Genetic Variants and Disease Mechanisms: Lessons from Monogenic Childhood Epilepsies

Abstract

The elucidation of the molecular basis of monogenic epilepsies is advancing rapidly. For clinicians, knowing not only the affected gene, but also the patient's exact genetic variant and gaining insight into its effect on RNA, protein, cell, and organism level is becoming increasingly important. As different variants in the same gene can lead to opposing functional effects, an understanding of their nature is crucial for informed treatment choices. Correctly counseling patients, parents, and families regarding the patient's prognosis and the risk to other family members of being affected or having an affected child is only possible with detailed knowledge of the genetic and functional alterations underlying the condition. This review aims to provide a comprehensive overview of genetic variants and their effects, following them from the DNA to the organism level. Protein-level outcomes, such as gain- and loss-of-function mechanisms as well as dominant-negative effects, will be illustrated using examples from monogenic epilepsies. Their downstream impact on cellular function and phenotype will be traced to shed light on the mechanisms by which different variants in the same gene can result in diverging clinical presentations. In doing so, we illustrate key genetic concepts relevant to clinical practice to help inform clinical interpretation of genetic variants and facilitate therapeutic decision-making.

Keywords

monogenic epilepsy - genetic variants - variant effects - gain-of-function - dominant-negative - loss-of-function

Introduction

Understanding the basis of monogenic epilepsies is crucial for accurately diagnosing, treating, and counseling patients and their families. Variants in genes implicated in epilepsy cause a wide array of different effects on the molecular, multimeric, cellular, and ultimately the organismal level. Understanding and correctly predicting such effects is one of the ongoing challenges of clinicians and scientists involved in human genetics and disciplines of fundamental science, but also for those treating patients with monogenic disorders in their respective specialties.[1] [2] The exact genetic alteration underlying a seizure phenotype and its functional effects determine the available treatment choices and prognosis of the patient, as well as the risk for other family members to be affected or to have an affected child. As research progresses and more targeted therapies become available, deciphering disease mechanisms will become increasingly important to guide treatment decisions.[3] [4]

Following Variant Effects from DNA to Organism Level

This review aims to trace variant effects from the DNA level to their consequences for protein function and interaction, as well as their downstream impact on the cell and organism, ultimately leading to disease ([Fig. 1]).[5] We will first highlight different types of variants on a DNA level, while addressing their effects on protein translation and the resulting phenotype in the subsequent sections. Well-characterized examples drawn from monogenic childhood epilepsies will be used to illustrate such functional consequences. While we aim to provide a comprehensive understanding of the most relevant and some specific molecular mechanisms involved, this review does not attempt to offer an exhaustive catalog of all monogenic epilepsies.[6] [7] [8] We will focus on variants affecting single genes associated with monogenic epilepsies, acknowledging that copy number variants encompassing several genes may also contribute to epilepsy phenotypes.

Fig. 1 Most common effects caused by heterozygous genetic variants from the DNA to the organism level. Variants on the DNA level can be single-nucleotide substitutions, which result in the RNA in the exchange of an amino acid codon (missense variant), in the alteration of a splice site (splice variant), or in the occurrence of a premature termination codon (PTC; nonsense variant). The deletion or insertion (indel) of one or several nucleotides (number indivisible by three) causes a frameshift in the three-base-pair reading frame, generally also resulting in a PTC (frameshift variant). An indel of a number of nucleotides divisible by three causes no frameshift but adds or removes one or several amino acid codons. Different protein level alterations can result: the lack of protein due to mRNA undergoing nonsense-mediated mRNA decay (NMD), a truncated protein (missing amino acids at the beginning or the end of the protein), a shortened or elongated protein due to the addition or deletion of several amino acids via splice effects or longer indels or a protein altered at one single amino acid position. Variants resulting in no protein production are also referred to as null variants. On the multimer/cell level, protein changes can cause a regular function, the loss- or gain-of-function (GoF), or dominant-negative effects, resulting either in a disease phenotype or in a healthy individual. Wider arrows indicate outcomes that are more likely. In rare cases, other outcomes are possible (e.g., a shortened protein with a GoF effect). (Figure was created with permission with BioRender and is available at: https://BioRender.com/7zcw9ia.)

Variant Types on the DNA and RNA Level

Identifying a variant on the DNA level is the first step to understanding its further effects on higher levels. In genetic nomenclature, a DNA-level variant can be identified by using genomic position (denoted by g.) and the nucleotide change (e.g., g.165374698C > A), in which nucleotides are numbered according to their position on each chromosome. More often, variants are labeled in reference to the coding sequence (denoted by c.) of a gene (e.g., c.3986C > A), which includes only the exonic regions of a gene and starts with the start codon (ATG) “A” as c.1.[9] As many genes undergo alternative splicing, they may have multiple coding sequences that differ in the number and arrangement of exons. Each varying coding sequence in a gene is assigned a transcript number (e.g., NM_001040142.2), which must necessarily be provided to unambiguously describe a coding position.[10] The protein change (denoted by p.) resulting from the DNA level variant is given by specifying the first amino acid of a gene that is altered and its number, followed by the alteration (alternate amino acid, stop codon, or description of a frameshift sequence), for example, p.(Ala263Val).[9] Stating the zygosity indicates whether a genetic change is found on one of two alleles (heterozygous), both alleles (homozygous), or on the single gene copy (hemizygous, e.g., for X-linked conditions).

Missense Variants

Missense variants cause alterations of the amino acid sequence due to single-nucleotide substitutions. Prediction of their effect on protein function is challenging, as the resulting protein may have a completely or partially reduced function or, conversely, an increased or altered protein function. In case of benign variation, the amino acid substitution might also have no obvious effect on protein function at all. The location of a missense variant in a gene (e.g., in a functionally important domain), as well as the exact amino acid alteration (e.g., exchanging two amino acids with very dissimilar properties), must be considered when assessing the effects of missense alterations.[11] [12] [13] [14] Also, a substitution at an amino acid position that is highly conserved, that is, unchanged in the reference sequence across homologous genes in different species, is more likely to have a more detrimental effect on protein function.[15] [16] An indication as to whether a given nucleotide alteration has a strong, and therefore, possibly deleterious, influence on the protein configuration can be gleaned from bioinformatic scores, such as CADD[17] and REVEL.[18] Newer prediction tools use artificial intelligence to help variant assessment, for example, AlphaMissense.[19] An evaluation of prediction tools in an epilepsy cohort found that REVEL and AlphaMissense performed particularly well in predicting pathogenicity.[20]

Nonsense and Frameshift Variants

Such variants result in premature termination codons of protein translation (PTCs): in nonsense variants, the PTC arises directly at the site of the variant, while in frameshift variants, the deletion or insertion of one or several nucleotides leads to a shift in the three-base-pair reading frame and results in a PTC downstream of the variant.

Splice Variants

Variants affecting splicing can have different effects on mRNA processing and the resulting protein product—they may disrupt canonical splice donor or acceptor sites, impeding the correct recognition of the beginning or the end of an exon. As a result, exons may be skipped, extended, or shortened, while introns may be retained either fully or partially.[21] [22] [23] [24] [25] [26] [27] [28] [29] Splicing alterations in which the reading frame is preserved can lead to the formation of a stable but altered protein, while, more commonly, changes in splicing may cause a frameshift effect that introduces a PTC. Additionally, deep intronic variants may result in the inclusion of so-called poison exons, which include a PTC. Such exons can have a physiological function in regulating gene expression, but certain (intronic) variants may cause a poison exon's aberrant, constitutive inclusion.[30] [31]

Deletions and Insertions (Indels)

The addition or loss of nucleotides in the genomic sequence can range from a single base-pair insertion or deletion to copy number changes involving parts of or entire chromosomes. Larger changes (>50 bp) are referred to as structural variants and can include complex rearrangements (e.g., duplications, inversions, or translocations). In intragenic indels, in-frame alterations may change the amino acid sequence without disrupting the reading frame, with small in-frame indels comprising only one or a few amino acids often acting similarly to missense variants. Out-of-frame changes (i.e., the addition or loss of a number of nucleotides indivisible by three) usually introduce a PTC.[32] [33] [34] [35]

Methylation

Aberrant methylation refers to genetic alterations on the epigenetic level without altering the nucleotide sequence by adding or removing methyl groups from cytosine nucleotides. This often results in gene silencing, as hypermethylation in regulatory regions can prevent normal gene expression.[36] [37]

Repeat Expansions

Repeat expansions refer to the increase in the number of repetitions of a motif, generally no more than two to six bases in length (short tandem repeat), although some disease-associated repeats consist of longer units.[38]

Variants in RNA Genes

While most known monogenic disorders are caused by DNA variants in protein-coding genes, some genetic conditions arise from variants in genes that encode RNAs. For example, variants in the gene RNU4-2 cause a severe neurodevelopmental disorder, often associated with a seizure phenotype.[39] The gene is transcribed into a small nuclear RNA (snRNA) that forms, together with proteins, small nuclear ribonucleoproteins (snRNPs) as part of the spliceosome. Variants in RNU4-2 convey pathogenicity by directly impairing the formation of snRNPs and the functional integrity of the spliceosome. This example highlights that while genetic variants are often assessed in terms of their impact on the protein, some alterations exert their effects primarily at the RNA level.

Variant Effects on the Protein Level

To understand the impact of DNA level variants and their RNA counterparts on the organism, it is crucial to observe the effects on protein function and interaction, ideally by assessing these effects experimentally. As this is not always possible, conclusions are often drawn based on the type of DNA variant to estimate its effect on the RNA and protein level ([Fig. 1]).

Loss-of-Function Effects

Loss-of-function (LoF) effects can refer to a complete LoF, as caused by an entirely absent, unstable, or severely impaired protein, or to a partial LoF, when an altered protein is able to fulfill part of its role. LoF effects on a protein level can thus be caused by different types of variants on the DNA level, including missense variants.

The overall impact of LoF depends on whether the genetic alteration is present on one or two alleles and, in a heterozygous state, whether the organism carrying the affected gene is sensitive to the loss of protein production from one allele (dosage effect). Some genes are tolerant toward heterozygous LoF variants, generally because the amount of protein produced from one allele is sufficient to sustain a correctly functioning organism.[40] [41] [42] An instrument for the measurement of dosage sensitivity (haploinsufficiency) is the “probability of LoF intolerance” score (pLi; ranges between 0 and 1, a higher score indicates intolerance) and the more refined “LoF observed/expected upper bound fraction” (LOEUF; ranges between 0 and approximately 2, a lower score points toward haploinsufficiency). For both parameters, the proportion of expected variants causing complete LoF in a gene (calculated based on the rate of mutation and length of the gene) is compared to the observed number. These metrics can be accessed on the gnomAD website (https://gnomad.broadinstitute.org/).

The function of genes that do not display haploinsufficiency may be impaired if pathogenic variants are present on both alleles or if the only copy physiologically present (e.g., for X-linked genes in males) is nonfunctional.[43] These conditions are referred to as autosomal-recessive (or X-linked) and are caused by homozygous (i.e., the identical variant on both alleles), compound-heterozygous (i.e., different variants on each allele), or hemizygous (one variant on the only copy present) variants. Monogenic epilepsies caused by LoF effects mainly arise from heterozygous variants. Some notable exceptions, which are due to biallelic variants, include inborn errors of metabolism associated with an epilepsy phenotype, such as pyridoxine- and pyridoxal phosphate-dependent epilepsies (variants in ALDH7A1, PNPO, and PLPBP).[44] [45] [46]

Complete LoF from a variant-carrying allele is routinely expected when a PTC is present, as is the case in nonsense, frameshift, and many splice variants, or when a deletion affects the entire gene or large parts of it. PTCs usually trigger nonsense-mediated mRNA-decay (NMD), that is, the complete degradation of mRNA, when they are recognized during the pioneer round of translation.[28] [29] Variants which result in the absence of any protein are also referred to as null variants. Missense variants resulting in the translation of an unstable protein, which is degraded before it can perform any function in the cell, may also be functional null variants.

Of note, PTCs that arise in the last exon or in the last approximately 55 bp of the second-to-last exon can escape NMD and lead to the production of a truncated protein. In rare instances, frameshifts in the last exon or variants altering the stop codon may also cause protein elongation. PTCs at the very start (probably within approximately 150 bp of the canonical start codon) may also evade NMD and produce a truncated protein, as the ribosome may initiate at a downstream alternative start codon.[24] [25] [26] [27] Predicting the occurrence of NMD as well as the effects of such truncated or, rarely, elongated protein is sometimes equally challenging as forecasting those of a missense variant.[47] [48]

Partial LoF caused by altered proteins (e.g., resulting from missense, truncating, or in-frame insertions or deletions, or reading-frame preserving splice variants) can also result in a phenotype if present in a gene sensitive to such changes. Depending on the degree of functional impairment, such partially deficient proteins may sometimes cause a milder phenotype than null variants leading to a complete LoF, and are then referred to as hypomorphic.[49] [50]

For example, the haploinsufficient gene SLC2A1, which encodes the glucose transporter type 1 (Glut1) responsible for transferring glucose across the blood–brain barrier, is susceptible to both heterozygous null and missense variants. Null variants causing complete LoF generally result in the more severe, classical Glut1DS phenotype with onset in infancy, characterized by mild to severe intellectual disability (ID) and epilepsy. Heterozygous missense variants in this gene resulting in a partial LoF are associated with a milder phenotype, which is sometimes not diagnosed until adolescence or adulthood and may include mild ID, speech impairment, movement disorders, and behavioral/psychiatric disturbances. Glut1 protein function, as measured by erythrocyte 3-O-methyl-D-Glucose (OMG) uptake, demonstrates a correlation between the clinical and biochemical phenotype and genotype.[51] In rare instances, biallelic missense variants have been shown to cause a severe phenotype, with erythrocyte OMG uptake below the level seen in heterozygotes for null variants.[52]

Heterozygous genetic variants causing LoF in SCN1A induce different epilepsy phenotypes: the milder generalized epilepsy with febrile seizures + (GEFS + ) and the more severe Dravet syndrome. While null variants resulting in complete LoF and missense variants in highly conserved functional domains are more likely to result in Dravet syndrome, missense variants with a smaller LoF effect are prone to giving rise to GEFS + . However, nonsense and frameshift variants located in the very beginning and end of the gene can cause GEFS + , consistent with a suspected NMD escape of PTCs at these positions and a likely partially retained function of the truncated protein.[53] Additionally, the inclusion of recurrent poison exons in SCN1A has been established as a mechanism leading to heterozygous LoF, thereby acting similarly to other null variants in the gene.[54] [55] [56] Accordingly, the reported phenotypes of SCN1A variants with poison exon inclusion include the phenotypes associated with SCN1A LoF variants, that is, GEFS+ and Dravet syndrome.[30] [56]

LoF can also be due to aberrant methylation: Angelman syndrome, an imprinting disorder which causes severe epilepsy in most patients, is caused by different molecular LoF mechanisms affecting the maternal copy of the chromosomal region 15q11–q13, for example, by deletions, uniparental disomy, or imprinting center mutations.[57] [58] This region includes the critical UBE3A gene as well as a gene cluster encoding different subunits of a GABA_A receptor.[59] [60] In a physiological state, some genes are only expressed from the maternal allele, while the paternal allele is physiologically methylated, and therefore silenced. Aberrant methylation additionally renders the only physiologically active (i.e., maternal) allele nonfunctional, similar to X-linked conditions in males, resulting in a complete lack of protein from the relevant region.

Repeat expansions in promoter regions may be another cause of LoF effects due to silencing of the gene via aberrant methylation: progressive myoclonic epilepsy type 1 (EPM1, Unverricht-Lundborg disease), which is characterized by, among other symptoms, progressive myoclonic jerks, generalized tonic-clonic seizures and progressive neurodegeneration, is caused by biallelic dodecamer repeat expansions or one repeat expansion in compound-heterozygosity with a pathogenic sequence variant in CSTB. It is one of the few monogenic epilepsies caused by repeat expansions with onset in childhood, generally between the ages of 6 and 15 years. The expanded repeat causes an increase in the distance between promoter elements and the transcription start site, and thereby leads to a LoF effect by decreasing the amount of CSTB mRNA and the corresponding protein. The loss of the CSTB-encoded cathepsin B, a cysteine protease inhibitor, results in increased neuronal death and a reduction in inhibitory GABAergic neurons, leading to hyperexcitability and seizures.[61] [62]

Dominant-Negative Effects

Dominant-negative effects result from heterozygous genetic alterations and occur when an altered, but stable protein produced from an allele with a missense variant is more detrimental to the cell or the entire organism than a lack of the protein from one allele, as would result from a heterozygous null variant.[63] Such effects result from altered protein interactions, for example, when a protein's changed configuration negatively impacts the formation or function of multimer units and/or impairs its interaction with the wild-type protein produced from the other allele or with other proteins.

Conditions arising from variants in the genes KCNQ2 and KCNQ3 demonstrate that in some genes, the complete LoF of one allele is better tolerated than heterozygous pathogenic missense variants. KCNQ2 and KCNQ3 encode subunits of a heterotetrameric potassium channel comprising two KCNQ2 and two KCNQ3 subunits.[64] The incorporation of mutant subunits appears to impair channel function more severely than the dosage effect resulting from the absence of protein produced from one allele: heterozygous null variants in KCNQ3 normally do not lead to a clinical phenotype, while heterozygous missense variants can cause self-limited neonatal epilepsy (SeLNE), a milder phenotype, and, more rarely, neonatal-onset seizures and neurodevelopmental disorder (NDD), via dominant-negative effects.[65] [66] [67] There seems to be no dosage effect resulting from the loss of protein from one allele, while the incorporation of mutant subunits decreases the function of the entire tetramer.[68] [69] Pathogenic variants in KCNQ2 can lead to SeLNE due to either null variants or missense variants with partial LoF effects, indicating haploinsufficiency. Missense variants with dominant-negative effects on the tetramers result in the more severe phenotype of neonatal-onset DEE (NEO-DEE).[70] [71] [72] [73]

Gain-of-Function Effects

Gain-of-function (GoF) effects are a common pathogenic mechanism of disease in monogenic epilepsies, encompassing heterogeneous ways in which a variant can enhance the protein's normal function. Because a functional protein must be produced to exert a GoF effect, such variants are typically missense changes, usually in a heterozygous state.

Many monogenic epilepsies result from missense variants in genes encoding ion channels, which often alter ion current properties in ways classified as a GoF.[74] [75] [76] [77] [78] Due to such GoF variants, channels can display, among others, hyperpolarizing shifts in activation, depolarizing shifts in inactivation, slower inactivation, faster recovery from inactivation, or increased persistent current ([Fig. 2])[5] causing, for example, an increased frequency of or prolonged action potentials. Other mechanisms implicated in GoF include features such as an increased surface expression or a mislocalization of the affected proteins. Depending on the type of current that is affected, this leads to different outcomes regarding the cell's excitability and, therefore, its propensity to contribute to a seizure phenotype.

Fig. 2 Gain-of-function (GoF) mechanisms due to heterozygous variants in voltage-gated ion channels. Different types of GoF mechanisms can be present in voltage-gated ion channels. (A, B) Voltage-dependent activation/Inactivation curves: A leftward shift in activation of the mutant channel means that the cell will open at more hyperpolarized potentials, that is, more easily. A rightward shift in inactivation from wild-type in comparison to the mutant channel means that the channel needs more depolarized voltages to inactivate and may be able to stay open longer. (C) Current traces of inward current following a depolarizing voltage step: Wild-type channels rapidly inactivate and return to the baseline, while mutant channels show slower fast inactivation and retain elevated current at the measurement time point, indicating increased persistent current and impaired inactivation. (D) Mutant channels show faster recovery from inactivation after a first pulse, enabling the cell to activate again more quickly. (Figure was created with permission with BioRender and is available at: https://BioRender.com/7zcw9ia.)

Different GoF mechanisms may contribute to the pathogenicity of variants in one gene alone: heterozygous pathogenic missense variants in KCNT1, which encodes a sodium-activated potassium channel, cause several different epilepsy phenotypes, among them: DEE with or without epilepsy of infancy with migrating focal seizures (EIMFS) and autosomal dominant or sporadic sleep-related hypermotor epilepsy ([AD]SHE).[79] These phenotypes are caused by GoF mechanisms, while the disease association of heterozygous null variants is not yet clear.[80] GoF mechanisms include an increased current amplitude by constitutive hyperactivation of the channel[81] [82] and an increased open probability due to greater sodium sensitivity.[83] Additionally, increased cooperativity of mutant KCNT1 channels was demonstrated to result in a higher peak current in vitro despite identical protein expression or even reduced unitary conductance.[84]

For ionotropic receptors, such as the heteropentameric GABA_A receptor, GoF effects can include increased GABA sensitivity (requiring lower concentrations of the neurotransmitter to activate), reduced desensitization (remaining responsive to agonist exposure for a longer period), or spontaneous channel activity in the absence of a ligand ([Fig. 3]).[5] This has been demonstrated, for example, for pathogenic variants in GABRB3 and GABRB2, which encode the β2- and β3-subunits of the GABA_A receptor.[85] [86] [87] In both cases, the GoF variants in these genes were linked to more severe phenotypes than their LoF counterparts.

Fig. 3 Gain-of-function mechanisms due to heterozygous variants in ligand-gated ion channels. (A) Alterations in ligand sensitivity: Normalized current of wild-type (black) and mutant (red) receptors in response to increasing ligand concentrations; the mutant receptor responds to lower ligand concentrations. (B) Alterations in the desensitization to a ligand: Schematic representation of desensitization in a ligand-gated receptor; with sustained ligand exposure, the mutant receptor maintains a larger current compared to the wild-type receptor, indicating reduced desensitization, which allows greater ion flow. (C) Channel activity with or without ligand: The mutant channel shows current activity before as well as after the end of ligand application, while the wild-type receptor shows current only after ligand application. (Figure was created with permission with BioRender and is available at: https://BioRender.com/7zcw9ia.)

Variants Displaying Mixed Gain- and Loss-of-Function Effects

Not all variants fit neatly into a LoF or GoF category—in some instances, one measurable property of a protein or protein complex may show a reduced functionality while a different parameter may demonstrate an increase in its effect over the wild type.[86] [88] [89] [90] [91] Pathogenic variants in the abovementioned GABRB2 can lead to phenotypes varying from febrile seizures to severe DEEs. In pathogenic missense variants, both gain- and LoF effects are known to be disease-causing. Some variants, however, display mixed GoF/LoF effects by simultaneously showing reduced current amplitudes but increased GABA sensitivity.[86]

Variant Effects on the Cell and Organism Level

Understanding a genetic variant on the DNA, RNA, and protein levels still does not enable a prediction of the resulting phenotype. In this section, we discuss how the function of the involved protein or channel, its cellular and subcellular location, and variants affecting different protein isoforms play a role in cell and organism outcomes.[78] [92] [93] [94]

The sodium channels encoded by SCN1A and SCN2A provide an illustrative example of how the type of affected channel and its location influence the excitability of a neuronal network. As mentioned previously, heterozygous LoF variants in SCN1A cause a GEFS+ or Dravet phenotype ([Fig. 4]).[5] On the cellular level, a decreased function of a sodium channel reduces its ability to depolarize the cell, and therefore, to generate an action potential. Sodium channel LoF variants therefore decrease the cell's excitability—however, due to the predominant expression of SCN1A in some types of inhibitory GABAergic interneurons, a reduced inhibition in turn causes hyperexcitability in the overall network[74] [75] [95] Additionally, for SCN1A-related Dravet syndrome, not only localization, but also timing of SCN1A expression contributes to an understanding of the phenotype: SCN1A expression in fetal brains is lower than that of other sodium channels and increases steadily after birth. This correlates with the age of onset of Dravet syndrome at the age of on average 6 months, while other sodium-channel-related conditions show earlier disease begin.[96] [97]

Fig. 4 Heterozygous SCN1A variants and their most common effects from the DNA to the organism level. Different types of DNA level variants (SNVs, small indels) can cause the introduction of a PTC (directly at the site of the SNV (nonsense), via frameshift or inclusion of a poison exon by splicing alterations) or an alteration of one or several amino acid codons. On the protein level, this results in NMD, in a truncated (shortened at the beginning or end) or in an altered protein (exchange/addition/loss of one or several amino acids). Most often, the complete absence of protein from one allele or a nonfunctioning altered protein causes Dravet syndrome, a severe type of DEE due to a reduced activity of inhibitory interneurons and an increase in excitatory pyramidal neuron excitability. A partial loss-of-function can be caused by a truncated or otherwise altered protein. A reduced interneuron inhibitory activity and subsequently mildly increased overall excitability most often leads to a GEFS+ phenotype. Missense variants (i.e., amino acid alterations that do not introduce a premature stop codon) can cause both loss and gain-of-function (GoF). Mild GoF effects also cause a severe, but clinically different, type of DEE, most likely via a homeostatic response that increases the excitability of excitatory pyramidal neurons in reaction to increased inhibition. An even stronger GoF most likely causes an increase in extracellular potassium due to heightened firing from inhibitory interneurons. This may lead to cortical spreading depolarization, an effect that is associated with migraine, but not seizures. The resulting phenotype is familial hemiplegic migraine type 3 (FHM 3). (Figure was created with permission with BioRender and is available at: https://BioRender.com/7zcw9ia.)

SCN2A, in contrast, is expressed mainly in excitatory neurons, and therefore, shows a different genotype-phenotype correlation than SCN1A. In SCN2A, GoF variants leading to hyperexcitability of excitatory neurons cause increased seizure activity and lead to self-limited familial neonatal-infantile epilepsy (SeLFNIE) or, with more pronounced GoF effects, to neonatal-onset DEE.[90] [98] Also, SCN2A shows a time-dependent expression of transcripts: exon 5 is expressed in a neonatal and a mature isoform, with a switch in the proportion of expression from a predominantly expressed neonatal form to a mainly expressed mature form soon after birth over a transition period that stretches from 24 weeks p.c. to the age of 6 years.[99] Despite great sequence homology between the neonatal and the mature exon, these isoforms can cause differences in their electrophysiologic properties and hence, in the effects these have on the phenotype of the individual. For example, mice expressing the neonatal SCN2A isoform showed reduced neuronal excitability in neurons and decreased seizure susceptibility in comparison to mice expressing only the adult form.[100] Some GoF variants causing early-onset DEE have been shown to support an increase in neuronal excitability when expressed in their neonatal isoform, for example, via a hyperpolarized shift in voltage-dependence of activation, which was not observable when comparing the adult isoforms with the wild type.[101] For several variants causing early-onset DEE, an overall predominance of several excitability-inducing GoF properties in the neonatally-expressed transcripts shifted to a mixed GoF/LoF set in the adult isoform.[88] However, shifts in isoform expression are not alone in mediating timing of disease onset: in one study, one of two examined variants leading to SeLNE showed GoF effects only in the neonatal transcript, the other one demonstrated GoF effects in mainly in the adult form; however, the expression of the adult isoform in the relevant axon initial segments was diminished in further development and replaced by the SCN8A-encoded Nav1.6, thereby possibly offsetting their impact.[102]

However, in both SCN1A and SCN2A, the mechanism presumably leading to reduced neuronal excitability (i.e., GoF variants in SCN1A, expressed in inhibitory neurons, and LoF variants for SCN2A, expressed in excitatory neurons) can also cause seizure phenotypes. For example, moderate GoF effects of certain variants in SCN1A lead to early-onset DEE,[103] and SCN2A LoF variants are associated with infant and childhood-onset DEE.[90] This demonstrates that a simplistic view equating enhanced inhibitory interneuron excitability or decreased excitatory neuron activity with a reduced seizure phenotype in an organism does not reflect the complexity of the brain. In SCN1A, it has been suggested that neuronal homeostatic plasticity restoring firing rates can contribute to an overall epilepsy phenotype[104]; however, the understanding of such network effects is still limited.

Similarly complex mechanisms play a role in GABA_A receptor-related disorders. GoF mechanisms in GABA receptors contribute to enhanced hyperpolarization of the neuron caused by increased chloride influx. Though this results in hypoexcitability of the cell, the overall effect of the GoF variant on epileptogenicity is dependent on the localization in the brain of the cells affected by the variant. While GABA_A receptors are expressed in both excitatory and inhibitory neurons, the decreased excitability of inhibitory neurons appears to underlie the severe epilepsy phenotype in patients with GoF variants.[78] As mentioned previously, some variants in GABA_A receptors display mixed GoF/LoF effects—rendering the prediction of variant effects on an organism level even more difficult.[86] While the authors surmised that a decreased current amplitude (LoF effect) might counteract the heightened GABA sensitivity (GoF effect) and result in an overall LoF phenotype in the patients, this did not align with the clinical picture in the affected individuals: patients carrying such variants with mixed LoF/GoF effects more closely resembled the more severe GoF phenotype. However, other types of mixed variant effects in GABRB2 or different genes may produce different phenotypic outcomes; this example serves to demonstrate that neither bioinformatic prediction nor functional measurements in overexpression systems are enough to reliably predict variant effects on an organism.

The latter highlights the difficulties faced when choosing adequate medication, be it a conventional anti-seizure medication targeting specific channel functions (e.g., sodium channel blockers) or a personalized therapy consisting of, for example, anti-sense oligonucleotides (ASO).[3] [4] [55] [89] [103] [105] [106] To tailor such therapy specifically to a condition, the underlying mechanism must be known.

Limitations of Studies Characterizing Functional Effects

While extensive work has been conducted to improve our understanding of variant effects, numerous caveats remain.

When assessing variant effects experimentally, several challenges complicate the goal of replicating human genetic variants in model organisms or cells. Even though endogenous expression systems or in vivo models would be ideal to study variant effects, many studies rely on measuring these in overexpression experiments, in which cells are transfected with a gene or genetic variant of interest and its electrophysiological properties are measured via methods such as patch-clamp. Such methods can provide insight into current-voltage relationships, activation curves, and response to ligands. However, as the cell overexpresses the transfected genes, this may not adequately reflect the effects of variants when endogenously expressed. Additionally, interacting proteins, such as the β-subunits of sodium channels, are often not considered. The physiological location of the protein in question, for example, the subcellular location within the neuron, and differing patterns of expression depending on the various types of neurons, cannot be represented in their complexity. Variants that were identified as disease-causing in a patient occasionally may demonstrate no abnormalities in their electrophysiologically measured properties. While for some variants this may be due to an erroneous attribution of pathogenicity, in some cases, disease-causing effects may go unnoticed in electrophysiological assays, as their pathogenicity could result from mislocalization, impaired trafficking, or altered interaction with other proteins that are not always accounted for in the experimental setup. Wider interactions within a neuronal network are equally hard to imitate in experimental conditions. Some variant effects may show altered electrophysiological findings depending on specific conditions in which the experiments are conducted (e.g., temperature[107]).

When considering multimeric channels, such as potassium channels (heterotetramers) and GABA_A receptors (heteropentamers), reproducing authentic stoichiometric ratios between subunits presents a challenge. To standardize experiments, such channels are sometimes modeled as concatenated constructs, that is, the subunits on a receptor are cloned onto a common sequence. This leads to the expression of multimers with fixed stoichiometry, for example, with one receptor containing always one mutant and one wild-type subunit, possibly along with different other subunits. While improving reproducibility and standardization, this does not necessarily reflect the distribution in a real organism and may neglect the interplay between two mutant subunits in one channel, as well as the effects that result from differentially composed channels in one cell. Co-transfecting a cell with wild-type and mutant subunits in certain ratios may lead to the formation of channels containing different distributions of subunits, but, especially in the context of overexpression, a reliable representation of physiological conditions is by no means guaranteed. In particular, dosage effects caused by haploinsufficiency of a gene encoding a subunit may be missed due to the artificially high protein abundance in overexpression systems.

Finally, a major limitation in variant interpretation remains the absence of experimental validation for most variants. Many variant effects are inferred from the type of variant on the DNA level, the phenotype, the location within the gene, the previously conducted research in similar variants, and bioinformatic modeling. Regarding the latter, some useful tools have emerged to approximate variant effects in cases where experimental studies are lacking (see following section), and efforts have been made to standardize the evaluation of electrophysiological findings in variant testing in channelopathies.[108]

Useful Tools

For (neuro)pediatricians, neurologists, and other medical professionals interested in a deeper understanding of the molecular basis underlying monogenic epilepsies, some of the following tools may be useful. For an overview of benign and pathogenic variants, as well as a gene's sensitivity to missense variants and haploinsufficiency or its different transcripts, the gnomAD website can be a valuable starting point (https://gnomad.broadinstitute.org/).[109] Prediction scores assessing whether a protein is affected by a particular variant can be obtained from prediction tools integrated into the UCSC Genome Browser (https://genome.ucsc.edu/)[110] or, for example, Franklin (https://franklin.genoox.com).

To understand whether pathogenic variants in a particular gene are known to be associated with one or several diseases, a first resource is often the OMIM website (https://www.omim.org/), which links genes with diseases and provides information on related publications.[111] However, some gene-disease associations are poorly characterized or outdated and not always marked as such. The more closely curated National Institutes of Health-funded Clinical Genome Resource (ClinGen; https://www.clinicalgenome.org/) ranks gene-disease validity from limited to definitive and provides, for some genes, expert recommendations for variant classification.[112] As very newly discovered disease genes may be found in neither resource, a review of the primary literature may be indispensable in some cases. Variants previously classified by laboratories as disease-causing or benign are collected in ClinVar (https://www.ncbi.nlm.nih.gov/clinvar/)[113]—though it is important to be aware that these annotations can be conflicting or incorrect. Gene panels associated with epilepsy or other disorders can be found using, for example, PanelApp (https://panelapp-aus.org/ or https://panelapp.genomicsengland.co.uk).[114]

Specific tools have been developed to aid in the prediction of variant effects in some epilepsy-associated genes, for example, to assess whether a variant is more likely to show loss- or GoF effects or to better infer a prognosis for the patient from a genetic test result. For example, the SCN1A prediction model (https://scn1a-prediction-model.broadinstitute.org/) provides estimates of whether an SCN1A variant is more likely to cause GEFS+ or Dravet syndrome based on the affected amino acid and the age of seizure onset.[115] Another tool has been developed for the variant effect prediction of potassium channels (https://cbosselmann.shinyapps.io/prefeKt/ [116] [117]). For other genes, repositories with functionally tested variants are found, for example, at: https://lal-portals.shinyapps.io/Neurogenetics-Gene-Portals/.

Conclusion

Dissecting the exact nature of a patient's underlying condition is a prerequisite for a correct choice of treatment. With the advent of precision genetic therapies, such as gene-replacement therapy or antisense-oligonucleotides, as well as for guiding the selection of traditional anti-seizure medications, it becomes ever more important to know not only the altered gene, but the exact genetic change, since different variant types can have opposite effects, requiring variant-specific therapeutic strategies.

As the complexities of clinical neuropediatrics, neurology, and medical genetics are difficult to grasp from literature study alone, we hold regular exchanges, for example, via interdisciplinary boards, to be an essential way of sharing the knowledge necessary to gain the full picture of a patient and their condition, thereby ensuring the best possible care and advice for the patient and their family.

Conflict of Interest

The authors declare that they have no conflict of interest.

• References

• 1 Helbig I, Heinzen EL, Mefford HC. International League Against Epilepsy Genetics Commission. Genetic literacy series: primer part 2-paradigm shifts in epilepsy genetics. Epilepsia 2018; 59 (06) 1138-1147
• 2 Helbig I, Heinzen EL, Mefford HC. ILAE Genetics Commission. Primer part 1-the building blocks of epilepsy genetics. Epilepsia 2016; 57 (06) 861-868
• 3 Zimmern V, Minassian B, Korff C. A review of targeted therapies for monogenic epilepsy syndromes. Front Neurol 2022; 13: 829116
• 4 Guerrini R, Balestrini S, Wirrell EC, Walker MC. Monogenic epilepsies: disease mechanisms, clinical phenotypes, and targeted therapies. Neurology 2021; 97 (17) 817-831
• 5 Created in BioRender. Oberlack, A. (2025). Accessed October 29, 2025 at: https://BioRender.com/7zcw9ia
• 6 Perucca P, Bahlo M, Berkovic SF. The genetics of epilepsy. Annu Rev Genomics Hum Genet 2020; 21: 205-230
• 7 Guerrini R, Conti V, Mantegazza M, Balestrini S, Galanopoulou AS, Benfenati F. Developmental and epileptic encephalopathies: from genetic heterogeneity to phenotypic continuum. Physiol Rev 2023; 103 (01) 433-513
• 8 Ng ACH, Chahine M, Scantlebury MH, Appendino JP. Channelopathies in epilepsy: an overview of clinical presentations, pathogenic mechanisms, and therapeutic insights. J Neurol 2024; 271 (06) 3063-3094
• 9 den Dunnen JT, Dalgleish R, Maglott DR. et al. HGVS recommendations for the description of sequence variants: 2016 update. Hum Mutat 2016; 37 (06) 564-569
• 10 Morales J, Pujar S, Loveland JE. et al. A joint NCBI and EMBL-EBI transcript set for clinical genomics and research. Nature 2022; 604 (7905) 310-315
• 11 Tavtigian SV, Greenblatt MS, Lesueur F, Byrnes GB. IARC Unclassified Genetic Variants Working Group. In silico analysis of missense substitutions using sequence-alignment based methods. Hum Mutat 2008; 29 (11) 1327-1336
• 12 Yampolsky LY, Stoltzfus A. The exchangeability of amino acids in proteins. Genetics 2005; 170 (04) 1459-1472
• 13 Pérez-Palma E, May P, Iqbal S. et al. Identification of pathogenic variant enriched regions across genes and gene families. Genome Res 2020; 30 (01) 62-71
• 14 Subramanian S, Kumar S. Evolutionary anatomies of positions and types of disease-associated and neutral amino acid mutations in the human genome. BMC Genomics 2006; 7 (01) 306
• 15 Malhis N, Jones SJM, Gsponer J. Improved measures for evolutionary conservation that exploit taxonomy distances. Nat Commun 2019; 10 (01) 1556
• 16 Capra JA, Singh M. Predicting functionally important residues from sequence conservation. Bioinformatics 2007; 23 (15) 1875-1882
• 17 Kircher M, Witten DM, Jain P, O'Roak BJ, Cooper GM, Shendure J. A general framework for estimating the relative pathogenicity of human genetic variants. Nat Genet 2014; 46 (03) 310-315
• 18 Ioannidis NM, Rothstein JH, Pejaver V. et al. REVEL: an ensemble method for predicting the pathogenicity of rare missense variants. Am J Hum Genet 2016; 99 (04) 877-885
• 19 Cheng J, Novati G, Pan J. et al. Accurate proteome-wide missense variant effect prediction with AlphaMissense. Science 2023; 381 (6664) eadg7492
• 20 Montanucci L, Brünger T, Boßelmann CM. et al. Evaluating novel in silico tools for accurate pathogenicity classification in epilepsy-associated genetic missense variants. Epilepsia 2024; 65 (12) 3655-3663
• 21 Truty R, Ouyang K, Rojahn S. et al. Spectrum of splicing variants in disease genes and the ability of RNA analysis to reduce uncertainty in clinical interpretation. Am J Hum Genet 2021; 108 (04) 696-708
• 22 Wang R, Helbig I, Edmondson AC, Lin L, Xing Y. Splicing defects in rare diseases: transcriptomics and machine learning strategies towards genetic diagnosis. Brief Bioinform 2023; 24 (05) 1-15
• 23 Daguenet E, Dujardin G, Valcárcel J. The pathogenicity of splicing defects: mechanistic insights into pre-mRNA processing inform novel therapeutic approaches. EMBO Rep 2015; 16 (12) 1640-1655
• 24 Lindeboom RGH, Supek F, Lehner B. The rules and impact of nonsense-mediated mRNA decay in human cancers. Nat Genet 2016; 48 (10) 1112-1118
• 25 Popp MW, Maquat LE. Leveraging rules of nonsense-mediated mRNA decay for genome engineering and personalized medicine. Cell 2016; 165 (06) 1319-1322
• 26 Torene RI, Guillen Sacoto MJ, Millan F. et al. Systematic analysis of variants escaping nonsense-mediated decay uncovers candidate Mendelian diseases. Am J Hum Genet 2024; 111 (01) 70-81
• 27 Lindeboom RGH, Vermeulen M, Lehner B, Supek F. The impact of nonsense-mediated mRNA decay on genetic disease, gene editing and cancer immunotherapy. Nat Genet 2019; 51 (11) 1645-1651
• 28 Khajavi M, Inoue K, Lupski JR. Nonsense-mediated mRNA decay modulates clinical outcome of genetic disease. Eur J Hum Genet 2006; 14 (10) 1074-1081
• 29 Hug N, Longman D, Cáceres JF. Mechanism and regulation of the nonsense-mediated decay pathway. Nucleic Acids Res 2016; 44 (04) 1483-1495
• 30 Carvill GL, Engel KL, Ramamurthy A. et al; EuroEPINOMICS Rare Epilepsy Syndrome, Myoclonic-Astatic Epilepsy, and Dravet Working Group. Aberrant inclusion of a poison exon causes dravet syndrome and related SCN1A-associated genetic epilepsies. Am J Hum Genet 2018; 103 (06) 1022-1029
• 31 Felker SA, Lawlor JMJ, Hiatt SM. et al. Poison exon annotations improve the yield of clinically relevant variants in genomic diagnostic testing. Genet Med 2023; 25 (08) 100884
• 32 Lin M, Whitmire S, Chen J, Farrel A, Shi X, Guo JT. Effects of short indels on protein structure and function in human genomes. Sci Rep 2017; 7 (01) 9313
• 33 Hurles ME, Dermitzakis ET, Tyler-Smith C. The functional impact of structural variation in humans. Trends Genet 2008; 24 (05) 238-245
• 34 Demidov G, Laurie S, Torella A. et al; Solve-RD consortium. Structural variant calling and clinical interpretation in 6224 unsolved rare disease exomes. Eur J Hum Genet 2024; 32 (08) 998-1004
• 35 Sanchis-Juan A, Stephens J, French CE. et al. Complex structural variants in Mendelian disorders: identification and breakpoint resolution using short- and long-read genome sequencing. Genome Med 2018; 10 (01) 95
• 36 Blake LE, Roux J, Hernando-Herraez I. et al. A comparison of gene expression and DNA methylation patterns across tissues and species. Genome Res 2020; 30 (02) 250-262
• 37 Liu R, Zhao E, Yu H, Yuan C, Abbas MN, Cui H. Methylation across the central dogma in health and diseases: new therapeutic strategies. Signal Transduct Target Ther 2023; 8 (01) 310
• 38 Chen Z, Morris HR, Polke J. et al. Repeat expansion disorders. Pract Neurol 2025; 25 (03) 204-216
• 39 Chen Y, Dawes R, Kim HC. et al. De novo variants in the RNU4-2 snRNA cause a frequent neurodevelopmental syndrome. Nature 2024; 632 (8026) 832-840
• 40 Fuller ZL, Berg JJ, Mostafavi H, Sella G, Przeworski M. Measuring intolerance to mutation in human genetics. Nat Genet 2019; 51 (05) 772-776
• 41 Lek M, Karczewski KJ, Minikel EV. et al; Exome Aggregation Consortium. Analysis of protein-coding genetic variation in 60,706 humans. Nature 2016; 536 (7616) 285-291
• 42 Karczewski KJ, Francioli LC, Tiao G. et al; Genome Aggregation Database Consortium. The mutational constraint spectrum quantified from variation in 141,456 humans. Nature 2020; 581 (7809) 434-443
• 43 Leitão E, Schröder C, Parenti I. et al. Systematic analysis and prediction of genes associated with monogenic disorders on human chromosome X. Nat Commun 2022; 13 (01) 6570
• 44 Wong-Kisiel L. Pyridoxine-dependent epilepsy. In: Epilepsy Case Studies: Pearls for Patient Care. Vol 9783319013. University of Washington, Seattle; 2014: 7-11
• 45 Al-Shekaili H, Ciapaite J, van Karnebeek C, Pena I. PLPBP Deficiency. GeneReviews. Published online February 16, 2023. Accessed August 2, 2025 at: https://www.ncbi.nlm.nih.gov/books/NBK589231/
• 46 Plecko B, Mills P. PNPO Deficiency. University of Washington, Seattle; 1993. . Accessed August 2, 2025 at: https://www.ncbi.nlm.nih.gov/books/NBK581452/
• 47 Kim YG, Kang H, Lee B. et al. A spectrum of nonsense-mediated mRNA decay efficiency along the degree of mutational constraint. Commun Biol 2024; 7 (01) 1461
• 48 Sato H, Singer RH. Cellular variability of nonsense-mediated mRNA decay. Nat Commun 2021; 12 (01) 7203
• 49 Benkirane M, Marelli C, Guissart C. et al. High rate of hypomorphic variants as the cause of inherited ataxia and related diseases: study of a cohort of 366 families. Genet Med 2021; 23 (11) 2160-2170
• 50 Thuma TBT, Procopio RA, Jimenez HJ, Gunton KB, Pulido JS. Hypomorphic variants in inherited retinal and ocular diseases: a review of the literature with clinical cases. Surv Ophthalmol 2024; 69 (03) 337-348
• 51 Yang H, Wang D, Engelstad K. et al. Glut1 deficiency syndrome and erythrocyte glucose uptake assay. Ann Neurol 2011; 70 (06) 996-1005
• 52 Rotstein M, Engelstad K, Yang H. et al. Glut1 deficiency: inheritance pattern determined by haploinsufficiency. Ann Neurol 2010; 68 (06) 955-958
• 53 Gallagher D, Pérez-Palma E, Bruenger T. et al. Genotype-phenotype associations in 1018 individuals with SCN1A-related epilepsies. Epilepsia 2024; 65 (04) 1046-1059
• 54 Voskobiynyk Y, Battu G, Felker SA. et al. Aberrant regulation of a poison exon caused by a non-coding variant in a mouse model of Scn1a-associated epileptic encephalopathy. PLoS Genet 2021; 17 (01) e1009195
• 55 Carvill GL, Matheny T, Hesselberth J, Demarest S. Haploinsufficiency, dominant negative, and gain-of-function mechanisms in epilepsy: matching therapeutic approach to the pathophysiology. Neurotherapeutics 2021; 18 (03) 1500-1514
• 56 Happ HC, Schneider PN, Hong JH. et al. Long-read sequencing and profiling of RNA-binding proteins reveals the pathogenic mechanism of aberrant splicing of an SCN1A poison exon in epilepsy.
• 57 Monk D, Mackay DJG, Eggermann T, Maher ER, Riccio A. Genomic imprinting disorders: lessons on how genome, epigenome and environment interact. Nat Rev Genet 2019; 20 (04) 235-248
• 58 Eggermann T, Monk D, de Nanclares GP. et al. Imprinting disorders. Nat Rev Dis Primers 2023; 9 (01) 33
• 59 Judson MCC, Wallace MLL, Sidorov MSS. et al. GABAergic neuron-specific loss of Ube3a causes Angelman syndrome-like EEG abnormalities and enhances seizure susceptibility. Neuron 2016; 90 (01) 56-69
• 60 Roden WH, Peugh LD, Jansen LA. Altered GABA(A) receptor subunit expression and pharmacology in human Angelman syndrome cortex. Neurosci Lett 2010; 483 (03) 167-172
• 61 Joensuu T, Lehesjoki AE, Kopra O. Molecular background of EPM1-Unverricht-Lundborg disease. Epilepsia 2008; 49 (04) 557-563
• 62 Singh S, Hämäläinen RH. The roles of cystatin B in the brain and pathophysiological mechanisms of progressive myoclonic epilepsy type 1. Cells 2024; 13 (02) 170
• 63 Zschocke J, Byers PH, Wilkie AOM. Mendelian inheritance revisited: dominance and recessiveness in medical genetics. Nat Rev Genet 2023; 24 (07) 442-463
• 64 Weckhuysen S, George AL. KCNQ2- and KCNQ3-Associated Epilepsy. Cambridge University Press; 2022.
• 65 Millevert C, Weckhuysen S. ILAE Genetics Commission. ILAE Genetic Literacy Series: Self-limited familial epilepsy syndromes with onset in neonatal age and infancy. Epileptic Disord 2023; 25 (04) 445-453
• 66 Adam MP, Feldman J, Mirzaa GM. KCNQ3-Related Disorders. Published online 2014:1993–2024
• 67 Miceli F, Soldovieri MV, Joshi N, Weckhuysen S, Cooper E, Taglialatela M. KCNQ2-Related Disorders. University of Washington, Seattle; 1993. . Accessed October 31, 2024 at: https://www.ncbi.nlm.nih.gov/books/NBK32534/
• 68 Miceli F, Striano P, Soldovieri MV. et al. A novel KCNQ3 mutation in familial epilepsy with focal seizures and intellectual disability. Epilepsia 2015; 56 (02) e15-e20
• 69 Sands TT, Balestri M, Bellini G. et al. Rapid and safe response to low-dose carbamazepine in neonatal epilepsy. Epilepsia 2016; 57 (12) 2019-2030
• 70 Soldovieri MV, Boutry-Kryza N, Milh M. et al. Novel KCNQ2 and KCNQ3 mutations in a large cohort of families with benign neonatal epilepsy: first evidence for an altered channel regulation by syntaxin-1A. Hum Mutat 2014; 35 (03) 356-367
• 71 Orhan G, Bock M, Schepers D. et al. Dominant-negative effects of KCNQ2 mutations are associated with epileptic encephalopathy. Ann Neurol 2014; 75 (03) 382-394
• 72 Weckhuysen S, Mandelstam S, Suls A. et al. KCNQ2 encephalopathy: emerging phenotype of a neonatal epileptic encephalopathy. Ann Neurol 2012; 71 (01) 15-25
• 73 Miceli F, Soldovieri MV, Ambrosino P. et al. Genotype-phenotype correlations in neonatal epilepsies caused by mutations in the voltage sensor of K(v)7.2 potassium channel subunits. Proc Natl Acad Sci U S A 2013; 110 (11) 4386-4391
• 74 Rusina E, Simonti M, Duprat F, Cestèle S, Mantegazza M. Voltage-gated sodium channels in genetic epilepsy: up and down of excitability. J Neurochem 2024; 168 (12) 3872-3890
• 75 Mantegazza M, Broccoli V. SCN1A/Na_V 1.1 channelopathies: mechanisms in expression systems, animal models, and human iPSC models. Epilepsia 2019; 60 (Suppl. 03) S25-S38
• 76 Scheffer IE, Zuberi S, Mefford HC, Guerrini R, McTague A. Developmental and epileptic encephalopathies. Nat Rev Dis Primers 2024; 10 (01) 1-19
• 77 Hedrich UBS, Lauxmann S, Lerche H. SCN2A channelopathies: mechanisms and models. Epilepsia 2019; 60 (Suppl. 03) S68-S76
• 78 Bryson A, Reid C, Petrou S. Fundamental neurochemistry review: GABA_A receptor neurotransmission and epilepsy: Principles, disease mechanisms and pharmacotherapy. J Neurochem 2023; 165 (01) 6-28
• 79 Bonardi CM, Heyne HO, Fiannacca M. et al. KCNT1-related epilepsies and epileptic encephalopathies: phenotypic and mutational spectrum. Brain 2021; 144 (12) 3635-3650
• 80 Hinckley CA, Zhu Z, Chu JH. et al. Functional evaluation of epilepsy-associated KCNT1 variants in multiple cellular systems reveals a predominant gain of function impact on channel properties. Epilepsia 2023; 64 (08) 2126-2136
• 81 Barcia G, Fleming MR, Deligniere A. et al. De novo gain-of-function KCNT1 channel mutations cause malignant migrating partial seizures of infancy. Nat Genet 2012; 44 (11) 1255-1259
• 82 McTague A, Nair U, Malhotra S. et al. Clinical and molecular characterization of KCNT1-related severe early-onset epilepsy. Neurology 2018; 90 (01) e55-e66
• 83 Tang QY, Zhang FF, Xu J. et al. Epilepsy-related slack channel mutants lead to channel over-activity by two different mechanisms. Cell Rep 2016; 14 (01) 129-139
• 84 Kim GE, Kronengold J, Barcia G. et al. Human slack potassium channel mutations increase positive cooperativity between individual channels. Cell Rep 2014; 9 (05) 1661-1672
• 85 Absalom NL, Liao VWY, Johannesen KMH. et al. Gain-of-function and loss-of-function GABRB3 variants lead to distinct clinical phenotypes in patients with developmental and epileptic encephalopathies. Nat Commun 2022; 13 (01) 1822
• 86 Mohammadi NA, Ahring PK, Yu Liao VW. et al. Distinct neurodevelopmental and epileptic phenotypes associated with gain- and loss-of-function GABRB2 variants. EBioMedicine 2024; 106: 105236
• 87 Lin SXN, Ahring PK, Keramidas A. et al. Correlations of receptor desensitization of gain-of-function GABRB3 variants with clinical severity. Brain 2024; 147 (01) 224-239
• 88 Thompson CH, Potet F, Abramova TV. et al. Epilepsy-associated SCN2A (NaV1.2) variants exhibit diverse and complex functional properties. J Gen Physiol 2023; 155 (10) e202313375
• 89 Kobayashi Takahashi Y, Tabata K, Baba S. et al. SCN1A gain of function effects in Dravet syndrome: Insights into clinical phenotypes and therapeutic implications. Epilepsia Open 2025; 10 (04) 1236-1243
• 90 Berg AT, Thompson CH, Myers LS. et al. Expanded clinical phenotype spectrum correlates with variant function in SCN2A-related disorders. Brain 2024; 147 (08) 2761-2774
• 91 Kan ASH, Kusay AS, Mohammadi NA. et al. Understanding paralogous epilepsy-associated GABA_A receptor variants: Clinical implications, mechanisms, and potential pitfalls. Proc Natl Acad Sci U S A 2024; 121 (50) e2413011121
• 92 Koch NA, Sonnenberg L, Hedrich UBS, Lauxmann S, Benda J. Loss or gain of function? Effects of ion channel mutations on neuronal firing depend on the neuron type. Front Neurol 2023; 14: 1194811
• 93 Meisler MH, Hill SF, Yu W. Sodium channelopathies in neurodevelopmental disorders. Nat Rev Neurosci 2021; 22 (03) 152-166
• 94 Bryson A, Petrou S. SCN1A channelopathies: Navigating from genotype to neural circuit dysfunction. Front Neurol 2023; 14: 1173460
• 95 Ogiwara I, Miyamoto H, Morita N. et al. Nav1.1 localizes to axons of parvalbumin-positive inhibitory interneurons: a circuit basis for epileptic seizures in mice carrying an Scn1a gene mutation. J Neurosci 2007; 27 (22) 5903-5914
• 96 Brunklaus A, Du J, Steckler F. et al. Biological concepts in human sodium channel epilepsies and their relevance in clinical practice. Epilepsia 2020; 61 (03) 387-399
• 97 Cheah CS, Westenbroek RE, Roden WH. et al. Correlations in timing of sodium channel expression, epilepsy, and sudden death in Dravet syndrome. Channels (Austin) 2013; 7 (06) 468-472
• 98 Berecki G, Howell KB, Heighway J. et al. Functional correlates of clinical phenotype and severity in recurrent SCN2A variants. Commun Biol 2022; 5 (01) 515
• 99 Liang L, Fazel Darbandi S, Pochareddy S. et al. Developmental dynamics of voltage-gated sodium channel isoform expression in the human and mouse brain. Genome Med 2021; 13 (01) 135
• 100 Gazina EV, Leaw BTW, Richards KL. et al. 'Neonatal' Nav1.2 reduces neuronal excitability and affects seizure susceptibility and behaviour. Hum Mol Genet 2015; 24 (05) 1457-1468
• 101 Thompson CH, Ben-Shalom R, Bender KJ, George AL. Alternative splicing potentiates dysfunction of early-onset epileptic encephalopathy SCN2A variants. J Gen Physiol 2020; 152 (03) e201912442
• 102 Liao Y, Deprez L, Maljevic S. et al. Molecular correlates of age-dependent seizures in an inherited neonatal-infantile epilepsy. Brain 2010; 133 (Pt 5): 1403-1414
• 103 Brunklaus A, Feng T, Brünger T. et al. Gene variant effects across sodium channelopathies predict function and guide precision therapy. Brain 2022; 145 (12) 4275-4286
• 104 Berecki G, Bryson A, Polster T, Petrou S. Biophysical characterization and modelling of SCN1A gain-of-function predicts interneuron hyperexcitability and a predisposition to network instability through homeostatic plasticity. Neurobiol Dis 2023; 179: 106059
• 105 Johannesen KM, Liu Y, Koko M. et al. Genotype-phenotype correlations in SCN8A-related disorders reveal prognostic and therapeutic implications. Brain 2022; 145 (09) 2991-3009
• 106 Brunklaus A, Brünger T, Feng T. et al. The gain of function SCN1A disorder spectrum: novel epilepsy phenotypes and therapeutic implications. Brain 2022; 145 (11) 3816-3831
• 107 Volkers L, Kahlig KM, Das JHG. et al. Febrile temperatures unmask biophysical defects in Nav1.1 epilepsy mutations supportive of seizure initiation. J Gen Physiol 2013; 142 (06) 641-653
• 108 Parthasarathy S, Cohen SR, Fitch E. et al. Optimizing clinical interpretability of functional evidence in epilepsy-related ion channel variants.
• 109 Chen S, Francioli LC, Goodrich JK. et al; Genome Aggregation Database Consortium. A genomic mutational constraint map using variation in 76,156 human genomes. Nature 2024; 625 (7993) 92-100
• 110 Perez G, Barber GP, Benet-Pages A. et al. The UCSC Genome Browser database: 2025 update. Nucleic Acids Res 2025; 53 (D1): D1243-D1249
• 111 Hamosh A, Scott AF, Amberger JS, Bocchini CA, McKusick VA. Online Mendelian Inheritance in Man (OMIM), a knowledgebase of human genes and genetic disorders. Nucleic Acids Res 2005; 33 (Database issue): D514-D517
• 112 Rehm HL, Berg JS, Brooks LD. et al; ClinGen. ClinGen–the clinical genome resource. N Engl J Med 2015; 372 (23) 2235-2242
• 113 Landrum MJ, Lee JM, Benson M. et al. ClinVar: public archive of interpretations of clinically relevant variants. Nucleic Acids Res 2016; 44 (D1): D862-D868
• 114 Martin AR, Williams E, Foulger RE. et al. PanelApp crowdsources expert knowledge to establish consensus diagnostic gene panels. Nat Genet 2019; 51 (11) 1560-1565
• 115 Brunklaus A, Pérez-Palma E, Ghanty I. et al. Development and validation of a prediction model for early diagnosis of SCN1A-related epilepsies. Neurology 2022; 98 (11) e1163-e1174
• 116 Boßelmann CM, Hedrich UBS, Lerche H, Pfeifer N. Predicting functional effects of ion channel variants using new phenotypic machine learning methods. PLOS Comput Biol 2023; 19 (03) e1010959
• 117 Boßelmann CM, Hedrich UBS, Müller P. et al. Predicting the functional effects of voltage-gated potassium channel missense variants with multi-task learning. EBioMedicine 2022; 81: 104115

Publication History

Received: 07 August 2025

Accepted: 17 October 2025

Accepted Manuscript online:
27 October 2025

Article published online:
06 November 2025