A novel SCN8A variant of unknown significance in pediatric epilepsy: a case report

Abstract

Variants in SCN8A are associated with several diseases, including developmental and epileptic encephalopathy, intermediate epilepsy or mild-to-moderate developmental and epileptic encephalopathy, self-limited familial infantile epilepsy, neurodevelopmental delays with generalized epilepsy, neurodevelopmental disorder without epilepsy, hypotonia, and movement disorders. Herein, we report an 8-year-old Moroccan boy with intermediate epilepsy of unknown origin, intellectual disability, autism spectrum disorder, and hyperactivity. The patient presented a normal 46, XY karyotype and a normal comparative genomic hybridization profile. Whole-exome sequencing was performed, and heterozygous variants were identified in KCNK4 and SCN8A. The SCN8A variant [c.4499C > T (p.Pro1500Leu)] was also detected in the healthy mother and was classified as a variant of uncertain clinical significance. This variant occurs in a highly conserved domain, which may affect the function of the encoded protein. More studies are needed to confirm the pathogenicity of this novel variant to establish the effective care, management, and genetic counselling of affected individuals.

Introduction

In electrically excitable cells, voltage-gated sodium channels (Nav) are key players in action potential initiation and spread. Of the nine Nav subtypes, Nav 1.6 is the most abundantly expressed and critical channel in the central nervous system. This protein is encoded by SCN8A, located in chromosome 12q13. ¹ Pathogenic SCN8A variants disrupt channel inactivation, alter neuronal excitability, and are reported to cause seizures and neuropsychological issues. They have also been associated with mobility problems, drug-resistant epilepsy, and even early death. ² Recently, SCN8A variants have been linked to a variety of epilepsy phenotypes, including self-limited familial infantile epilepsy (OMIM #617080), developmental and epileptic encephalopathy (OMIM #614558), intermediate epilepsy or mild-to-moderate developmental and epileptic encephalopathy (mild/modDEE), and neurodevelopmental delays with generalized epilepsy. ³ Furthermore, some cases of neurodevelopmental disorder without epilepsy, as well as of movement disorders such as ataxia and hypotonia, have also been reported.^3,4

Nine years ago, Veeramah et al. ⁵ reported the first epileptic patient with a pathogenic SCN8A variant. ⁵ With advances in genome sequencing technology, and particularly whole-exome sequencing, the number of cases of epilepsy linked with Nav1.6 variants has expanded considerably. More than 300 SCN8A variants have been identified, and 200 potential mutational hot spots have been found. De novo variants are the most prevalent; just 22 inherited variants have been reported. Currently, over 100 reports linking epilepsy with Nav1.6 variants have been published. ¹

Nav 1.6 is highly conserved and is one of the least tolerant proteins in humans. ⁶ SCN8A sequence alterations can cause a loss or gain of function, thus triggering a variety of clinical manifestations that depend on the nature and location of the variant. The majority of SCN8A variants are missense and cause a gain of function, inducing developmental and epileptic encephalopathy.^7–11 Other phenotypes, particularly those of mild/modDEE and self-limited familial infantile epilepsy, are relatively underrepresented. ¹² Loss-of-function variants have been less reported, and primarily cause neurodevelopmental delays with generalized epilepsy, neurodevelopmental disorder without epilepsy, intellectual disability, autism spectrum disorder, and mobility disorders. ³

It has been noted that people carrying the same variant may display considerably different clinical severities. This implies that secondary variants or other factors may have an impact on the effects of the main mutation. ¹³ Herein, we report the identification of novel SCN8A and KCNK4 variants in a patient with epilepsy and a strong family history.

Case report

We report the case of an 8-year-old Moroccan boy born to a non-consanguineous couple of Berber origin. After a normal pregnancy, he was delivered at term. He presented epilepsy, intellectual disability, autism spectrum disorder, hyperactivity, language impairments, hypertrichosis, and dental abnormalities without gingival overgrowth. The patient showed a mild psychomotor delay (walking at 18 months but speaking just a few words at 8 years). Myoclonic nocturnal seizures started at 2 years old, and his first tonic-myoclonic generalized epileptic seizure occurred at 3 years of age. Although the patient did not respond to valproic acid (Depakine, Microdepakine) or lamotrigine (Lamictal) treatments, symptom amelioration was noted after switching to ethosuximide (Zarontin) and levetiracetam (Keppra). A maternal family history of benign epilepsy (the patient’s brother, uncle, and two cousins) was reported (Figure 1). The patient’s brother had a seizure at the age of 6 months with no other symptoms. Epilepsy ceased after 2 years of treatment with valproic acid. The clinical presentations of the other affected family members were the same as that of the patient’s brother. The patient’s methylmalonic acid dosage, brain magnetic resonance imaging, constitutional karyotype, and comparative genome hybridization array results were normal.

Figure 1.

Family pedigree.

Informed consent was obtained from the legally authorized representative of the patient for his treatment and the publication of this case report. This manuscript adheres to the applicable CARE (for CAse REports) guidelines. ¹⁴

After obtaining the parents’ consent, the child and his parents underwent whole-exome sequencing. The DNA was sequenced using an Illumina platform with 20× coverage depth for >98% of the targeted bases, and was analyzed using alignment to the GRch37/hg 19 genome assembly. Segregation of the rest of the family, including the other affected members, was unable to be performed for various reasons. Following whole-exome sequencing, the following online tools were used to examine how the amino acid replacements affected protein structure and function: PolyPhen-2 (http://genetics.bwh.harvard.edu/pph2/), Align-GVGD (http://agvgd.hci.utah.edu/about.php), SIFT (Sorting Intolerant From Tolerant; https://sift.bii.a-star.edu.sg/), and MutationTaster (https://www.mutationtaster.org/). In addition, we used STRING (Search Tool for the Retrieval of Interacting Genes/Proteins) version 11.5 (https://string-db.org/) to examine the protein–protein interactions between SCN8A and KCNK4 gene products.

The whole-exome sequencing revealed two missense heterozygous variants: NM_001317090.1: c.512C > G (p.Ser171Trp) in KCNK4 and NM_001330260.1: c.4499C > T (p.Pro1500Leu) in SCN8A. A subsequent parental analysis identified the c.512C > G variant in the father. According to its American College of Medical Genetics classification, this variant is likely benign. In silico analysis revealed that the variant occurs in a highly conserved area; it is predicted to be probably damaging by PolyPhen-2 and disease causing by MutationTaster. However, it was predicted to be tolerated by SIFT. In addition, the patient’s SCN8A variant was detected in the mother and has an ACMG classification as a variant of uncertain significance. This variant that was inherited from the mother, with incomplete penetrance, was predicted to be probably damaging by PolyPhen-2 and disease causing by MutationTaster (Table 1). Furthermore, it occurs in a highly conserved domain, which might affect the function of the protein.

Table 1.

Evaluation of the possible impact of the c.512C > G (p. Ser171Trp) variant of KCNK4 and the c.4499C>T (p. Pro1500Leu) variant of SCN8A using different bioinformatic prediction tools.

Variant	PolyPhen-2	Align-GVGD	SIFT	MutationTaster
KCNK4p.(Ser171Trp)	Probably damaging	C15	Tolerated	Disease causing
SCN8Ap.(Pro1500Leu)	Probably damaging	C0	Deleterious	Disease causing

SIFT, Sorting Intolerant From Tolerant.

To the best of our knowledge, neither of these two variants have been reported in the literature; they were not included in the ExAC (Exome Aggregation Consortium) or gnomAD (Genome Aggregation Database) databases. Using STRING software, we generated a node network to display the indirect protein–protein interactions between SCN8A and KCNK4. The results indicate that the two variants may have a combined influence on the phenotype of our patient (Figure 2).

Figure 2.

Interaction network generated by STRING (Search Tool for the Retrieval of Interacting Genes/Proteins) version 11.5, showing the indirect protein–protein interactions between SCN8A and KCNK4 gene products.

Discussion

Epilepsy is characterized by clinical variability and molecular complexity, which makes diagnosis and treatment difficult. In the current report, we have described a patient with infantile intermediate epilepsy, intellectual disability, and other developmental and psychiatric disorders, who carries novel variants in SCN8A and KCNK4. The phenotype of our patient is particularly severe compared with his affected family members, who have a benign form of epilepsy.

Clinically, epilepsy rarely occurs alone; comorbidities are common. More than half of all individuals with epilepsy have one or more additional medical problems. The phenotype of our patient reflects the common clinical features reported in patients with intermediate epilepsy or mild/modDEE related to SCN8A variants: intellectual disability, developmental delay, neurodevelopmental regressions such as autism spectrum disorder and language impairment, a partial response to treatment, and a late age of onset. ⁴ Moreover, our proband presented some differences, such as a resistance to sodium blocker channel drugs, and additional clinical manifestations including hypertrichosis and dental abnormalities.

The identified missense substitution (p.Pro1500Leu) in the Nav 1.6 protein occurs within the highly conserved inactivation gate domain. This domain is part of a region in which 10% of pathogenic missense variants are located, along with the transmembrane segments and the proximal two-thirds of the C-terminal domain.^6,7 The inactivation gate, a highly functional protein component, is located between domains III and VI. It inactivates the channel by closing the pore with the hydrophobic IMF (isoleucine–phenylalanine–methionine) motif. This closure is stabilized by the glycine and proline flanking the IMF motif. ¹⁵ When a deleterious variant occurs within the inactivation gate, its function is impaired, resulting in hyperactivation of the sodium channel and persistent current in neurons. ¹⁵ Such gain-of-function variants are dominant because only a few hyperactive Nav 1.6 channels are required to produce a phenotype. ¹⁶ The number of hyperactive sodium channels that are expressed can also modulate the penetrance, leading to the concept of incomplete penetrance. Pavone et al. ¹⁷ have discussed the incomplete penetrance that typically occurs in dominantly inherited epilepsies, as well as the implications of genetic variants in the complexity of phenotypes. Furthermore, incomplete penetrance can be linked to the variant itself. Wagnon et al. ¹⁸ investigated the N1768D variant in heterozygous mice, who presented seizures, ataxia, and sudden unexpected death in epilepsy, with incomplete penetrance of just 50%. ¹⁸ Moreover, Bunton-Stasyshyn et al. ¹⁹ reported that heterozygous mice carrying the R1872W variant showed complete penetrance and had a more severe phenotype than those with the N1768D variant.

To assess the functional effect of the amino acid substitution in SCN8A (p.Pro1500Leu), we used the SCN (Voltage-Gated Sodium) viewer prediction tool that was developed by Brunklaus et al. ⁹ The Pro1500 position was aligned with other known variants of SCN1A and SCN5A. These variants have been reported to cause a loss of function and are associated with epilepsy and/or neurodevelopmental disorders and Brugada syndrome, respectively. However, using the funNCion (functional variant prediction in voltage-gated Na⁺/Ca²⁺ ion channels) in silico tool (https://funnc.shinyapps.io/shinyappweb/), which was developed to predict the pathogenicity and functional consequences of missense variants in both sodium and calcium ion channels, the p.Pro1500Leu variant was predicted to be pathogenic with a probability of 0.74, and to cause a gain of function with a probability of 0.54. This finding confirms the functional effects of variants within the inactivation gate.

A strong association between ‘”epilepsy, intellectual disability, poor speech, seizures, and generalized hypertrichosis”’ and KCNK4 variants was noted in three patients by Bauer et al. in 2018. The observed phenotype in this previous report was typically caused by de novo gain‐of‐function variants. Two patients carried the likely pathogenic mutation c.515C > A (p. Ala172Glu). However, the absence of gingival overgrowth in our patient raises the possible clinical heterogeneity of this syndrome. ²⁰

Notably, seizure control has been achieved in nearly half of all patients with SCN8A variants using sodium channel blockers. ⁶ This was not the case in our patient, who did not respond to valproic acid, and whose seizures were exacerbated by lamotrigine; however, levetiracetam resulted in good seizure control. Although levetiracetam usually aggravates seizures/developmental regression in SCN8A-related epilepsy, rare cases have been reported to respond favorably to levetiracetam.^10,11,21 Indeed, patients with SCN8A variants may present variable drug responses. ⁶

The co-existence of variants in both SCN8A and KCNK4 might explain the severe phenotype of our patient. Indeed, the products of SCN8A and KCNK4 interact indirectly via mutual genes such as KCNQ2 and KCNQ3, as well as via global interactions between Nav, voltage-gated potassium channels, voltage-insensitive potassium channel proteins, and other regulator proteins such as ankyrin-3 (Figure 2). The complexity of protein–protein interactions between the Nav and potassium channel families reveals the possible double effect of both variants on the phenotype of our patient. Kearney et al. ²² produced mice with heterozygous mutations in both SCN2A and KCNQ2 to assess the outcome of combining epilepsy-related variants at two different loci. Although single mutant SCN2A mice experienced brief partial seizures that started in adulthood and single mutant KCNQ2 mice did not experience spontaneous seizures, double heterozygotes developed a substantially more severe phenotype, with early-onset disease and high mortality. The authors concluded that the changes in protein interactions generated by a combination of variants in sodium and potassium channel genes can significantly impact the clinical presentation of epilepsy. ²² In accordance with this previous study, Ohba et al. ¹⁶ reported two patients with missense variants in the inactivation gate portion of SCN8A who had different ages of onset and severity, thus confirming the heterogeneous clinical presentation of SCN8A mutations. The authors then addressed the possibility that the co-existence of other genetic factors might change the phenotype, such as interactions between SCN8A and SCN1A or SCN1A and SCN9A. ¹⁶

To further investigate the effect of the present SCN8A variant of uncertain significance, we identified transcription factor binding sites in the targeted nucleotide sequence by conducting a predictive analysis using the PROMO online tool (http://alggen.lsi.upc.es/cgi-bin/promo_v3/promo/promoinit.cgi?dirDB=TF_8.3) with data from the TRANSFAC database version 8.3 (http://gene-regulation.com/pub/databases.html#transfac).^13,23 The results indicate that the genomic position of the present variant of uncertain significance is altered, and the binding site for enkephalin transcription factor 1 is removed. Additionally, the impact of non-coding regions of SCN8A, as well as epigenetic factors, might explain the observed clinical symptoms of our patient. Yu et al. ²⁴ demonstrated that adding a homozygous hypomorphic variant of GABRA2 to SCN8A mutant mice causes an aggravated prognosis and earlier age of seizure onset, whereas wild-type GABRA2 is a protective factor in SCN8A mutant mice. Similarly, Martin et al. demonstrated that SCN8A acts as a protective modifier in a mouse model of severe myoclonic epilepsy of infancy by restoring normal seizure thresholds and preventing lethal effects in SCN1A mutant mice. ²⁵

Conclusion

Pediatric epilepsy caused by SCN8A variants can have a heterogeneous clinical appearance, making molecular diagnosis difficult. This is especially true when the variant in question is classified as a variant of unknown significance. It is difficult to demonstrate the degree of functional impairment of an altered protein, as well as to dissect its interactions with other proteins. More studies, particularly functional studies, are needed to confirm the pathogenicity of the variants identified in the present report.

Data availability statement

The datasets used and/or analyzed during this case report are available from the corresponding author upon reasonable request.

Ethics statement

Written informed consent was obtained from the legally authorized representative of the patient for his anonymized information to be published in this article. This case report contains no identifiable information about the patient. Ethics committee approval is not applicable for this paper because case reports do not need to be approved by a review board.