Familial Hemiplegic Migraine and Alternating Hemiplegia of Childhood Panel

Familial hemiplegic migraine (FHM) is a rare, episodic neurological disorder characterized by migraine with aura in conjunction with hemiparesis (weakness of half the body) (Gasparini et al. 2013. PubMed ID: 24403849; Jen. 2015. PubMed ID: 20301562). The neurological symptoms of FHM aura commonly include visual disturbance, sensory loss, and/or dysphasia that may last hours to days following the initial migraine. Other common symptoms include nausea, vomiting, and/or extreme sensitivity to light or sound. Age of onset is variable, often occurring within the first or second decade of life, with attack frequency decreasing with age. There are three established types of familial hemiplegic migraine: FHM1, caused by defects in CACN1A1; FHM2, caused by defects in ATP1A2; and FHM3, caused by defects in SCN1A (Jen. 2015. PubMed ID: 20301562). Rarely, pathogenic variants in PRRT2 have also been shown to cause hemiplegic migraine with or without other PRRT2-related conditions (Riant et al. 2012. PubMed ID: 23077016).

Although isolated migraine is common in the population (with an estimated cumulative lifetime incidence of migraine in 43% of women and 18% of men), familial hemiplegic migraine is considerably rarer (Stewart et al. 2008. PubMed ID: 18644028). One study in Denmark estimated the prevalence of clinically-diagnosed FHM at 0.01%; the actual incidence is unknown at this time (Thomsen et al. 2002. PubMed ID: 12023326). Additionally, FHM commonly overlaps with other conditions, depending on the gene involved. Individuals with pathogenic variants in CACNA1A typically present with cerebellar signs such as nystagmus, while patients with defects in ATP1A2 may present with epilepsy (Yabe et al. 2008. PubMed ID: 18670797; Russell and Ducros. 2011. PubMed ID: 21458376; Deprez et al. 2008. PubMed ID: 18028407). Patients harboring PRRT2 defects often display symptoms of other PRRT2-related disorders, such as paroxysmal kinesigenic dyskinesia (PKD) and epilepsy (Riant et al. 2012. PubMed ID: 23077016).

Several other genes (ATP1A3, NOTCH3, COL4A1, and SLC2A1) have been included in this panel; these genes are all associated with disorders that have significant diagnostic overlap with FHM. One common diagnostic differential for familial hemiplegic migraine is alternating hemiplegia of childhood (AHC), a disease caused by pathogenic variants in ATP1A3 or ATP1A2. Alternating hemiplegia of childhood is initially characterized by alternating hemiparesis or dystonia, quadriparesis, seizure-like episodes, oculomotor abnormalities, and occasionally migraines (Brashear et al. 2018. PubMed ID: 20301294; Carecchio et al. 2018. PubMed ID: 29291920). The clinical presentation for cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL), caused by pathogenic variants in NOTCH3, may include hemiplegic migraine, mood disturbances, and cognitive dysfunction; one study estimated that ~38% of patients present with migraine, and it is often one of the earliest symptoms of disease (Rossi and Shambhu et al. 2018. PubMed ID: 30027023; Liem et al. 2010. PubMed ID: 21038489). Patients with pathogenic variants in COL4A1 may initially present with migraine with aura and hemiparesis; however, neuroimaging often reveals that these individuals have features of brain small-vessel disease, and they may also eventually exhibit seizures, dementia, intracerebral hemorrhage, and ischemic stroke (Plaisier and Ronco. 2016. PubMed ID: 20301768). Lastly, pathogenic variants in SLC2A1 have been rarely implicated in hemiplegic migraine with alternating hemiplegia of childhood; pathogenic variants in this gene are routinely associated with autosomal dominant or recessive GLUT1 deficiency syndrome, which is classically characterized by infantile seizures, dysarthria, acquired microcephaly, ataxia, dystonia, and chorea (Weller et al. 2015. PubMed ID: 24824604; Wang et al. 2018. PubMed ID: 20301603). However, non-classic symptoms of this disorder, which account for ~10% of affected individuals, may include alternating hemiplegia and migraine (Wang et al. 2018. PubMed ID: 20301603).

An accurate molecular diagnosis for individuals with features of FHM or AHC is often useful, as several conditions do have treatment options available. For SLC2A1 disorders, switching to a ketogenic diet may substantially alleviate the symptoms of disease (Wang et al. 2018. PubMed ID: 20301603). Antiepileptic medications may help control seizures in several of the disorders mentioned, and migraine prophylactic drugs (including tricyclic antidepressants, beta blockers, or calcium channel blockers) may be utilized for cases of FHM in individuals with frequent attacks (Jen. 2015. PubMed ID: 20301562). In addition, an accurate diagnosis may help guide appropriate screening and preparation for future complications, especially in the case of COL4A1 and NOTCH3-related disorders.

Support and information for individuals with CACNA1A-related disease and their families is available through the CACNA1A Foundation. This organization can be contacted through their website: http://www.cacna1a.org.

Familial hemiplegic migraine (FHM) is typically inherited as an autosomal dominant disease caused by defective regulation of ions across neuronal and glial cell membranes (Gasparini et al. 2013. PubMed ID: 24403849). Four genes have been associated to date with FHM: CACNA1A, ATP1A2, SCN1A, and PRRT2 (Gasparini et al. 2013. PubMed ID: 24403849; Riant et al. 2012. PubMed ID: 23077016). Very rarely, pathogenic variants in PRRT2 may be inherited in an autosomal recessive manner (Ebrahimi-Fakhari et al. 2018. PubMed ID: 29334453). In addition, de novo pathogenic variants in any of these genes may lead to a condition referred to as sporadic hemiplegic migraine (SHM) (Riant et al. 2010. PubMed ID: 20837964). Penetrance for certain pathogenic variants, including those of CACNA1A or ATP1A2, may be variable (between ~64%-90%) (Thomsen et al. 2007. PubMed ID: 17142831).

Of the known causes of FHM, defects in CACNA1A or ATP1A2 appear to be the most common, while SCN1A or PRRT2 defects are less frequently implicated in this disorder (Jen et al. 2011. PubMed ID: 20301562; Dichgans et al. 2005. PubMed ID: 16054936; Riant et al. 2012. PubMed ID: 23077016). Pathogenic variants in the other genes in this panel (ATP1A3, NOTCH3, COL4A1, and SLC2A1) are also associated with autosomal dominant disorders, although SLC2A1 is rarely linked to autosomal recessive disease (Wang et al. 2018. PubMed ID: 20301603).

The CACNA1A gene encodes the alpha-1A subunit of the voltage-dependent P/Q-type calcium channel (Ophoff et al. 1996. PubMed ID: 8898206). Over 200 causative variants have been reported in CACNA1A. Approximately 30 of these variants have been associated with FHM, while the remaining variants are causative for other CACNA1A-related diseases such as episodic ataxia type 2 and spinocerebellar ataxia type 6 (Stenson et al. 2014. PubMed ID: 24077912; Human Gene Mutation Database). While the majority of FHM-related CACNA1A defects are missense variants, splicing variants and small deletions have also been documented, as well as gross deletions (Ophoff et al. 1996. PubMed ID: 8898206; Stenson et al. 2014. PubMed ID: 24077912; Garcia Segarra et al. 2014. PubMed ID: 24836863; Grieco et al. 2018. PubMed ID: 30167989). Although de novo variants have been reported in this gene, the proportion of cases caused by de novo variants is currently unknown (Jen. 2015. PubMed ID: 20301562; Riant et al. 2010. PubMed ID: 20837964).

The ATP1A2 gene encodes the alpha-2 isoform of a sodium/potassium transporting ATPase responsible for maintaining sodium/potassium gradients across the plasma membrane (Shull et al. 1989. PubMed ID: 2477373). Over 80 pathogenic variants have been linked to ATP1A2-related FHM. While the majority of these variants are missense or small insertion/deletions, several nonsense variants, one splicing variant, and one single exon duplication have also been described (Stenson et al. 2014. PubMed ID: 24077912; Gagliardi et al. 2017. PubMed ID: 28593511; Human Gene Mutation Database). Although de novo variants have been reported in this gene, the proportion of cases caused by de novo variants is currently unknown (Jen. 2015. PubMed ID: 20301562; Vanmolkot et al. 2006. PubMed ID: 16538223).

The SCN1A gene encodes the alpha subunit of a neuronal type I sodium channel (Dichgans et al. 2005. PubMed ID: 16054936). SCN1A defects may cause a wide variety of conditions, including Dravet syndrome and generalized epilepsy (Akiyama et al. 2012. PubMed ID: 23093055). To date, over 1,500 pathogenic sequence variants have been reported in SCN1A (Stenson et al. 2014. PubMed ID: 24077912; Human Gene Mutation Database). Missense, nonsense, and splicing variants have been documented, in addition to several complex rearrangements and a number of small to large deletion/insertion events. A fraction of these defects (~10) have been linked to FHM (Dichgans et al. 2005. PubMed ID: 16054936; Cestele et al. 2013. PubMed ID: 23398611; Stenson et al. 2014. PubMed ID: 24077912). The proportion of cases caused by a pathogenic de novo variant in SCN1A appears to vary by phenotype, and may be more common in patients who present with features of Dravet syndrome (Miller and Menezes. 2019. PubMed ID: 20301494).

The PRRT2 gene encodes a proline-rich transmembrane protein of unknown function that is highly expressed in the cerebral cortex and basal ganglia (Heron et al. 2012. PubMed ID: 22243967). PRRT2 defects are typically associated with a wide variety of different childhood-onset episodic disorders, such as paroxysmal kinesigenic dyskinesia (PKD) or benign familial infantile epilepsy. Over 90 pathogenic variants have been reported for this gene to date (Human Gene Mutation Database). Approximately 10% of PRRT2 pathogenic variants are de novo (Ebrahimi-Fakhari et al. 2018. PubMed ID: 29334453). Several missense variants, at least one splicing variant, and several small deletions have been linked to PRRT2-related FHM (Ebrahimi-Fakhari et al. 2015. PubMed ID: 26598493; Stenson et al. 2014. PubMed ID: 24077912; Riant et al. 2010. PubMed ID: 20837964).

The ATP1A3 gene encodes the alpha₃-subunit of sodium-potassium ATPase (Rosewich et al. 2012. PubMed ID: 22850527). Over 115 pathogenic variants have been reported for this gene to date (Human Gene Mutation Database). The overwhelming majority of reported variants are missense, although one nonsense variant, several splicing variants, and several small insertions/deletion events have also been reported. The proportion of cases caused by de novo events varies depending on phenotype; in AHC, de novo variants are much more prevalent and likely account for the majority of cases (Brashear et al. 2018. PubMed ID: 20301294). Mosaicism has also been described in asymptomatic parents of proband cases, as well as in individuals with a milder phenotype (Yang et al. 2019. PubMed ID: 30891744).

The NOTCH3 gene encodes a single pass transmembrane protein that functions as part of a signaling pathway for transcription activation (Monet-Lepretre et al. 2009. PubMed ID: 19293235). Over 320 causative variants have been reported for this gene to date (Human Gene Mutation Database). Nearly 300 of these variants are missense; the remainders are nonsense variants, several splicing variants, and several small- to moderate-sized insertion/deletion events. De novo variants are a rare cause of NOTCH3-related disease, and have only been described in a few cases; the vast majority of pathogenic variants in this gene are inherited from an affected family member (Hack et al. 2019. PubMed ID: 20301673).

The COL4A1 gene encodes for the alpha-2 chain of type IV collagen (Yoneda et al. 2012. PubMed ID: 22209246). Over 125 causative variants have been reported for this gene to date (Human Gene Mutation Database). Nearly 100 of these variants are missense, although nonsense variants, several splicing variants, several regulatory variants, several small insertion/deletion events, and two large duplications have also been described. The proportion of cases that can be attributed to de novo events is estimated to be at least 27% (Plaisier and Ronco. 2016. PubMed ID: 20301768).

The SLC2A1 gene encodes for the primary glucose transporter in the brain and erythrocytes (Leen et al. 2010. PubMed ID: 20129935). Over 270 causative variants have been reported for this gene to date (Human Gene Mutation Database). Over 100 of these variants are missense, although at least 30 nonsense variants have been reported, in addition to numerous splicing variants, small- to moderate-sized insertion/deletion events, large multi-exonic or full gene deletions, and one regulatory variant. Approximately 90% of pathogenic variants in SLC2A1 arise de novo, while the remaining 10% are inherited from an affected parent (Wang et al. 2018. PubMed ID: 20301603).

See individual gene summaries for additional information about molecular biology of gene products and spectra of pathogenic variants.

Clinical sensitivity for familial hemiplegic migraine (FHM) appears to be highly variable depending on the specific population assessed and the diagnostic criteria of the study.

Previous studies have shown that 4% to 94% of FHM cases may be caused by defects in CACNA1A. For example, in a study conducted in the United States, 15 of 16 probands of families affected by hemiplegic migraine and cerebellar signs harbored likely causative sequence variants in the CACNA1A gene (94%; Ducros et al. 2001. PubMed ID: 11439943). On the other hand, an Australian-based study reported that seven out of 168 patients with FHM alone tested positive for pathogenic CACNA1A sequence variants (4%; Stuart et al. 2012. PubMed ID: 22784462). In a screen for pathogenic variants in CACNA1A, ATP1A2, or SCN1A in a Danish cohort of 147 Familial Hemiplegic Migraine (FHM) patients from 42 different families, affected individuals in six families harbored a pathogenic variant in CACNA1A or ATP1A2 (Thomsen et al. 2007. PubMed ID: 17142831).

Previous studies have shown that 11% to 56% of FHM cases may be caused by defects in ATP1A2. In a study involving 18 Spanish patients with hemiplegic migraine, two cases were determined to harbor sequence variants in the ATP1A2 gene (11%; Carreno et al. 2013. PubMed ID: 24498617). In another study involving 25 French patients with hemiplegic migraine, 14 harbored pathogenic sequence variants in the ATP1A2 gene (56%; Riant et al. 2010. PubMed ID: 20837964).

In an investigation involving 3 families (a total of 18 affected individuals) of European origin with FHM without causative sequence variants in the CACNA1A and ATP1A2 genes, all individuals harbored pathogenic variants in the SCN1A gene (100%; Dichgans et al. 2005. PubMed ID: 16054936). In a similar study of 10 Dutch families with FHM without pathogenic variants in the CACNA1A and ATP1A2 genes, one family was determined to carry causative sequence variants in the SCN1A gene (10%; Vanmolkot et al. 2007. PubMed ID: 17397047).

PRRT2 defects are a rare cause of FHM. In a study involving 128 patients with hemiplegic migraine, one patient harbored a pathogenic variant in the PRRT2 gene (Gardiner et al. 2012. PubMed ID: 23077024).

The vast majority of cases of alternating hemiplegic of childhood (AHC) are likely caused by pathogenic variants in ATP1A3. In one cohort of patients who presented with AHC, pathogenic variants in ATP1A3 were identified in 98 out of 105 patients; the majority of these variants (90/93) were confirmed as de novo (Yang et al. 2019. PubMed ID: 30891744). In the same study, four asymptomatic parents and one proband with a milder phenotype were identified as harboring mosaic pathogenic variants in multiple tissue types.

This test is performed using Next-Gen sequencing with additional Sanger sequencing as necessary.

This panel provides 100% coverage of all coding exons of the genes plus 10 bases of flanking noncoding DNA in all available transcripts along with other non-coding regions in which pathogenic variants have been identified at PreventionGenetics or reported elsewhere. We define coverage as ≥20X NGS reads or Sanger sequencing. PGnome panels typically provide slightly increased coverage over the PGxome equivalent. PGnome sequencing panels have the added benefit of additional analysis and reporting of deep intronic regions (where applicable).

Dependent on the sequencing backbone selected for this testing, discounted reflex testing to any other similar backbone-based test is available (i.e., PGxome panel to whole PGxome; PGnome panel to whole PGnome).

This test is currently not validated to identify repeat expansions in the CAG repeat region of CACNA1A; CAG repeat expansions in this region have been primarily associated with spinocerebellar ataxia 6 (Ishikawa et al. 1997. PubMed ID: 9311738).