Abstract Recent single-cell RNA-sequencing analyses have revealed a strong association between glioma-neuronal circuit remodeling and regional immunosuppression. Subsequent preclinical investigations using syngeneic tumor models demonstrated that genetic disruption of circuit assembly genes, including Thrombospondin-1 (TSP-1/Thbs1) in GBM cells suppressed synaptogenesis, excitatory neurotransmitter ligand and receptor density in addition to neuronal activities while simultaneously restoring inflammatory responses, indicating glioma-neuron-immune cross talk. However, the underlying mechanisms and directions of causality remain unclear. Here we test the hypothesis that tumor-derived synaptogenic factors directly reprogram the tumor immune microenvironment (“model 1”). Alternatively, tumor intrinsic neuronal activity may indirectly regulate adaptive immunity (“model 2”). In this study, we performed in-vitro bone-marrow-derived macrophage (BMDM) reprogramming assays evaluating various stimulation conditions and in-vivo experiments evaluating the effect of neuronal activity-oriented therapy on the tumor-associated macrophages (TAMs). Our in-vitro assays demonstrated that, secreted factors from mouse neonatal cortical neurons (not tumor-derived synaptogenic factors) induced an anti-inflammatory, M2-like phenotype of BMDMs, corroborating the proposed model 2. This M2-inducing effect was inhibited when the neuronal activities were suppressed pharmacologically, using perampanel (PER), a noncompetitive inhibitor of neuronal AMPA-type glutamate receptors. Direct treatment of BMDMs with PER did not affect the polarization patterns, further supporting the M2-inducing effect of glutamatergic excitatory neuronal signals. In in-vivo syngeneic orthotopic tumor models, treatment with PER significantly prolonged survival compared to the non-treated group (27.5 [PER] vs. 23 days [control], p = 0.01). Flow cytometric analysis of brain-infiltrating leukocytes showed that TAMs isolated from PER-treated mice exhibited a more proinflammatory M1-like polarization compared to the control group (median M1/M2 ratio, 1.2 [PER] vs. 0.7 [control], p = 0.004), which explained the improved prognosis. Collectively, these findings unveil novel immunosuppression mechanisms induced by glioma-neuronal circuit remodeling and the resulting excitatory neuronal signals, and potential therapeutic vulnerabilities that could be targeted in combination with cancer immunotherapy.