Brainstem Disinhibition Destabilizes Breathing in Dravet Syndrome,Leading to Sudden Unexpected Death in Epilepsy (SUDEP)

Commentary

Changes in neuronal circuit excitability are essential for the normal functioning of the brain. Alterations in synaptic excitability are vital for information processing and storage. These changes that underlie learning and memory, collectively referred to as plasticity, include habituation, sensitization, and long and short-term potentiation, to name a few. ¹ However, an imbalance of excitation and inhibition can also be pathological, such as during seizures and epilepsy. Multiple etiological factors can perturb this balance, including genetic mutations. Many forms of such monogenic epilepsies arise during childhood, highlighting the vulnerability of the developing brain to seizures. ¹

One such pediatric epilepsy is Dravet Syndrome (DS), a severe form of childhood epilepsy where febrile seizures are common before the first year of life. ² DS falls under the category of developmental epileptic encephalopathies, and many patients with DS are refractory to treatment. Along with poor seizure control, they face developmental delays and other comorbidities that severely impair their quality of life. ² In 70%–95% of cases, DS can be attributed to a de novo mutation in the SCN1A gene, causing haploinsufficiency of the voltage-gated sodium channel (VGSC) Nav1.1. Early in development, NaV1.1 is primarily expressed by GABAergic interneurons. Loss-of-function mutations in this VGSC lead to disinhibition, or impaired excitability of inhibitory neurons, allowing larger brain circuitry to become hyperexcitable and ultimately produce seizures. ³

Sudden unexpected death in epilepsy (SUDEP) is a major concern for DS patients: the risk for SUDEP is estimated to be 15 times higher than in other childhood epilepsies, resulting in SUDEP as the cause of half of all deaths in DS. ⁴ Perhaps unsurprisingly, DS mouse models have been commonly utilized in SUDEP research, due to their relevance, availability, high mortality rates, and inducible seizures via hyperthermia. The first study examining SUDEP in a mouse model of DS proposed cardiac dysfunction as the primary driver of death. ⁵ However, shortly thereafter, a major clinical study examining witnessed cases of SUDEP was published, called MORTality in Epilepsy Monitoring Units (MORTEMUS). ⁶ The MORTEMUS study examined data from patients in Epilepsy Monitoring Units and identified 9 fatalities in which there was sufficient monitoring of cardiac and respiratory activity to determine that the primary cause of death was terminal apnea, followed later by terminal asystole. Upon reexamination, another preclinical study using a DS mouse model found that the primary cause of death was terminal apnea that was followed by progressive cardiac impairment and eventual asystole, ⁷ similar to what was observed in MORTEMUS.

This brings us to the present day, where Milla et al ⁸ examined the mechanisms underlying breathing impairment in a DS model. Breathing is driven primarily by 2 sources: descending inputs from forebrain structures and CO₂-sensitive automatic respiratory neurons in the brainstem. ⁹ When descending input is reduced—as during sleep, anesthesia, or certain seizures—ventilation depends almost entirely on the brainstem's ability to detect and respond to rising CO₂ levels. ¹⁰ This CO₂-sensitive automatic drive is known as the hypercapnic ventilatory reflex (HCVR). Milla et al ⁸ investigated respiratory dysfunction in DS by assessing how loss of Scn1a can impair the HCVR and neuronal function in the retrotrapezoid nucleus (RTN). Neurons in the RTN are central chemoreceptors essential for regulating breathing in response to changes in tissue CO₂ levels and integrating multiple inputs to maintain respiratory homeostasis. ¹¹

Milla et al ⁸ first examined breathing in 2-week-old DS mice—a time point known to precede seizures in this mouse line. The DS mice exhibited normal respiratory activity comparable to wild-type controls; however, their HCVR was diminished at low CO₂ levels but, interestingly, not at higher CO₂ levels, suggesting a shift in the CO₂-sensing response curve. Indeed, the authors characterized the CO₂ sensitivity of RTN neurons in 2-week-old DS mice and found that their CO₂ responsiveness was unaltered. Still, they exhibited a higher baseline action potential firing rate than wild-type littermate mice. This was determined to be due not to intrinsic properties of the RTN CO₂-sensitive neurons, but to reduced synaptic input from local inhibitory neurons. The authors found that the deficits could be rescued by pharmacological activation of NaV1.1 with the neurotoxin Heteroscodratoxin-1 (Hm1a), which selectively binds NaV1.1. In vitro, the application of Hm1a, which presumably affects only inhibitory neurons in the region, normalized RTN neuronal activity, including decreased baseline firing and increased ability to sustain firing during current injection.

In the DS model used by Milla et al, ⁸ the mice begin having seizures in the 4th week of life. At this time point, the authors observe a more severe respiratory phenotype, at least in mice that subsequently die from SUDEP. The authors once again measure the HCVR and also the ventilatory response to hypoxia. In this case, breathing responses are blunted specifically in those that die of SUDEP; DS mice that did not go on to die from SUDEP had ventilatory responses similar to those measured in wild-type mice. Whether these responses are due to inhibitory neuronal inactivity or to the progression of seizures at this developmental time point remains unclear.

In summary, just as loss of inhibition, or disinhibition, of forebrain circuitry is believed to cause the seizure phenotype of Scn1a haploinsufficiency, loss of inhibition in brainstem circuits appears to disrupt homeostatic breathing, which could tip the scales toward fatal seizure and lead to SUDEP. In addition, as the mice age and progress toward SUDEP, breathing responses to both hypercapnia and hypoxia are impaired. Since these forms of breathing impairment are clinically measurable, the authors suggest that these respiratory reflexes are a biomarker for SUDEP in DS patients. Although clinical HCVR is not a standard clinical test, it is noninvasive and clinically accessible. Even if clinical implementation poses some challenges, the effort to overcome these would be justified by the dire outcome of SUDEP.

Of course, many questions remain. What is the mechanism underlying HCVR suppression in mice that progress to SUDEP? Perhaps it is due to seizure-related activity, as shown in a previous study of DS mice. ¹² Furthermore, is this mechanism specific to DS, or could it be generalized to other forms of epilepsy? Considering the SUDEP-specific respiratory phenotype may not be directly related to NaV1.1 expression, this could be the case. However, this is not yet clear. Future experiments that can reconstitute brainstem inhibitory function to resolve the HCVR and prevent SUDEP may help answer some of these questions. Finally, if we can identify those at high risk of SUDEP, what interventions can we offer? Such interventions, of course, are not currently available, which underscores the need for more mechanistic work examining respiratory failure in SUDEP models.