Decomposing SV2A Function and Dysfunction one Neuron at a Time

Abstract

Synaptic Vesicle Glycoprotein 2A Knockout in Parvalbumin and Somatostatin Interneurons Drives Seizures in the Postnatal Mouse Brain

Bartholome O, Neirinckx V, De La Brassinne O, Desloovere J, Van Den Ackerveken P, Raedt R, Rogister B. J Neurosci. 2025 Feb 19;45(8):e1169242024. doi:10.1523/JNEUROSCI.1169-24.2024. PMID: 39753304; PMCID: PMC11841765.

Synaptic vesicle glycoprotein 2A (SV2A) is a presynaptic protein targeted by the anti-seizure drug levetiracetam. One or more of the three SV2 genes is expressed in all neurons and is essential to normal neurotransmission. Loss of SV2A results in a seizure phenotype in mice and mutations in humans are also linked to congenital seizures. How SV2A action impacts the epileptic phenotype remains unclear, especially among the diverse neuronal populations that regulate network excitability. This study explored how brain structure and function are affected by SV2A conditional knockout (SV2A-cKO) in specific neural cell subtypes. We show that SV2A-cKO in all neurons of the postnatal brain triggers lethal seizures, suggesting that the seizures observed in earlier knockout models were not due to aberrant brain development. Similar lethal seizures are detected in mice in which the loss of SV2A is limited to GABAergic neurons, whereas loss in excitatory neurons produces no noticeable phenotype. No apparent gender difference was ever observed. Further investigation revealed that SV2A-cKO in different GABAergic interneuron populations induces seizure, with variable timescales and severity. Most notably SV2A-cKO in parvalbumin interneurons (PV⁺) leads to lethal seizures in young animals, while SV2A-cKO in somatostatin (SST) inhibitory neurons results in seizures that were scarcely observed only in adult mice. These results support the crucial role SV2A plays in PV and SST interneurons and suggest that the action of levetiracetam may be due largely to effects on a subset of GABAergic interneurons.

Commentary

If Synaptic Vesicle Glycoprotein 2A (SV2A) rings a bell to epilepsy clinicians and researchers, it's likely because it is the target of one of the most prescribed anti-seizure medications, levetiracetam (LEV). ¹ SV2A knockout mice also have early onset seizures and preweaning lethality,^2,3 and at least five dominant or recessive variants have been described in humans with severe childhood epilepsy.^4,5 Despite these strong links to epilepsy, SV2A remains somewhat of a mystery. It is known to be localized on synaptic vesicles at the presynaptic terminal, but its core molecular function is still unclear. ⁶ Physiological studies in SV2A knockout models have generally found alterations in inhibitory neurotransmission,^3,7 which may be consistent with epilepsy, but how this relates to the efficacy of LEV is puzzling. A recent paper by Bartholome et al ⁸ utilized a preclinical conditional knockout model to narrow down the neuron types that link loss of SV2A to epilepsy.

In this study, the researchers selectively deleted SV2A in specific neuronal subtypes—including all neurons, excitatory neurons, inhibitory neurons, parvalbumin-positive (PV) interneurons, and somatostatin-positive (SST) interneurons. They assessed SV2A expression levels, body weight, and seizure frequency across these groups to pinpoint which neuronal population plays a critical role in seizure generation. The novelty of the study lies in its cell type-specific approach, allowing the identification of the neuronal subtype most responsible for driving seizure pathology.

To investigate the role of SV2A in the mature brain, Bartholome et al used tamoxifen-inducible Ubiquitin C–Cre^ERT2 mice to delete SV2A at postnatal day 60 (P60), after the completion of neurodevelopment. Following tamoxifen administration, SV2A conditional knockout (cKO) mice developed severe, persistent seizures and died within approximately 2 months. Importantly, the expression levels of the other SV2 isoforms, SV2B and SV2C, did not increase in response to SV2A loss, underscoring its unique and essential role in maintaining neuronal stability in adulthood. In parallel, the authors generated Nestin-Cre; SV2A-cKO mice to delete SV2A throughout the entire nervous system from early development. These animals began seizing around P12 and died within 2-5 days thereafter, confirming that SV2A is critical not only in the mature brain but also during developmental stages for normal seizure thresholds and survival.

Together, these experiments establish that SV2A is essential for preventing aberrant hyperexcitability whether it is lost during development or after maturation. The next phase of the study employed cell-type-specific Cre drivers to delete SV2A selectively in excitatory neurons, inhibitory neurons, PV interneurons, and SST interneurons. Bartholome et al first compared the effects of SV2A deficiency in the brain's two major neuronal populations: glutamatergic and GABAergic neurons. In Nex-Cre;SV2A-cKO mice, where SV2A was selectively deleted in excitatory neurons, there were no seizures or changes in life expectancy, despite successful recombination and a marked reduction in SV2A expression within the targeted cells. In contrast, Dlx-Cre;SV2A-cKO mice—where SV2A was deleted in most GABAergic interneurons—began exhibiting seizures around P15 and died within 3-6 days. Interestingly, brain lysates from Dlx-Cre mice did not show a significant overall reduction in SV2A protein levels. The authors attributed this to the relatively small proportion of interneurons compared to the total neuronal population, which may have masked cell type-specific decreases in SV2A expression. These findings highlight the critical role of SV2A in inhibitory neurons for maintaining network stability and preventing seizures.

To further dissect the role of SV2A within specific subtypes of GABAergic interneurons, they crossed SV2A-cKO mice with PV-Cre, SST-Cre, and VIP-Cre driver lines to induce targeted deletion of SV2A in PV, SST, and VIP interneurons, respectively. PV-Cre;SV2A-cKO mice developed spontaneous seizures between P16 and P20, accompanied by reduced body weight and shortened lifespan. Electroencephalogram (EEG) recordings revealed frequent seizures and epileptic spike activity, indicating a severe disruption in network stability resulting from SV2A loss in PV interneurons. In contrast, VIP-Cre;SV2A-cKO mice exhibited normal behavior, seizure-free EEG patterns, and typical life expectancy, suggesting SV2A is dispensable in this interneuron subtype. SST-Cre;SV2A-cKO mice showed intermediate phenotypes: while they did not exhibit overt behavioral seizures, EEG recordings revealed epileptiform spikes, indicating subclinical network hyperexcitability. These findings underscore the particularly critical role of SV2A in PV interneurons for maintaining inhibitory control and preventing seizure activity. By sequentially analyzing SV2A deletion in all neurons, major neuronal populations (glutamatergic and GABAergic), and then specific inhibitory subtypes, the study demonstrated that the loss of SV2A in PV interneurons alone is sufficient to induce spontaneous seizures and early mortality. However, detailed quantification of seizure type and frequency or the duration of EEG recordings was lacking. Including such data would have strengthened the interpretation of seizure severity or the type of seizures in each model. Nevertheless, the findings underscore the pivotal role of SV2A in PV cells for maintaining inhibitory control and preventing epileptic activity.

Bartholome et al's work demonstrates complete loss of SV2A in PV interneurons is most closely tied to epilepsy. However, mechanistic links between SV2A's putative role in presynaptic neurotransmission and epilepsy were unexplored. A previous paper used a knockin rat model ⁹ to demonstrate that an SV2A mutation selectively impaired depolarization-induced GABA release in the hippocampus while depolarization-induced glutamate release remained unaffected. These findings support the idea that SV2A plays a specialized role in regulating inhibitory neurotransmission. Building on this, an important next step would be to analyze synaptic transmission from PV-, SST-, VIP-expressing and glutamatergic neurons through paired whole cell recordings or optogenetic activation to fully investigate presynaptic mechanisms. Such experiments would determine whether PV neuron function is more sensitive to SV2A loss than other neurons and enhance our understanding of the cellular mechanisms underlying SV2A-related epileptogenesis.

If loss of SV2A in PV neurons is sufficient for seizures in a mouse model, does that point to PV neurons as the key cell type mediating the effects of LEV? There is evidence that LEV inhibits SV2A, ¹⁰ which, in the most straightforward interpretation, suggests that it may also inhibit PV neuron function. This would be a surprising mechanism of seizure suppression. This apparent paradox raises important questions about the difference between pharmacological inhibition and complete loss of protein function, as partial inhibition may induce different effects than complete loss. Alternatively, other cell types may mediate the effects of LEV, for example, if it reduces glutamate release due to more subtle mechanisms of action than is currently known. SV2A may also have yet undiscovered functions in other cellular processes, as is the case for other proteins originally identified at the presynapse. Ultimately, answers to these questions will deepen our understanding of LEV and its target, and determine if the story is more than the sum of its parts.