Introduction

Synaptic neurotransmission is essential for neuronal communication and circuit function. However, neurons can also release neurotransmitters outside of conventional synapses. The significance of this non-synaptic release remains unclear, although growing evidence suggests an important role in neuron-glia communication. For example, myelination of axons by oligodendrocytes in the CNS is regulated by neuronal activity and neurotransmitter signalling. Yet, the mechanisms underlying axon-myelin communication via neurotransmitters in vivo remain poorly understood due to the technical challenge of visualising neurotransmission in intact neural circuits.

 

Aims/Objectives

Using larval zebrafish (Danio rerio) as an in vivo model, we employed genetically encoded fluorescent reporters to visualise neurotransmission and define the spatial and molecular mechanisms of non-synaptic release along myelinated axons.

 

Methods

We optimised two genetically encoded fluorescent reporters: SypHy to detect synaptic vesicle fusion¹ and iGluSnFR4s to analyse extracellular glutamate signalling². To identify axonal domains associated with vesicle release, we generated a novel CRISPR knock-in line to label the heminodal/nodal marker Neurofascin a (Nfasca)³.

 

Imaging was performed on 4-5 days post-fertilisation larvae (immobilised with mivacurium chloride:1.5mg/ml) using a Zeiss LSM880 Airyscan to capture discrete vesicle fusion and glutamate transients along axons. All experiments were conducted with UK Home Office approval.

 

Results

We observed that neurotransmitter vesicle release occurs along myelinated axons at a frequency comparable to that observed at conventional synaptic sites (axon: 45.3 ± 34.2 vs collateral: 56.4 ± 37.5 events/100µm/hr, n=20 for both; paired t-test, P = 0.025). Despite similar release frequencies, axonal glutamate transients exhibited distinct kinetics compared to synapses, characterised by lower peak amplitudes but larger event areas (axonal vs collateral max dF/F0: 0.67 ± 0.26 vs 2.45 ± 1.18; n=5; paired t-test, P = 0.024: Event area: 11.23 ± 3.36 vs 7.46 ± 3.74; P = 0.004). These findings establish that axonal release is frequent and kinetically distinct from synaptic transmission but do not address whether events are spatially organised or stochastic.

 

We previously demonstrated that non-synaptic axonal vesicle fusion emerges with the onset of myelination and becomes enriched at axonal sites near to the growing ends of myelin sheaths⁴. To determine whether this spatial enrichment corresponds to heminodal and nodal domains, we labelled endogenous levels of Nfasca. Co-localisation analysis revealed a significant enrichment of vesicle events at Nfasca-positive domains (expected: 1.03 ± events/100µm/h vs observed: 8.06 ± 8.01 observed frequency; n=16; Wilcoxon matched-pairs rank test, P = 0.004). Despite this spatial association, genetic loss of Nfasca did not significantly alter axonal vesicle release frequency across genotypes (control: 39.1 ± 31.7 vs Het: 32.5 ± 25.1 vs Hom: 42.7 ± 22.3; One-way ANOVA, P = 0.11), indicating that while these domains mark sites of release, Nfasca itself is not required for vesicle fusion to occur.

 

Conclusions

This work shows that non-synaptic neurotransmitter release is a frequent and spatially organised feature of myelinated axons. Although axonal vesicle fusion occurs at rates comparable to synapses, the resulting glutamate signals are kinetically distinct, indicating mechanistic differences between synaptic and axonal release. The enrichment of release events at Nfasca-positive domains suggests that axonal neurotransmission is patterned by specialised membrane regions rather than occurring randomly.