Abstract Glioblastoma, the most common primary brain tumor in adults, is a major cause of neurological morbidity and mortality with no effective therapies. Its proliferation and invasion are regulated by direct and paracrine-mediated neuronal activity. In this study, we uncover molecular targets and protein-protein signaling driving activity dependent proliferation. To identify candidate molecular pathways, we used high-density electrode arrays in vivo to record human local field potentials. Single-cell RNA sequencing identified Netrin-G1 (NTNG1) and thrombospondin-1 (TSP1) elevation in glioblastoma clonal population that proliferates in response to the presence of neurons. The amino acid sequence of NTNG1 and the TSR1 domain of TSP1 were utilized to predict the 3D interaction surface of NTNG1 and TSP1. Compared to structure predictions of NTNG1 and its known binding partner NGL1, NTNG1 and TSP1 complex predictions showed agreement between machine learning predicted models and lower predicted aligned error in interdomain regions of the complex, suggesting higher confidence in the relative position of NTNG1-TSP1 as compared to NTNG1-NGL1. We performed 45 protein structure predictions and two molecular dynamics simulations of mutant and WT TSP1 in complex with NTNG1. Each simulation box consisted of >263,000 atoms, each run for 50 million 2 femtosecond timesteps for a total simulation time of 200 ns. These calculations confirmed a stable complex in the WT condition and significant conformational change in the R46A condition, emphasizing the importance of this residue in mediating NTNG1-TSP1 binding. Computational affinity models between NTNG1 and TSP1 were validated by pull-down assays and immunofluorescence labelling using cerebral organoid and human glioma-neuronal coculture models. Finally, we design small molecule and fusion-protein inhibitors disrupting this interaction. Here we advance the development of targeted precision-medicine therapies to treat glioblastoma proliferation through activity-dependent mechanisms.