Current Neurology and Neuroscience Reports

, 13:410

Genetics in Dystonia: An Update


  • Tania Fuchs
    • Department of Genetics and Genomic SciencesIchan School of Medicine at Mount Sinai
    • Department of Genetics and Genomic SciencesIchan School of Medicine at Mount Sinai
    • Department of NeurologyIchan School of Medicine at Mount Sinai
Movement Disorders (SA Factor, Section Editor)

DOI: 10.1007/s11910-013-0410-z

Cite this article as:
Fuchs, T. & Ozelius, L.J. Curr Neurol Neurosci Rep (2013) 13: 410. doi:10.1007/s11910-013-0410-z
Part of the following topical collections:
  1. Topical Collection on Movement Disorders


The past year has been extremely successful with regard to the genetics of dystonia, with the identification of four new dystonia genes (CIZ1, ANO3, GNAL, and TUBB4A). This progress was primarily achieved because of the application of a new technology, next-generation DNA sequencing, which allows rapid and comprehensive assessment of a patient’s genome. In addition, a combination of next-generation and traditional Sanger sequencing has expanded the phenotypic spectrum associated with some of the dystonia plus (ATP1A3) and paroxysmal (PRRT2) loci. This article reviews the newly identified genes and phenotypes and discusses the future applications of next-generation sequencing to dystonia research.


DystoniaGeneticsNext-generation sequencingExomeCIZ1ANO3GNALTUBB4APhenotypic spectrum


Dystonia is a movement disorder characterized by sustained muscle contractions, causing twisting, repetitive movements, and abnormal postures [1]. A revised definition has recently been put forth that expands on this definition and contains other aspects that are essential and unique elements of dystonia, including that “dystonic movements are typically patterned, twisting and may be tremulous and are often initiated or worsened by voluntary action and associated with overflow muscle activation” [2]. This article also proposes a new classification system based on the consideration of two axes: clinical characteristics (including age at onset, body distribution, temporal pattern, coexistence of another movement disorder and other neurological manifestations) and etiology (including nervous system disease and inheritance). In this new scheme, primary dystonia would be referred to as isolated dystonia and include other clinical descriptors such as early-onset generalized dystonia or segmental dystonia with onset in adulthood, whereas dystonia plus forms would be considered combined dystonia [2]. In this review we will use the current terminology but indicate the newer classification when possible.

Although most forms of monogenic dystonia are inherited as autosomal dominant traits with reduced penetrance, there are also autosomal recessive and X-linked forms. Among the dominant forms, both genetic and environmental factors are believed to contribute to the reduced penetrance [3]. At the end of 2011, 19 dystonia loci were distinguished genetically; however, the genes for only ten of these were known. In 2012, four genes for primary dystonia (termed “isolated dystonia” in the new classification system) were identified using next-generation sequencing and a fifth DYT locus has been called into question. In addition, a combination of next-generation and traditional Sanger sequencing has expanded the phenotypic spectrum associated with some of the dystonia plus and paroxysmal loci (now both fall under the term “combined dystonia”). As the main focus of this review is the impact of next-generation sequencing on dystonia genetics, it is necessary to provide a brief background on previous genetic techniques and an introduction to the technology of whole exome sequencing.

Most of the dystonia genes identified thus far have been found using a traditional gene discovery approach which included linkage analysis and positional mapping of the disease gene (reviewed in [4]), a process that required large pedigrees with multiple affected members. However, application of this method to dystonia gene discovery was limited by locus heterogeneity of the disease, late age of onset (in focal forms), and reduced penetrance, resulting in families with only a few affected members. The novel technology of next-generation DNA sequencing allows rapid, cost-effective sequencing of either the entire human genome or only its protein-coding part (exome), and has significantly accelerated gene discovery [5]. The human exome is the collection of all exons in the human genome, and consists of approximately 164,000 exons distributed across the genome, and comprises a total of 27.9 Mb of sequence. Whole exome sequencing detects all exon variants in an individual’s genome, in an unbiased manner, including potentially causative variants. As a result, unlike linkage analysis, whole exome sequencing has no prerequisite for the number of affected individuals in a family and could even be used in cases with sporadic disease [5]. Since most Mendelian disorders are caused by rare mutations in the coding region or at splice junctions, exome sequencing readily captures these types of mutations.

Primary Dystonia (Isolated Dystonia)

Primary torsion dystonia (PTD) is defined as a syndrome in which dystonia is the only clinical sign (except for tremor), and there is no evidence of neuronal degeneration or an acquired cause by history or routine laboratory assessment. The phenotypic spectrum associated with PTD is broad-ranging, from early-onset generalized dystonia to adult-onset focal dystonia. The prevalence of PTD has been estimated at between 330 per million [6] to 152 per million [7], with focal cases constituting the majority. However, these studies are likely to underestimate the true frequency of PTD, because a significant proportion of disease is not diagnosed [8, 9]. Only two PTD genes, TOR1A (DYT1) for early-onset generalized dystonia and THAP1 (DYT6) for adolescent-onset dystonia with prominent craniocervical and laryngeal involvement were known before last year. The introduction of next-generation sequencing together with exome capture techniques tremendously facilitated gene discovery in PTD. Four novel genes, CIZ1, ANO3, GNAL, and TUBB4A were identified in 2012 using this approach [10, 11••, 12••, 13••, 14••]. All of these genes, except TUBB4A, were found for previously unknown loci. In addition, a combination of genetic methods, including exome sequencing, has called the DYT7 locus into question [15].

CIZ1 (DYT23)

CIZ1 was detected as a causative dystonia gene in a multigenerational Caucasian pedigree from the USA with predominantly cervical dystonia (CD) that was previously described [16, 17]. The combination of linkage analysis and exome sequencing revealed a p.S264G mutation in CIZ1 that segregated in this family. Functional studies suggest that this variant alters the splicing of CIZ1 and subcellular localization of the protein. Screening of an additional 308 CD patients for mutations in CIZ1 identified two potentially pathogenic variants (p.P47S and p.R672M) [10], but there were no other family members to test segregation. As a large number of missense variants have been reported in CIZ1 in variant databases of supposedly healthy controls and because only a single family has been shown to segregate a mutation, it has been suggested that validation in a larger population of dystonia patients and functional tests are needed to prove pathogenicity of this locus [18].

CIZ1 encodes Cip1-interacting zinc finger protein 1, which is expressed in the brain [10]. The protein interacts with p21Cip1/Waf1 [19] and participates in DNA synthesis and cell-cycle control [20, 21].

ANO3 (DYT24)

Another dystonia gene, ANO3, was recently identified by a combination of linkage analysis and exome sequencing in a family from the UK with autosomal dominant craniocervical dystonia [11••]. Linkage analysis revealed five peaks with logarithm of odds (LOD) scores of 2.01 on chromosomes 4, 5, 6, 7, and 11. Exome sequencing of two affected individuals from the family revealed a heterozygous missense mutation—c.1480A > T (p.R494W)—in the ANO3 gene which resided in one of the linked regions and completely co-segregated with dystonia in this family. Screening of an additional 110 familial and 78 sporadic patients with CD with or without dystonic upper-limb tremor revealed five additional missense mutations in different exons of ANO3, including one in the 5′ untranslated region. Two of these variants, c.1470G > C (p.W490C) and c.2053A > G (p.S685G), showed co-segregation with dystonia when tested in additional family members [11••].

The dystonia phenotype associated with ANO3 mutations comprises mostly a focal or segmental distribution. All of the mutation carriers had dystonia involving the craniocervical region, but some also had laryngeal dystonia, blepharospasm, and/or limb tremor. The time of the disease onset ranged between early childhood and 40 years [11••]. It is interesting that a patient with an ANO3 mutation was previously diagnosed with essential tremor, raising the question of whether mutations in ANO3 are also involved in tremor associated with dystonia.

ANO3 codes for the anoctamin 3 protein, which belongs to a family of genes encoding Ca2+-activated Cl- ion channels [22]. ANO3 is increasingly expressed during brain development from the early fetal stage to adolescence in the striatum, where the highest level of expression was observed, but is also found in the neocortex, amygdala, and hippocampus [11••]. The precise function of anoctamin 3 is not yet known, but it is apparently targeted at the endoplasmic reticulum (ER) rather than at the cell surface like for other anoctamins [22]. Patient fibroblasts expressing the p.W490C mutation exhibited reduced Ca2+ signaling, most likely resulting from a smaller calcium pool within the endoplasmic reticulum [11••]. Therefore, it is possible that mutations in ANO3 lead to abnormal striatal neuron excitability, which clinically manifests itself as dystonic movements.


GNAL was detected as a dystonia causative gene in two moderately sized PTD families with mostly segmental dystonia with neck onset [12••]. The combination of linkage analysis and exome sequencing revealed a nonsense p.S293X mutation in GNAL resulting in a premature stop codon in one family and a missense p.V137M mutation in the other. Screening the coding region of GNAL in 39 additional multiplex PTD families revealed novel mutations in six additional families.

The phenotype in GNAL carriers is characterized by an average age of onset of 31.3 years, with a range from 7 to 54 years. Most carriers (82 %) had onset in the neck, and 93 % had neck involvement at the final examination; however, most progressed to have dystonia at other sites, and only 46 % had focal dystonia at the last examination. Further, cranial involvement was present in 57 % of carriers, and 44 % had speech involvement. Brachial onset was not observed, and eventual arm involvement was seen in only 32 %, distinguishing GNAL carriers from THAP1 carriers [12••].

GNAL encodes the stimulatory alpha subunit of heterotrimeric G protein, Gαolf, which was initially identified as a G-protein-mediating odorant signaling in the olfactory epithelium [23]. However, later it was recognized as the main G-protein alpha subunit in the striatum, where it couples dopamine type 1 receptors of the direct pathway and adenosine A2A receptors of the indirect pathway to the activation of adenylyl cyclase type 5 in the medium spiny neurons [2426]. Heterozygote GNAL mice demonstrate a muted response to short-term psychostimulant and caffeine exposure [27]. Homozygote null mice are anosmic, hyperactive at the baseline, and fail to respond to short-term psychostimulant exposure, but do not manifest an overt movement disorder [27, 28]. Gαolf has been physiologically linked to levodopa-induced dyskinesias [29, 30] and has been genetically linked to bipolar disorder, schizophrenia, and attention deficit–hyperactivity disorder [3133].

Most recently, using exome sequencing, Vemula et al. [34] identified a novel GNAL missense mutation in an African-American family, with most members affected with segmental dystonia. The follow-up screening of 760 patients with familial and sporadic forms of CD revealed three additional mutations in mainly focal CD cases. Most of the GNAL carriers had focal CD (seven of 11), and the average age of onset in GNAL carriers was later than previously reported (45 years), most probably owing to the high prevalence of focal dystonia in this cohort. Interestingly, related to the role of GNAL as a mediator of olfactory signaling, microsomia was very prominent in members of the African-American family affected with dystonia; however, it was absent or variable in the Caucasian GNAL carriers [34].

Identification of dystonia causative mutations in GNAL points to primary abnormalities of dopamine type 1 receptor and/or adenosine A2A receptor transmission as possibly leading to dystonia and calls for thorough investigation of the role of this pathway in dystonia pathophysiology.


A single large Australian family with dystonia symptoms ranging from focal to generalized but with prominent “whispering dysphonia” has been designated as having DYT4 [35, 36]. The disease is inherited in an autosomal dominant manner with reduced penetrance and has a wide range of age at onset (13–37 years), although most cases begin when the affected individuals are in their 20s. Of the 191 identified family members [37], more than 30 affected individuals were reported as typically presenting with laryngeal dysphonia which progresses to a generalized dystonia with a characteristic “hobby horse” ataxic gait. Recent reevaluation of the phenotype in this family identified clinical nonmotor features, such as thin face and body habitus, which co-segregated with the DYT4 phenotype [37]. DYT4 dystonia has long remained one of the few on the DYT list missing both locus assignment and gene identification. Recently, two research groups independently identified the TUBB4A gene on chromosome 19 as causative for DYT4 dystonia using a combination of linkage analysis and exome sequencing [13••, 14••].

The missense variant c.4C > G, p.R2G in exon 1 of the TUBB4A gene completely co-segregated with dystonia in the family and was found neither in the public databases nor in the healthy control chromosomes screened [13••, 14••]. Screening of 394 unrelated dystonia patients for mutation in TUBB4A revealed an additional missense variant (p.A271T) in a 71-year-old female patient with spasmodic dysphonia, oromandibular dystonia, and dyskinesia [14••]. Although she had a family history of dystonia, other family members were not available for analysis.

TUBB4A encodes β-tubulin, the main constituent of microtubules, a major cytoskeletal component [38]. The protein is mostly expressed in the brain; however, lower levels are also detected in other tissues, such as testes, cardiomyocytes, and blood [13••]. The p.R2G mutation affects an extremely conserved tetrapeptide, MREI, at the N-terminal part of the protein involved in the autoregulation of the β-tubulin messenger RNA [39]. Site-directed mutagenesis studies demonstrated the loss of autoregulatory capability in Arg2 mutants, resulting in increases in TUBB4A transcript levels [40]. In contrast, the expression of TUBB4A was significantly reduced in various cell types from the p.R2G mutation carrier [14••]. The reason for this discrepancy is not clear, and further studies are required to determine the functional consequences of TUBB4A mutations.

A new twist was recently added to the TUBB4A story when a de novo mutation in this gene was detected in 11 unrelated patients with hypomyelination and atrophy of the basal ganglia and cerebellum (H-ABS) [41], a rare leukodystrophy characterized by developmental delay, extrapyramidal movement disorders (dystonia, choreoathetosis, rigidity, opisthotonus, and oculogyric crises), progressive spastic tetraplegia, ataxia, and rarely seizures. In contrast to DYT4 (TUBB4A) patients who have no pathological findings on MRI [37], MRI of H-ABS patients shows hypomyelination, cerebellar atrophy, and the absence or disappearance of the putamen. The de novo p.D249N mutation occurs in a functional domain of TUBB4A which is responsible for interactions with other tubulins. In addition, Asp249 forms a salt bridge with the Arg2 residue mutated in DYT4. Further research is required to establish genotype–phenotype relationships in TUBB4A mutation carriers and explain how different mutations can result in variability in phenotypic features. Another interesting aspect is the finding of p.D249N mutation mosaicism in the asymptomatic mother of one of the H-ABS TUBB4A carriers [41], suggesting that rare de novo mutations that are initially phenotypically neutral in a mosaic individual can be disease-causing in the subsequent generation if they are inherited. This finding raises the question of whether apparently sporadic cases of focal dystonia could result as the subsequent generation of a de novo mutation carrier. The answer to this question can inform and transform the genetic study design and the analytical pipelines for exome sequence analysis in focal dystonia cases.


The DYT7 locus was mapped to chromosome arm 18p in a large German family. All affected members had adult-onset CD (mean 43 years; range 28–70 years), although some also had brachial and cranial involvement [42]. This family was recently clinically reevaluated, and six individuals were documented as definitely affected with focal dystonia [15]. Another individual was identified with an alteration of gait and possible CD but without the DYT7 risk haplotype. In addition, 12 other family members had a diagnosis of questionable CD. Four genes on chromosome arm 18p were sequenced on the basis of an incidental finding of a small deletion of this region in an unrelated CD patient, but no mutations were found (one of these was the GNAL gene; see earlier). Exome sequencing of two definitely affected patients revealed no causative variants in other genes located on chromosome arm 18p. In addition, copy number variations were ruled out using a SNP array, and the CIZ1 gene (CD phenotype, see earlier) was sequenced, with no mutations being identified. Taken together, these data suggest that the original linkage data may have been mistaken and call the DYT7 locus on chromosome arm 18p into question [15].

Phenotypic Spectrum

Exome sequencing can also be used to define the phenotypic spectrum of a disease. Two recent examples related to dystonia involve the ATP1A3 gene, originally described as a cause for rapid-onset dystonia parkinsonism (RDP; DYT12), and PRRK2, a gene implicated in paroxysmal kinesigenic dyskinesia (PKD; DYT10/DYT19). RDP is classified as a dystonia plus form, and PKD is a paroxysmal dystonia in the current classification system, but both would fall under “combined dystonia” in the new scheme, with RDP being combined persistent dystonia and PKD being combined paroxysmal dystonia [2]

ATP1A3 (DYT12)

ATP1A3 encodes the catalytic subunit of the sodium pump that uses ATP hydrolysis to exchange Na+ and K+ across the cell membrane to maintain ionic gradients. In 2004, mutations in the ATP1A3 gene were first described in cases of RPD (DYT12) [43]. The disease phenotype is usually characterized by onset of dystonia with parkinsonism (bradykinesia and postural instability) over hours to weeks, often triggered by physical or emotional stress with prominent bulbar dystonia that typically follows a rostrocaudal (face to arm to leg) gradient [44]. The typical age of onset is usually in the teens or early 20s, but recently two infants with motor delay and ataxia were reported with ATP1A3 mutations [45]. In addition, nonmotor manifestations are also found in RDP patients, including an elevated prevalence of mood disorders (50 %) and psychosis (19 %) compared with familial controls without ATP1A3 mutations [46].

Recently, using exome sequencing in trios, several groups identified de novo mutations in ATP1A3 as a cause of alternating hemiplegia of childhood (AHC) [47••, 48••, 49]. The diagnostic criteria for AHC include (1) paroxysmal episodes of hemiplegia, (2) episodes of bilateral hemiplegia, or quadriplegia, (3) other paroxysmal manifestations, such as abnormal eye movements, nystagmus, strabismus, dystonia, tonic spells, or autonomic disturbance which can occur during hemiplegia or as isolated events, (4) evidence of permanent neurological dysfunction which can manifest itself as mental retardation, seizures, ataxia disorders, choreoathetosis, developmental delay, and/or persistent motor deficits such as spastic diplegia/quadriplegia or hypotonia, (5) sleep during a paroxysmal attack relieves symptoms, although attacks may resume soon after awakening, (6) the first signs of dysfunction occur prior to the age of 18 months, and (7) it not being attributable to other disorders [50].

A total of 11 novel mutations (nine missense mutations, a 3-bp in-frame deletion, and a 3-bp in-frame insertion) have been reported in 20 RDP families, including eight de novo cases, eight that occur in families, and four with unknown origin of the mutation [45, 51, 52]. In AHC patients, a total of 27 different ATP1A3 mutations have been reported, including ten that are recurrent and two (D801N and E815K) that explain more than 50 % of AHC patients [47••, 48••, 49]. For the most part, none of the mutations overlap between RDP and AHC although the same amino acid is mutated to distinct residues in several cases (i.e., D801Y in RDP and D801N in AHC). However, there is one multiplex family where the same mutation (D923N) leads to both RDP (one typical and one atypical case) and AHC (one typical and three AHC-like cases) phenotypes [52]. There are several other examples of patients with intermediate phenotypes [46, 48••], suggesting that these disorders are different manifestations along a clinical spectrum [53]. Studies to determine exactly how these mutations disrupt the Na+/K+ pump may shed light on these phenotypic differences, but other factors, including the timing of the expression of the mutation and environmental factors, may also play a role. Moreover, it would not be surprising if a wider range of phenotypes associated with ATP1A3 mutations become apparent as next-generation sequencing becomes more widely used in clinical practice.


PKD is characterized by brief (less than 1 min) frequent (up to 100 times per day) attacks of choreoathetoid and/or dystonic movements induced by sudden voluntary movements or startle [54]. Patients remain conscious, and the neurological examination findings are normal. The disease is inherited in autosomal dominant manner, and the prevalence is estimated at one in 150,000 individuals [54]. Two thirds of PKD patients have a family history, and sporadic cases are reported more frequently in males. Two loci for PKD (DYT10 and DYT19) were previously mapped on chromosome 16 very close to each other [55, 56]. Recently, exome sequencing in PKD families identified mutations in proline-rich transmembrane protein 2 (PRRT2) as a cause of PKD in these families [5759] as well as in a PKD cohort with infantile convulsions [60]. Presently, PRRT2 is the major gene associated with PKD and is responsible for both DYT10 and DYT19 [61]. In the Chinese population, three different studies identified PRRT2 mutations in a total of 16 of 17 families and ten of 29 sporadic cases [5759]. In a European population of 34 PKD index cases selected on the basis of the criteria defined by Bruno et al. [54], PRRT2 mutations were reported in 65 %, including 13 of 14 familial cases (93 %) and nine of 20 sporadic cases (45 %) [62]. In an ethnically diverse cohort of PKD/infantile convulsion families, 24 of 25 clinically well documented families had a PRRT2 mutation, whereas 28 of 78 less well clinically characterized families also possessed a mutation [60]. In all of these studies, most of the PKD patients harbor the same insertion mutation (c.649_650insC), resulting in truncation of the encoded protein [5762]. The age of onset in PRRT2 mutation carriers is younger (range 6–14 years) than in non-PRRT2 PKD patients (range 9.5-19 years). A small number of patients with the typical PKD phenotype do not harbor mutations in PRRT2, suggesting existence of at least one more PKD causative gene. PRRT2 is highly expressed in the basal ganglia, and disrupted neurotransmitter release has been suggested as a possible pathogenic mechanism [5760].

Mutations in the PRRT2 gene have also been found in patients with a wide array of phenotypes (reviewed in [63]), including:
  • Benign familial infantile seizures: an autosomal dominant disorder with nonfebrile convulsions that begins between 2 and 12 months of age, responds well to anticonvulsant treatment, and remits by 2 years of age.

  • Infantile convulsions with choreoathetosis with or without PKD: an autosomal dominant disorder characterized by infantile seizures, paroxysmal dyskinesia or both.

  • Episodic ataxia: a rare disorder characterized by attacks of ataxia and usually caused by mutations in KCNA1 or CACNA1A; half of the individuals who did not have mutation in these two genes had mutation in PRRT2.

  • Hemiplegic migraine: a rare subtype of migraine with aura and attacks associated with transient weakness or hemiparesis.

  • Benign paroxysmal torticollis of infancy.

No obvious genotype–phenotype correlation has been observed yet, since phenotypes differ not only between families but also between carriers of the same mutation within the same family [64]. However, the shared feature of all the above-mentioned phenotypes seems to be the paroxysmal nature of the symptoms. It is worth noting that in a family carrying two copies of a PRRT2 mutation, homozygous mutation carriers had a more severe phenotype than their heterozygous carrier relatives, suggesting that dose can play a role in the severity of the phenotype [65].


Exome sequencing is a useful research tool for dystonia gene discovery which has allowed geneticists to more than double the number of genes responsible for primary dystonia in the past year. Both the phenotypic variability associated with mutations in the same gene (ATP1A3, PRRT2, TUBB4A) and the overlapping clinical phenotypes among the primary dystonia genes (DYT1, DYT6, CIZ1, ANO3, GNAL) make molecular diagnosis of inherited dystonia challenging and expensive. The price of Sanger sequencing for the currently identified dystonia genes will soon be more expensive than clinical exome sequencing. However, clinical exome sequencing in a single dystonia patient also has drawbacks as sorting through the novel variants to determine which one is pathogenic is currently very difficult and will lead to many “variants of unknown significance.” Instead, the community should work towards a database of novel variants identified in dystonia patients coupled to detailed clinical phenotypes that will allow investigators to identify rare variants that are present in multiple patients or groups of rare variants that concentrate in particular genes. Moreover, detailed phenotypic information will be useful to more fully describe the associated clinical features in those patients who possess novel variants in known dystonia genes.


Laurie J. Ozelius has received grant support from the National Institutes of Health (NS075881; for exome sequencing to identify dystonia genes). Tania Fuchs has received grant support from the Dystonia Medical Research Foundation (DMRF).

Compliance with Ethics Guidelines

Conflict of Interest

Tania Fuchs declares that she has no conflict of interest.

Laurie J. Ozelius has received royalties from Athena Diagnostics for patents related to DYT1 and DYT6 diagnostics.

Human and Animal Rights and Informed Consent

This article does not contain any studies with human or animal subjects performed by any of the authors.

Copyright information

© Springer Science+Business Media New York 2013