Stochastic tissue environment expands spatial limits of synaptic and astrocytic glutamate actions

The point precision of excitatory transmission relies on high-affinity astrocytic transporters that rapidly buffer glutamate escaping from the synaptic cleft. While numerous theoretical models have supported this concept, experimental findings indicate that glutamate can diffuse farther than predicted, at times reaching receptors at neighbouring synapses. This challenges the classical view of neural networks as operating through strictly one-to-one synaptic connections. Even less is understood about the fate of glutamate released by astrocytes, as it must bypass the buffering effect of astrocytic transporters located in the immediate vicinity. We hypothesized that this mismatch between experiment and theory stems, at least in part, from a key assumption in existing biophysical models, that excitatory synapses are uniformly surrounded by transporter-rich neuropil. In reality, synapses are embedded within a complex environment composed of cellular structures with highly variable shapes and positions, of which only a fraction are astrocytic processes. To address this, we developed and empirically constrained a conceptually novel in silico model of synaptic neuropil, in which glutamate diffuses within a stochastically generated tissue environment partially occupied by transporter-expressing astrocytic processes. Our simulations predict that glutamate released from either synapses or astrocytes can potentially activate high-affinity receptors, and desensitize AMPA receptors, beyond one micron from the release site, thus influencing dozens of nearby synapses. While this prediction helps to resolve the long-standing debate, experimental studies of glutamate’s actions in the intact brain are needed to establish the precise physiological relevance of these findings.