Abstract
Developmental and epileptic encephalopathies (DEEs) are among the most severe and difficult to treat epilepsies. Two broad strategies for understanding the etiology and impacts of DEEs include genetic and complex adaptive systems approaches. This review, inspired by the 2024 Merritt-Putnam Symposium, describes current perspectives of DEE, identifies limitations of current views, and discusses potential novel ways forward. First, we discuss the rationale for a reevaluation of the role of seizures in the pathogenesis of cognitive and behavioral impairments in DEE. Second, we discuss newly emerging methods employing neural organoids to study brain development and DEE in vitro. Third, we present recent precision therapy approaches for the clinical treatment of DEE. Lastly, we discuss computational systems approaches to understanding the genetic landscape of DEE. The severe and multifaceted impacts of DEE and associated comorbidities underscore the necessity of novel interdisciplinary approaches to produce an improved understanding of etiology and more effective treatment strategies.
Introduction
Developmental and epileptic encephalopathy (DEE) is a concept that has its origins in the 1960s when Gastaut introduced the idea that epileptic activity per se contributes to adverse cognitive and behavioral outcomes in children with epilepsy. In 2001, the concept had become more mainstream and was therefore included in the 2001 ILAE classification scheme for epilepsies. 1 The recognition that there was a developmental component was recognized by 2020 and the ILAE terminology moved from epileptic encephalopathy to developmental and epileptic encephalopathy. 2 The developmental aspect refers to the cognitive and behavioral impacts of etiology, independently of epileptic activity, either clinical or in the EEG. DEEs are among the most severe and most difficult to treat epilepsies.
There are currently two broad approaches to understanding developmental aspects of DEE. The first is a genetic approach in which specific genes are identified in children with DEE. The downstream implication is that precision therapy, based on pathophysiological knowledge of specific genes will have major effects. The second is a complex adaptive systems approach in which interactions between elements such as genes, proteins, or neurons are viewed as system-level mechanisms that are targetable in their own right.
In this review, inspired by the presentations in the Merritt-Putnam Symposium at the 2024 meeting of the American Epilepsy Society, we will describe current perspectives of DEE, identify limitations of current views, and discuss potential novel ways forward.
Seizures in the Pathogenesis of Cognitive and Behavioral Impairments in DEE—Time to Reevaluate?
There is a clear correlation, in a subset of children, between having epileptic seizures or subclinical discharges in the EEG and having cognitive impairments and/or psychiatric comorbidities.
3
This observation in conjunction with some biological plausibility arguments has been construed as a causative relationship in which the epileptic activity is causing harm with resultant pervasive cognitive impairments that are present even outside of the time that discharges are occurring. The consequence has been very aggressive treatment of seizures. Unfortunately, this has not had a major clinical impact on cognitive and behavioral impairments. It is therefore appropriate that the idea of causation is re-evaluated. There are several lines of evidence that do
Seizures Induced in Animal Models
There have been many attempts to show that seizures induced in early life are associated with significant cognitive impacts when animals are evaluated in adulthood. When rats or mice are exposed to frequent flurothyl or pentylenetetrazol-induced seizures during the first few weeks of life, those animals show mild cognitive impairments that are not in line with the magnitude of impairment observed in children with DEE.4,5 Subclinical flurothyl-induced status epilepticus for 10 days has no important cognitive impact. 6 Importantly, the induction of multiple early life seizures in a model of malformation of cortical development has no additional effect on the large cognitive impairment that is a direct function of the malformation. 7 Cortical infusion of tetrodotoxin (TTX) for 4 weeks causes infantile spasms and cognitive impairments, but only the animals that had spasms were included in the behavioral analyses. Previous work in this model suggests only 30% of the animals develop spasms. Unfortunately, the effect of TTX in animals without spasms was not reported and thus the direct effect of TTX cannot be ascertained. 8 Therefore, showing that ictal activity per se is a major cause of cognitive and behavioral impairments in DEE has not been convincingly shown using animal models.
Longitudinal Follow-up in Individuals
Many cross-sectional studies have shown a relationship between the duration of epilepsy and failure to make appropriate cognitive gains. Epilepsy duration has been conflated with seizure burden despite some children having few seizures over many years and others having many seizures over few years. This is difficult to manage in a cross-sectional study and therefore, it is more informative to identify changes in cognitive trajectories in individuals. Such longitudinal follow-up fails to identify progressive changes in cognitive and behavioral performance. 9 In addition, the most important predictor of later cognitive performance is current cognitive performance with little to no relationship between an individual's seizure burden or seizure type. 10 This constellation of findings is most consistent with the etiology of epilepsy (the D in DEE) being the major driver of nonseizure outcomes.
Treatment Outcomes
There have been many drug trials that have included children with DEE. Many have shown a statistically significant reduction in seizure burden but almost universally, there have been no significant impact on cognition or behavior. A possible exception is the use of fenfluramine, which may have a small impact on executive function. However, in clinical practice, this has not resulted in important improvements in quality of life. Outcomes from surgical series suggest that there is an improvement in cognitive outcomes over 10–15 years that is not observed within the first few years of becoming seizure free. Seizure freedom could lead to weaning of medications, better access to school, and more access to enriching activities, all of which could positively impact cognition. Thus, it is not obvious that the reduction in seizure burden is an important driver of cognitive outcomes.11-13
Seizure Prevention
The EPISTOP 14 and PReVENT 15 trials evaluated whether prevention of infantile spasms with early treatment of children with tuberous sclerosis. Although the strategy was effective in reducing the number of children that had infantile spasms, there was no positive impact on cognitive and behavioral outcomes at 24 months of age. Overall, these arguments suggest that the role of seizures in the pathophysiology of cognitive and behavioral disorders has been heavily overstated and that it is now time to concentrate on the developmental component of DEE with the view that this has a far greater chance of improving outcomes in DEE than treatment of seizures.
Approaches, Promises, and Challenges of Neural Organoids to Study Human Development and Disease
Understanding the effect of genes on cells, networks, and molecular pathways is crucial for developing effective treatments for DEEs. While animal models like mice and zebrafish are invaluable for studying seizure mechanisms and discovering treatments, they face limitations due to species-specific differences and incomplete replication of human neurodevelopment. Human induced pluripotent stem cell (hiPSC)-based models address these limitations by enabling the study of patient-derived cells and tissues in a human-relevant context. Recent examples of disease-specific iPSCs include Rett syndrome, tuberous sclerosis complex, DEPDC5-related epilepsy, Angelman syndrome, CDKL5 encephalopathy, STXBP1 encephalopathy, Dravet syndrome, and focal cortical dysplasia type II.16-23 Traditionally, researchers used 2D neuronal differentiation to model these diseases, but this approach lacks the complex tissue architecture, cell–cell interactions, and diverse cell types necessary for fully capturing disease mechanisms.
What are Organoids Good for?
Organoids are self-organizing, 3D structures derived from stem cells that recapitulate key features of brain development. These models are valuable for studying neural differentiation, maturation, neurodevelopmental disorders, network formation, and electrophysiological properties. Compared to traditional models, organoids offer several advantages. Derived from human cells, organoids bypass interspecies differences in brain structure and gene expression. There are also ethical advantages as they avoid the ethical concerns associated with higher-order animal models. Furthermore, there is the advantage of scalability as organoids are suitable for screening for therapeutic compounds.
In the Hsieh lab, Vanesa Nieto Estevez and Parul Varma are leading research on DEEs using neural organoids, specifically cortical organoids (COs) and ganglionic eminence organoids (GEOs), to study excitatory and inhibitory neurons. Focusing on ARX (Aristaless Related Homeobox), a gene critical for GABAergic interneuron development and linked to epilepsy and lissencephaly, we used iPSC lines from ARX patients with poly-alanine tract expansion and healthy controls. 24 In COs, early stages (30 DIV) showed increased ARX expression and radial glial cells, implicating cell cycle regulation in cortical progenitors. 25 At later stages (120 DIV), cortical neuron differentiation decreased, suggesting altered neurogenic potential. In GEOs, ARX expression was reduced which was associated with decreases in GABAergic progenitors and inhibitory neurons. Another method of generating excitatory and inhibitory neuron circuits is to culture CO and GEOs separately and then fuse them to generate “assembloids.” 26 In ARX mutant organoids, interneurons migrated faster and farther compared to controls. This was linked to elevated expression of CXCR4, a chemokine receptor critical for interneuron migration and a direct ARX target. Treatment with the CXCR4 inhibitor AMD3100 normalized interneuron migration, confirming CXCR4's role. Electrophysiological assessment using multielectrode arrays (MEA) showed hyperactive networks in ARX mutant organoids, with increased active electrodes, firing rates, and synchronized bursting activity.
Current Limitations and Future Directions of Organoid Models
Organoids enable the modeling of gene disruptions (eg, ARX) and their effects on differentiation, migration, and network activity, as well as complex processes, such as interneuron migration using assembloids. They hold promise for therapeutic screening in DEEs but have limitations. Organoids remain small, variable, and fail to fully replicate the human brain's complexity. Many protocols fail to generate nonneural components like vasculature or microglia. Another issue is standardization. There is extensive batch-to-batch variability and scalability remains a challenge for experimental reproducibility. Despite these limitations, improvements in differentiation protocols, bioengineering techniques, and long-term culture conditions are on the horizon. Organoids are transforming our ability to study early brain development and discover therapeutic interventions for DEEs. They provide a human-relevant platform to investigate disease mechanisms at molecular, cellular, and network levels and test gene-editing strategies and drug candidates in a preclinical setting.
Precision Therapy Approaches for DEE in the Clinic
Over the past three decades, there has been an explosion in genetic discovery with over 1000 genes now identified in patients with epilepsy, 27 the majority in the DEEs. In 2025, the yield of genetic investigations in a patient with DEE can reach 50%, with diagnostic rates highest in the neonatal and infantile onset DEEs and in specific DEEs such as Dravet syndrome.28,29 Identification of variants in genes encoding a diverse range of processes including ion and solute channels and transporters, synaptic signaling and scaffolding proteins, cell growth and metabolism and regulation of gene expression have expanded our understanding of the pathophysiology of the DEEs.28,30 This era of discovery has provided the opportunity to use targeted or precision treatments directed at the genetic etiology.
Currently available antiseizure medications (ASMs) such as sodium channel blockers (eg phenytoin or carbamazepine) can effectively treat seizures in some patients with gain of function early onset sodium and potassium channelopathies.31,32 Early identification of a metabolic etiology such as glucose transporter deficiency due to pathogenic variant in SLC2A1 can allow timely introduction of the ketogenic diet with positive impacts on outcomes. 33 Understanding the genetic basis of the DEE allows development of in vitro high throughput assays to test both novel small molecules and repurposed therapies. Drug repurposing refers to the practice of using an existing drug for a medical condition or indication different from the original approval. 34 This approach is attractive in the DEEs due to the protracted timeframe of novel therapy development. Examples include repurposing of fluoxetine in KCNT1-DEE, 35 memantine 36 or radiprodil (Phase 1B trial, NCT05818943) as NMDA antagonists in gain of function GRIN disorders and L-serine as an alternative NMDA agonist in loss of function GRIN-related DEEs. 37 For the SLC6A1-related neurodevelopmental disorder due to loss of function of the GABA transporter, phenylbutyrate, a medication used in urea cycle disorders, improves GABA transporter trafficking to the cell surface and reduces EEG spiking EEG in a mouse model. 38 In C. elegans and murine neuronal models of STXBP1-neurodevelopmental disorder, phenylbutyrate rescued abnormal protein aggregation. 39 In an open label pilot study (NCT04937062), phenylbutyrate is under assessment for both SLC6A1 and STXBP1-related disorders. 40 Development of novel small molecules exploits knowledge of the effect of pathogenic variants on structure and function of the affected protein or channel, such as targeting persistent sodium current in SCN2A and SCN8A DEE (NCT05818553). 41
Another approach is to directly target the genetic perturbation at a DNA or RNA level. RNA therapies alter mRNA processes to modify protein production, utilizing either synthetic short nucleotide sequences, or oligonucleotides, or small RNA molecules, such as small interfering RNA (siRNA) or microRNAs (miRNAs), to silence the expression of specific genes. 42 These therapies currently require intrathecal delivery due to their rapid degradation. In DEEs due to toxic gain of function variants such as SCN2A, SCN8A, and KCNT1 DEEs, ASOs which reduce the total amount of protein production in a nonallele specific manner have been shown to rescue epilepsy phenotypes in murine models,43-45 with an SCN2A ASO Elsunersen in current trials (NCT05737784). Interestingly, an ASO targeting Nav1.6 also improved survival in a Dravet mouse model, while decreasing KCNT1 expression with an ASO rescued survival in Scn1a and Scn8a mouse models,44,46 suggesting there may be a role for broader use of gene-targeting ASOs across a number of DEEs. For loss of function disorders such as Dravet syndrome, alternative approaches to upregulate gene expression or protein production are required. Poison exons are RNA sequences naturally occurring in some transcripts that, when included in mRNA, result in nonsense mediated decay. This process can be targeted to convert nonproductive to productive transcripts using a short ASO which excludes the poison exon and leads to more protein production. In a Dravet mouse model this approach reduced seizures and SUDEP,47,48 leading to first in-human safety and tolerability studies of STK-001 (NCT04442295 and ISRCTN99651026). For loss of function disorders such as Dravet syndrome, classical AAV gene therapy is precluded by the size of Nav1.1 which exceeds the AAV packaging limit. An alternative approach, using AAV9 targeted to interneurons to deliver a transcription factor which upregulates Nav1.1 expression rescued seizures and survival in a mouse model 49 and has recently gone into Phase 1/2 trials (NCT06283212; NCT06112275). A number of other approaches focused on RNA, genome and epigenome editing are under exploration in preclinical studies and hold great promise for the DEEs. 50
In conclusion, genetic discovery has paved the way for a number of potentially disease-modifying genetic etiology-directed treatments, which in turn will teach us a great deal about the link between the genetic variant and the complex, evolving system of the developing brain.
Systems Approaches to Investigate DEE Pathophysiology
Our genetic understanding of DEEs is rapidly evolving. To date, variants in around 900 genes have been associated with DEEs, 28 each a perturbation that can result in epilepsy and neuropsychiatric comorbidities. This wealth of findings bodes well for developing novel therapies. Gene-trait associations for a drug target are among the strongest predictors of drug trial success. 51 Moreover, precision therapies often generalize to phenotypically related diseases. 52 To this end, a systems-level picture of the genetic landscape of DEEs can help us organize novel treatment hypotheses and approaches. Systems biology is a paradigm for reasoning about biological systems that emphasizes networks of interactions (among genes, cells, circuits, etc) and the emergent phenomena resulting from the collective dynamics of these networks. No gene functions in isolation but instead is embedded in a molecular network that dynamically modulates the expression of all genes through complex feedback loops. In the systems framework, the DEE genetic landscape is beginning to show important biological themes that expose the limitations of our existing pharmacopeia.
DEE-associated genes are enriched for a broad variety of neurological processes, including development, plasticity, and neural membrane excitability. 28 Current ASMs are dominated by neuron-targeted drugs that regulate membrane excitability, but these therapies have limited efficacy in DEEs. 28 The wealth of genetic associations outside of membrane excitability strongly argues for an expansion of the target landscape to these other processes. However, prioritizing among possible targets remains a significant challenge.
One path forward is to recognize that clinical outcomes in DEEs are strongly influenced by polygenic effects of common polymorphisms. 28 For example, one recent study showed that patients with Dravet syndrome have higher polygenic risk for low IQ and lower polygenic risk for shorter lifespan, suggesting that common variation affecting cognitive and longevity phenotypes may be contributing significantly to DEE phenotypes. 53 Corroborating these human findings, mouse studies of Dravet syndrome have shown that mouse outcomes can vary from no overt phenotype to catastrophic epilepsy with early-life mortality as a function of genetic background. 54 These findings suggest that genetically mapping resilience to severe outcomes is a powerful tool for identifying novel therapeutic targets.
It is important to stress, however, that polygenic effects are driven by hundreds or thousands of polymorphisms, each with very small phenotypic effect. Pritchard and colleagues have posited the “omnigenic model” of complex traits wherein the vast majority of heritable variation is driven by extremely small effect regulatory polymorphisms that indirectly influence the expression of “core” trait-driving genes through a complex network of molecular interactions.55,56 While still controversial in some details, 57 the omnigenic picture correctly emphasizes a key role for detecting the aggregate genetic effect on trait-driving gene expression programs.
The systems genetics paradigm integrates genetic analysis of clinical traits with high-throughput molecular assays to resolve genetic effects at the molecular level. 58 Systems genetics approaches have been applied to rare and common epilepsies with tantalizing mechanistic findings, including the identification of SESN3 as a proconvulsive gene regulating neural inflammation and convergent gene networks driving of cognitive and neurodevelopmental phenotypes.59,60 Systems genetics studies in human patients are limited by the availability of relevant tissues for molecular analysis, but genetic diversity resources in mice 61 and human iPSC platforms 62 make experimental systems genetics analyses of DEEs possible. Preclinical work in these systems has the potential to dramatically improve our understanding of mechanisms of resilience and focus pharmaceutical development efforts on novel pathways.
Beyond pharmaceuticals, stimulation-based therapy is an attractive option to exogenously modulate brain activity in a way that the malformed brain will not do autonomously. While stimulation is now used to control seizures, emerging evidence suggests that it is also possible to target cognition in patients with epilepsy.63,64 Stimulation-based therapies act at a higher level of organization than pharmaceuticals but are complementary in that they directly modulate brain activity toward normal, potentially activating plasticity mechanisms en masse. Such approaches are in their infancy and require extensive further research but could become a powerful adjunct to pharmaceutical interventions.
The complexity of DEEs precludes simple therapeutic hypotheses for curative treatment. The systems-level view of genetic landscape of DEEs suggests that pathophysiology must be addressed at multiple spatial and temporal scales, with a view to how these different scales interact as a system. Both pharmaceutical approaches targeting mechanisms of resilience and stimulation approaches augmenting homeostatic brain functions may be required for optimized care.
Concluding Remarks
DEE and its associated comorbidities are both severe and multifaceted. This complexity underscores the necessity of applying novel interdisciplinary approaches to the problem. Continued integration of advances in precision genetics with new model systems, advanced analytical and computational tools, and appreciation of the interacting and distinct developmental and epileptic factors of DEEs promises to improve treatment of children with epilepsy.
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
