SCN2A - Associated Episodic and Persistent Ataxia with Cerebellar Atrophy: A Case Report

Abstract

SCN2A, a gene that codes for a sodium channel highly expressed in the cerebellum, has been linked to a heterogeneous phenotype, including episodic ataxia (EA) and epilepsy, among other symptoms¹. Given the rarity of SCN2A-associated EA and its recent description, it is important the genotype-phenotype relationship of SCN2A-associated EA be better defined for prognosis and optimizing future management. Thus, we describe a 2-year-old boy with a SCN2A variant causing an initial prolonged episode of profound ataxia lasting 4 months, cerebellar atrophy, and persistent mild ataxia with episodic exacerbations. Due to the patient’s lack of early epilepsy, prolonged initial episode of ataxia, and cerebellar atrophy, this case broadens the scope of the SCN2A variant phenotype. SCN2A should be considered as a cause of early onset ataxia in children with targeted testing or as part of Whole Exome Sequencing (WES) in patients with new onset persistent or EA with or without seizures.

Introduction

Over the last three decades, many genetic causes for episodic ataxia (EA) have been discovered.² EAs are characterized by “attacks” of ataxia of variable duration and frequency, lasting from seconds to several weeks, followed by asymptomatic periods.³ These attacks typically manifest with ataxia, nystagmus, and dysarthria.¹ Dystonia is also a feature of these attacks at times. Differentiating between EA with an underlying baseline ataxia and chronic hereditary ataxia with exacerbations is a continuing issue in solidifying a phenotype for many EAs, especially when complicated by potential side effects of anti-epileptic drugs (AEDs).⁴ There are currently at least 9 known genetic causes for EA.⁵

SCN2A encodes the alpha subunit of sodium channels that are found highly concentrated in the cerebellum.^1,5-9 Specifically, SCN2A is crucial for the neuronal Na_V1.2 voltage gated sodium channel that functions in initiation and conduction of action potential in the brain.¹⁰ Variants in this gene have been associated with a broad range of neurological phenotypes from benign familial neonatal infantile seizures (BFNIS) to intractable epilepsy or from mild neuro-cognitive symptoms to autism and intellectual disability.⁶ SCN2A variants have more recently been associated with EA. The first case of SCN2A-associated EA was published in 2010.⁵

In the largest series to date, 21 patients with SCN2A variants and EA, the majority of patients had a co-occurrence of seizures and EA.⁵ Several categories of SCN2A- associated epilepsy have been described.⁹ Divergence from these categories is often found and many focus on the nature of the epilepsy rather than other accompanying symptoms.^7,9 Most similar to our case, a case report by Leach et al, described a patient with mild cerebellar atrophy and EA with episodes lasting 5-10 days and possible seizure secondary to a SCN2A variant.⁸ Some forms of EA have been responsive to acetazolamide, while SCN2A-associated EA has had mixed results when treated with this same medication.⁵

Case

A 2-year-old boy presented to the Emergency Department after abdominal pain, altered mental status, and abnormal body movements. During the episode, he would draw his limbs in and arch his back, “roll his eyes” into the back of his head, and then become somnolent. This was followed by 15 minutes of altered mental status characterized by severe irritability and inability to speak. He had several of these episodes. In the emergency department, his mental status did not return to baseline after the episodes and abdominal imaging did not suggest any acute cause for abdominal pain. He was admitted to the hospital to further assess the etiology of his altered mental status.

In the hospital, it was noted that he had episodes of nystagmus, stiffening of limbs and trunk, and altered mental status. These episodes were described as intermittent and lasting only a few minutes during which the patient was unresponsive to stimulation like snapping. He was typically fussy after an episode. He had markedly decreased strength in upper extremities, was unable to reach against gravity, and was unable to maintain head position when sitting upright. He was unable to maintain static or dynamic balance without assistance and had markedly decreased truncal and head control. Long term EEG monitoring identified no seizures, but identified generalized slowing with a disorganized background, suggesting encephalopathy. An initial MRI was completed at this time (Figure 1 and Figure 2). Extensive workup for a broad differential of possible causes of encephalopathy, including infectious, autoimmune, and metabolic, was completed but unremarkable. Of particular note, urine porphyrins were evaluated and unremarkable. The differential diagnosis for his initial spells at that time included seizure and dystonia. He was empirically treated with 3-day IV pulse steroids for possible autoimmune encephalitis. He showed mild but not marked improvement of his ataxia during his hospital course, with no clear response to his steroid infusion. As no definite seizures were identified, he was not discharged on any medication for seizure prophylaxis. He had persistent moderate ataxia at the time of discharge. No episodic worsening of his ataxia was noted during his 2-week hospitalization. He was discharged without a clear etiology for his presentation. The differential diagnosis for his persistent ataxia and overall course was still extremely broad at the time of discharge but focused on autoimmune and genetic causes of ataxia.

Figure 1.

Left: Patient’s T1 weighted sagittal MRI of the brain at 2 years and 9 months on initial presentation to the hospital showing cerebellar volume within the range of normal. Right: Patient’s T1 weighted sagittal MRI of the brain at 3 years and 9 months demonstrating midline cerebellar vermian atrophy.

Figure 2.

Left: Patient’s coronal T2 FLAIR MRI at 2 years and 9 months on initial presentation to the hospital showing normal cerebellar volume without FLAIR hyperintensity. Right: Patient’s coronal T2 FLAIR MRI of the brain at 3 years and 9 months demonstrating hyperintensity in the superior more than inferior bilateral portions of the cerebellar hemispheres.

The patient had persistent, but slowly improving, dysmetria, intention tremor, ataxia, and dysarthria for the following 4 months. At 4 month follow up, the parents first noted that his tremor and ataxia would worsen with illness or emotional stress. Whole exome sequencing (WES) demonstrated a likely pathogenic de novo SNC2A variant, c4651 G > C p.E1551Q. This was identified as a missense mutation that resulted in a change from glutamic acid to glutamine in exon 26 of the SCN2A gene. This was felt to be consistent with his presentation at that time and to explain his subsequent course. This exact mutation has also been reported in a SCN2A Variant database by University of California San Francisco, however, it is known that this patient has contributed to this database.¹¹

The patient was readmitted to the hospital approximately one year after the initial incident [3 years old] due to a first focal seizure followed by post-ictal paralysis characterized by transient contralateral hemiparesis and facial droop. He was subsequently placed on oxcarbazepine for seizure treatment. Other than this seizure, the patient had been stable clinically and on exam since his 4-month visit. However, repeat MRI at this time was concerning for cerebellar volume loss, especially in the vermis, with widening of the sulci and an increased size of the 4^th ventricle and the cisterna magna compared to his baseline scan from his initial presentation (Figure 1). Imaging also identified diffuse cerebellar cortical and deep white matter T2/FLAIR hyperintense signal in the cerebellum at that time (Figure 2). The pediatric neurology and neurosurgery teams agreed that there was no role for surgical intervention in this case as hydrocephalus was not a concern.

In subsequent follow up, the patient was placed on Acetazolamide 125 mg BID. The parents subjectively reported that his speech and possibly his EA with tremor were improved by the medication, however no change was seen in the patient’s persistent baseline ataxia. Repeat imaging at 6 months [4 years old] and 18 months [5 years old] after the identified cerebellar volume loss showed that the T2 hyperintensities and the volume loss in the cerebellum were stable. At the last follow up 3 years after initial presentation [5 years old], he had persistent mild ataxia without clear progression and rare EA but no further seizures.

Discussion/Conclusion

This case further broadens the phenotypic expression of SCN2A variants. As per the largest published series of SCN2A-associated EA, most patients had episodes of ataxia that lasted minutes to hours, with the longest episodes in that cohort lasting 3 weeks.⁵ Our patient demonstrated that episodes of ataxia associated with SCN2A may last weeks to months. Beyond that, it demonstrates that SCN2A variants can cause persistent ataxia after an initial attack as indicated by this patient’s persistent dysmetria and intention tremor. The series performed by Schwarz et al excluded patients that had persistent or baseline ataxia due to concern that it could be an adverse side effect of AEDs.⁵ Because this patient had dysmetria from presentation without prior exposure to an AED or prior seizures, this indicates that SCN2A can cause persistent ataxia from onset and should be considered in patients with such a presentation.

In addition to the unique persistent ataxia seen in this patient, the cerebellar atrophy is also a rarely reported complication of SCN2A variant. Although a prior case report demonstrated cerebellar atrophy, our patient had more prominent atrophy than in that case and additional persistent ataxia which has not previously been reported.⁸ Additionally, the patient in that case report had a second variant that also effected neuronal excitability and could have had an epistatic interaction with the SCN2A gene variant.⁸ The patient detailed in our case supports that SCN2A alone can cause cerebellar atrophy, persistent ataxia, and EA since WES in this patient did not demonstrate any other variants. Howell et al. reported one patient with a SCN2A variant and cerebellar atrophy; however, that patient presented with epilepsy alone and was not reported to have EA.⁷

Furthermore, our patient’s presentation further contributes to attempts to establish a relationship between mutation type, effect on the Na_V1.2 channel, and clinical phenotype. According to the WES report, our patient has a missense mutation and the in-silica analysis predicted that it is likely to have a deleterious effect on the protein. According to Wolff et all, many of the SCN2A mutations, whether they are gain- of- function or loss-of-function, are correlated with aspects of the expected clinical phenotype. Wolff described that loss-of-function mutations are typically associated with seizure disorders in which the onset is after 3 months of age. This includes disorders such as West Syndrome, Lennox Gastaut Syndrome, and focal epilepsies with features of electrical status epilepticus during slow sleep. This is considered late onset compared to the gain-of-function mutations, which were correlated with seizure onset before 3 months of age. This study found that there was a correlation between loss or gain of function mutation, response to AEDs, and certain demographic features, such as age of onset of seizures.¹⁰ No identified studies have addressed the genotype-phenotype correlation between SCN2A mutations and the development of episodic ataxia.

In contrast to the proposed correlation of Wolff et al, a study done by Shi et al, concluded that no consistent relationship could be established between genotype and phenotype in SCN2A- associated epilepsy. They found that several different studies determined that loss-of-function mutations and gain-of-function mutations both played a role in the development of BFNIS, which has been linked to pathogenic SCN2A mutations. Furthermore, they explain that some mutations in the gene not only alter the channel function, they also change the channel surface expression, which complicated determining the net effect of the mutation on the channel function.¹²

Although our patient is concerning for a loss of function mutation of SCN2A, his presentation mainly with ataxia and only mild, easily controlled epilepsy broadens the phenotype that may be expected with a missense mutation that has a deleterious effect on this protein product.

Finally, from a diagnostic perspective, patients with EA, including those that have baseline ataxia after an episode, should be evaluated for a potential SCN2A variant. With this broader phenotype in mind, clinicians should consider ordering a genetic panel that includes SCN2A in patients with episodes of ataxia as well as those with new onset persistent ataxia, with or without seizures. Patients with SCN2A- associated EA can be seizure-free or have very mild epilepsy as our patient demonstrates. For a more in-depth discussion of SCN2A associated epilepsy please see reference 9. Also, our patient demonstrates that no family history is necessary, as these can be de novo variants. Furthermore, the patient’s case above suggests that patients with extended episodes of ataxia or persistent ataxia, in the context of an SCN2A variant, should have repeated imaging to monitor the progression of potential cerebellar atrophy. Ensuring that the patient’s atrophy has stabilized could be an important prognostic factor for baseline ataxia progression when working with these patients.