Abstract
Background and Purpose
Mutations in KCNQ3 have classically been associated with benign familial neonatal and infantile seizures and more recently identified in patients with neurodevelopmental disorders and abnormal electroencephalogram (EEG) findings. We present 4 affected patients from a family with a pathogenic mutation in KCNQ3 with a unique constellation of clinical findings.
Methods
A family of 3 affected siblings and mother sharing a KCNQ3 pathogenic variant are described, including clinical history, genetic results, and EEG and magnetic resonance imaging (MRI) findings.
Results
This family shows a variety of clinical manifestations, including neonatal seizures, developmental delays, autism spectrum disorder, and anxiety. One child developed absence epilepsy, 2 children have infrequent convulsive seizures that have persisted into childhood, and their parent developed adult-onset epilepsy. An underlying c.1091G>A (R364H) variant in KCNQ3 was found in all affected individuals.
Conclusions
The phenotypic variability of KCNQ3 channelopathies continues to expand as more individuals and families are described, and the variant identified in this family adds to the understanding of the manifestations of KCNQ3-related disorders.
Introduction
Epilepsy, intellectual disability, and autism spectrum disorder belong to a complex group of conditions with both multifactorial and monogenic etiologies. Genetic workup is challenged by the clinical overlap between heterogeneous causes, pleiotropy, and inter- and intrafamilial phenotypic variability.
KCNQ2 and KCQN3 together form a heterotetramer as part of a voltage-gated potassium channel superfamily. 1 The phenotypic spectrum of KCNQ2-related disorders has been robustly described, including benign familial neonatal seizures and developmental and epileptic encephalopathy. Literature continues to emerge regarding the clinical manifestations of KCNQ3-related disorders. Classically, heterozygous pathogenic variants in KCNQ3 were associated with benign familial neonatal seizures in children with a typical neurodevelopmental trajectory. Recently, gain-of-function variants in KCNQ3 were found in patients with global developmental delay, autism spectrum disorder, and sleep-activated epileptiform discharges with limited seizure history. 2
Here, we present a family with developmental delays, autism spectrum disorder, neonatal seizures, and epilepsy of varied subtypes in which a heterozygous KCNQ3 variant (R364H) segregated with affected members. Furthermore, we evaluate the pathogenesis of the identified variant. The family was identified via the Central Ohio Registry for Autism (CORA)3,4 research project and through the clinical neurology and genetics services at Nationwide Children's Hospital.
Patient Descriptions
Proband (Patient 1; Pedigree Nomenclature II-2)
This 7-year-old girl was born at term after a pregnancy complicated by gestational hypertension, macrosomia, and polyhydramnios. Labor was induced at 39 weeks given suspected macrosomia, and delivery was complicated by brief shoulder dystocia.
Neonatal seizures started on day of life 3, which were multifocal and both electrographic and electroclinical. Brain magnetic resonance imaging (MRI) showed mild right hemispheric periventricular white matter injury with focal germinal matrix hemorrhage (Figure 1).
Patient 1 magnetic resonance imaging (MRI) scans from day of life 7 demonstrating mild right periventricular white matter injury on axial T1 sequence (left, indicated by red arrow) and hemosiderin deposition within the right lateral ventricle on axial SWI sequence (right, indicated by blue arrow) reflecting germinal matrix hemorrhage at this location.
Seizures responded to levetiracetam and phenobarbital, which were weaned following several normal EEGs in infancy. She developed prolonged convulsive seizures at 10 months of age and was treated with combinations of topiramate, oxcarbazepine, zonisamide, perampanel, and clobazam over several years, many of which were abandoned because of intolerable side effects. Other suspicious events of staring, blinking, or sudden falls were determined to be nonepileptic based on the absence of ictal EEG correlate with long-term video EEG monitoring. Multiple EEG recordings have shown a normal background without epileptiform discharges. She continues to have rare convulsive seizures.
The patient had early developmental delays. She has spastic diplegic cerebral palsy with decreased strength in bilateral lower extremities, and she uses a wheelchair for distance travel. She had no significant speech until age 2 years and was diagnosed with autism spectrum disorder. The Stanford-Binet Intelligence Scales–Fifth Edition indicated an unevenly developed cognitive profile with a full-scale Intelligence Quotient (IQ) of 42. Presumably, some degree of motor and cognitive delays may be attributable to early white matter injury, though seizure intractability prompted further evaluation with genetic testing.
Single-nucleotide polymorphism microarray was nondiagnostic (Tables 1 and 2). Copy number losses involving 14q32.1 and 2p16.3 include noncoding portions of the FOXN3 and NRXN1 genes, respectively.
Copy Number Variant Coordinates.
Copy Number Variants Detected by Microarray in Family.
Microarray not performed.
A next-generation sequencing–based comprehensive epilepsy panel revealed a missense variant in KCNQ3 (NM_004519.4:c.1091G>A, p.R364H) (ClinVar ID 934997). The panel also revealed 3 heterozygous variants in SLC46A1, ALG3, and ALG9 associated with autosomal recessive conditions unrelated to the phenotype. Research-based exome sequencing and pursuant clinical exome sequencing showed that the KCNQ3 variant was maternally inherited, and the 3 additional heterozygous variants were paternally inherited.
Mother (Patient 2; I-2)
Clinical and research exome sequencing determined that the mother shares the KCNQ3 variant with her daughters. This patient, who is of European and Cherokee descent, has a history of multiple spontaneous first trimester miscarriages and early speech delay in childhood. She has not undergone formal neuropsychological evaluation, though she does not appear to have a deficit in her intelligence. She developed convulsive seizures in adulthood and, for a period of time, was treated with antiseizure medications, which have since been discontinued without seizure recurrence. She has anxiety and headaches. She also had single-nucleotide polymorphism microarray (Tables 1 and 2).
Sister (Patient 3; II-3)
The 6-year-old sister was born by term vaginal delivery after a pregnancy complicated by gestational diabetes and hypertension. Labor was induced at 37 weeks’ gestation, and delivery was reportedly uncomplicated. She developed seizures on day of life 1, which were multifocal and both electrographic and electroclinical. Brain MRI showed hemorrhagic foci over the occipital poles secondary to birth trauma, which resolved on subsequent imaging (Figure 2).
Patient 3 magnetic resonance imaging (MRI) scans from day of life 7 (left, middle) demonstrating hemorrhagic foci over the occipital Poles and along posterior fossa suspected to be birth-related on axial T1 sequence (left, indicated by red arrow) and axial SWI sequence (middle, indicated by blue arrow); subsequent normal MRI at age 4 (right) with axial T2 sequence shown.
Neonatal seizures were managed with levetiracetam and phenobarbital. She developed bilateral tonic-clonic seizures in early infancy and continues to experience them several times per year. She was treated with various combinations of topiramate, oxcarbazepine, zonisamide, perampanel, and clobazam over several years, the majority of which were weaned because of intolerable side effects. Other events, including staring, blinking, and cyanotic breath-holding spells, were determined to be nonepileptic given the lack of ictal EEG correlate with long-term video EEG monitoring. Interictal EEG showed sleep-activated multifocal epileptiform discharges and diffuse background slowing.
She has mild hypotonia and first walked at 22 months of age—now ambulatory with frequent falls. She has a vocabulary of 10 words. A neuropsychological evaluation at age 2 was diagnostic of autism spectrum disorder and global developmental delay. She receives therapies and is in a modified academic curriculum.
Single-nucleotide polymorphism microarray was nondiagnostic (Tables 1 and 2). Research-based exome sequencing showed the familial KCNQ3 variant.
Sister (Patient 4; II-4)
The 13-year-old maternal half-sister was born at term and developed seizures on day of life 3. Brain MRI was normal. Seizures were managed with levetiracetam and phenobarbital, which were weaned by age 2 years following a normal EEG. At age 4 years, she had a single convulsive seizure in the setting of fever. At age 7 years, she developed absence seizures with EEG correlate of 3 Hz generalized spike and slow-wave discharges. She also had several unprovoked bilateral tonic-clonic seizures during early childhood. Several medications were tried, and the combination of ethosuximide and lamotrigine was effective and tolerable. Shortly following absence seizure onset, she began having more prolonged staring episodes lasting hours exacerbated by periods of stress. These events were determined to be nonepileptic given the lack of ictal EEG correlate with long-term video EEG monitoring and normal interictal EEG background. Multiple subsequent EEG recordings have been normal; no further seizures occurred since age 8 years.
She had an early speech delay, and neuropsychological evaluation was diagnostic of mild intellectual disability with a full-scale IQ of 65. She has anxiety, aggressive behaviors, attention deficit hyperactivity disorder, and headaches. Motor development was appropriate. Targeted testing confirmed the familial KCNQ3 variant. Clinical information for this patient's biological father is not available.
Brother (Patient 5; II-I)
An 11-year-old brother was diagnosed with autism spectrum disorder at age 3 years. He underwent evaluation because of speech delays and behavioral concerns, including tantrums and purposeful separation of himself from others during play. As part of his evaluation, single-nucleotide polymorphism microarray, and Fragile X testing were ordered, both of which were undiagnostic (Tables 1 and 2). He does not have seizures. Targeted testing of the familial KCNQ3 variant was negative.
Father (Patient 6; I-1)
The father of the proband is of Caucasian descent. He has a history of high cholesterol and high blood pressure, but otherwise is healthy. He does not have a history of seizures, autism, or developmental delay. He had exome sequencing testing as part of his daughter’s (patient 1’s) exome sequencing and was found to harbor the 3 heterozygous variants in SLC46A1, ALG3, and ALG9 along with his daughter. These variants are associated with autosomal recessive conditions unrelated to the phenotypes of the child and father. He also had single-nucleotide polymorphism microarray showing the gain in Xp22.13 (Tables 1 and 2).
Additional Family History
Mother's twin sister is healthy and unaffected. Maternal grandmother had seizures as a child and has mental health concerns. A maternal second cousin to the proband had a febrile seizure as a child. These individuals have not been tested for the KCNQ3 variant. See Figure 3 for pedigree.
Family pedigree demonstrating clinical diagnoses within the family. Shading indicates patients who share the KCNQ3 variant c.1091G>A (p.Arg364His).
Discussion
This family has a combination of clinical features not previously described in KCNQ3 channelopathies, including neonatal seizures, global developmental delay, autism spectrum disorder, and anxiety. One child developed childhood absence epilepsy, 2 children have ongoing infrequent convulsive seizures, and their mother has adult-onset epilepsy. The proband also has spastic diplegic cerebral palsy, presumably secondary to early white matter injury. No child included in the series had evidence of cortical injury on head imaging. Intrafamilial variability in other monogenic early-onset epileptic encephalopathies has been described in KCNQ2, SLC2A1, CLCN4, and SCN2A.5,6 The mechanism of the variability is speculated to be attributable to additional germline or somatic variant or environmental factors. Recent studies support the contributions of common variants to the predisposition to generalized epilepsies with different weights from different variants based on the epilepsy subtypes.7,8 In addition, there is some evidence that the same set of common variants can affect the clinical presentation of monogenic and non‐monogenic neurodevelopmental disorders. 9 Thus, we cannot exclude the possibility of the polygenic risk contribution to the intrafamilial variability. Furthermore, there may be other contributing factors underlying autism spectrum disorder in the family in light of the autism spectrum disorder phenotype in patient 5, who does not share the familial KCNQ3 variant.
Copy number losses involving 14q32.1 and 2p16.3 include noncoding portions of the FOXN3 and NRXN1 genes, respectively. Deletions of intronic regions of FOXN3 have not been reported in association with disease and are not likely contributory to the phenotype. Rare deletions of NRXN1 are reported in the context of neurodevelopmental or neuropsychiatric disorders and generally involve pathogenic exonic deletions. The role of intronic deletions is less well understood; most cluster within intron 5 (NM_001135659.1) and are almost always inherited, prevalent in controls, and therefore unlikely to be clinically significant.10,11 The NRXN1 deletion observed in our study includes a 68- to 80-kb region of intron 6 of the gene and is classified as a variant of uncertain significance.
With regard to the duplication in Xp22.13, this region encompasses the ARX gene, mutations in which have been well described in X-linked intellectual disability in males with intra- and interfamilial pleitropy.12,13 However, Xp22.13 duplication has also been seen in patients with normal intelligence. 14 Notably, in our cohort, the father carrying this duplication has normal intelligence, and one affected child (II-4) is not biologically related to this individual.
The KCNQ3 (c.1091G>A, p.R364H) variant identified in this family was classified as likely pathogenic based on ACMG/APM variant classification guidelines. 15 This variant affects a highly conserved amino acid residue, is very rare in large control population databases, including gnomAD with an allele count of 1, and is predicted to be deleterious by multiple computational algorithms including BayesDel, FATHMM, LRT, SIFT, and Mutation Taster (ACMG/AMP: PM2, PP3). Additionally, in this particular family, the variant segregates with multiple affected individuals (ACMG/AMP: PP1). The R364H variant has been reported as a de novo variant in a patient with benign infantile epilepsy with unilateral centrotemporal spikes on EEG (ACMG/AMP: PM6). 16 A different missense alteration at the same amino acid position (R364C) has been reported in individuals with neurodevelopmental disorders.17,18
The KCNQ3 channels have similar topology with other KCNQ channels, with each subunit of the tetramer consisting of 6 transmembrane domains (S1-S6) and amino and carboxy termini. Although several of the pathogenic KCNQ3 variants to date have been reported to occur in the transmembrane and pore domains of the channel, the R364H variant observed here is located in the intracellular, proximal portion of the C-terminal region, adjacent to the S6 transmembrane domain of the KCNQ3 protein. This region is reported to enable functional interactions between KCNQ2 and KCNQ3, channel assembly, and surface expression.1,19–22 Further characterization of the R364H variant to determine mechanism of action on channel pore activity and assembly remains to be performed. Other variants that have been reported in this region of the protein (A381V, N468S) are of uncertain significance and associated with reduced penetrance.23,24
Variants reported in typical familial forms of benign epilepsies are commonly missense alterations located within the pore region (S5-S6 intervening loop),25,26 resulting in loss of function (LOF) effect on channel activity.23,27–30 More recently, de novo missense alterations located outside of the pore region (R227G, R230C/S/H) have been reported in multiple individuals with global developmental delay, autism spectrum disorder, and sleep-activated epileptiform discharges on EEG. 2 Functional analysis using voltage-gated clamp recordings revealed that these variants stabilized the activated state of the channel and result in gain-of-function (GOF) effects on channel activity. In contrast to patients with LOF variants in KCNQ3, seizures were not a common finding in patients with GOF variants.
Biallelic alterations in this gene have been reported in at least 1 case of early-onset epileptic encephalopathy inherited as compound heterozygous missense alterations (V359L and D542N located in the C-terminal region) in KCNQ3 from asymptomatic parents. Functional analysis demonstrated that the channel current density was only reduced when both variants were coexpressed, 31 supporting an additive effect for each variant on channel disruption.
Clinical management of seizures in patients with KCNQ2- and KCNQ3-related epilepsies follows the standard therapeutic options adopted for non-monogenic seizures. There may be increased responsiveness to pharmacologic agents that act via sodium channel inhibition given co‐localization of the sodium and potassium channels at the cell membrane,32,33 though evidence of superior efficacy is lacking. Agents that act to increase the open state of the neuronal KCNQ channel are not currently commercially available.
Conclusions
Our series supports the previously documented pleiotropy of the KCNQ3 gene that includes intellectual disability and autism spectrum disorder with inter- and intrafamilial variability. This series expands the phenotypic spectrum associated with KCNQ3 channelopathies. Polygenic risk may also contribute to phenotypic variability among family members. Identification of a KCNQ3 disease-causing variant in a family will lead to crucial genetic counseling and management implications.
Footnotes
Acknowledgments
We are grateful to the families that participated in the CORA registry.
Author Contributions
KA conceptualized the study and methodology, curated data, drafted the original manuscript, and edited and reviewed the final manuscript. MM, MTM contributed to the methodology, curated data, drafted the original manuscript, and edited and reviewed the final manuscript. VJ drafted the original manuscript and edited and reviewed the final manuscript. GEH, EHK, and CM curated data and edited and reviewed the final manuscript. DB and AS curated data and reviewed the final manuscript.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Ethical Approval
This study was approved by the Institutional Review Board of Nationwide Children's Hospital (STUDY00001751), and the participants provided written authorization for inclusion in the series.
