Human Brain Organoids for Studying Malignant Cell States and Intercellular Communication in Human Glioma

Abstract

Human glioma is an incurable cancer of the central nervous system. Recent advances in glioma biology have highlighted the heterogeneity and inter-cellular communication within these tumors; it follows that appropriate models for studying glioma behavior and progression – and, in turn, therapeutic avenues – must adequately recapitulate these key features. Indeed, patientderived tumor xenografts (PDXs) are, for this reason, attractive and widely used in vivo model systems for studying gliomas, despite their limitations. The development of complementary (and currently non-existent) in vitro glioma models that better capture the molecular and phenotypic spectrum of the corresponding human tumor would enable reliable disease modeling and therapeutic testing at unprecedented scale and spatiotemporal resolution, potentially leading to much-needed breakthroughs for the field.

The compartmentalization and emergent phenotypes of human gliomas are determined, in large part, by cooperative interactions between the intrinsic features of malignant cells and the tumor microenvironment. In this regard, a fundamental limitation of current in vitro glioma models (e.g., gliomaspheres) is the lack of appropriate environmental cues, leading to a prohibitively reductionist or skewed representation of the disease. In recent years, human brain organoids have emerged as promising 3D, in vitro model systems for partially recreating the cellular composition and function of the human brain. In the context of this research, human brain organoids represent a potential construct through which to provide 3D, human-specific environmental cues to patient-derived glioma cells, at once addressing a significant limitation of current in vitro glioma models.

In this thesis, we describe our efforts to develop glioma-brain organoid models for a variety of glioma subtypes and applications. In the first section, we show technological feasibility of growing pediatric and adult glioma-brain organoid models, including those involving otherwise intractable IDH-mutant gliomas. In the second section, we focus on IDH-WT glioma and show that human brain organoids induce a spectrum of malignant cell states that are more faithful to human glioma than matched gliomasphere models. Finally, in the third section, we show evidence of intercellular communication between malignant and non-malignant cells in gliomabrain organoid models, demonstrating a functionally integrated tumor microenvironment. Collectively, this thesis represents a major advance in the in vitro modeling of human glioma for eventual therapeutic testing.