functional, molecular and drug profiles of paediatric CNS cell lines, we construct a system to prioritize investigation of novel therapeutic target-biomarkers pairs in specific CNS tumour types.