TOR2A Variants in Blepharospasm

Introduction

Blepharospasm (BSP) is a focal dystonia, typically adult-onset, characterized by the occurrence of involuntary spasms of the orbicularis oculi muscles, which usually manifests bilaterally and demonstrates synchrony and symmetry [1]. Over time, BSP often extends to affect neighboring craniocervical regions, encompassing the lower face, masticatory muscles, and neck, leading to the development of segmental craniocervical dystonia [2, 3]. The term “BSP-plus” (BSP+) is employed to describe individuals with BSP who exhibit further expansion of signs into these contiguous anatomical segments [4]. Many individuals with BSP have a first-degree relative with dystonia and penetrance is roughly 20% in BSP pedigrees [5, 6]. BSP alone or in combination with dystonia in other anatomical segments has been reported in patients with deleterious variants in TOR1A, THAP1 and GNAL. In biorepositories of isolated dystonia that include subjects with BSP, approximately 10% of participants report having a dystonia-affected relative [2, 7, 8].

TOR2A, encoding torsin2A, is a member of the AAA+ superfamily of ATPases and primarily associated with endoplasmic reticulum (ER) function [9]. The TOR2A homologue TOR1A was the first gene convincingly linked to isolated dystonia [10]. Mutations in TOR2A have been implicated in various dystonic phenotypes, including both isolated blepharospasm and more complex forms of dystonia [11, 12, 13]. In particular, a highly deleterious TOR2A variant was found to co-segregate in a large multiplex family with BSP (NM_001085347.3:c.568C>T, p. Arg190Cys) [11].

Methods

The genetic material utilized in this investigation was sourced from specimens collected by the Dystonia Coalition (DC) and obtained from the Coriell Institute for Medical Research, located in Camden, New Jersey, USA. The DC operates as an integral component of the Rare Diseases Clinical Research Network, which receives funding support from the National Institutes of Health and is administered by the National Center for Advancing Translational Sciences (NCATS). This funding is provided through a collaborative grant (U54NS116025) involving NCATS and the National Institute of Neurological Disorders and Stroke. The ethical clearance for the DNA analyses conducted in this study was granted by the Institutional Review Board at the University of Memphis.

The cohort under investigation comprised 307 individuals with BSP phenotypes (N = 200) or BSP+ phenotypes (N = 107), which included a range of combinations involving lower facial, oromandibular, and cervical dystonia. The demographic breakdown included 224 females and 83 males, with ages spanning from 19 to 87 years. The median age at data acquisition was 63.1 ± 11.0 years. The racial composition consisted of 259 whites, 1 Native American, 1 Pacific Islander, 11 Asians, 20 blacks, 5 individuals with mixed racial backgrounds, and 10 individuals with undisclosed or unknown race.

For our genetic analyses, we employed the GRCh38.p14 reference genome assembly. Primers were designed to comprehensively cover TOR2A coding regions, including exon-intron boundaries (Table 1). Our sequencing efforts extended to proximal intergenic regions both 5’ and 3’ to TOR2A. We conducted unidirectional Sanger sequencing for the entire cohort, followed by bidirectional Sanger sequencing to validate the identified variants.

To identify previously reported TOR2A variants, we conducted systematic searches of ClinVar [14] and PubMed. PubMed searches were refined using specific search terms, including dystonia, blepharospasm, gene, genetics, mutation, genetic variant, Meige, and TOR2A. Additionally, we leveraged the gnomAD V3.1.2 database to gauge the population prevalence of these variants [15]. To assess the potential deleteriousness of variants, we utilized CADD Phred-scores [16, 17], MetaLR [18, 19], and REVEL [19]. Our classification of pathogenicity adhered to the established criteria of the American College of Medical Genetics and Genomics [20], which takes into account a variety of factors, including population data, variant databases, co-segregation, disease databases, and the variant’s location within established functional domains of the encoded protein. Variants were categorized using recommended terminology, including ‘pathogenic,’ ‘likely pathogenic,’ ‘uncertain significance,’ ‘likely benign,’ and ‘benign.’ Furthermore, we examined the gnomAD v3.1.2 dataset to identify putative loss-of-function (pLoF) variants.

Results

No highly deleterious TOR2A variants were identified in our cohort of 307 subjects (Table 2). The major TOR2A isoform (NM_001085347.3, transcript variant 1) harbors 5 exons. Other RefSeq and Consensus Coding Sequence isoforms contain 2 to 4 exons. As seen in Table 2, several variants were identified but none of these showed notable differences in allele frequency when compared to the gnomAD v3.1.2 database. One variant (NC_000009.12:g.127733354C>G) is present in both coding (NM_130459.4:c.624G>C, p.Trp208Cys) and non-coding regions (NM_001085347.3:c.593+31G>C).

TOR2A is included in 62 ClinVar submissions. Of these, 40 are pathogenic structural variants affecting multiple genes. There are 17 missense variants of uncertain significance (Table 3) and several of these are deleterious and rare or absent from gnomAD v3.1.2.

Two independent studies screened patients with mainly BSP for TOR2A variants [12, 13]. No highly deleterious (CADD > 25) variants were identified in their cohorts (Table 3). Co-segregation was not performed in either study and no family history information is provided for identified variants. A variant of undetermined significance was identified in a single subject with BSP (NM_001085347.3:c.593+36del; NM_130459.4:c.629del, p.Gly210AlafsTer60). This single nucleotide deletion results in a frameshift and likely nonsense mediated decay within transcript 2 (NP_569726.2:p.Gly210AlafsTer60). Another variant in the same manuscript appears to be incorrectly assigned to an intronic location (NM_001085347.3:c.289insGGCTGGACCGGC/c.299delC). The c.299C cannot be validated and is a likely annotation error due to misinterpretation of the electropherogram. The insertion is actually located in Exon 2 (NM_001085347.3:c.277_288dup, p.Gly93_Gly96dup) with a total allele count of 11/152226 and East Asian allele count of 10/5188 in gnomAD v3.1.2 and CADD-Phred score of 21.5.

Discussion

Given our previous work showing co-segregation of a highly deleterious variant in a multiplex pedigree and the close similarity to TOR1A, we undertook a comprehensive analysis of TOR2A in BSP/BSP+. In our cohort of 307 subjects with BSP/BSP+, there were no highly deleterious TOR2A variants. Moreover, common single nucleotide polymorphisms showed no association with BSP/BSP+ when compared to a population cohort (gnomAD v3.1.2). Unified in silico analysis of two other screening studies of BSP identified several variants of uncertain significance.

Like TOR1A, TOR2A is located at 9q34.11. Also, like TorsinA, deletion of Torsin2A increases nuclear envelope blebbing [21]. Online Inheritance in Man does not yet link TOR2A to a human medical disorder. In gnomAD v3.1.2, there are a total of 11 unflagged putative loss-of-function (pLoF) coding variants. Numerous deleterious single nucleotide variants are reported in ClinVar but no trio analyses are included in the individual reports and most of these variants are present in normal populations. It is unlikely that these variants could cause severe early-onset phenotypes in Mendelian fashion but could contribute to oligogenic burden [22].

The previously published TOR2A variant (NM_130459.3:c.568C>T, p.Arg190Cys) variant [11] was not found in the DC cohort or two other screening studies [12, 13] Importantly, this variant is predicted to be deleterious by various in silico tools and co-segregated with BSP in a four-generation pedigree. At the time of analysis, penetrance was 43% in this pedigree (3 affected, 4 carriers). Co-segregation was not assessed in the two other published screening studies.

The are several limitations to our work. Most importantly, we only examined patients with BSP/BSP+. It is possible that TOR2A plays a larger role in generalized and other anatomical distributions of dystonia. We only used Sanger sequencing which can fail to detect exonic deletions and larger structural variants. Our sample size was modest. For illustration, power analysis indicates that a sample size of 545 would provide 80% power (α = 0.05) to detect a pathogenic TOR2A allele in a BSP cohort if TOR2A contributes to 1% of BSP/BSP+ cases with a penetrance of 20%. Finally, we did not assess the biological effects of any variant with a read-out such as nuclear bleb formation.

In conclusion, highly deleterious variants in TOR2A are rare in BSP/BSP+ phenotypes. Future studies should include younger patients and other anatomical distributions of dystonia.

Ethics And Consent

Our analysis of DNA acquired from Coriell was approved by the University of Memphis Institutional Review Board. Samples from Coriell were not associated with personally identifiable information.

Funding Information

Dr. LeDoux was funded by the National Institutes of Health (R21NS123827, R56NS123059), Department of Defense, Dystonia Medical Research Foundation, and Benign Essential Blepharospasm Research Foundation.

Competing Interests

The authors have no competing interests to declare.

Author Contributions

Dr. LeDoux designed the experiments, analyzed data, and wrote the manuscript. Dr. Saeirad sequenced DNA, analyzed data, and wrote the manuscript.