Investigating circuit and spatial heterogeneity of human brain tumors with patient-derived organoid models

Malignant gliomas, including glioblastoma (GBM), are aggressive adult primary brain cancers with dismal prognosis. Both malignant cell heterogeneity and the influence of myriad cell types present in the tumor microenvironment, such as neurons, contribute to glioma progression. This dissertation represents a multi-pronged effort to unravel GBM heterogeneity by leveraging patient-derived GBM organoids (GBOs), a 3-dimensional model inspired by human brain organoids. In the first sections of this dissertation, I describe the application of circuit neuroscience techniques to study diverse tumor-neuron circuitry. Previous studies discovered a direct synaptic input to glioma cells via glutamatergic neurons, although their locations and whether other neuronal types may project to glioma remain unclear. We leveraged rabies virus and herpes simplex virus (HSV)-based transsynaptic tracing techniques in the rodent brain to define a whole-brain connectome of GBM-projecting neurons. We found that, remarkably, GBM receives inputs from all over the brain and are innervated by diverse neural subtypes, such as modulatory neurons. We demonstrated the existence of a cholinergic neuron-to-glioma synapse functionally and morphologically. Acetylcholine augments diffuse tumor invasion, and inhibition of the metabotropic receptor CHRM3 represents a potential therapeutic strategy for GBM. We also established an induced pluripotent stem cell (iPSC)-derived co-culture model with GBOs, and we provide a detailed methodological primer for applying circuit tracing techniques to the study for cancer. In the second branch of the dissertation, I discuss the development of a next-generation giant GBO (gGBO) model to study the spatial heterogeneity of GBM. While profiling studies have delineated a set of cellular states within GBM that exist in recurring spatial niches, the ability to model these architectures is not possible with current approaches. We demonstrate that gGBOs capture the spatial structure of primary GBM, that presence of a hypoxia gradient is necessary and sufficient to organize these states along a neurodevelopmental axis, and finally that gGBOs represent an emerging model to study therapeutic responses. Overall, our works illustrate how GBM co-opts neuronal circuits and neurodevelopmental niche mechanisms to drive progression.

Author dc:creator

• Sun, Yusha

Advisor dc:contributor.advisor

• Song, Hongjun

• Neuroscience and Neurobiology

Language dc:language.iso

en

Repository record dc:identifier.uri

https://repository.upenn.edu/handle/20.500.14332/62340

OAI identifier oai:identifier

oai:repository.upenn.edu:20.500.14332/62340