The review’s central premise rests on a conceptual shift that has been gathering momentum in neuroscience: epilepsy is not a neuron-proprietary disorder. Astrocytes actively shape brain function through what researchers now call the active milieu, an integrated cellular network encompassing neurons, glia, vasculature, and the extracellular matrix. Within this framework, astrocytes perform two homeostatic functions that are directly relevant to seizure generation. First, they regulate extracellular potassium concentrations through the inwardly rectifying potassium channel Kir4.1. Because the extracellular space is limited, intense neuronal activity can drive extracellular potassium to 10 to 12 millimolar, levels characteristic of epileptiform activity. Astrocytes counteract this through spatial potassium buffering, a process in which potassium ions enter astrocytes at sites of high concentration and exit at distant sites through the coupled glial syncytium, effectively redistributing the ionic load across the tissue.

Second, astrocytes clear glutamate, the brain’s principal excitatory neurotransmitter, from the synaptic cleft. Under physiological conditions, glutamate released during synaptic transmission is taken up by astrocytes, converted to glutamine by the enzyme glutamine synthetase, and recycled back to neurons in what is known as the glutamate-glutamine cycle. This uptake is electrogenic and enhanced at hyperpolarized membrane potentials, meaning it depends critically on the astrocyte’s own electrical state. When either of these homeostatic mechanisms fails, the consequences are immediate and severe. Elevated extracellular potassium depolarizes neuronal resting membrane potentials and lowers the seizure threshold, while impaired glutamate reuptake allows the neurotransmitter to overstimulate neurons and trigger seizures. Animal studies have shown that Kir4.1 is essential for astrocytic potassium buffering, and astrocyte-specific deletion of Kcnj10, the gene encoding Kir4.1, disrupts both potassium and glutamate uptake, leading to increased seizure activity and premature death.

The review also highlights a more nuanced picture of astrocytic involvement in epilepsy, one that resists simple categorization. Reactive astrogliosis, the transformation of astrocytes into an activated state following neurological insults such as traumatic brain injury, stroke, status epilepticus, or viral infection, is a hallmark of the epileptic focus. Reactive astrocytes release pro-inflammatory cytokines, including interleukin-1β and tumor necrosis factor-α, which reinforce epileptogenic inflammatory signaling and play a key role in seizure initiation. Recurrent excitatory loops between astrocytes and neurons, driven by astrocytic calcium elevation, correlate with the initiation and sustenance of ictal discharges. Yet in a striking counterpoint, a recent study demonstrated that optogenetic activation of channelrhodopsin-2-expressing astrocytes exerts an anti-seizure effect on kainic acid-induced neocortical seizures, driven by stimulation of astrocytic Na⁺-K⁺-ATPase through sodium influx, independent of calcium signaling. Astrocytes, in other words, exhibit marked functional heterogeneity and can either promote or suppress seizure activity depending on the pathological context.

Where the review makes its most provocative contribution is in connecting these astrocytic mechanisms to the respiratory dysfunction that precedes SUDEP. Hypoxia occupies a central place in this connection. Mild hypoxia increases ventilation through carotid body activation, and brainstem astrocytes contribute to this hypoxic ventilatory augmentation by directly sensing low oxygen and enhancing respiratory output via gliotransmission. Severe hypoxia, however, tells a different story. After an initial augmentation, severe hypoxia inhibits the brainstem respiratory circuit, causing diminished ventilation, exacerbated hypoxia, and global brain suppression. This deepens ventilatory depression in a vicious cycle that culminates in respiratory arrest and death. The authors’ own experimental work provides a compelling piece of evidence: pretreatment with arundic acid, an inhibitor of astrocytic activation, delays the onset of severe hypoxia-induced seizures and respiratory arrest in mice. This finding suggests that excessive astrocytic activation during severe hypoxia may promote neuronal hyperexcitability and hypersynchrony, facilitating seizure generation and potentially seizure-induced respiratory arrest.

The review proposes a sequential activation model in which microglia, the brain’s resident immune cells, act first. Severe hypoxia activates microglia, which release ATP that binds to P2Y1 receptors on neighboring astrocytes, stimulating them in a paracrine manner. Activated microglia lower the seizure threshold by promoting pro-inflammatory cytokines and astrocytic glutamate release. Thus, hypoxia-enhanced microglial activation in brain regions critical for epileptogenesis drives seizure onset and the subsequent astrocytic activation, which in turn precipitates postictal ventilatory suppression. This microglia-first, astrocyte-second sequence has been demonstrated in epileptic models and represents a key mechanism linking hypoxia to both epileptogenesis and respiratory compromise.

On the respiratory side, the review assembles evidence that brainstem structures, including the pre-Bötzinger complex, the parafacial respiratory group and retrotrapezoid nucleus, and the nucleus tractus solitarius, are essential for maintaining adequate respiratory regulation. Neuropathological studies have consistently demonstrated volume loss of the ventrolateral medulla in SUDEP victims, accompanied by reductions in both neurons and astrocytes. This volume loss deepens as epilepsy progresses, implicating chronic brainstem dysfunction in postictal respiratory compromise. Seizures originate in the cortex, but lethality involves disruption of autonomic and respiratory function at the brainstem level. Brainstem respiratory centers maintain reciprocal connections with multiple cortical and subcortical structures, providing an anatomical substrate through which seizure activity can propagate and suppress respiratory output. Specifically, hyperexcitability of GABAergic neurons in the nucleus tractus solitarius reduces the threshold for spreading depolarization, a self-propagating wave of near-complete neuronal excitation accompanied by ionic and neurotransmitter homeostasis disruption that can facilitate acute respiratory and cardiac shutdown.

The review also addresses the controversial role of postictal generalized EEG suppression, or PGES, which has been proposed as a biomarker of SUDEP risk. Studies have shown that the occurrence and duration of PGES vary across seizures within individual patients, suggesting it may represent a seizure-specific, transient phenomenon rather than a stable trait-like marker. The authors argue that the pathophysiological processes underlying SUDEP likely extend beyond PGES itself and involve a rapid failure of cardiorespiratory and autonomic control, manifested by early postictal respiratory dysfunction and terminal ventricular arrhythmias. This perspective underscores the limitations of PGES as a standalone predictor and highlights the need to consider the broader context of postictal cardiorespiratory dysfunction.

Synthesizing these threads, the authors propose an integrative model of SUDEP pathogenesis. SUDEP is frequently initiated by terminal generalized tonic-clonic seizures. Upon seizure onset, activity propagates to the brainstem, sparking astrocytic dysfunction. Compromised astrocytic potassium buffering impedes clearance of elevated extracellular potassium, sustaining a lowered seizure threshold. Astrocytic impairment also attenuates carbon dioxide sensitivity and the hypercapnic ventilatory response while suppressing the release of gliotransmitters such as ATP, collectively destabilizing respiratory regulation. Seizure propagation within the brainstem disrupts central autonomic networks, triggering central and obstructive apneas and hypoventilation, which lead to acute ventilatory failure. Intertwined astrocytic and respiratory disturbances contribute to hypoxia and hypercapnia, precipitating cardiorespiratory collapse. In this model, astrocytic malfunction and breathing deficit trap the body in a harmful feedback loop that pushes epilepsy toward sudden death.

The review acknowledges significant open questions. It remains unresolved whether astrocytic epileptic alterations are the driving pathogenic force or secondary responses to recurrent seizures. It is also unclear whether epilepsy-associated genes expressed by neurons or astrocytes within brainstem respiratory networks contribute directly to respiratory susceptibility to SUDEP, and how dysregulated astrocytic potassium buffering, glutamate clearance, and gliotransmission influence seizure propagation into the brainstem requires further clarification. Nevertheless, the therapeutic implications are substantial. Restoring astrocytic homeostasis, including potassium buffering, glutamate clearance, gliotransmission, and the regulation of neuroinflammation, opens novel avenues for preventing seizure progression and reducing SUDEP risk. Elucidating bidirectional communication between astrocytes and microglia may identify additional interventions. As the authors conclude, a more comprehensive understanding of neuron-glia interactions within the active neuronal milieu will not only advance understanding of epileptic pathophysiology but also accelerate the development of targeted preventative and therapeutic measures for one of epilepsy’s most devastating outcomes.

Subject of Research: The role of astrocyte dysfunction in peri-ictal breathing impairment and sudden unexpected death in epilepsy

Article Title: Astrocytes at the interface of peri-ictal breathing dysfunction and SUDEP

Article References: Fukushi, I., Ozaki, I., Pokorski, M., & Okada, Y. (2026). Astrocytes at the interface of peri-ictal breathing dysfunction and SUDEP. Frontiers in Network Physiology, 6, Article 1955271. https://doi.org/10.3389/fnetp.2026.1955271

Image Credits: AI Generated

DOI: 10.3389/fnetp.2026.1955271

Keywords: SUDEP, astrocytes, epilepsy, hypoxia, respiratory failure, brainstem, Kir4.1, microglia, gliotransmission, spreading depolarization, neuroinflammation, pre-Bötzinger complex

The logic behind the approach builds on a decade of epilepsy gene therapy. Most experimental genetic treatments for seizures work by boosting the potassium conductance of neurons, effectively adding brake pedals to cells that fire too readily. Adeno-associated virus vectors, the workhorses of modern neurological gene therapy, are used to smuggle engineered ion channel genes into a circumscribed patch of brain tissue, where they render hyperexcitable neurons less likely to ignite pathological activity. Because seizures and spreading depolarization frequently co-occur, the Moscow group, led by Evgeny Nikitin and Lyudmila Vinogradova, asked a question that had not been systematically addressed: what happens to spreading depolarization when you give neurons extra potassium conductance?

The channel they chose was KCa3.1, encoded by the gene KCNN4. Unlike voltage-gated potassium channels, KCa3.1 opens in response to intracellular calcium, which surges whenever a neuron fires vigorously. This makes the channel a self-regulating safety valve: the harder a cell works, the more the channel opens, letting potassium ions flow out and pulling the membrane potential back toward rest. Earlier work by the same team had shown that overexpressing KCNN4 in principal neurons produces an anti-seizure effect without degrading the cells’ ability to encode information, a crucial consideration for any therapy that must leave normal brain function intact. Whether the same manipulation would influence the massive, synchronized ionic collapse of spreading depolarization was an open question.

To find out, the researchers injected an AAV vector carrying KCNN4 into the neocortex of mice and then attacked the problem from three technical directions. Patch-clamp electrophysiology allowed them to record the membrane potential of individual neurons as a depolarization wave swept past. Functional imaging in acute brain slices, using voltage- and calcium-sensitive indicators, revealed how the wave propagated through the cortical network. And direct-current-coupled field potential recordings captured the full signature of spreading depolarization in vivo, including in awake, non-anesthetized animals, a demanding preparation that avoids the confounding effects of anesthetics on cortical excitability.

The results were striking. KCNN4 expression reduced the amplitude of the spreading depolarization several-fold, an effect documented both in slices and in the intact cortex of freely moving mice. The wave also traveled more slowly through treated tissue, as measured by the timing of its arrival at recording sites in the slices. Slower propagation matters clinically: the damage inflicted by spreading depolarization after stroke or traumatic brain injury is thought to scale with the depth and duration of the ionic disruption, so a wave that is smaller and slower should, in principle, be less injurious. In patients with malignant stroke or decompressive craniectomy, spreading depolarizations detected with subdural electrodes are associated with worse outcomes, making amplitude and propagation speed attractive therapeutic targets.

The gene therapy also reshaped the aftermath of the wave. Spreading depolarization imposes a period of profound neuronal silencing, during which epileptiform activity is temporarily abolished. Paradoxically, in the KCNN4-treated slices this suppression ended sooner, with seizure-like events reappearing earlier than in control tissue. The finding highlights a subtle and clinically important trade-off. Some evidence suggests that spreading depolarization itself can act as an innate antiseizure mechanism, terminating seizures by short-circuiting electrical propagation across the cortex. A therapy that shrinks the depolarization wave might therefore, in some circumstances, shorten the very silencing that ends a seizure. The authors flag this as a side effect that any potassium-conductance-enhancing gene therapy for epilepsy will need to weigh.

Delving into the mechanism, the team documented a calcium rebound that follows spreading depolarization when epileptiform activity resumes. As neurons recover and begin firing again, intracellular calcium climbs steeply, and this rebound can feed back into the intensity of the recovered activity. In KCNN4-expressing neurons, the channel’s calcium sensitivity turns this rebound to advantage: the rising calcium opens the overexpressed KCa3.1 channels, which generate a potassium-mediated afterhyperpolarization that counterbalances the depolarizing drive. In effect, the therapy converts a vicious cycle, in which calcium influx amplifies excitability, into a negative feedback loop that restrains it. This cell-autonomous quality, where each neuron polices its own excitability, distinguishes the approach from drugs that act globally and indiscriminately.

The technical execution relied on a mature toolkit. AAV vectors have become the dominant platform for nervous system gene therapy because they are relatively safe, transduce neurons efficiently, and support long-term expression. The team’s vector design and cloning were performed in-house, and the imaging data were processed with supervised calcium event detection tools developed by the group. The in vivo experiments in awake animals are particularly notable, since spreading depolarization amplitude and threshold can be altered by anesthesia, and recordings from non-anesthetized mice provide the most physiologically relevant readout of how the therapy would behave in a clinical setting.

The broader significance lies in the intersection of two therapeutic frontiers. Gene therapy for epilepsy is advancing rapidly, with engineered potassium channels, optogenetic tools, and on-demand dynorphin-based systems all showing efficacy in rodent models, and at least one company has raised substantial funding to develop a single-dose gene therapy for focal refractory epilepsy. Spreading depolarization, meanwhile, has emerged as a therapeutic target in its own right, with ketamine trials in acute brain injury aiming to suppress the waves pharmacologically. The new study connects these threads, demonstrating that a single genetic manipulation can simultaneously dampen seizures and attenuate spreading depolarization, while also revealing the unexpected interaction between the two phenomena.

Caveats remain before any translation to patients. The work is a proof of principle in rodents, and the authors themselves note that reducing spreading depolarization via increased potassium conductance may carry the paradoxical risk of prolonging seizure activity by removing a natural braking mechanism. The dosing of channel expression, the durability of the effect, and the behavior of the therapy in injury models of stroke and trauma all await further study. Still, the demonstration that a targeted viral gene delivery can measurably shrink one of the brain’s most violent electrical events opens a concrete path toward treating disorders, from migraine to stroke, that are complicated by spreading depolarization, using nothing more than the neurons’ own molecular machinery.

Subject of Research: Viral gene therapy targeting spreading depolarization via KCa3.1 potassium channel expression

Article Title: Local targeted suppression of neocortical spreading depolarization amplitude by viral expression of KCa3.1 channels

Article References: Oblasov, I. A., Smirnova, M. P., Borodinova, A. A., Zuzina, A. B., Smirnov, I. A., Balaban, P. M., Vinogradova, L. V., & Nikitin, E. S. (2026). Local targeted suppression of neocortical spreading depolarization amplitude by viral expression of KCa3.1 channels. Gene Therapy. https://doi.org/10.1038/s41434-026-00645-x

Image Credits: AI Generated

DOI: 10.1038/s41434-026-00645-x

Keywords: spreading depolarization, gene therapy, KCa3.1, KCNN4, adeno-associated virus, epilepsy, migraine, stroke, potassium channels, neocortex, electrophysiology, afterhyperpolarization