Microglia-Mediated Disinhibition: Taking the Brakes off Epileptic Networks

Commentary

The study by Chen et al ¹ is a novel mechanistic investigation of inhibitory synapse elimination by microglia in a mouse model of temporal lobe epilepsy (TLE). Microglia play a unique role in the brain as activity-sensitive regulators of neuronal networks. Their role in synapse removal has perhaps been best described in the developing cortex, where complement proteins from neurons and glia tag weak synapses for pruning. ² Beyond physiological synapse elimination, a reactivation of complement-mediated synapse elimination has been implicated in excitatory synapse loss in neurodegenerative disease ³ and epilepsy. ⁴ However, less is known about the pruning of inhibitory synapses or how this process may be altered in pathological states.

In this study, the authors establish a maladaptive role for microglia, demonstrating selective elimination of inhibitory synapses in the kainic acid (KA) mouse model of TLE. In this model, status epilepticus is induced by systemic intraperitoneal injection of KA. The period of generalized seizures was associated with a substantial increase in extracellular GABA and GABA-mediated microglia activation. In as little as 3 h, this GABA-dependent activation of microglia increased the association of microglia with inhibitory synapses. Elimination of inhibitory synapses occurred on the order of days, with significant loss first detected 5 days post-KA with a further decline at later time points. The observation that inhibitory synapse elimination is driven by GABA is particularly interesting and may reflect an aberrant activation of homeostatic programs in the epileptic brain. Functionally, the loss of inhibitory synapses was found to correspond to a decrease in the frequency of spontaneous and miniature inhibitory postsynaptic currents (mIPSCs), which aligns with a shift in the excitation/inhibition balance. However, neuronal circuit function and epileptogenesis reflect more than a simple shift in excitation versus inhibition. Accordingly, this inhibitory synapse loss should be viewed as 1 component of a broader and highly dynamic process of epileptic circuit remodeling rather than as a singular driver of network hyperexcitability.

The authors next demonstrate through a compelling series of experiments that microglia activation via GABA_B receptors and astrocytic C3 expression are both necessary for inhibitory synapse loss in this model. Aligned with their mouse model work, they also identified a subpopulation of activated microglia that was associated with high expression of GABA_B and synapse elimination-related genes, using snRNA-seq from tissue resected from patients with TLE. They further confirmed an association between C3 expression, microglia, and inhibitory synapses and a functional reduction in mIPSC frequency in individuals with TLE. This body of work robustly outlines a role for glial-mediated inhibitory synapse removal in TLE, underscoring glia as dynamic regulators of neuronal network function. Interestingly, it was found that synapse removal was specific to inhibitory synapses, with excitatory synapses being spared, and even further that the loss was selective for parvalbumin-positive cells. An open question remaining is to understand how this process is regulated to promote the selective elimination of certain inhibitory synapses.

Supporting the pathological relevance of this pathway, genetic or pharmacological blockade of either microglial-GABA_B or astrocytic C3 partially reduced chronic seizures and epileptiform activity. The most parsimonious explanation for the subtle effects of preventing the loss of inhibitory synapses on epileptiform activity is that such a loss is just 1 feature in a broader landscape of circuit remodeling. Alternatively, the spared synapses may have already been functionally compromised and therefore less effective at maintaining inhibitory restraint. In this context, the sustained association of microglia with inhibitory synapses in the chronic phase of epilepsy observed in this study is noteworthy. The prolonged period of inhibitory synapse removal may reflect the extended time necessary for microglia to remove damaged synapses when there is a large insult, such as KA-induced status epilepticus. ⁵ In this case, preserving synapses would not be predicted to have a large therapeutic effect. Conversely, if the elimination of functionally active inhibitory synapses was ongoing, we might predict a progressive worsening of seizures over time. The extent to which seizure burden changes over time in KA models of TLE remains variable across studies.^6,7 Thus, it would be of interest to compare rates of seizures in a longitudinal manner after blockage of microglia-mediated elimination of inhibitory synapses to test for effects on disease progression. This would be particularly impactful when considering whether modulating chronic inhibitory synapse loss could be developed as a potential therapeutic strategy.

In conclusion, this study by Chen et al ¹ is a thoughtful and comprehensive investigation into inhibitory synapse elimination in TLE and offers a novel framework for understanding the contributions of microglia as active modulators of inhibitory dysfunction in epilepsy.