Glioblastomas are invasive brain tumors with high therapeutic resistance. Neuron-to-glioma synapses have been shown to promote glioblastoma progression. However, a characterization of tumor-connected neurons has been hampered by a lack of technologies. Here, we adapted retrograde tracing using rabies viruses to investigate and manipulate neuron-tumor networks. Glioblastoma rapidly integrated into neural circuits across the brain, engaging in widespread functional communication, with cholinergic neurons driving glioblastoma invasion. We uncovered patient-specific and tumor-cell-state-dependent differences in synaptogenic gene expression associated with neuron-tumor connectivity and subsequent invasiveness. Importantly, radiotherapy enhanced neuron-tumor connectivity by increased neuronal activity. In turn, simultaneous neuronal activity inhibition and radiotherapy showed increased therapeutic effects, indicative of a role for neuron-to-glioma synapses in contributing to therapeutic resistance. Lastly, rabies-mediated genetic ablation of tumor-connected neurons halted glioblastoma progression, offering a viral strategy to tackle glioblastoma. Together, this study provides a framework to comprehensively characterize neuron-tumor networks and target glioblastoma.
Glioblastomas are incurable tumors infiltrating the brain. A subpopulation of glioblastoma cells forms a functional and therapy-resistant tumor cell network interconnected by tumor microtubes (TMs). Other subpopulations appear unconnected, and their biological role remains unclear. Here, we demonstrate that whole-brain colonization is fueled by glioblastoma cells that lack connections with other tumor cells and astrocytes yet receive synaptic input from neurons. This subpopulation corresponds to neuronal and neural-progenitor-like tumor cell states, as defined by single-cell transcriptomics, both in mouse models and in the human disease. Tumor cell invasion resembled neuronal migration mechanisms and adopted a Lévy-like movement pattern of probing the environment. Neuronal activity induced complex calcium signals in glioblastoma cells followed by the de novo formation of TMs and increased invasion speed. Collectively, superimposing molecular and functional single-cell data revealed that neuronal mechanisms govern glioblastoma cell invasion on multiple levels. This explains how glioblastoma’s dissemination and cellular heterogeneity are closely interlinked.
Abstract Incurable gliomas are characterized by their infiltration into the whole brain. Recently, we described tumor microtubes as a novel structure contributing to glioma cell invasion and uncovered synaptic contacts on glioma cells that drive brain tumour progression. However, the exact effects of neuronal activity on glioma cell motility are yet unclear. Here, we show how a recently described neuronal-like cellular transcription state of glioblastoma cells is correlated to glioma cell invasion in vivo. To unravel the details of neuronal features of glioma invasion in space and time, we established a novel approach of intravital imaging for brain tumor cells with a membrane-bound GFP combined with deep learning algorithms that are used to track glioma cell processes with a high temporal resolution over several hours. This approach uncovers how invading tumor microtubes use Levy-like movement patterns indicative of efficient search patterns often employed by animal predators searching for scarce resources such as food. Neuronal activity is able to accelerate the tumor microtube dynamics, accelerate the Levy-like movement patterns and increase the overall invasion speed of glioma cells. These processes are mediated by local calcium transients in glioma cell somata and tumor microtubes. In accordance, genetic manipulation and pharmacological perturbation of AMPA receptors reduces tumor microtube length, number and branching points by interfering with intracellular calcium transients. All in all, the work here uncovers novel neuronal activity-mediated mechanisms of glioma cell invasion, a hallmark of this yet fatal disease.