BIOL-07. DISTINCTIVE FEATURES OF HIGH-GRADE GLIOMA MOUSE MODELS REVEALED BY SINGLE-NUCLEUS RNA-SEQUENCING GUIDE PRE-CLINICAL MODEL SELECTION

Abstract

Abstract Glioma is the most common pediatric central nervous system tumor, with high-grade gliomas (HGG) having one of the worst prognoses of all human cancers. In order to develop better diagnostics and therapies, it is essential to use faithful disease models. Currently, highly passaged patient-derived xenograft (PDX) models are most widely used in preclinical studies, although alternative immunocompetent models are becoming increasingly available. Here, we compare several in vivo glioma models on a single-cell level and investigate their similarity to primary human tumors. Single-nucleus sequencing was used to analyze >125,000 nuclei of primary patient samples, xenografts, autochthonous mouse tumors and related allografts including early and late in vivo passages – with a focus on MET-fusion-driven as well as H3 K27M-mutant HGG. Transcriptomic profiles of single tumor cells and associated stromal/immune components reveal insights into model-specific intratumoral heterogeneity, tumor evolution, and similarities between mouse models and patient samples. In addition, matched tumors engrafted into immunocompromised and immunocompetent animals are used to examine tumor-immune crosstalk and the modifying role of the tumor microenvironment. This improved understanding of how the analyzed model systems evolve over time and how they reflect patient’s tumor compositions will allow to prioritize and refine modern preclinical trials.