Parkinson's disease (PD), a progressive movement disorder, is the second most common neurodegenerative disorder after Alzheimer's disease. PD is characterized by motor symptoms, including bradykinesia, resting tremor, rigidity, and postural instability, as well as a wide variety of non-motor features, such as olfactory dysfunction, sleep disorders, and dysautonomia (Bloem et al., 2021). As a complex multifactorial disease, PD is caused by a combination of advancing age, genetic, environmental, and lifestyle factors, most of which have not yet been clearly identified (Lim et al., 2019; Bandres-Ciga et al., 2020; Blauwendraat et al., 2020). As another kind of movement disorder, dystonia is characterized by sustained or intermittent muscle contractions causing abnormal, often repetitive, movements, postures, or both. Dystonia can occur in isolation, in combination with other movement disorders, or coexist with a variety of other neurological or systemic manifestations (Albanese et al., 2013). In some conditions, dystonia is more likely a descriptive term rather than a specific diagnosis. The etiology of many forms of dystonia is recognized to be due to a large number of different causes and is still not fully elucidated, but genetic factors are involved unquestionably (Lohmann and Klein, 2013; Balint et al., 2018).

Parkinson's disease and dystonia have overlaps in clinical phenotypes and are both clinically and genetically heterogeneous conditions. Studies estimated that 30% or more of the patients suffering from PD may experience dystonia as a symptom or as a complication of treatment (Tolosa and Compta, 2006; Wickremaratchi et al., 2011). Dystonia is a common early symptom of young-onset PD and sometimes can precede overt parkinsonism, but it can also appear in the middle of advanced stages of PD (Shetty et al., 2019). Conversely, patients suffering from dopa-responsive dystonia (DRD), a specific type of dystonia characterized by lower limb dystonia in childhood with an excellent response to low doses of levodopa, frequently present with parkinsonism. Some DRD patients can even present prominent parkinsonism, leading to clinical difficulty to differentiate between DRD and young-onset PD (Wijemanne and Jankovic, 2015). Especially, dystonia is a prominent symptom for PD patients caused by several genes including PRKN, PINK1, DJ-1, and PLA2G6 (Kasten et al., 2018; Niemann and Jankovic, 2019). Analogously, parkinsonism is commonly seen in subtypes of dystonia caused by genes such as GCH1, TH, TAF1, ATP1A3, and PRKRA (Phukan et al., 2011; Diez-Fairen et al., 2021).

These observed phenomena indicate a potential correlation between PD and dystonia and suggest the coexistence might be due to overlaps in the genetic background. Therefore, a possible role for dystonia-related genes (abbreviated as the “DYT genes” in this article) in PD is needed to be explored. Our previous research, consistent with other studies from different cohorts, had proven the role of GCH1, the most common cause of DRD, in the pathogenesis of PD (Mencacci et al., 2014; Guella et al., 2015; Chen et al., 2016; Pan et al., 2020). However, genetic evidence for the involvement of other DYT genes in PD has not been fully studied. To clarify the correlation between DYT genes and PD, we designed to comprehensively analyze the rare variants of the DYT genes in a large Chinese cohort of patients with PD and healthy controls by mining the whole-exome sequencing (WES) and whole-genome sequencing (WGS) data.

Within two decades, the role of genetics in PD has been emphasized and better understood, but many gaps remain to be filled. We observed similarities in many aspects of PD and dystonia, which sparked our curiosity about the genetic overlaps in these two movement disorders. Consequently, we have comprehensively analyzed the rare variants in coding regions of DYT genes in a large Chinese cohort of PD patients and healthy controls by employing unbiased whole-exome and whole-genome approaches.

To our knowledge, this is the first and largest-ever study to explore the association between rare variants of DYT genes and PD. Firstly, we identified five patients with PD who carried potentially pathogenic biallelic variants in recessive dystonia-related genes, including COL6A3 and TH. Secondly, we detected a total of 180 deleterious variants in dominant dystonia-related genes, four of which were considered potentially pathogenic variants, including p.W591X and p.G820S in ANO3, p.R678H in ADCY5, and p.R458Q in SLC2A1. Thirdly, the burden analysis revealed an increased burden of variant subgroups in TH, SQSTM1, THAP1, and ADCY5 in sEOPD, as well as COL6A3 in sLOPD, although statistical significance was not achieved after Bonferroni correction. Overall, our study provides evidence that several DYT genes contribute to the pathogenesis of PD, supported by a large sample size and robust methodologies.

The COL6A3 gene encodes the alpha-3 chain of type VI collagen, which is an extracellular matrix (ECM) protein, playing a pivotal role in the central nervous system (CNS) and participating in various aspects of neurodevelopment and neurodegeneration by regulating autophagy and synaptic plasticity (Dityatev et al., 2010; Neill et al., 2021). Although the physiological role of Col6a3 in the CNS remains poorly understood, emerging data suggest its neuroprotective potential (Cheng et al., 2011; Cescon et al., 2016). Col6a3 was widely expressed throughout the adult mouse brain with the highest mRNA levels in the brainstem and midbrain (Zech et al., 2015). Biallelic loss-of-function mutations of COL6A3 can cause early-onset segmental isolated dystonia (DYT27) by affecting the EMC in the CNS (Zech et al., 2015). The genetic study of COL6A3 in PD or other neurodegenerative diseases was insufficient. A previous study conducted in China reported a PD patient without dystonia symptoms carrying compound heterozygous variants in COL6A3 gene. The study also observed an elevated aggregate variant burden of the COL6A3 gene in 173 PD patients compared to 200 controls, thereby proposing a potential role of COL6A3 in PD. However, it is important to note that this evidence is still preliminary (Jin et al., 2021). Here, we reported four patients with PD harboring potentially pathogenic biallelic variants in COL6A3, and all the patients showed typical manifestations of PD without dystonia or muscular dystrophy. None of the variants were located in previously reported hotspots for mutation in dystonia (exons 41 and 42), indicating variants in specific regions of COL6A3 may have an association with PD (Domingo et al., 2016; Jin et al., 2021). We also observed an association between rare damaging variants of the COL6A3 gene and an increased risk for sLOPD, which replicated the conclusion of the previous study with a larger sample size. As a result, the role of COL6A3 in PD has been highlighted, although further evidence and mechanism explorations are needed.

TH is a well-established causative gene for DRD similar to GCH1, and mutations of TH and GCH1, resulting in deficiencies of the enzymes involved in the dopaminergic synthesis pathway, can present with dystonia and parkinsonism (Lee and Jeon, 2014; Wijemanne and Jankovic, 2015). The association between GCH1 and PD has been proven previously, which was validated in the Chinese population (Pan et al., 2020). By contrast, the relationship between the TH gene and PD is not so clear, due to the limited number of research studies and inconsistencies across the studies (Bademci et al., 2012; Chen et al., 2020; Kawahata and Fukunaga, 2020). Previous studies were mainly focused on common variants of the TH gene and utilized a small sample size (Sutherland et al., 2008; Punia et al., 2010). With the rapid advances in genomic technology, subsequent studies performed genetic research between rare variants of TH and PD; however, the studies conducted are trivial due to their extremely rare presence with a small sample size (Hertz et al., 2006; Rengmark et al., 2016; Yan et al., 2018). In our cohort, we found a potentially pathogenic homozygous variant of TH in a patient with PD without symptoms of DRD, and we also provide suggestive evidence that loss-of-function variants in the TH gene might contribute to an increased risk for sEOPD, which help us fully understand the genetic risk for PD conferred by the TH gene.

ANO3 encodes anoctamin 3, which is highly expressed in the brain tissue, specifically in the striatum (Charlesworth et al., 2012). It is a transmembrane protein that belongs to a family of calcium-activated chloride channels and thus may play a role in signal transduction (Pedemonte and Galietta, 2014). Additionally, a previous study has reported an infantile-onset patient manifested dystonia, parkinsonism, and developmental regression, caused by a de novo missense variant in ANO3 (Nelin et al., 2018). However, the exact role of anoctamin 3 remains uncertain, and the two variants found in our patients have not been fully researched in our study and have no other evidence to support the effects yet.

ADCY5 encodes isoform 5 of adenylyl cyclase, which is highly expressed in medium spiny neurons (MSNs) of the striatum. Adenylyl cyclase is responsible for the conversion of adenosine triphosphate (ATP) to cyclic adenosine-3′,5′-monophosphate (cAMP) and is involved in dopaminergic signaling. Dysfunction of the cAMP pathway can contribute to postsynaptic movement disorders such as dystonia, chorea, and parkinsonism (Abela and Kurian, 2018; Ferrini et al., 2021). A novel variant of ADCY5 evaluated to be potentially pathogenic was identified in our cohort, and the burden of rare variants in ADCY5 was increased in sEOPD, which suggested the involvement of ADCY5 in the pathogenesis of PD, whereas research of the underlying mechanisms is further needed.

Additionally, we identified a variant in SLC2A1 (p.R458Q) in one PD patient, which was evaluated to be potentially pathogenic. The variant has been recorded as a candidate causal variant of an individual with ataxic gait, global developmental delay, and disrupted sleep, which was included in the ClinVar database. Additionally, a missense variant occurring in the same codon (p.R458W) has been reported in individuals with features of SLC2A1-related disorder and has been demonstrated to result in a marked reduction in glucose transport, supporting the functional importance of this position in the protein (Arsov et al., 2012; Tzadok et al., 2014). Therefore, further studies are needed to confirm whether p.R458Q plays a role in PD onset or modification.

The burden analysis also showed the enrichment of extremely rare variants of the SQSTM1 gene in sEOPD. The SQSTM1 gene encodes p62, a prototype autophagy receptor, which is commonly found in protein aggregates associated with major neurodegenerative diseases (Deng et al., 2020). It was confirmed to play an essential role in the PINK1/Parkin-mediated mitophagy involved in the pathogenesis of PD (Geisler et al., 2010; Chu, 2019). Mutations in SQSTM1 contribute to neurodegeneration in amyotrophic lateral sclerosis and frontotemporal dementia (Goode et al., 2016). We have completed an initial exploration and obtained a suggestive but encouraging result about the genetic association between SQSTM1 and PD, which deserves further investigation. Additionally, THAP1 was also nominated in the burden analysis, contributing to increased risk for sEOPD. THAP1 encodes a transcription factor involved in the regulation of gene expression in the nervous system. The reported dysregulated genes by loss of Thap1 are functionally related to neurodevelopment, lysosomal lipid metabolism, myelin, cytoskeleton, synaptic transmission, and gliogenesis (Frederick et al., 2019; Domingo et al., 2021). Our results suggested the involvement of these genes in PD, but much remains to be understood about the underlying mechanisms.

As an exploratory study, limitations are inevitable. First, the DYT genes included in the study had been widely collected by researchers at the preliminary stage of the study, inevitably omitting some latest discovered genes and at the same time, a few controversial genes were incorporated, such as CACNA1B and CIZ1. The correlation between the two genes and DYT23 was pending confirmation, therefore, the pathogenicity of the variants located in the questioned genes was taken with caution. Besides, we initially detected a large number of variants under AD inheritance, whereas most of them were considered to be uncertain significance due to many aspects such as the reliability and the mechanisms of genes. The conservative interpretation of the variants led to the neglect of certain variants that merit further investigation. Second, we only focused on rare coding SNVs/Indels in this study and did not consider more complex forms of variants, such as CNVs, as well as non-coding variants. The main reason for this decision was the undisclosed role of the DYT genes in PD, which we aimed to explore in the current study. We prioritized SNVs/Indels in coding regions as they were more likely to reveal possible genetic associations. While other types of variants are also important, their complex and potentially indirect mechanisms are currently challenging to elucidate. We will give more attention to the analysis and interpretation of non-coding and complex variants in future studies. Third, the pathogenicity of the discovered genes and variants still requires further validation through family segregation analysis and functional experiments, which have not been comprehensively investigated in our study. Family segregation analysis is a powerful tool for further classification of the identified candidate variants. In our study, we attempted to conduct family segregation analysis using Sanger sequencing in affected or unaffected parents and siblings of the proband whenever feasible. However, we encountered a limitation that most patients had no family history of PD or other neurodegenerative diseases, and unfortunately, DNA samples of their family members were not available. In addition, although it is the largest genetic study of the DYT genes in PD up to now, the sample size was insufficient for rare variants analysis, which may be the cause of little statistical significance detected.

In conclusion, our findings indicated that rare variants in several DYT genes are associated with PD, and the role of COL6A3 and TH genes in PD is highlighted, although further evidence and mechanism explorations are still needed.

The data analyzed in this study is subject to the following licenses/restrictions: The data that support the findings of this study are available from the corresponding author upon reasonable request. Requests to access these datasets should be directed to liuzhenhua@csu.edu.cn.

The studies involving human participants were reviewed and approved by Medical Ethics Committee of Xiangya Hospital, Central South University. The patients/participants provided their written informed consent to participate in this study.

YW: conceptualization, data curation, formal analysis, methodology, and writing—original draft. YZ: data curation and writing—review and editing. HP and QZ: data curation and methodology. XZ, YX, and ZZ: data curation. QX, QS, JT, and XY: data curation and funding acquisition. JL, JG, and BT: funding acquisition and writing—review and editing. ZL and QY: project administration, supervision, funding acquisition, and writing—review and editing. All authors contributed to the article and approved the submitted version.