Ⅰ. INTRODUCTION
Tourette syndrome, also known as Gilles de la Tourette syndrome (GTS), is a childhood-onset neuropsychiatric disorder characterized by the combination of persistent motor and vocal tics, affecting 1~3% of the pediatric population [1]. The core features of symptoms are repetitive involuntary movements (motor tics) and vocalization (vocal tics) following a waxing and waning course. The majority of children experience a gradual improvement and complete recovery of symptoms by early adulthood [2]. However, moderate to severe tics are continuously observed in approximately 20% of patients [3]. Although GTS is defined by tics, patients up to 90% are reported to exhibit other comorbidities such as obsessive-compulsive disorder or behaviors (OCD/OCB) and attention-deficit/hyper-activity disorder (ADHD) [4].
Several studies have shown the abnormalities in the brain of GTS patients including the one demonstrating the dysfunction of neurotransmitter circuitry-regulating basal ganglia [5]. Moreover, among various types of neurotransmitters, gamma-aminobutyric acid (GABA), the inhibitory neurotransmitter for the dopaminergic pathway is known to be associated with TS [6, 7]. Recently, several candidate genes for GTS susceptibility have been suggested including inner mitochondrial membrane peptidase 2 like gene (IMMP2L), X-linked gene neuroligin4 (NLGN4X), histidine decarboxylase (HDC) and contactin-associated protein 2 (CNTNAP2) [8-10]. However, the evidence for a causal relationship between these mutations and GTS should be proved to provide insights practical for the application on GTS cases [11, 12]. Environmental factors have also been suggested as the risk factor in the pathogenesis of GTS, since perinatal ischemic events, maternal smoking or stress and low birth weight may lead to the advent of tics [13].
The treatment option for GTS is still limited and complicated because of the issues in insufficient efficacy and adverse effects of currently available drugs as well as the presence comorbid psychiatric diseases [14]. So far, the most potent drug for tics is antipsychotics, the antagonist for D2 dopamine receptor. Antipsychotics are reported to suppress the frequency and severity of tics up to 60~80% in most patients, however, severe side effects in motor functions have been reported [15]. Other drugs including α2-Adrenergic agonists, GABA agonists and cholinergic agents, have been reported to provide potential benefits for the treatment of GTS, but their efficacy is lesser or still controversial [15-17].
In line with the genetic and environmental hypotheses for GTS pathogenesis and investigation of pharmacological candidates for GTS treatment, several groups attempted to establish animal models as well as cell-based platforms recapitulating disease phenotype and mechanism. In this review, we provide an integrative overview on current knowledge reporting the genetics, immunology and relevant dysfunction in GTS, as well as the platform for the investigation of pathogenesis and candidate drugs using GTS-recapitulating animal models.
Ⅱ. GENETIC ABNORMALITIES UNDERLYING GTS
Similar to other neuropsychiatric diseases, the pathophysiology of GTS appears to be highly polygenic, influenced by multiple risk variants throughout the entire genome. Furthermore, its heritability is estimated up to 77% based upon a recent review [18]. Among the putative genes responsible for the development of GTS, many candidates belong to protein families that play a critical role in regulating synaptic transmission and structural integrity. We will further discuss a few examples of these candidate genes below for their potential roles in the nervous system and recent updates regarding their association with GTS (Fig. 1).
1. Neurexins and Neuroligins
Neurexins (NRXNs) are expressed primarily by neurons and localized mostly to synapses [19]. Neurexins constitute a protein family that functions as presynaptic cell-adhesion molecules, thus affecting the function of synapses and mediate the conduction of nerve signals. Each neurexin subfamily varies in its postsynaptic ligand, including neuroligins [20] and latrophilins [21] (Fig. 1A). Specific combinations of presynaptic neurexin and its postsynaptic partner thus determine the modes of action for each complex, exerting diverse regulatory functions. The clinical significance of these versatile protein complexes in relation to GTS was first brought to our attention when a familial case of GTS and OCD was investigated for genetic abnormalities in the affected family [8]. In this study, the affected father and two children were found to carry a common genomic translocation affecting the same parts of chromosomes 2 and 7, within which CNTNAP2 encoding a member of the neurexin family protein was affected.
Recent studies have revealed a strong association of altered NRXNs, especially NRXN1, with various neurodevelopmental diseases other than GTS, including autism spectrum disorders (ASD). Indeed, dysfunction of NRXN1 has been repetitively mapped as a monogenic risk factor responsible for the occurrence and phenotypes of ASD in both genomic analyses of human subjects [22, 23] and experimental settings using animal models [24]. While these findings have shifted the focus of NRXN-related investigations toward ASD, recent reports based upon genomic analyses have demonstrated that rare copy number variations (CNVs) at the NRXN1 locus are also strongly implicated in the development of GTS [25, 26], thus suggesting the shared molecular abnormalities underlying multiple neurodevelopmental diseases.
In addition to NRXNs, neuroligins (NLGNs) appear to be associated with GTS as well as ASD. NLGNs are expressed in postsynaptic neurons that interact with presynaptic NRXNs and implicated in the maturation and function of neuronal synapses [27] (Fig. 1A). Multiple deletions within NLGN4 exons have been reported in a familial study, in which the affected family displayed an array of neuropsychiatric illnesses, including ASD, attention deficit hyperactivity disorder (ADHD), learning disability and GTS [10]. Interestingly, this finding is in line with the idea that alterations in NRXNs, and perhaps their binding to NLGNs, are commonly associated with multiple neuropsychiatric and neurodevelopmental disorders [18].
2. Contactins (CNTNs)
Similar to NRXNs and NLGNs, genomic alterations of CNTNs, including CNTN6, have been implicated in the array of comorbid neuropsychiatric disorders, including ASD, intellectual disability, schizophrenia, bipolar disorder and GTS [18, 28]. CNTNs consist of six subclasses of the immunoglobulin CAM superfamily, each of which plays crucial roles in neural and glial networks. CNTN1 and CNTN2, two prototypical members of CNTN family, function at the neuron-glia interface and in the formation of the nodes of Ranvier [29, 30] (Fig. 1B), along with its roles in regulating neuronal migration and axon [31]. While most studies point toward a stronger association between CNTNs and ASD [28], a recent study has addressed a novel insight into CNVs at the locus of CNTN6 as a risk factor for GTS [26]. In this study, 2434 cases of GTS and 4093 controls were genotyped and screened for CNVs. As a result, exonic duplications were detected within CNTN6 on chromosome 3p26, with the value of OR at 10.1, suggesting a strong association between extra copies of CNTN6 and GTS cases.
3. Slit- and Trk-like family proteins (SLITRKs)
SLITRK family is a group of transmembrane proteins that are highly expressed in the central nervous system, where they interact with presynaptic leukocyte common antigen-related receptor protein tyrosine phosphatases (LAR-RPTP) [32] (Fig. 1A). They are also thought to participate in the process of synapse formation from the postsynaptic compartments [33]. Their association with GTS was first reported when rare sequence variants within SLITRK1 gene on chromosome 13q31.1 were mapped from 3 out of 174 unrelated probands of GTS, including a frameshift mutation and two independent and identical variants [34]. A similar result was deduced from a subsequent study with Taiwanese subjects [35], further confirming the association between SLITRK1 and GTS. These sequence variants in SLITRKs were capable of inducing loss-of-function phenotypes in vitro, with reduced formation of synapses and neurite outgrowth in cultured hippocampal neurons [36]. Taken together, these results are in line with the idea that altered function or expression of synaptic proteins is often responsible for the development of GTS.
4. Netrin-4
Netrins are a family of extracellular proteins that play an important role in axon migration during embryogenesis, via their interaction with a set of receptors, including DCC, DSCAM and UNC-5 homolog families [37] (Fig. 1A). Netrin-4 (NTN4) is one of the key molecules that participate in the process of axon guidance in the developing striatum by its interaction with DCC, where it interacts SLIT and WNT at the tip of the migrating axons [38, 39]. A recent GWAS study based upon 610 control and 609 GTS subjects has revealed repetitive single nucleotide polymorphisms (SNPs) at rs2060546 on chromosome 12q22 near NTN4, reaching the OR value of 2.41 [40]. The identical SNPs at ts2060546 was also evident in a large cohort consisting of 1316 GTS patients and 5023 controls from the GTS GWAS Replication Initiative (GGRI), with the comparative OR of 3.74 [41]. Since this specific SNP lies in the intergenic area just upstream to NTN4, it is currently unclear whether it could affect the function of NTN4 in vivo. Further functional analyses should be performed to delineate the causative gene(s) at rs2060436 locus.
5. L-histidine decarboxylase (HDC)
HDC is a rate-limiting enzyme for the biosynthesis of histamine [12]. The association between HDC/histamine biosynthesis and GTS was first proposed by animal model studies. For example, HDC knockout mice displayed progressive augmentation of stereotypic behaviors following repetitive exposures to dopamine agonists [42] and evoked stereotypic movements such as rearing, sniffing and biting [43], as well as facilitated anxiety [44]. A linkage analysis of a two-generation pedigree with GTS has revealed rare variants in HDC in the affected family, confirming its association with GTS [12]. Further target re-sequencing candidate genes for GTS using a set of 382 GTS patients successfully verified deleterious rare variants in HDC [45]. In addition, tic-like stereotypic behaviors in HDC KO mice were ameliorated by the application of dopamine D2 antagonist haloperidol, paralleling the effect of the drug used for treating human GTS patients [46], suggesting histamine-dopamine interaction as a core target of GTS treatments.
6. Other candidates potentially linked to GTS
In addition to the genes described above, multiple risk variants have been uncovered from cohort studies in different scales. One such example is arylacetamide deacetylase (AADAC), a protein proposed to participate in lipolysis, detoxification and drug metabolism [47]. Although its function in the nervous system remains poorly understood, a cohort study based upon 1181 GTS patients and 118,730 controls revealed a significant association between AADAC and GTS with the OR of 1.9 [48]. In addition, de novo coding variants presumably associated with GTS in a recent study include WW and C2 domain containing 1 (WWC), Catherine EGF LAG seven-pass G-type receptor 3 (CELSR3), nipped-B-like and Fibronectin 1 [49].
Recent genome-wide sequencing effort targeted for identifying chromosomal rearrangements has also yielded a number of potential candidates associated with GTS. For instance, a balanced translocation between chromosomes 6 and 22 in the cohorts affected by GTS and OCD pointed toward a breakpoint in chromosome 6 [50]. The genes affected at this locus include G protein-coupled receptor 63 (GRP63) encoding a sphingosine 1 phosphate receptor, NADH Dehydrogenase (Ubiquinone) 1 Alpha Subcomplex 4 (NDUFA4) expressed in mitochondria and Kelch-like 32, of which function remains elusive. Another case-control study using 460 GTS and 1131 control subjects has indicated multiple genes disrupted in histaminergic pathways in GTS as well as ASD patients [51].
Taken together, these results demonstrate the presence of multiple risk variants spreading throughout the genome, with presumably different contributions to the development of GTS, and suggest unknown candidate genes that are yet to be discovered. It should be noted that the experimental evidence is mostly lacking for elucidating the functional consequences of these rarely mapped variants. Whether experimental emulations of these variations are capable of producing noticeable changes in the structure and function of the nervous system and subsequent behavioral alterations awaits further investigations.
7. Candidate genes in complex clinical settings: co-morbidity among multiple neurodevelopmental and neuropsychiatric disorders
The main impediment that hinders experimental validation of candidate genes responsible for GTS is the findings that similar patterns of genetic alterations can also be detected in a number of other neuropsychiatric and neurodevelopmental disorders, including OCD, ASD and ADHD [4, 18]. These comorbid disorders often remain persistent throughout the clinical course of GTS, leading to lower self-esteem and poor life quality of the affected individuals [52, 53]. Therefore, it is critical to understand molecular mechanisms that portrait genetic background specific to individual neuropsychiatric illnesses in order to develop effective treatment remedies suitable for each clinical situation, i.e., GTS with or without comorbid disorders. Then, what makes up genetic similarities or differences in these cases? One of the current standing hypotheses concerns the extent to which each variation or chromosomal rearrangement detected in GTS comorbid with other disorders overlaps the genes of interest. In addition, the shared genetic abnormalities among GTS and other neuropsychiatric disorders can differ for minor genetic variants or risk factors elsewhere in the genome. Alternatively, the identical genes may be implicated in the pathophysiology of multiple illnesses including GTS, with different functional alterations introduced to the genes of interest due to the individual sites of mutations. Future studies will be needed to properly examine the validity of these diverse, but not exclusive hypotheses applicable to GTS and comorbid disorders.
Ⅲ. TOURETTE SYNDROME-ASSOCIATED IMMUNE DISORDERS
The therapeutic approaches to treat GTS include pharmacological treatment targeting dopaminergic or noradrenergic pathways, behavioral training, and neurosurgery (high frequency deep brain stimulation) [18, 54, 55]. In part, these complexities of disease management are reflected by the heterogeneous etiology of GTS. Although it is hard to figure out the relevant mutations responsible for the onset of GTS, GTS has evident pathophysiological genetic background and family history [56, 57]. Aside from the genetic basis of GTS, a mutual relation between immune system and the neuropsychiatric disorders underlying GTS has been suggested by recent studies. Despite the recent promising evidence, the mechanistic details of pathophysiologic contributions of immune dysregulation in GTS are remaining unclear. Nevertheless, the increasing bodies of evidence support the existence of a common molecular pathway for immune systems and behavioral anomalies in GTS patients and therefore it is largely believed that the dysregulation of immune responses is causes and consequences of GTS.
The involvement of abnormal immune mechanism in neuropsychiatric disease is relatively clear in the syndrome known as Pediatric Autoimmune Neuropsychiatric Disorders Associated with Streptococcal Infections (PANDAS) [58]. In case of PANDAS, the onset of obsessive-compulsive disorder (OCD) and tic symptoms are observed in childhood after a streptococcal infection [59]. As one of the post-infectious autoimmune disorders, PANDAS is known to be triggered by infection of Group A Streptococcus (GAS) which accompanied by the production of brain-reactive autoantibodies [58, 59]. With analogy to PANDAS, GAS infection has been extensively suggested as a major immunogenic trigger underlying GTS development. Anti-basal ganglia antibodies (ABGA) are prominently observed in PANDAS and similarly, ABGA are also seen in approximately 25% of GTS patients [60]. The potential linkage between GAS infections and GTS development is substantially proved by animal models. In rats, the direct intracerebral infusion of anti-neuronal antibodies obtained from GTS patients induced behavioral abnormalities analogous to involuntary movements of GTS patients [61, 62]. GTS typically begins at childhood and often improved in adolescence. Importantly, the intracerebral infusion induced abnormal behaviors of rats were only sustained during infusion period and gradually disappeared after the end of infusion [60, 61].
The innate and adaptive immune responses of the brain are regulated by brain resident phagocytes, microglia. Basically, in neurological diseases, abnormal activation of microglia has been understood in terms of brain inflammation and neural damage/degeneration. In addition to its role in brain immune system, novel roles for microglia in the maturation/ selection of synaptic junction during development and in the plasticity of synaptic contacts during adult life have emerged [63]. In keeping with our changed concept of microglia’s role in modulating normal brain function, recent researches have focused on the pathophysiologic contribution of microglial dysfunction in the development of GTS. The direct evidence of abnormal microglial function in GTS patients was reported by recent postmortem analyses of basal ganglia biopsies [64]. Of note, this study showed the increased number of CD45+ microglial cells in basal ganglia of GTS patients. These cells also had elevated expression of inflammatory cytokines, an indicative of neurotoxic activation of microglial cells. Recent imaging techniques for microglial activation in vivo have made a monumental advance and provide another line of evidence suggesting microglial activation and GTS. In 2015, Kumar et al. have shown the increased binding of 11C-[R]-PK11195 (PK) to transporter protein (TSPO) in GTS children using positron emission tomography (PET) technique [65]. Considering that PK is a ligand for TSPO expressed by activated microglial cells, the increased PK binding signifies the microglial activation in GTS. Recently, Frick et al. reported distinct abnormal microglia activation in Hdc-KO mice [66]. In basal condition, Hdc-KO mice have little changes in microglia activation and also have normal expression of inflammatory cytokines. However, in inflammatory challenged condition, the microglia activation was markedly enhanced in Hdc-KO mice compared with that of control mice [66]. These findings strongly suggest the microglia hyperactivity-mediated onset of GTS in the presence of environmental challenge.
In addition to brain inflammation and immune activation, patients with GTS show a hyperactive tendency in systemic immune responses (Table. 1). The baseline levels of pro-inflammatory cytokines including tumor necrosis factor-a (TNF-a) and interleukin-12 (IL-12) were found to be increased in children with GTS compare to age-matched healthy children [67]. In 2012, Cheng et al. also reported the raised plasma levels of IL-6, IL-1, and IL-17 in GTS patients [68]. The changes in numbers of the immune cell population in GTS patients also support the over-activated immune responses in GTS context. GTS patients had reduced numbers of CD4+ CD25+ T-cells (T-reg) compared to healthy children [69]. In general, T-reg cells take part in normal surveillance of selfantigens and therefore prevent potential autoimmune responses. Hence, it is possible that the declined numbers of T-reg cells in case of GTS make GTS patients more susceptible to autoimmune disease.
Ⅳ. ANIMAL MODELING FOR GTS
1. Validation of GTS animal models
In the field of psychiatry, the reliability of the animal model is evaluated in three categories: face, construct and predictive validity [70, 71]. In brief, face validity indicates phenotypic similarity between patients and animal models in terms of clinical symptoms, including behavioral abnormality. Construct validity refers to a pathophysiological correlation with psychiatric disorders such as a neural deficit within the target brain region. Predictive validity of animal models is determined by similarity of responses to certain tasks and medications compared to patients. Construct validity enables us to verify the mechanistic hypothesis and understand the etiology of the disease, while others contribute primarily to establishing the therapeutic strategy to alleviate symptoms [72].
Up to date, various animal models have been proposed as GTS models. The motor tics and/or stereotypic behaviors (e.g., repetitive grooming, circling, excessive biting and jerk-like movement) represent the face validity of GTS models in general [73]. However, the presence of premonitory urges, another common clinical manifestation preceding the tics, remain elusive in animals although assessment of sensory- motor gating system via prepulse inhibition (PPI) task implies that some GTS models also experience these premonitory sensations [74]. In addition, since GTS often accompanies comorbid symptoms of other psychiatric disorders such as ADHD and OCD [75, 76], dissection of GTSassociated behaviors from these comorbid conditions as well as primary motor stereotypies should be proceeded to determine the face validity of GTS animal models.
Construct validity of GTS modeling is mainly dependent on the current understandings of GTS etiology for geneticand neurophysiologic factors. Indeed, analysis of biological specimens, neuroimaging data and post-mortem studies have revealed that neural network is impaired in the cortico-basal ganglia circuit with abnormal dopaminergic activity in GTS status [77-79]. Based on these findings, various GTS animal models have been developed through the manipulation of specific neural subpopulations to impede dopamine pathway within the striatum [80, 81]. Therefore, comprehensive investigations of GTS neuropathologic mechanism addressing controversial issues regarding the contribution of other brain regions (e.g., cortex and thalamus) or neurotransmitters (e.g., GABA, serotonin and glutamate) would improve the GTS animal modeling technique with construct validity. Predictive validation of GTS animal models is based on the animal response towards the medications used for GTS treatment, although the absence of specific GTS therapy hinders the model evaluation. In practice, animals with abnormal dopamine pathway are considered to be the best-described GTS model with predictive validity due to their apparent responsiveness towards dopaminergic agents as reported in clinical cases [73, 82].
2. GTS modeling strategy
GTS modeling methods can be classified broadly into three categories: genetic, pharmacological and immunological approaches to the disorder. This review will provide a brief description of each concept and representative animal models (Table. 2).
3. Transgenic GTS models
Recent research has proven that GTS is not a single geneassociated genetic disorder; however, genealogy studies including twin cases have still suggested that some unknown genetic factors would contribute to GTS development[83, 84]. Mutant models for several candidate genes implicated in familial GTS, such as SLITRK1 (SLIT and TRK like family member1), CNTNAP2 (contactin-associated protein 2) and HDC (histidine decarboxylase), have been studied to recapitulate GTS status[8, 46, 85, 86]. While these models represent some behavioral changes due to the neurological deficit, they often fail to reproduce the clinical signs of GTS. For example, SLITRK1 mutant mice show the high level of anxiety instead of tic-like behavior which is the key to assess face validity of GTS animals[85, 87]. These findings emphasize the importance of considering multiple genetic factors in GTS development.
In addition, transgenic animals with disrupted neurotransmitter system can be applied to GTS modeling. Both D1CT-7(D1 receptor cholera toxin) transgenic mice and DAT (dopamine transporter) knockdown mice display a hyper-active dopaminergic pathway with abnormal stereotypic behaviors as observed in GTS patients [75, 88, 89].
4. Pharmacological GTS models
Chemically induced GTS models are based on psychostimulation or inhibition by pharmacological agents, and manipulation of dopamine pathway is a major strategy in these models[82]. Several studies have reported that systemic- and local administration of dopaminergic agonist such as amphetamine induces tic-like behaviors in animals[90-92]. Modification of GABA activity can be applied to GTS modeling, since GABA is one of the major inhibitory neurotransmitters that regulate the dopaminergic system[93]. Indeed, microinjection of GABA antagonist, bicuculline or picrotoxin, into the basal ganglia leads to acute dysfunction of the sensory-motor neural circuit, followed by sudden, repetitive movements in both primate and rodent models[80, 94-96]. It is also reported that serotonergic- or adrenergic agents can induce PPI deficits with some behavioral abnormalities in rodents, although underlying mechanisms have not been explained yet[97-99]. Using pharmacological GTS models, researchers can evaluate the correlation between severity of clinical symptoms and neurotransmitter activity by regulating the injection dose and duration. In addition, these models are useful for identifying involved brain regions in GTS development as they represent specific patterns of behavioral change according to the injection site of pharmacological agents.
5. Immune-mediated GTS models
Based on current reports with post-infectious Tourette-like syndromes, immune dysfunction has been suggested as a contributing environmental factor for GTS[100, 101]. In particular, some of the patients acquire autoimmune disorder accompanied by rapid-onset stereotypic tics after Group A β-hemolytic streptococcal (GABHS) infection, and this phenomenon is part of hypothesis entitled ‘pediatric autoimmune neuropsychiatric disorder associated with a streptococcal infection (PANDAS)’ [102]. The main idea of PANDAS hypothesis is that antibodies for GABHS can eventually target the basal ganglia to impair neuronal connectivity. Since the contribution of autoimmunity to GTS has not been proven yet and the presence of autoantibodies in the brain is still under debate[103, 104], several attempts have been made to mimic immune-mediated GTS status using animal models. Active immunization with immunogens(e.g., GABHS antigens, LPS and poly I:C)[105, 106], passive immunization with serum obtained from patients or GTS animals[62, 107] and infusion of general immune mediators such as IL-2, IL-6 and TNF-β1 [108, 109] are commonly used for model induction. Animal models represent a variety of symptoms ranging from asymptomatic to complex psychiatric behavioral changes, making it difficult to validate the immunologic GTS models.
Ⅴ. CONCLUSION
GTS is not regarded as a degenerative, life-threatening disease in general; however, uncontrolled behavioral and psychological complications can have a devastating impact on the quality of life in affected individuals. GTS management is mainly aimed at suppressing abnormal behaviors with pharmacological approaches accompanied by supportive treatment such as psychoeducation, while surgical intervention can be applied in some severe cases. Up to date, several drugs are available for GTS treatment and most of the medications are neuromodulatory agents which target dopaminergic- and/or serotonergic pathways. However, these non-specific options are hard to estimate the therapeutic efficacy and possible side effects, impeding the treatment with precision the individual need of GTS patients. Therefore, understanding of GTS etiology should precede to establish an effective personalized therapeutic strategy. It is noted that candidate gene studies mainly based on family case analysis have suggested some genetic burden on disease development, although no causative genomic change has been uncovered yet in most of GTS cases. In this aspect, deep width genetic analysis such as GWAS and SNPs/CNV studies should be conducted across world-wide research centers to harbor novel susceptible genes and chromosomal abnormalities. In addition, growing evidence suggests the potential role of immune dysfunction associated with infection and autoimmune reaction in GTS physiopathology. In the term of PANDAS, auto- reactive antibody production after streptococcal infections seems to play a key role in the development of neuropsychiatric symptoms. Hyperactive immune responses of both innate and adaptive system are also observed in GTS-affected patients. Since the immune system is a complicated communication network throughout the body, causal relationship between immune dysfunction and GTS manifestation should be investigated to develop novel immune therapies. Moreover, large-scale epidemiologic studies regarding infection-mediated neuropsychiatric disorders will provide a better understating towards GTS etiology. All of these future approaches, combined with advanced GTS animal modeling technique to confirm the novel findings, can contribute to open new therapeutic paradigms for GTS.