Context Matters for GluN1 Variants in Epilepsy

Abstract

Context-Dependent NMDA Receptor Dysfunction Predicts Seizure Treatment in Mice with Human GluN1 Variant

Venkatesan S, Nazarkina D, Sullivan MT, Tan YF, Qu S, Ramsey AJ, Lambe EK.. iScience. 2025;29(1):114301. doi: 10.1016/j.isci.2025.114301. PMID: 41602908; PMCID: PMC12834114.

Mutations in N-methyl-D-aspartate receptors (NMDARs) cause epilepsy and profound cognitive impairment, though the underlying subunit-specific vulnerabilities remain unclear. We investigate the impact of a severe human variant in the lurcher motif of obligate GluN1 NMDAR subunit using transgenic mice, revealing unexpected context-dependent phenotypes. We show that the GluN1 Y647S variant significantly reduces current flow through pharmacologically isolated synaptic NMDARs in prefrontal neurons. Yet in intact local circuits, this loss-of-function paradoxically extends NMDAR-dependent dendritic integration, causing prolonged circuit-wide excitation that promotes seizures. Mutant receptors appear deficient in engaging opposing dendritic ion channels that normally curtail NMDAR-dependent excitation. Boosting SK channel activity normalizes dendritic integration, whereas slight decreases in extracellular magnesium further extend abnormally prolonged integration in mutant mice. We find that magnesium supplementation successfully treats seizures in vivo in the transgenic mice, despite loss-of-function of NMDARs. Overall, we disentangle a GluN1 variant's receptor-level effects and its dendritic impact to treat seizures effectively.

Commentary

N-methyl-D-aspartate (NMDA) receptors are glutamate-gated ion channels that play a central role in excitatory synaptic transmission, synaptic plasticity, and activity-dependent circuit development. ¹ As a subtype of ionotropic glutamate receptors, NMDA receptors are uniquely characterized by their voltage-dependent magnesium block and high permeability to calcium, allowing them to act as molecular coincidence detectors that couple synaptic activity to intracellular signaling pathways. In epilepsy, excessive or dysregulated NMDA receptor activation contributes to neuronal hyperexcitability, network synchronization, and excitotoxic injury. ² Enhanced NMDA receptor-mediated currents can lower seizure threshold, promote seizure propagation, and drive activity-dependent neuronal death through calcium-dependent signaling cascades. Both genetic alterations in NMDA receptor subunits (eg, GRIN genes) and acquired changes in receptor expression or phosphorylation have been implicated in epileptogenesis, making NMDA receptors key mechanistic mediators and therapeutic targets in epilepsy research. ³ Nevertheless, translating these receptor- and synapse-level mechanisms into an integrated understanding of circuit-level dysfunction has remained a challenging task.

Venkatesan et al ⁴ approached this challenge by studying the Y647S variant of the NMDA receptor GluN1 subunit, which is associated with severe epilepsy and developmental delay in patients and mouse models. Teasing apart the functional consequences of Y647S and other GluN1 variants has not been straightforward because the receptor-level impact does not always match the holistic phenotype. Therefore, Venkatesan et al took an integrative approach to investigate the effects of Y647S at multiple scales—from receptor to circuit to an in vivo mouse model—which ultimately revealed how functional context is essential to understand a variant's pathogenicity and develop rational treatment strategies.

Initial measurements of excitatory postsynaptic currents in principal neurons of prefrontal brain slices from mice demonstrated that Y647S reduced NMDAR currents, while leaving α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid currents intact. This suggested that the variant confers a loss-of-function effect at the receptor level. However, NMDAR plateau potentials—dendritic phenomena that occur when extra-synaptic NMDARs are activated by glutamate spillover during repetitive stimulation—were significantly prolonged in mutant neurons despite the reduced NMDAR currents observed at the synapse. Furthermore, simultaneous patch-clamp electrophysiology and Ca²⁺  imaging revealed a delayed onset of Ca²⁺ influx followed by a prolonged elevation, which led to sustained dendritic excitation in the absence of alterations in dendritic morphology. These results showed that the Y647S variant can have opposing loss- and gain-of-function effects at different levels of NMDAR signaling.

Shifting to the circuit level, Venkatesan et al observed extended population activity in mutant cortical neuronal ensembles using wide field calcium imaging. By blocking gamma-aminobutyric acid receptors, they were able to appreciate widespread and persistent excitation which led them to conclude that the Y647S variant results in prolonged cortical network excitation that could transition to epileptiform activity. This is significant because it offers an ex vivo model demonstrating the conditions that can initiate seizures in the context of a clinically relevant GRIN1 patient variant. It also further highlights the context dependence of NMDAR variants, where there is a seemingly paradoxical reduction in synaptic NMDAR current yet extended dendritic excitation resulting in cortical epileptiform activity.

To investigate mechanisms that may indirectly prolong excitation in Y647S neurons, Venkatesan et al targeted 2 potential causes: small conductance calcium-activated potassium channel (SK channel) mediated negative feedback and magnesium blockade of NMDARs. Under normal conditions, Ca²⁺ influx through NMDARs activates small conductance SK calcium-activated potassium channels that hyperpolarize the membrane, which promotes NMDAR Mg²⁺ blockade and terminates further NMDAR activation. Venkatesan et al hypothesized that Y647S neurons may experience insufficient SK channel activation, due to reduced Ca²⁺ influx at the synapse, which in turn could cause prolonged dendritic plateau potentials. By using an SK channel activator, they were able to successfully reduce the prolonged plateau depolarization in mutant neurons, suggesting mutant neurons have impaired negative feedback through SK channels. In a second experiment, they lowered extracellular Mg²⁺ concentrations, which caused a disproportionate increase in current amplitude only in mutant neurons, further suggesting that the impaired SK channel response prevents normal NMDAR Mg²⁺ blockade, thus increasing vulnerability to NMDAR hyperexcitation.

Magnesium supplementation has garnered some interest as an antiseizure treatment, but its bioavailability in the brain is limited.^5,6 Therefore, Venkatesan et al asked whether magnesium L-threonate, which has improved bioavailability in the brain following oral administration, may be able to prevent or reverse seizure activity in Y647S mice. ⁷ They treated mice continuously via the drinking water and observed no seizures developed at all in mutant mice until 10 weeks when a few breakthrough seizures occurred, though they were less severe than those observed in untreated mice. In a second experiment, they waited until after seizures were established to begin treatment. Y647S mice that started treatment showed an immediate improvement, including a complete elimination of the most severe seizure types after 2 weeks, suggesting that magnesium L-threonate could be a viable treatment approach.

This work is significant in that it serves as a reminder of the biological complexity involved in disentangling a variant's circuit-wide effects. It is not always sufficient to interpret a variant's pathogenicity based on knowledge of receptor or synapse level effects alone. In the case of Y647S, magnesium supplementation would not have been considered as a treatment based on the loss of NMDAR current observed at the synapse. However, by obtaining an integrated view of the variant's effects at multiple levels, it became clear that seizure-promoting excitability arises from indirect effects that are sensitive to extracellular magnesium changes. This unexpected finding, made possible by multiscale analyses, is what brought magnesium L-threonate into consideration as a rational treatment strategy. Following this lead, future studies could extend this work to investigate Y647S's impact on synaptic plasticity or interneuron function to obtain an even more comprehensive understanding of the effects on brain excitability. Taken together, this article argues for considering variant effects at multiple levels to develop better treatments.