Search tips
Search criteria 


Logo of neurologyNeurologyAmerican Academy of Neurology
Neurology. 2010 August 10; 75(6): 508–512.
PMCID: PMC2918477

Analysis of GWAS-linked loci in Parkinson disease reaffirms PARK16 as a susceptibility locus

E.-K. Tan, MD,* H.-K. Kwok, BSc,* L.C. Tan, MD, W.-T. Zhao, MSc, K.M. Prakash, MD, W.-L. Au, MD, R. Pavanni, MD, Y.-Y. Ng, BSc, W. Satake, MD, PhD, Y. Zhao, MD, PhD, T. Toda, MD, PhD, and J.-J. Liu, PhD



A genome-wide association study (GWAS) in the Japanese population identified 2 new Parkinson disease (PD) susceptibility loci on 1q32 (PARK16) (OMIM 613164) and BST1. We analyzed single nucleotide polymorphism (SNPs) located at the GWAS-linked loci (PARK16, PARK8, PARK1, and BST1) in a Chinese population and also conducted a meta-analysis in Asians by pooling 2 independent replication studies from Japan.


We conducted an analysis of 13 SNPs associated with PD GWAS-linked loci in 2 case-control cohorts comprised of 1,349 ethnic Chinese subjects.


PARK16, PARK8, and PARK1 loci but not BST1 were found to be associated with PD. PARK16 SNPs were associated with a decreased risk while PARK1 and PARK8 SNPs were associated with an increased risk of PD. A pooled analysis of our Chinese cohorts and 2 Japanese replication cohorts involving 1,366 subjects with PD and 16,669 controls revealed robust association with these 3 loci and also BST1. There was a trend toward a stronger protective effect of SNPs at the PARK16 locus in sporadic PD compared to familial cases and in older compared to younger subjects.


Our study reaffirms the role of GWAS-linked loci in PD in Asian subjects and the strength of association is similar between Chinese and Japanese subjects. Efforts to elucidate the associated gene within PARK16 locus are warranted.


= genome-wide association study;
= Parkinson disease;
= single nucleotide polymorphism.

Parkinson disease (PD) (OMIM168600), a neurodegenerative disorder, is characterized by loss of dopaminergic neurons in the pars compacta of the substantia nigra. In recent years, several causative genes have been associated with PD for both familial and sporadic forms of the disease.1 However, these mutations probably account for a small percentage of PD cases in most populations. Therefore the search for genetic susceptibility risk factors in the vast majority of PD continues to be of scientific interest. Specific to PD, genetic variants involving pathogenic genes (LRRK2 [PARK8], OMIM 607060, SYN [PARK1], OMIM 168601) and specific candidate genes have been shown to associate with the disease.2–9 To date, there have been a few PD genome-wide association studies (GWAS) in the Caucasian population.2,10,11 However, their findings have not been consistently replicated.12,13 Among the many reasons, sample size and population stratification are some probable limitations. Recently, a GWAS study identified 2 new susceptibility loci on 1q32 (PARK16) (OMIM 613164) and BST1 (bone marrow stromal cell antigen 1) (OMIM 004334) and also associations with known pathogenic genes involved in autosomal dominant forms of parkinsonism (PARK1 on 4q22 and PARK8 on 12q12) in the Japanese population.14 PARK1 have also been implicated as a genetic risk factor in another GWAS study using samples from subjects of European ancestry.

Moreover, the disease associations at PARK16 and PARK8 were replicated in a Caucasian replication study, using individuals of European ancestry.15 Since Japanese and Chinese are of close Asian ancestry, we conducted a replication study of the GWAS-linked loci (PARK16, PARK8, PARK1, and BST1) in a Chinese population and also conducted a meta-analysis in Asian subjects by pooling 2 independent replication studies from Japan.


Ethnic Han Chinese subjects diagnosed with idiopathic PD by movement disorders neurologists at 2 different centers in Singapore (Singapore General Hospital and National Neuroscience Institute) were included. The diagnosis of PD was based on the UK Parkinson's Disease Society Brain Bank clinical diagnostic criteria.16 Sporadic PD was defined as PD without a family history of disease. Controls of similar race, gender, and age from the same region as the patients with PD were also included. For controls, the age was matched ±5 years to the age at onset of PD cases.

Patient consent.

The study received approval from each institutional ethics committee and all the study subjects gave written informed consent for their DNA to be used for genetic research. PD samples that had previously screened positive for pathogenic mutations in [alpha]-synuclein, Parkin, DJ-1, LRRK2, and PINK1 were not included.

Selection of SNPs.

Of the 20 SNPs from 4 genes that were reported in the GWAS study,14 some SNPs are closely correlated (r2 >0.8). Therefore, we only selected 13 noncorrelated SNPs (r2 <0.8) (figure) for analysis in this study.

figure znl0301079160001
Figure Selection of genome-wide association study single nucleotide polymorphisms (SNPs)

Genetic analysis.

Genotyping was carried out with MALDI-TOF mass spectrometry using the Sequenom MassARRAY™ system (San Diego, CA). Multiplex genotyping assays were designed using the Sequenom DESIGNER software (San Diego, CA). PCR (5 ng of genomic DNA) and primer extension reactions were carried out initially according to the Sequenom genotyping assay iPLEX™ protocol. Confirmation of the variants with sequence analysis was carried out for random samples in ABI 3730 automated DNA sequencer (Applied Biosystems).

Statistical analysis.

Statistical analyses were performed using R 2.10.1 software. χ2 and Student t tests were used for comparing the categorical and continuous variables. We assessed each variant for departure from Hardy-Weinberg equilibrium. The results from pooled datasets were analyzed together and individual and pooled odds ratios and associated 95% confidence intervals were tabulated. We estimated the per-allele odds ratios using logistic regression and used Wald test to test whether the coefficients are significant. To test for association of allele frequency at each SNP with PD, we used χ2 test with 1 degree of freedom. A multivariate logistic regression analysis adjusted for age and gender was performed. Stratified analysis by family history and age at onset was also carried out. Meta-analysis was performed to combine our Chinese study with 2 Japanese studies. Heterogeneity among sample sets was accessed using Woolf test. The meta-analysis was conducted using the Mantel-Haenszel method. As this is a replication study involving 4 independent loci, we made a modest correction for multiple comparisons with statistical significance defined at p < 0.01.



We studied a total of 1,349 ethnic Chinese subjects comprised of 2 case-control cohorts including a total of 433 patients with PD and 916 controls from 2 independent centers in Singapore. None of the patients with PD were from consanguineous families and about 3% reported a positive family history. The demographics of the study subjects are summarized in table 1.

Table thumbnail
Table 1 Demographics of subjects with Parkinson disease and controls (total n = 1,349)

Genotyping data.

Thirteen SNPs located within the PARK16, PARK8, PARK1, and BST1 loci, which were associated with PD in the GWAS study,14 were analyzed. The 2 case-control cohorts were analyzed together to improve the power of analysis. The frequency of all the SNPs in the studied sample followed Hardy-Weinberg equilibrium, with the exception of rs4698412 (BST1 locus), which showed a slight deviation (p = 0.022).

Ten SNPs belonging to PARK16, PARK8, and PARK1 loci were found to be associated with PD (table 2). PARK16 SNPs were associated with a decreased risk while PARK1 and PARK8 SNPs were associated with an increased risk of PD. A multivariate logistic regression analysis with disease/control group as the outcome measure and adjusting for age and gender revealed significant association with PARK16, PARK8, and PARK1 but not the BST1 locus (table 3). Stratification by family history revealed a trend toward a stronger protective effect of SNPs at the PARK16 locus in sporadic PD compared to familial cases and in the older compared to younger subjects (table e-1 on the Neurology® Web site at At the PARK8 locus, there was a trend toward a higher risk in familial PD compared to sporadic PD and in the older age group compared to the younger ones (stratified at age at onset <50 years or <55 years) (table e-2, A and B).

Table thumbnail
Table 2 SNP frequency and summary of analysis in subjects with Parkinson disease and controls
Table thumbnail
Table 3 Multivariate analysis adjusted for age and gender

Pooled analysis of Asian subjects.

There was no heterogeneity among the Chinese and Japanese datasets except rs11931532 (BST1) (table e-3). As the frequency of the studied SNPs was similar in the Japanese and Chinese control populations, a combined analysis of previously published 2 case-control replication cohorts in Japanese14 and our cohorts was carried out (1,366 subjects with PD and 16,669 controls). Robust association with PARK16, PARK8, and PARK1 was observed in addition to BST1 (table e-4).


A PD GWAS study in an American population identified 11 SNPs using a family-based design in tier 1 and a case-control design in tier 2.2 A consortium from 14 centers which pooled 5,526 patients with PD and 6,682 controls was unable to replicate any significant association with the PD-associated SNPs.12

More recently, in a PD GWAS study conducted in Asia, investigators14 identified 2 new susceptibility loci (PARK16, BST1) and also strong associations at PARK1 and PARK8, 2 known loci implicated in autosomal dominant forms of parkinsonism. In the same study, the findings were replicated in 2 Japanese cohorts. The signal at the PARK16 locus was less robust in the GWAS study in Caucasians15 and did not surpass correction for multiple testing. However, the investigators subsequently conducted an analysis in their replication sample and found an association of SNP (rs823128) at the PARK16 locus. The minor allele frequency (3%–4%) of the implicated PARK16 SNP in Caucasians is low and this probably accounts for the relatively weaker association.

We conducted a replication study in the Chinese population. In the combined analysis of our 2 cohorts, we were able to independently demonstrate an association of PARK16, PARK8, and PARK1 loci, except for BST1. While not all the SNPs located in PARK 16 reached significant association, the trend and the effect size difference (odds ratio 0.8 to 0.9) between cases and controls was largely similar between our Chinese subjects and the Japanese subjects in the discovery GWAS sample and 2 replication Japanese cohorts. The frequency of the 13 SNPs (10%–50%) was also similar in these 2 Asian races. We observed a more robust association of the SNPs at the PARK8 and PARK1 loci. As there was no evidence of genetic heterogeneity between Chinese and Japanese at these GWAS-linked loci, we conducted a pooled evaluation of our cohorts and the 2 replication Japanese cohorts to increase the power of analysis. The pooled analysis further reaffirms the results across each independent dataset and the discovery dataset. As we did not observe any positive trend with the BST1 locus in our cohort, replication in an independent Chinese cohort is advised. Though the median age of the controls in our study was 4 years younger compared to median age at onset of 60 years in subjects with PD, multivariate analysis after taking into account the effects of age and gender revealed significant associations similar to those observed in the univariate analysis. Furthermore, comparison with a selected set of 428 controls with median age at onset of 60 years showed similar results (data not shown).

Consistent and independent replication of genetic association studies remains the litmus test of the validity of the findings. While identification of susceptibility alleles is an important step, unraveling the biologic basis of their actions, if any, remains a challenging task as many of these variants are not located in the coding regions. Among the identified loci, PARK16 is interesting as it contains a few candidate genes (SLC41A1, RAB7L1, NUCKS1). rs947211 is associated with transcript level of NUCKS1,14 though Simón-Sánchez et al.15 did not find any association of the SNPs and expression levels of the genes at PARK16, PARK8, and PARK1. The association of a lower risk with PD suggests that protective gene variants or genes are located within this region or they need interaction with unknown genes to exert their effect. It is possible that these SNPs are not the actual causative variants based on their biologic plausibility and the varied strength of their association. The original GWAS study14 did not analyze the strength of association between familial and sporadic PD and between older and younger subjects. Despite the limitation of a smaller sample size in the subset analysis, our data revealed a consistent trend toward a stronger protective effect of SNPs at the PARK16 locus in sporadic PD compared to familial cases and in the older compared to younger subjects (table e-1). The significance of this is unclear, though it suggests that common variants at PARK16 locus may be more relevant to a general group of patients with PD. There was also a consistent trend toward a higher risk in familial PD compared to sporadic PD and in the older age group compared to the younger ones at the PARK8 locus (table e-2). As PARK8 is responsible for autosomal dominant parkinsonism and is associated with late-onset disease, this observation may not be unexpected. Since these SNPs are located from intron 2 of SLC2A13 to the upstream region of LRRK2, our finding strengthens the suggestion that these are likely to be risk variants of the LRRK2 gene.

While current evidence suggests that common genetic variants play a role in the etiology of typical PD, GWAS studies by their inherent design may not be able to detect rare variants.9,17 It is also possible that cases selected for GWAS studies may not be particularly enriched with genetic susceptibility alleles, and other compounding factors like reduced penetrance of PD genes and gene-environmental interaction were unaccounted for. Thus multiple approaches including linkage analysis, sequencing, and sibpair analysis would be needed to uncover additional variants/causative genes and susceptibility loci.


The authors thank Singapore General Hospital, National Neuroscience Institute, Duke-NUS Graduate Medical School, Singapore Millennium Foundation, National Medical Research Council, and Dr. Yuen Yih for support. The authors also thank Mitsutoshi Yamamoto, Nobutaka Hattori, and Miho Murata for the data published in reference 14.


Dr. E.-K. Tan has received funding or speaker honoraria from Boehringer Ingelheim and Novartis; serves as editor of Annals, Academy of Medicine, Singapore, and associate editor of European Journal of Neurology and Parkinsonism Related Disorders; and receives research support from Eisai Inc., Schering-Plough Corp., Allergan Inc., the National Medical Research Council, Singapore, and the Singapore Millennium Foundation. H.-K. Kwok reports no disclosures. Dr. L. Tan has received funding for travel or speaker honoraria from Boehringer Ingelheim, Novartis, GlaxoSmith Kline, and Medtronic, Inc.; serves on the editorial board of Movement Disorders; and receives research support from Eisai Inc., Schering-Plough Corp., Allergan Inc., the National Medical Research Council, Singapore, and the Singapore Millennium Foundation. W.-T. Zhao, Dr. Kumar, Dr. Au, Dr. Pavanni, Y.-Y. Ng, Dr. Satake, and Dr. Zhao report no disclosures. Dr. Tatsushi serves as an Associate Editor for the Journal of Human Genetics and receives research support from the Japan Science and Technology Agency, the Ministry of Education, Culture, Sports, Science and Technology of Japan, and the Ministry of Health, Labor and Welfare of Japan. Dr. Liu serves as an Associate Editor for BMC Genetics.

Supplementary Material

[Data Supplement]


Address correspondence and reprint requests to Dr. Eng-King Tan, Department of Neurology, Singapore General Hospital, Singapore 169108 gs.moc.hgs@ketrng; or Dr. Jian-Jun Liu, Human Genetics, Genome Institute of Singapore, A*STAR, 60 Biopolis Street, 138672, Singapore gs.ude.rats-a.sig@3juil.

Supplemental data at

*These authors contributed equally to this work.

Disclosure: Author disclosures are provided at the end of the article.

Received February 10, 2010. Accepted in final form April 26, 2010.


1. Tan EK, Skipper LM. Pathogenic mutations in Parkinson disease. Hum Mutat 2007;28:641–653. [PubMed]
2. Maraganore DM, de Andrade M, Lesnick TG, et al. High-resolution whole-genome association study of Parkinson disease. Am J Hum Genet 2005;77:685–693. [PubMed]
3. Maraganore DM, de Andrade M, Elbaz A, et al. Collaborative analysis of alpha-synuclein gene promoter variability and Parkinson disease. JAMA 2006;296:661–670. [PubMed]
4. Tan EK, Zhao Y, Skipper L, et al. The LRRK2 Gly2385Arg variant is associated with Parkinson's disease: genetic and functional evidence. Hum Genet 2007;120:857–863. [PubMed]
5. Ross OA, Wu YR, Lee MC, et al. Analysis of Lrrk2 R1628P as a risk factor for Parkinson's disease. Ann Neurol 2008;64:88–92. [PubMed]
6. Skipper L, Li Y, Bonnard C. Comprehensive evaluation of common genetic variation within LRRK2 reveals evidence for association with sporadic Parkinson's disease. Hum Mol Genet 2005;14:3549–3556. [PubMed]
7. Zabetian CP, Hutter CM, Factor SA, et al. Association analysis of MAPT H1 haplotype and subhaplotypes in Parkinson's disease. Ann Neurol 2007;62:137–144. [PMC free article] [PubMed]
8. Sidransky E, Nalls MA, Aasly JO, et al. Multicenter analysis of glucocerebrosidase mutations in Parkinson's disease. N Engl J Med 2009;361:1651–1661. [PMC free article] [PubMed]
9. Pankratz N, Wilk JB, Latourelle JC, et al. Genomewide association study for susceptibility genes contributing to familial Parkinson disease. Hum Genet 2009;124:593–605. [PMC free article] [PubMed]
10. Fung HC, Scholz S, Matarin M, et al. Genomewide genotyping in Parkinson's disease and neurologically normal controls: first stage analysis and public release of data. Lancet Neurol 2006;5:911–916. [PubMed]
11. Tan EK. The role of common genetic risk variants in Parkinson disease. Clin Genet 2007;72:387–393. [PubMed]
12. Elbaz A, Nelson LM, Payami H, et al. Lack of replication of thirteen single-nucleotide polymorphisms implicated in Parkinson's disease: a large-scale international study. Lancet Neurol 2006;5:917–923. [PubMed]
13. Evangelou E, Maraganore DM, Annesi G, et al. Non-replication of association for six polymorphisms from meta-analysis of genome-wide association studies of Parkinson's disease: large-scale collaborative study. Am J Med Genet B Neuropsychiatr Genet 2010;153B:220–228. [PubMed]
14. Satake W, Nakabayashi Y, Mizuta I, et al. Genome-wide association study identifies common variants at four loci as genetic risk factors for Parkinson's disease. Nat Genet 2009;41:1303–1307. [PubMed]
15. Simón-Sánchez J, Schulte C, Bras JM, et al. Genome-wide association study reveals genetic risk underlying Parkinson's disease. Nat Genet 2009;41:1308–1312. [PMC free article] [PubMed]
16. Hughes AJ, Daniel SE, Kilford L, Lees AJ, Hughes AJ. Accuracy of clinical diagnosis of idiopathic Parkinson's disease: a clinico-pathological study of 100 cases. J Neurol Neurosurg Psychiatry 1992;55:181–184. [PMC free article] [PubMed]
17. Bodmer W, Bonilla C. Common and rare variants in multifactorial susceptibility to common diseases. Nat Genet 2008;40:695–701. [PMC free article] [PubMed]

Articles from Neurology are provided here courtesy of American Academy of Neurology