Challenges in Essential Tremor Genetics

Abstract

The field of essential tremor (ET) genetics remains extremely challenging. The relative lack of progress in understanding the genetic etiology of ET; however, does not reflect the lack of a genetic contribution, but rather, the presence of substantial phenotypic and genotypic heterogeneity. A meticulous approach to phenotyping is important for genetic research in ET. The only tool for phenotyping is the clinical history and examination. There is currently no ET-specific serum or imaging biomarker or defining neuropathological feature (e.g., a protein aggregate specific to ET) that can be used for phenotyping, and there is considerable clinical overlap with other disorders such as Parkinson’s disease (PD) and dystonia. These issues greatly complicate phenotyping; thus, in some studies, as many as 30 – 50% of cases labeled as “ET” have later been found to carry other diagnoses (e.g., dystonia, PD) rather than ET. A cursory approach to phenotyping (e.g., merely defining ET as an “action tremor”) is likely a major issue in some family studies of ET, and this as well as lack of standardized phenotyping across studies and patient centers is likely to be a major contributor to the relative lack of success of genome wide association studies (GWAS). To dissect the genetic architecture of ET, whole genome sequencing (WGS) in carefully characterized and well-phenotyped discovery and replication datasets of large case-control and familial cohorts will likely be of value. This will allow specific hypotheses about the mode of inheritance and genetic architecture to be tested. There are a number of approaches that still remain unexplored in ET genetics, including the contribution of copy number variants (CNVs), ‘uncommon’ moderate effect alleles, ‘rare’ variant large effect alleles (including Mendelian and complex/polygenic modes of inheritance), de novo and gonadal mosaicism, epigenetic changes and non-coding variation. Using these approaches is likely to yield new ET genes.

Introduction

Essential tremor (ET) is a chronic, progressive neurologic disease [1]. Its hallmark motor feature is a 4 –12 Hz kinetic tremor. Thus, tremor occurs during voluntary movements such as writing, eating, and drinking. At disease onset, the tremor affects the hands and arms, but it may also eventually spread to involve the head (i.e., neck), voice, jaw and other body regions [2]. Given the presence of etiological, clinical, pharmacological response profile and pathological heterogeneity, there is growing support for the idea that ET may be a family of diseases whose central defining feature is kinetic tremor of the arms, and therefore, it might more appropriately be referred to as “the essential tremors” [3].

The presence of both clinical as well as etiological heterogeneity has hindered gene identification. With a resultant lack of progress in the field of ET genetics, there are no appropriate animal models, and this has greatly hindered progress in the area of experimental therapeutics. Indeed, despite the high prevalence of ET, which is considered among most common movement disorders, and the highly familial nature of the disease, with an apparent Mendelian autosomal dominant pattern of inheritance, the identification of ET genes has remained elusive. The discovery of ET genes would have a major impact on the ET field, and the search for genes is an area of intense research. In this viewpoint article, we summarize the current literature and progress in ET genetics, the current challenges in ET genetics, and novel approaches in ET gene identification.

Etiology of ET – Genetic Vs. Environmental Factors

Both genetic and environmental (i.e., toxic) factors are likely contributors to disease etiology. Many large kindreds show an autosomal dominant pattern of inheritance [4], and in a familial aggregation study, first-degree relatives of ET patients were approximately five times more likely to develop the disease than were members of the general population, and ten times more likely if the proband’s tremor began at an early age [5]. Twin studies provide strong evidence for a genetic contribution to ET with heritability estimates ranging from 45 – 90% in monozygotic twins [6, 7]. However, the existence of sporadic cases, variability in age at onset in familial cases, and lack of complete disease concordance in monozygotic twins also argues for nongenetic (i.e., environmental) causes as well [8]. A number of environmental toxins are under investigation, including β-carboline alkaloids (e.g., harmane, a dietary toxin) and lead [8].

Genetics - modes of inheritance and transmission in ET

Introduction

Historically, most studies have assumed a Mendelian inheritance pattern because of the high heritability and aggregation of ET in families [4, 6, 7, 9]. While other modes of inheritance, including autosomal recessive and complex inheritance patterns are possible, published family and linkage studies suggest an autosomal dominant mode of inheritance with reduced penetrance [10–13]. Echoing this, our own experience with ET pedigrees strongly suggests an autosomal dominant mode of inheritance with reduced penetrance because:

• ET is inherited or transmitted from father to son, which suggests that ET is not an X-linked trait,

• ET is observed and inherited in ≥3 consecutive generations, which is never observed with recessive traits,

• Males and females are affected with approximately the same probability,

• In some instances, a family member has clearly transmitted the allele to offspring but does not express the disease, suggesting incomplete penetrance.

In reality; however, the genetic architecture of familial and sporadic ET is likely to be explained by several modes of inheritance and transmission including both Mendelian and complex disease patterns. These modes of inheritance and transmission are unlikely to be mutually exclusive, as we have learned from other common complex diseases, including Parkinson’s disease (PD) [14], Alzheimer’s disease (AD) [15] and schizophrenia [16], and it is likely that in ET both Mendelian/monogenic and complex disease patterns contribute to the genetic architecture.

In the paragraphs below we summarize the studies to date. Family and linkage studies have been largely unsuccessful at identifying ET genes [10, 11, 13], which may reflect the technological and analytic limitations of these studies, which were performed in 1997 and 2001 before the era of whole exome sequencing (WES) and whole genome sequencing (WGS). We also suspect that it reflects, in large measure, the challenges in phenotyping.

Genome wide association studies (GWAS) do not overwhelmingly support the common disease common variant (CDCV) hypothesis, with only two significant genes identified from two published GWAS [17, 18]. To date, a complex disease rare variant (CDRV) model has not been tested in ET studies.

Mendelian Inheritance and Monogenic Genes: Family Studies and Linkage

Many studies indicate that on the order of 30 – 70% of ET patients have a family history, with the vast majority (>80%) of young-onset (≤40 years old) cases reporting ≥1 affected first-degree relative [19]. To date, only three published genome wide linkage scans have been performed, all in north American or Icelandic ET families [10, 11, 13]. These studies led to the identification of genetic loci harboring ET genes on chromosomes 3q13 (ETM1; OMIM:190300) [10], 2p22-p25 (ETM2; OMIM:602134) [11], and 6p23 (ETM3; OMIM: 611456) [13]. Several studies have attempted to replicate linkage to ETM1 [12, 20, 21], ETM2 [12, 22, 23] and ETM3, without success (no LOD score >2.0), and the genes and causal mutations for these loci (ETM1, ETM2 and ETM3) have yet to be identified despite the reporting of linkage over a decade ago. Recently, using a linkage and WES approach, a p.Q290X mutation in the fused in sarcoma/translated in liposarcoma (FUS/TLS) gene was identified as the cause of ET in a large Quebec family [24]. Subsequent studies [25–27], including our own, suggest that mutations in FUS are an extremely rare or family-specific cause of ET, and without functional studies, the pathogenicity of mutations identified so far (p.Q290X [24] and R377W reported in 1 patient with family history of ET [28]) is unknown. More recently, in a six-generation consanguinous Turkish kindred with both ET and PD, the mitochondrial serine protease HTRA2 p.G399S variant was shown to segregate with both phenotypes (PD and ET). The authors pointed out that all of the patients with the mixed phenotype (ET+PD) had severe ET and that the ET had been present for many years prior to the onset of PD. This makes it unlikely that the action tremor (“ET”) was merely an early motor sign of an evolving PD diagnosis. It also makes it less likely that the diagnosis was tremor-predominant PD rather than ET+PD. The diagnosis of PD required at least two cardinal features, which makes it less likely that these were merely longstanding ET cases who had developed isolated rest tremor. All affected individuals in the family were either heterozygous or homozygous for the HTRA2 variant and homozygosity was associated with earlier age at onset of tremor (p<0.0001), more severe postural tremor (p<0.0001), and more severe kinetic tremor (p = 0.0019) [29]. Follow up studies in ET family and case-control studies will be needed to determine whether HTRA2 represents a major ET susceptibility gene.

Complex Disease Inheritance Pattern: CDCV Hypothesis

Candidate Gene Studies

Since 2006, numerous genes have been evaluated as candidates for ET [30–53], based on either localization to linkage intervals or function (reviewed in Testa 2013 [30] and Jiménez-Jiménez 2013 [31]). These genes (table 1) provide at best weak evidence of association or no association at all (odds ratio [OR] range = 0.6 –1.5; p range = 0.94 – 0.01). We and others have also evaluated additional genes that are associated with other neurodegenerative disease such as PD, dystonia, spinocerebellar ataxias and Fragile X Tremor Ataxia Syndrome. We did not observe an association with the PD genes synuclein, alpha (non A4 component of amyloid precursor) (SNCA) [52], Leucine-rick repeat kinase 2 (LRRK2) [53], Glucosidase, beta, acid (GBA) [53] or Microtubule-associated protein tau (MAPT) [42], nor did we identify pathogenic repeat expansions in the 10 common spinocerebellar ataxia loci (SCA-1 (ATXN1), SCA-2 (ATXN2), SCA-3 (ATXN3), SCA-6 (CACNA1A), SCA-7 (ATXN7), SCA-8 (ATXN8OS), SCA-10 (ATXN10), SCA-12 (PPP2R2B), SCA-17 (TBP) and DRPLA (ATN1) or FRAXA (unpublished results).

Table 1.

Candidate Genes for ET

Gene	Published Citation
Dopamine receptor D3 (DRD3)	21, 32–35
HS1 binding protein (HS1BP)	36
solute carrier family 1 (glial high affinity glutamate transporter) member 2 (SLC1A2)	18, 37 – 39
microtubule associated protein 2 (MAPT)	40–42
methylenetetrahydrofolate reductase (NAD(P)H)(MTFHR)	43
alpha-2 macroglobulin (A2M)	44
Cytochrome P450, family 2, subfamily C, polypeptide 19 (CYP2C19)	45
Cytochrome P450, family 2, subfamily C, polypeptide 9 (CYP2C9)	46
Cytochrome P450, family 2, subfamily C, polypeptide 8 (CYP2C8)	46
alcohol dehydrogenase 4 (class II) pi polypeptide (ADH2)	47
glutathione S-transferase pi 1 (GSTP1), paraoxonase 1 (PON1)	48
gamma-aminobutyric acid (GABA) receptor genes	49–51
fused in sarcoma/translated in liposarcoma (FUS/TLS)	24–27
synuclein, alpha (non A4 component of amyloid precursor) (SNCA)	52
Leucine-rick repeat kinase 2 (LRRK2),	53
Glucosidase, beta, acid (GBA)	53
Microtubule-associated protein tau (MAPT)	42

Genome Wide Association Studies (GWAS)

There are two published GWAS, which variably identify single nucleotide polymorphisms (SNPs) in the Leucine rich repeat and Ig domain containing 1 (LINGO1) gene or an intronic variant in the SLC1A2 gene, with increased risk for ET (see below).

LINGO1

A genome-wide SNP association study of ET in an Icelandic population identified an association with a marker in the LINGO1 gene [17]. Since the initial report, numerous studies, including our own, have replicated the association in independent ET case-control samples worldwide [54–64]. Collectively, these data suggest that the LINGO1 SNP rs9652490 confers modest risk, with ORs in the range of 1.2 – 1.7, across different studies and populations. Although the majority of studies positively replicate the LINGO1 SNP rs9652490 association, some studies did not observe an association [59, 60, 62, 64]. One explanation for this lack of association may be allelic heterogeneity, and that rs9652490 does not confer risk in these populations, and that other variants in LINGO1 are risk factors. Alternatively, clinical and genetic heterogeneity in ET may explain the lack of association. We and others have also shown that a highly related LINGO1 family member (61% amino acid identity), the leucine-rich repeat (LRR) and Ig domain containing two genes (LINGO2), might also be a risk factor for ET and PD, providing further evidence that the LRR gene pathway may be perturbed in ET pathogenesis [55]. To date, SNPs in LINGO1 provide the strongest evidence for association with ET; however, data are not completely consistent.

SLC1A2

More recently, the SNP, rs3794087, located in SLC1A2, was identified in a two-stage European GWAS in a total of 990 ET cases and 1,537 controls, with an OR of ~1.4 [18]. To date, four studies have attempted to replicate the association of rs3794087 with ET in different populations, with conflicting results, some of which are positive (Chinese, Taiwanese) and others of which are negative (Spanish, North American) [37–39]. While further studies in different populations are needed to confirm the role of SLC1A2 in ET, rs3794087 is unlikely to represent a major risk factor for ET.

To summarize, there is a lack of consistent and robust associations from candidate gene studies and GWAS.

Complex Disease Inheritance Pattern: CDRV Hypothesis

To date, there is only one published study, as described above, which performed WES (combined with linkage); the investigators studied a single large ET family from Quebec and identified a nonsense mutation in the FUS/TLS gene. There are no published studies that have performed WGS in ET families.

Current challenges in ET genetics

Why has the field of ET genetics made so little progress? Despite significant efforts to identify genes for ET, the field has made little progress. There are a number of possible explanations, some of which we highlight below.

Phenotyping

We would rank this as perhaps the top problem. A meticulous approach to phenotyping is critical for genetic research in ET. The ET diagnosis relies on clinical evaluation. The only tool for phenotyping is clinical. Thus, there is currently no serum/imaging biomarker or defining neuropathological feature (e.g., a protein aggregate specific to ET or a distinctive imaging finding) that can be used for diagnosis, and there is clinical overlap with other disorders such as PD and dystonia. These issues greatly complicate the diagnosis of ET; thus, in some studies, as many as 30 – 50% of cases labeled as “ET” have later been found to carry other diagnoses (e.g., dystonia, PD, and other disorders) rather than ET [65]. Poor attention to phenotyping (e.g., merely defining ET as an “action tremor”) is likely a major issue in some family studies of ET, and this as well as lack of standardized phenotyping across studies and patient centers is likely to be the major contributor to the lack of success of GWAS. We have focused a great deal of attention on detailing the phenotypic characteristics that set ET apart. To begin with, in ET, the tremor itself is not merely a featureless, nondescript action tremor (although it is often viewed as such, and consequently there is a tendency to label many non-ET tremors as “ET”). Studies show that 30–50% of “ET” cases do not have ET [65–67]. Close inspection indicates that the tremor of ET is characterized by a definable and specific pattern of clinical features:

• kinetic tremor is nearly always greater than postural tremor [68, 69],

• the tremor involves movement at particular joints in specific directions [70, 71],

• the tremor is regularly recurrent and without directionality,

• the arm tremor is generally slightly asymmetric [72, 73],

• when drawing a spiral, a single tremor orientation axis is identifiable rather than multiple axes [74],

• intention tremor of the arms is seen on finger-nose-finger maneuver in approximately 50% of cases [75–77],

• rest tremor occurs in the arms (however, not the legs) in as many as 20% of cases [78],

• arm tremor precedes cranial tremor [79, 80],

• head (i.e., neck) tremor, unless very severe, generally resolves while supine [81],

• patients are often unaware of their head tremor, esp. if it is mild [82],

• the prevalence of neck tremor is greater than that of other cranial tremors [82, 83],

• and there is a tendency for tremor severity to increase over time [3, 84, 85].

These clinical features distinguish the tremor of ET from that of PD, dystonia, and drug-induced tremor [85].

Dystonia requires some additional discussion, as there are unresolved issues and some of the boundaries between ET and dystonia are unclear. The presence of mild dystonic movement or postures in the hands/arms (i.e., slight dystonic features) is not uncommon in ET patients, and it is questionable whether this physical finding merits any change in diagnosis, especially when in the presence of advanced, long-standing ET. By contrast, the presence of more marked dystonic features in the hands/arms or the presence of cranial dystonias (blepharospasm, torticollis) is considered an indicator of a mixed phenotype (ET + dystonia) [86]. The presence of more marked dystonia is even interpreted by some as an indicator that the patient has only dystonia, and that the “ET” is actually dystonic tremor [87, 88]. For genetic studies, the safest course is to separate patients with ANY dystonia from those with pure ET.

A related issue is the possibility, as discussed above, is that ET may represent a family of diseases rather than a single clinical-pathological entity.

Mode of inheritance

Mendelian and complex disease inheritance patterns may play a role in ET. While most studies have assumed an autosomal dominant mode of inheritance with reduced penetrance, other modes of inheritance including complex disease inheritance patterns may be important.

Genetic heterogeneity and sample size

Linkage and GWAS provide evidence for genetic heterogeneity in ET. Genetic heterogeneity has been observed in other neurodegenerative disorders such as PD, but does not preclude the identification of ET gene(s). One method to deal with the effects of genetic heterogeneity is to increase the sample size.

Molecular methods and statistical analysis of rare variants

A major roadblock to gene discovery, until recently, has been the lack of technological and analytical advances in genome wide sequencing and statistical models for multiple rare variants. Recent advances in Next generation sequencing (NGS) techniques, particularly for WGS, sequence data processing, decreased cost of WGS and faster computational power, together with new analytical methods to study rare variants, means that WGS and rare variant analysis is now feasible. While WES identifies on the order of ~20,000 coding variants per exome sequenced, WGS provides more uniform coverage and captures a range of genetic variation to allow analysis of both coding and noncoding variants in addition to indels and structural variation, with on average of ~4,000,000 variants per genome.

Novel approaches to ET gene identification

In understanding the genetic architecture of ET and novel approaches that can be used in gene identification, we turn to approaches being used to identify genes in other common complex diseases such as neuropsychiatric and neurodegenerative disease (AD and PD), and child neurodevelopmental disorders (e.g., autism).

To dissect the genetic architecture of ET, WGS in carefully characterized and well-phenotyped discovery and replication datasets of large case-control and familial cohorts is needed. This will allow specific hypotheses about the mode of inheritance and genetic architecture to be tested. As we outline in the following paragraphs, there are a number of approaches that still remain unexplored in ET genetics, including copy number variants (CNVs), the contribution of ‘uncommon’ moderate effect alleles, ‘rare’ variant large effect alleles (including Mendelian and complex/polygenic modes of inheritance), de novo and gonadal mosaicism, epigenetic changes and the contribution of non-coding variation. These approaches are likely to yield new ET genes.

Copy Number Variants

To date there are no published studies that have evaluated the contribution of CNVs to ET case-control studies or familial studies. As in other neurodegenerative disease [89, 90] and autism [91], CNVs are likely to play a significant role and be key contributors to the ET disease phenotype.

Uncommon Moderate Effect Alleles

GWAS of ET to date and the small number of significant loci (LINGO1 and SLC1A2) do not explain the fraction of heritability of ET. GWAS designs and SNP genotyping arrays typically evaluate ‘common’ SNPs but exclude SNPs with MAF <5%. Therefore SNPs with MAF in the range 1–5% representing ‘uncommon’ moderate effect alleles are excluded from analysis. Re-evaluation of existing ET GWAS datasets together with rigorous clinical phenotyping of ET cases may yield new genome wide significant genes.

Rare Variant Large Effect Alleles

Mendelian

As we outline in the section ‘Mendelian Inheritance and Monogenic Genes: Family Studies and Linkage’, family studies and linkage approaches have been largely unsuccessful in identifying ET genes. As we note in the previous section, only two genes, FUS and HTRA2, have been identified in single ET families and, although further studies are needed to determine the contribution of HTRA2 to ET, several studies including our own suggest FUS mutations are extremely rare and possibly isolated to this single Quebec family. In fact, our own experience with linkage and exome analysis of 50 early onset ET families, representing the largest familial and exome study to date, and that of others (including Dr. Coro Paisan Ruiz pers. commun.), suggests a substantial amount of genetic heterogeneity with no single mutation or different mutations within the same gene being shared across families. Our data suggest that either private mutations contribute to ET in these families or that a different mode of inheritance plays a role. To demonstrate that private mutations in single ET families do indeed contribute to the ET phenotype, functional studies will be needed to determine pathogenicity.

Complex/Polygenic Modes of Inheritance

As with ET, common genetic variants do not contribute to the genetic etiology of autism spectrum disorders (ASD), and recent studies have focused on the role of rare causal variants and de novo mutations [92, 93]. For ET, this approach remains largely unexplored. Like ET, ASD is a highly heritable disorder with a high monozygotic twin concordance rate for ASD at 70 to 90%. There are two possible explanations for this high heritability: a polygenic model involving several genes and environmental factors or de novo mutations with limited transmission. In the complex/polygenic model rare variant ‘burden’ analysis assessing the contribution of both inherited and de novo rare variants in genes in families may help in the identification of new genes. Most de novo mutations will occur in affected individuals and will not be transmitted, but exceptions include maternal transmission of mutations on the X chromosome to male offspring or gonadal mosaicism, which together would result in higher sibling recurrence rate. Pathway and gene network analysis of genes with rare variants may identify the functional impact, gene sets and connected pathways.

Epigenetic Changes

Epigenetic changes are likely to play a role in sporadic ET. Environmental exposures in both prenatal and postnatal life can lead to epigenetic changes. Although the number of published studies linking epigenetic changes to neurodegenerative disease such as PD and AD is rapidly increasing, a recent search of NCBI Pubmed found no published papers in this area in ET. Exposure to environmental factors that have already been shown to effect and alter brain gene expression, such as dietary factors (Vitamin B-12 and myelination of the nervous system), metal exposure (lead, aluminium, zinc, iron, copper and mercury have all previously been linked to neurodegenerative disease), pesticides (organophosphates, organochlorines and carbamates), exercise, smoking and drug and chemical exposure (caffeine, alcohol, harmane etc.) may all contribute to epigenetic changes in ET. Studying differences in patterns of histone acetylation and DNA methylation in ET case-control studies may identify genes that are subject to epigenetic changes in ET.

Contribution of non-coding Variation

One area of complex disease genetics that is gaining popularity, as we exhaust the findings in coding regions of the genome together with completion of the ENCODE project, is the identification of causal and risk variants outside the protein coding exome that affect gene transcription (promoter, enhancer elements), transcription of coding and non-coding RNAs or RNA functions and processing. This will be another important area of focus in ET genetics which, again, to date remains largely unexplored.

Conclusion

The field of ET genetics remains extremely challenging. The lack of progress; however, does not reflect the lack of genetic contribution, but rather, substantial phenotypic and genotypic heterogeneity. Dissecting the genetic architecture of ET using novel approaches such as WGS, CNVs and epigenetics is likely to help unravel the puzzle of ET genetics and identify genes and pathways that can be targeted for therapeutic intervention.