Bidirectional crosstalk between the nervous system and the tumour microenvironment: mechanisms, feedback loops and therapeutic opportunities.

The nervous system is increasingly recognized as an active and integral component of the tumour microenvironment (TME), rather than a passive bystander affected by tumour invasion. Emerging evidence indicates that neural inputs shape tumour behaviour through both direct and indirect mechanisms. Neurotransmitters, neuropeptides and neurotrophic factors act on tumour cells, stromal cells, endothelial cells and immune cells to regulate proliferation, invasion, metastasis, angiogenesis, metabolic reprogramming and immune evasion. In parallel, the TME feeds back to the nervous system through inflammatory mediators, extracellular vesicles, axon guidance molecules and metabolic signals, thereby driving axonogenesis, tumour innervation, Schwann cell reprogramming, neuronal hyperexcitability and synaptic remodelling. These reciprocal interactions establish dynamic neuro-immune-metabolic feedback loops that sustain tumour progression and therapeutic resistance. Particularly in glioma and other highly innervated malignancies, activity-dependent neuron-tumour communication further highlights the functional integration between neural circuits and cancer. In this Review, we summarize the structural and molecular basis of neural components within the TME, discuss neurotransmitter receptor-mediated signalling and indirect regulation of immune, vascular, stromal and metabolic niches, and outline how tumour-derived signals remodel peripheral and central neural systems. We further highlight emerging therapeutic opportunities targeting β-adrenergic signalling, neurotrophin pathways, extracellular vesicle-mediated tumour innervation, Schwann cell-associated perineural invasion circuits, and neuron-tumour synaptic coupling. Finally, we discuss current translational challenges, including tumour-type heterogeneity, context-dependent neural effects, evidence-level heterogeneity and the need for spatially resolved biomarkers, and propose that incorporating the neural dimension into future mechanism-guided studies may inform biomarker-stratified trials and symptom-oriented interventions, with the long-term goal of improving both tumour control and neurological outcomes.