Definition

Chronic tic disorders (CTDs) are a family of neuropsychiatric conditions that are defined by the presence of motor and verbal “tics” that have lasted for over 1 year before 18 years of age. Tics are very generally defined as unpredictable, uncontrolled movement or sounds produced by patients. Patients will often experience a “premonitory urge” before the tic, which in many cases the tic relieves. The most severe diagnosis characterized by chronic tics, and the one most recognizable to the public by name, is Tourette syndrome (TS). To formally receive a TS diagnosis, a patient must have a CTD that features a combination of 2 or more motor tics and at least 1 verbal tic. There are many patients with long-lasting tics that are either motor or verbal in nature, but not both. These patients receive diagnoses of CTD, specifically motor CTD or verbal CTD. The diagnosis of provisional tic disorder (PTD) is also a member of the greater tic disorder (TD) family, and distinguishes itself from CTD by being present in the patient for less than 12 months. Many patients with PTD will see their symptoms subside over time, and will not see their diagnosis get reclassified as CTD or TS, but there are some whose symptoms do not disappear over time and instead may worsen. This review will describe what we have learned about genetic contributions to TS and CTD inclusively. For a review of the history of TS/CTD diagnosis, which is beyond scope of this review, please refer to an earlier study.

Characterization of Tics that Define Tic Disorder

It is important to better define tics as they are observed in TD and to establish what types of tics are most common among patients with TD. Tics in TD can be defined as either simple or complex. Simple motor tics will often involve only one group of muscles, with examples including blinking or sudden movement of a specific body part, while complex motor tics will involve coordinated activity between several muscle groups that mirrors learned motor activity. ^, Analogous to this, simple verbal tics will involve the production of sounds like sniffling and throat-clearing, while complex verbal tics will frequently involve uncontrolled, unwanted production of full words or phrases. ^, As highlighted before, whether a patient has both 2 or more sustained motor and at least 1 sustained verbal tic as opposed to just one or the other will control whether they receive a formal diagnosis of TS versus CTD. In spite of this, there is a great deal of variance in tic severity from patient to patient, and the TS diagnostic classification does not properly portray the spectrum nature of TD as a whole. This is why experts in treating TS and CTD now often consider these diagnoses as being on the more extreme end of symptom severity within the larger spectrum of TD, where individuals with lower symptom severity may never receive diagnoses. ^,

Prevalence of Tourette Syndrome and Chronic Tic Disorder

The prevalence of TD on a whole depends on the minimum diagnosis required for inclusion. A meta-analysis from Scharf and colleagues produced a pooled population prevalence estimate of 0.52% for TS. A greater number of individuals experience CTD, with Scahill and colleagues producing prevalence estimates in this range for motor CTD and verbal CTD separately. The prevalence of PTD is harder to estimate, with the general occurrence of transient tics in childhood estimated to be as high as 20%. There is a wide spectrum to not just tic type and severity inherent to TD, but also tic cessation. Around 20% of childhood cases with TS or CTD go on to have moderate-to-severe symptoms in adulthood. ^, A greater understanding of risk factors for TD will aid in better treatment not only for those in adulthood who continue to have TS or CTD, but also a sizable portion of children who experience tics at some point during development.

Twin Studies

Twin studies comprise the longest-standing formal evidence for genetic contributions to TS/CTD risk. The first sizable twin study, from Price and colleagues, focused on TS and featured 30 MZ twins and 13 same-sex dizygotic (DZ) twins, deriving a TS diagnostic concordance of 53% and 8%, respectively. Given the relatively low prevalence of TS diagnoses in the general population, such high concordances are unlikely to occur by chance. Later studies leveraged the power of Scandinavian national registries to identify MZ twins with TS/CTD diagnoses on a nation-wide scale. Lichtenstein and colleagues used the Swedish registry to assess twin-based diagnostic concordance for neurodevelopmental conditions, including TS/CTD, and estimated a total contribution to TS/CTD risk of 56% from genetics, in line with the concordance rate derived from Price and colleagues. An even larger scale twin study from Polderman and colleagues several years later produced twin concordance estimates for 17,804 traits using all published twin study statistics that had been published at time of writing, and among other traits and conditions, reported an MZ and DZ TS concordance rate of 0.63 across 2658 pairs and 0.34 across 3780 pairs, respectively. Collectively, these results are consistent with these twins both inheriting genetic risk factors for TS/CTD that make them more likely to receive a formal diagnosis.

Population-Based Registry Studies

Access to national registries has also made it possible to perform well-powered analyses of TS diagnoses in extended families relative to those that lack an in-family diagnosis. Mataix-Cols and colleagues utilized the Swedish registry to determine if individuals related to those with TS/CTD diagnoses were more likely to have such formal diagnoses themselves. Consistent with twin studies, Mataix-Cols and colleagues reported that first-degree relatives of cases were far more likely to be cases themselves than those who lacked first-degree relatives with TS/CTD (OR = 18.69). Those with second-degree and third-degree relatives with TS/CTD were also more likely to be cases themselves (OR = 4.58 and OR = 3.07, respectively). A similar study from Browne and colleagues was published that focused on TS/CTD and obsessive-compulsive disorder (OCD) diagnostic statistics from the Danish national registry. As with Mataix-Cols and colleagues, siblings and children of TS/CTD diagnoses were far more likely to have TS/CTD themselves (relative recurrence risk of 18.63 and 61.02, respectively). These findings line up well with those from twin studies and support a substantial amount of risk for TS being heritable.

Recent population-based registry studies are able to estimate contributions to TS risk beyond standard additive genetic liability. The best example of this can be found in Mahjani and colleagues in 2023, where authors used registry data to estimate contributions to TS/CTD risk from maternal effects. These maternal effects are defined as the effect of the mother’s phenotypes on those of the offspring and can be partitioned into genetic maternal and environmental maternal effects. The study followed the procedure utilized in a prior OCD-focused study, deriving sibling relative risks specific to maternal half siblings (mHS) and paternal half siblings (pHS), and found that mHS pairs had notably higher recurrence risk than pHS pairs (7.1 vs 3.6). It went on to quantify the total amount of phenotype variance explained by genetic effect as 60.7% (in line with prior estimates), 4.8% from genetic maternal effects, and 0.5% from environment maternal effects. These results still support the large majority of phenotypic variance being drawn from genetic effects that are detectable via formal association studies.

A First-Generation Genome-Wide Association Study of Tourette Syndrome

The first single sizable genome-wide association study (GWAS) of common genetic variation in TS, by Scharf and colleagues, featured 1285 cases and 4964 controls. Though it failed to implicate any specific loci with TS risk, it did note that top-ranking independent single nucleotide polymorphisms (SNPs) were enriched for overlap with expression and methylation quantitative trait loci specific to brain, but not for similar loci in lymphoblastoid cell lines. This would be consistent with at least some of these top-ranking SNPs conferring TS risk and acting by altering gene expression in brain tissue.

Contemporary Genome-Wide Association Study Meta-Analyses

A more recent large-scale analysis by Yu and colleagues of TS genetic data from 4819 cases compared to 9488 controls offered far more compelling evidence for a contribution to TS risk from common variation. This GWAS identified one genome-wide significant locus within the FLT3 gene that did not replicate in a separate cohort. More importantly, it estimated a total heritability from common variation at 0.21 (standard error = 0.024), meaning that around 21% of the variance in case status can be explained by common genetic variation. It also utilized polygenic risk score (PRS) calculation to assess the relationship among family history, tic severity, and TS genetic burden, and it found that both family history and tic severity are associated with an increased TS PRS.

The largest and most recent GWAS was published by Tsetsos and colleagues, featured 6133 cases and 13,565 controls in their full meta-analysis and implicated a novel locus with TS at genome-wide significance. The locus is 350 kilobases downstream of NR2F1-AS1 , a gene that produces a long noncoding RNA. Authors took the summary statistics from their unique cases and controls (n = 1438 and n = 4356, respectively) and found that in comparing these statistics with those from Yu and colleagues described previously that the genetic correlation between the statistics was high (r _g = 0.95, p = 6 × 10 ⁻⁸ ), meaning that both have consistent results pointing to some true set of common risk variants for TS. Tsetsos and colleagues also noted a significant genetic correlation with OCD (r _g = 0.39) that survived correction for multiple testing.

Polygenic Risk Scores and Clinical Features

TS genetic study has benefited from methodological development that uses common variant genotypes across the genome to produce PRS for different traits. These approaches have allowed researchers to quantify patients total genetic risk for TS and for other psychiatric conditions, and while their accuracy remains low, they are informative enough to be able to discern TS cases from unaffected controls. This was demonstrated in Yu and colleagues and in Tsetsos and colleagues. Both publications also derived further insights using these PRS values. In Tsetsos and colleagues, a negative association is described between TS PRS and putamen volume within the brain. Even more critical to understanding the condition, in Yu and colleagues, patients with TS are shown to have higher PRS than both patients with CTD and unaffected controls, while patients with CTD have lower PRS than TS but higher than controls. This is in line with the contemporary viewpoint that TS is on the phenotypic extreme end of the TS/CTD family which CTD is a part of. ^, Further in line with this, Yu and colleagues also found that TS cases from multiplex families had higher PRS than those from simplex families that lacked family histories. In line with this, a recent genetic study of an extended pedigree spanning 6 generations and featuring extensive amounts of TS/CTD diagnoses described a substantially higher TS PRS in affected family members and versus unaffected controls. There was no difference observed in unaffected family members versus these same controls. The data from these studies currently support a model where TS is on the extreme end of the TD spectrum, where genetic load is predictive of symptom severity, and where family history is predictive of the amount of common risk variant load inherited.

Rare Copy Number Variant Studies

Copy number variants (CNVs) are deletions and duplications of DNA that may only be present in a subset of individuals within a population. CNVs can be detected from genotype array data and from sequence data, but since much genetic data generated thus far are via genotype array, many of the studies so far have focused on these already-generated data. CNVs that contribute to neuropsychiatric risk are known to often be large (>30,000 base pairs) and rare (found in <1% of a population).

There are 2 TS-focused CNV studies worth highlighting, and both support a contribution from large, rare CNVs to risk. The first is focused on array-based CNV calls from 1086 TS and 1613 OCD cases relative to 1789 unaffected controls. Though this study only utilized particularly large CNVs (>500,000 base pairs in size), it noted an excess of neurodevelopmental deletions in the combined OCD and TS cohort relative to controls. It failed to observe a generalized increase in CNV rate beyond these loci. In a CNV study by Huang and colleagues specifically focused on TS cases (n = 2434) relative to unaffected controls (n = 4093), a global excess of large (>30,000 base pairs) rare (<0.01 frequency) CNVs was noted in cases which centered on protein-coding genes and, in particular, featured an excess of very large CNVs (>1,000,000 base pairs). Based on statistics provided in the publication, the rate of these risk CNVs could be estimated at 0.035 per TS patient. Furthermore, Huang and colleagues identified 2 specific types of risk CNVs for TS: Neurexin 1 ( NRXN1 ) gene deletions and contactin 6 ( CNTN6 ) gene duplications. While CNTN6 gene duplications have not been implicated with risk for other neurodevelopmental conditions, NRXN1 gene deletions are well-known risk factors for other neuropsychiatric conditions. This study in particular makes it clear that rare, large CNVs are a component of overall genetic risk for TS.

Exome Sequencing Studies of Tourette Syndrome

Unlike whole genome sequencing (WES), where sequence data across the entirety of the human genome is generated, WES specifically generates data for the 1% of the genome that codes for protein, focusing the study on the loci most likely to harbor disease-causing variants while also cutting sequencing and data storage costs. All relevant WES studies have focused on TS and relied on trio study designs, where samples contributed by a proband and both parents undergo WES. Many trio WES study designs (including the ones described here) specifically concentrate on protein-coding small variants that are de novo in origin, or absent in both parents and present in the proband. These coding de novo mutations (DNMs) occur at a rate of around 1 per proband and can be explicitly modeled based on sequence content and the number of trios in the study. DNM burden can also be compared in case trios versus unaffected control trios.

WES studies have revealed that these coding DNMs contribute to TS risk and have led to the identification of specific risk genes that when perturbed greatly increase chances of being a TS case. ^, The first major TS WES study that utilized the trio design included 511 trios total. It noted an excess of damaging DNMs in TS trio probands relative to unaffected control trios, consistent with the contribution of these mutations to TS risk. It also had a sufficient power to identify 4 TS risk genes: Cadherin EGF LAG Seven-Pass G-Type Receptor 3 ( CELSR3 ), Nipped-B-like ( NIPBL ), Fibronectin 1 ( FN1 ), and WW and C2 Domain Containing 1 ( WWC1 ). The gene WWC1 , the only ‘high-confidence’ TS risk gene identified in this study, codes for a cytoplasmic phosphoprotein expressed in both kidney and brain. ^, The second major TS WES study involved many of the same authors as the first, and expanded the original study cohort with an additional 291 trios for a total of 802 trios analyzed. An excess of damaging DNMs (including for the first time, de novo CNVs) was also observed in this study, in trio probands relative to control trios. Authors estimated that around 10.5% of TS cases have a damaging DNM that contributes to risk. This study identified 6 TS risk genes including the 4 already mentioned plus Fibrillin 2 ( FBN2 ) and OPA1 mitochondiral dynamin like GTPase ( OPA1 ). It also recategorized CELSR3 as a high-confidence risk gene alongside WWC1 . CELSR3 codes for an atypical cadherin protein, which is expressed in brain tissue and thought to be involved in definition of cell polarity. ^, Since these articles were released, the largest WES trio study of note to be published since then featured 100 TS trios and focuses on a DNM and reported overtransmission of coding variation to the known neurodevelopmental gene ASH1L . ^{, (p28),}

Additional WES studies focused on phenotypes that are comorbid or thought to be relevant have also yielded TS-relevant results. OCD, often comorbid with TS, has been the focus of several WES trio studies, the largest of which described significant enrichment of damaging DNMs in OCD probands within the 6 TS risk genes total described in Wang and colleagues. ^, This result is consistent with the high TS/OCD common variant genetic correlation noted by Tsetsos and colleagues earlier, and indicates that there are risk genes shared between OCD and TS. Another more recent WES trio study focused on stereotypies, which are disorders centered more on repetitive, rhythmic movement, and can occur alongside TS, described enrichment of overlap with genes carrying damaging DNMs in this cohort with genes carrying similar DNMs in TS. In general, these studies suggest that there are risk genes that when perturbed may increase the risk for TS alongside other conditions like OCD, stereotypies, and attention-deficit/hyperactivity disorder (ADHD) that are highly comorbid with TS diagnosis.

Rare Variant Burden and Clinical Features

Limited analyses of overlap between rare variant burden and clinical features for patients with TS have been conducted in the patients described, with results overlapping those from GWAS. Wang and colleagues compared the burden of damaging DNMs in probands from simplex families versus multiplex families and found that simplex probands had a higher burden of these variants. This aligns well with the GWASes described where cases from multiplex families have higher inherited common variant burden and suggests that simplex cases may be more likely to carry damaging DNMs that explain more of the phenotype.

At least one recent study from Wang and colleagues has used also WES data from male-biased diagnoses like TS, autistic spectrum disorder (ASD), and ADHD to determine if there is at least a partial genetic explanation for this vulnerability. They specifically concentrated on the variation in the X chromosome in male subjects which using trio data, are determined to be newly hemizygous. This led to the identification of one novel neurodevelopmental risk gene, MAGE Family Member C3 ( MAGEC3 ), and noted that in general TS, ASD, and ADHD are both enriched for damaging newly hemizygous coding variants. This result is consistent with known male genetic vulnerability for TS and provides at least one mechanistic reason for this occurrence.