Abstract
Parkinson’s disease (PD) is a multifactorial movement disorder characterized by progressive neurodegeneration. Genome-wide association studies (GWAS) have nominated over fifteen distinct loci associated with risk of PD, however the biological mechanisms by which these loci influence disease risk are mostly unknown. GWAS are only the first step in the identification of disease genes: the specific causal variants responsible for the risk within the associated loci and the interactions between them must be identified to fully comprehend their impact on the development of PD. In the present study, we first attempted to replicate the association signals of 17 PD GWAS loci in our series of 1381 patients with PD and 1328 controls. BST1, SNCA, HLA-DRA, CCDC62/HIP1R and MAPT all showed a significant association with PD under different models of inheritance and LRRK2 showed a suggestive association. We then examined the role of coding LRRK2 variants in the GWAS association signal for that gene. The previously identified LRRK2 risk mutant p.M1646T and protective haplotype p.N551K-R1398H-K1423K did not explain the association signal of LRRK2 in our series. Finally, we investigated the gene-gene interaction between PARK16 and LRRK2 that has previously been proposed. We observed no interaction between PARK16 and LRRK2 GWAS variants, but did observe a non-significant trend toward interaction between PARK16 and LRRK2 variants within the protective haplotype. Identification of causal variants and the interactions between them is the crucial next step in making biological sense of the massive amount of data generated by GWAS studies. Future studies combining larger sample sizes will undoubtedly shed light on the complex molecular interplay leading to the development of PD.
Keywords: Association studies in genetics, Parkinson’s disease/Parkinsonism
Introduction
Genome-wide association studies (GWAS) heralded a new era in the resolution of common genetic risk factors in disease. Parkinson’s disease (PD) was long considered the archetypal age-related sporadic disorder with a minimal genetic component. Early GWAS in PD highlighted a known risk locus, SNCA, with other candidates failing to replicate [1-3]. However, as the field has progressed with improved analysis platforms and larger patient-control series, a number of novel candidate loci have been proposed including two other familial parkinsonism genes, MAPT and LRRK2 [4-7].
In a recent study by the International Parkinson Disease Genomics Consortium (IPDGC), six previously reported and five novel loci were nominated for association with PD via a meta-analysis of five datasets from published GWAS in US and European populations [8]. Subsequent meta-analytical studies have nominated a number of additional loci [9-12]. For example Lill and colleagues used data collected via the PDgene.org website, while Do et al. performed a GWAS study using information collected through the 23andMe personal genomics company [10,11]. Further investigation and replication of the original IPDGC dataset and combined analysis with Do et al. nominated a further five loci [9].
Nomination of genomic regions of association (confirmed by independent replication) is the first step in disease-related gene identification. However, the regions of association are generally large and it is critical that the causal variant(s) responsible for the association signal is identified to provide diagnostic biomarkers, mechanistic insights and rational drug targets. Our recent investigation of the LRRK2 locus highlighted one coding variant conferring risk to PD (p.M1646T) in Caucasians as well as a protective haplotype p.N551K-R1398H-K1423K [13]. In addition, it is important to not only identify risk loci but also to understand the joint effects and interaction of the individual associations [14]. Therefore herein, we attempt to replicate single variant association from GWAS loci, we investigate whether the previously nominated functional LRRK2 coding variants account for a proportion of the LRRK2 GWAS signal, and finally we examine if LRRK2 variation interacts with the associated PARK16 SNP as recently reported [15].
Methods
Study subjects
A total of 1,381 patients with PD and 1,328 controls from a US series (674 patients, 724 controls), an Irish series (362 patients, 370 controls), and a Polish series (345 patients, 234 controls) were included in this case-control study. Characteristics of patients with PD and controls are summarized in Table 1 for each series. Patients were diagnosed with PD using standard criteria. Controls were individuals free of PD or a related movement disorder at the time of examination. All subjects were unrelated within and between diagnosis groups. The Mayo Clinic Institutional Review Board approved the study, each individual site received local IRB approval, and all subjects provided informed consent.
Table 1.
Patient characteristics
| Variable | Patients with PD | Controls |
|---|---|---|
| Irish series | N=362 | N=370 |
| Age | 58±12 (32 - 87) | 66±22 (17 - 97) |
| Gender | ||
| Male | 203 (56%) | 134 (36%) |
| Female | 159 (44%) | 236 (64%) |
| Age at PD onset | 59±11 (18 - 87) | N/A |
| US series | N=674 | N=724 |
| Age | 69±11 (33 - 97) | 65±13 (18 - 88) |
| Gender | ||
| Male | 427 (63%) | 300 (41%) |
| Female | 247 (37%) | 424 (59%) |
| Age at PD onset | 64±12 (28 - 94) | N/A |
| Polish series | N=345 | N=234 |
| Age | 65±11 (29 - 88) | 56±15 (19 - 96) |
| Gender | ||
| Male | 216 (63%) | 126 (54%) |
| Female | 129 (37%) | 108 (46%) |
| Age at PD onset | 58±11 (25 - 81) | NA |
| Combined series | N=1,381 | N=1,328 |
| Age | 65±12 (29 - 97) | 64±17 (17 - 97) |
| Gender | ||
| Male | 846 (61%) | 560 (42%) |
| Female | 535 (39%) | 768 (58%) |
| Age at PD onset | 61±12 (18 - 94) | NA |
The sample mean ± SD (minimum - maximum) is given for age and age at PD onset. Information was unavailable regarding age at PD onset for 94 patients with PD in the Irish series and for 25 patients with PD in the Polish series.
Variant and genotype information
Seventeen variants that have been previously nominated for association with PD in GWAS were genotyped in this study [6,8-12]; a summary of these variants is provided in Table 2. Genomic DNA extracted from peripheral blood lymphocytes using the Autogen FlexStar (Holliston, MA.) was used for genotyping. GWAS loci genotyping was performed using the Sequenom iPlex platform and data acquisition was obtained using Typer 4.0 software (Sequenom, San Diego, CA). LRRK2 coding variants were genotyped using a combination of the Sequenom iPlex platform and bidirectional Sanger sequencing approaches; primer sequences are available upon request.
Table 2.
Variants included due to previous associations with PD in GWAS
| Variant | Chromosome | Position (bp)a | Candidate Gene | MAF in current study |
|---|---|---|---|---|
| rs2230288 | 1q22 | 155206167 | GBA | 1.7% |
| rs34372695b | 1q22 | 156030037 | SYT11 | 1.8% |
| rs708723 | 1q32 | 205739266 | PARK16 | 32.3% |
| rs10928513b | 2q21 | 135456759 | ACMSD | 44.8% |
| rs2102808b | 2q24 | 169117025 | STK39 | 13.7% |
| rs11711441b | 3q27 | 182821275 | MCCC1/LAMP3 | 12.2% |
| rs6599388b | 4p16 | 939087 | GAK | 30.7% |
| rs11724635b | 4p15 | 15737101 | BST1 | 44.4% |
| rs6812193 | 4q21 | 77198986 | STBD1/SCARB2 | 36.5% |
| rs356219b | 4q22 | 90637601 | SNCA | 40.2% |
| rs3129882b,c | 6p21 | 32409530 | HLA | 42.0% |
| rs156429 | 7p15 | 23306020 | GPNMB | 39.0% |
| rs7077361 | 10p13 | 15561543 | ITGA8 | 12.5% |
| rs1491942b | 12q12 | 40620808 | LRRK2 | 21.3% |
| rs10847864b | 12q24 | 123326598 | CCDC62/HIP1R | 35.1% |
| rs2942168b | 17q21 | 43714850 | MAPT | 19.4% |
| rs12456492 | 18q12 | 40673380 | RIT2 | 33.0% |
Chromosomal positions are based on the February 2009 (GRCH37/hg19) genome assembly.
MAF=minor allele frequency.
Indicates a variant examined in a previous larger study by Sharma et al.16 that included many of the same subjects utilized in the current study.
The association of HLA rs3129882 with PD has been previously reported in essentially the same patient-control group (~97% overlap) utilized in the current study17; these results are reported again in the current study in order to display a more complete replication of GWAS risk factors for PD.
Genotype data for 11 of these variants was included in a previous consortium replication effort of 8,750 patients and 8,955 controls16 that involved many of the same subjects included in the current study; the analysis of these 11 variants in our study differs from this previous study in that dominant and recessive statistical models were utilized in addition to the additive models utilized in the previous study (see Statistical analysis section). One of these 11 variants, HLA rs3129882, has previously been genotyped and assessed for association with PD in essentially the same patient-control group (~97% overlap) utilized in the current study17; these results are reported again in the current study in order to display a more complete replication of GWAS risk factors for PD. The remaining 6 GWAS PD risk factors have not been previously reported in the patients or controls utilized in this study.
LRRK2 variants in the p.N551K-R1398H-K1423K protective haplotype were also genotyped, as was the LRRK2 risk substitution p.M1646T; the majority of the patients and controls utilized in the current study were also included in the aforementioned larger original study nominating these LRRK2 variants for association with PD13. In each of the three individual series, all genotype call rates were >95% and there was no evidence of any departures from Hardy Weinberg Equilibrium in study controls (all P≥0.05 after Bonferroni correction).
Statistical analysis
For each individual series and the combined series, associations of GWAS-nominated variants with PD were evaluated using logistic regression models adjusted for age, gender, and series (combined series only). Odds ratios (ORs) and 95% confidence intervals (CIs) were estimated. We considered each GWAS-nominated variant under an additive model (effect of each additional minor allele), a dominant model (presence vs. absence of the minor allele), and a recessive model (presence vs. absence of two copies of the minor allele). Associations of LRRK2 p.N551K, p.R1398H, p.K1423K, and p.M1646T with PD in the combined series were evaluated using logistic regression models adjusted for age, gender, and series, where each variant was considered under a dominant model owing to the small number of homozygotes of the minor allele. Haplotype analysis was performed using a score test for association [16].
In the combined series, we examined the degree of linkage disequilibrium of LRRK2 rs1491942 with the three variants in the protective LRRK2 p.N551K-R1398H-K1423K haplotype and also with LRRK2 p.M1646T by estimating r2 values in study controls; the association of LRRK2 rs1491942 with PD while adjusting for p.N551K, p.R1398H, p.K1423K, and p.M1646T in logistic regression analysis was also investigated in order to evaluate the independence of associations with PD. Interactions of PARK16 rs708723 with LRRK2 rs1491942 and the three aforementioned LRRK2 variants in the protective haplotype in the combined series were evaluated using logistic regression models adjusted for age, gender, and series. LRRK2 p.M1646T was not evaluated in any interaction analysis owing to its lower frequency.
A relatively large number of statistical tests were performed in evaluation of associations with PD for variants previously nominated via GWAS (48 - 49 tests per series with 17 variants and 3 potential statistical models). In order to adjust for multiple testing and control the family-wise error rate at 5% in this primary analysis, we employed the single-step minP method [17] separately for each series with 10,000 permutations of patient and control labels, after which p-values ≤0.0018 (Irish series), ≤0.0015 (US series), ≤0.0019 (Polish series), ≤0.0016 (combined series) were considered as statistically significant. P-values ≤0.05 were considered as statistically significant in all remaining analysis. All statistical analyses were performed using R Statistical Software (version 2.14.0; R Foundation for Statistical Computing, Vienna, Austria).
Results
Replication of GWAS loci
To confirm the effect of the PD GWAS loci in disease risk, we attempted replication in our PD series. Single variant associations with PD according to each model of inheritance are displayed in Table 3A for the combined series, while association results for individual Irish, US, and Polish series are presented in Table 3B, 3C and 3D. In the combined group of 1,381 patients with PD and 1,328 controls, variants that were significantly associated with PD after correction for multiple testing included BST1 rs11724635 under an additive model (OR: 0.83, P=0.0006) and a recessive model (OR: 0.67, P=5.9 x 10-5), SNCA rs356219 under an additive model (OR: 1.44, P=1.8 x 10-10), a dominant model (OR: 1.67, P=6.3 x 10-10), and a recessive model (OR: 1.54, P=4.9 x 10-5), HLA-DRA rs3129882 (as previously reported17) under a recessive model (OR: 0.70, P=0.0008), CCDC62/HIP1R rs10847864 under a dominant model (OR: 1.34, P=0.0004), and MAPT rs2942168 under an additive model (OR: 0.72, P=2.9 x 10-6) and a dominant model (OR: 1.44, P=4.4 x 10-6). Additionally, though not quite statistically significant after multiple testing correction, LRRK2 rs1491942 was associated with PD under a recessive model (OR: 1.82, P=0.0022). All of these associations were relatively consistent in magnitude across the three individual series’ except those involving LRRK2 rs1491942 and MAPT rs2942168. For LRRK2 rs1491942, the association with PD was observed in the Irish series (OR: 1.87, P=0.11) and the US series (OR: 2.19, P=0.005) but not in the Polish series (OR: 0.94, P=0.89). For MAPT rs2942168, associations under an additive model were also observed in the Irish series (OR: 0.63, P=0.0009) and the US series (OR: 0.66, P=1.6 x 10-5) but not the Polish series (OR: 1.03, P=0.85); results regarding MAPT rs2942168 were similar under a dominant model.
Table 3A.
Single variant associations with PD in the combined series (1,381 patients with PD, 1,328 controls)
| Additive model | Dominant model | Recessive model | ||||||
|---|---|---|---|---|---|---|---|---|
|
|
||||||||
| Variant | Minor allele | MAF | OR (95% CI) | P-value | OR (95% CI) | P-value | OR (95% CI) | P-value |
| rs2230288 | T | 1.7% | 1.75 (1.12 - 2.74) | 0.014 | 1.75 (1.12 - 2.74) | 0.014 | N/A | N/A |
| rs34372695 | T | 1.8% | 1.25 (0.82 - 1.90) | 0.30 | 1.25 (0.82 - 1.90) | 0.30 | N/A | N/A |
| rs708723 | C | 32.3% | 0.88 (0.79 - 0.98) | 0.023 | 0.83 (0.70 - 0.98) | 0.028 | 0.86 (0.71 - 1.05) | 0.15 |
| rs10928513 | T | 44.8% | 1.12 (1.00 - 1.25) | 0.046 | 1.16 (0.98 - 1.37) | 0.089 | 1.17 (0.96 - 1.41) | 0.12 |
| rs2102808 | T | 13.7% | 1.24 (1.06 - 1.46) | 0.0079 | 1.24 (1.03 - 1.47) | 0.020 | 1.90 (1.01 - 3.54) | 0.045 |
| rs11711441 | A | 12.2% | 0.84 (0.71 - 0.99) | 0.036 | 0.84 (0.70 - 1.01) | 0.072 | 0.56 (0.29 - 1.08) | 0.083 |
| rs6599388 | T | 30.7% | 1.06 (0.94 - 1.19) | 0.35 | 0.98 (0.84 - 1.15) | 0.83 | 1.38 (1.06 - 1.80) | 0.015 |
| rs11724635 | C | 44.4% | 0.83 (0.74 - 0.92) | 0.0006 | 0.86 (0.73 - 1.02) | 0.082 | 0.67 (0.55 - 0.82) | 5.9 x 10-5 |
| rs6812193 | T | 36.5% | 0.85 (0.76 - 0.96) | 0.0065 | 0.84 (0.72 - 0.99) | 0.034 | 0.76 (0.60 - 0.95) | 0.017 |
| rs356219 | G | 40.2% | 1.44 (1.29 - 1.61) | 1.8 x 10-10 | 1.67 (1.42 - 1.96) | 6.3 x 10-10 | 1.54 (1.25 - 1.90) | 4.9 x 10-5 |
| rs3129882a | G | 42.0% | 0.93 (0.83 - 1.04) | 0.21 | 1.07 (0.91 - 1.27) | 0.40 | 0.70 (0.57 - 0.86) | 0.0008 |
| rs156429 | G | 39.0% | 1.00 (0.90 - 1.12) | 0.93 | 1.00 (0.85 - 1.17) | 0.99 | 1.02 (0.82 - 1.26) | 0.85 |
| rs7077361 | C | 12.5% | 0.87 (0.74 - 1.03) | 0.10 | 0.87 (0.72 - 1.04) | 0.14 | 0.71 (0.38 - 1.33) | 0.29 |
| rs1491942 | C | 21.3% | 1.17 (1.02 - 1.34) | 0.023 | 1.12 (0.96 - 1.31) | 0.16 | 1.82 (1.23 - 2.68) | 0.0022 |
| rs10847864 | T | 35.1% | 1.20 (1.07 - 1.35) | 0.0023 | 1.34 (1.14 - 1.56) | 0.0004 | 1.12 (0.88 - 1.42) | 0.36 |
| rs2942168 | A | 19.4% | 0.72 (0.63 - 0.83) | 2.9 x 10-6 | 0.68 (0.58 - 0.80) | 4.4 x 10-6 | 0.62 (0.43 - 0.91) | 0.014 |
| rs12456492 | G | 33.0% | 1.03 (0.92 - 1.16) | 0.58 | 1.07 (0.91 - 1.25) | 0.40 | 0.98 (0.76 - 1.26) | 0.85 |
ORs, 95% CIs, and p-values result from logistic regression models adjusted for age, gender, and series. ORs correspond to an additional minor allele (additive models), presence of the minor allele (dominant models), and presence of two copies of the minor allele (recessive models). N/A is given for variants with no homozygotes of the minor allele in either patients with PD or controls.
The association of HLA rs3129882 with PD has been previously reported in essentially the same patient-control group (~97% overlap) utilized in the current study17; these results are reported again in the current study in order to display a more complete replication of GWAS risk factors for PD.
MAF=minor allele frequency; OR=odds ratio; CI=confidence interval.
Table 3B.
Single variant associations with PD in the Irish series (362 patients with PD, 370 controls)
| Additive model | Dominant model | Recessive model | ||||||
|---|---|---|---|---|---|---|---|---|
|
|
||||||||
| Variant | Minor allele | MAF | OR (95% CI) | P-value | OR (95% CI) | P-value | OR (95% CI) | P-value |
| rs2230288 | T | 1.8% | 4.29 (1.58 - 11.67) | 0.0043 | 4.29 (1.58 - 11.67) | 0.0043 | N/A | N/A |
| rs34372695 | T | 1.5% | 2.42 (0.90 - 6.51) | 0.079 | 2.42 (0.90 - 6.51) | 0.079 | N/A | N/A |
| rs708723 | C | 44.7% | 0.82 (0.66 - 1.01) | 0.066 | 0.68 (0.49 - 0.95) | 0.024 | 0.88 (0.60 - 1.29) | 0.52 |
| rs10928513 | T | 42.5% | 0.84 (0.68 - 1.04) | 0.10 | 0.86 (0.63 - 1.19) | 0.37 | 0.69 (0.47 - 1.02) | 0.060 |
| rs2102808 | T | 14.0% | 1.18 (0.87 - 1.59) | 0.30 | 1.17 (0.83 - 1.66) | 0.37 | 1.53 (0.56 - 4.17) | 0.40 |
| rs11711441 | A | 10.8% | 0.63 (0.45 - 0.90) | 0.011 | 0.65 (0.44 - 0.95) | 0.028 | 0.20 (0.04 - 0.96) | 0.045 |
| rs6599388 | T | 30.7% | 1.02 (0.81 - 1.28) | 0.85 | 0.89 (0.66 - 1.21) | 0.46 | 1.53 (0.92 - 2.54) | 0.098 |
| rs11724635 | C | 44.4% | 0.78 (0.63 - 0.96) | 0.019 | 0.68 (0.49 - 0.94) | 0.021 | 0.75 (0.51 - 1.09) | 0.13 |
| rs6812193 | T | 38.2% | 0.88 (0.71 - 1.09) | 0.25 | 0.91 (0.67 - 1.25) | 0.57 | 0.73 (0.48 - 1.12) | 0.15 |
| rs356219 | G | 42.1% | 1.79 (1.44 - 2.24) | 2.7 x 10-7 | 2.38 (1.71 - 3.31) | 2.7 x 10-7 | 1.90 (1.27 - 2.83) | 0.0016 |
| rs3129882 | G | 38.2% | 0.87 (0.69 - 1.09) | 0.23 | 1.02 (0.75 - 1.41) | 0.88 | 0.55 (0.35 - 0.87) | 0.011 |
| rs156429 | G | 37.3% | 0.97 (0.77 - 1.21) | 0.78 | 1.06 (0.77 - 1.44) | 0.73 | 0.79 (0.50 - 1.23) | 0.30 |
| rs7077361 | C | 12.1% | 1.15 (0.82 - 1.61) | 0.43 | 1.17 (0.82 - 1.67) | 0.40 | 0.93 (0.17 – 5.00) | 0.93 |
| rs1491942 | C | 20.8% | 1.21 (0.93 - 1.58) | 0.15 | 1.18 (0.86 - 1.61) | 0.30 | 1.87 (0.87 - 4.04) | 0.11 |
| rs10847864 | T | 36.9% | 1.08 (0.86 - 1.35) | 0.51 | 1.11 (0.81 - 1.52) | 0.51 | 1.09 (0.70 - 1.71) | 0.70 |
| rs2942168 | A | 19.1% | 0.63 (0.48 - 0.83) | 0.0009 | 0.59 (0.43 - 0.82) | 0.0016 | 0.46 (0.22 – 1.00) | 0.050 |
| rs12456492 | G | 32.3% | 0.93 (0.74 - 1.18) | 0.54 | 1.03 (0.75 - 1.40) | 0.86 | 0.66 (0.39 - 1.10) | 0.11 |
ORs, 95% CIs, and p-values result from logistic regression models adjusted for age and gender. ORs correspond to an additional minor allele (additive models), presence of the minor allele (dominant models), and presence of two copies of the minor allele (recessive models). N/A is given for variants with no homozygotes of the minor allele in either patients with PD or controls. MAF=minor allele frequency; OR=odds ratio; CI=confidence interval.
Table 3C.
Single variant associations with PD in the US series (674 patients with PD, 724 controls)
| Additive model | Dominant model | Recessive model | ||||||
|---|---|---|---|---|---|---|---|---|
|
|
||||||||
| Variant | Minor allele | MAF | OR (95% CI) | P-value | OR (95% CI) | P-value | OR (95% CI) | P-value |
| rs2230288 | T | 1.9% | 1.16 (0.65 - 2.06) | 0.62 | 1.16 (0.65 - 2.06) | 0.62 | N/A | N/A |
| rs34372695 | T | 2.2% | 0.94 (0.55 - 1.61) | 0.83 | 0.94 (0.55 - 1.61) | 0.83 | N/A | N/A |
| rs708723 | C | 44.2% | 0.99 (0.85 - 1.16) | 0.92 | 1.00 (0.79 - 1.26) | 0.97 | 0.98 (0.75 - 1.29) | 0.90 |
| rs10928513 | T | 43.7% | 1.12 (0.96 - 1.30) | 0.15 | 1.13 (0.90 - 1.43) | 0.29 | 1.20 (0.92 - 1.57) | 0.19 |
| rs2102808 | T | 13.3% | 1.28 (1.02 - 1.60) | 0.036 | 1.28 (0.99 - 1.64) | 0.056 | 1.82 (0.77 - 4.31) | 0.18 |
| rs11711441 | A | 13.2% | 0.93 (0.74 - 1.17) | 0.53 | 0.92 (0.71 - 1.18) | 0.51 | 0.96 (0.41 - 2.25) | 0.92 |
| rs6599388 | T | 29.3% | 1.08 (0.91 - 1.27) | 0.39 | 1.04 (0.83 - 1.29) | 0.74 | 1.32 (0.90 - 1.93) | 0.16 |
| rs11724635 | C | 44.8% | 0.85 (0.72 - 0.99) | 0.036 | 0.86 (0.68 - 1.09) | 0.22 | 0.73 (0.55 - 0.96) | 0.024 |
| rs6812193 | T | 36.1% | 0.85 (0.73 – 1.00) | 0.057 | 0.85 (0.68 - 1.06) | 0.15 | 0.75 (0.54 - 1.04) | 0.088 |
| rs356219 | G | 39.6% | 1.29 (1.10 - 1.51) | 0.0014 | 1.50 (1.19 - 1.88) | 0.0005 | 1.25 (0.93 - 1.68) | 0.13 |
| rs3129882 | G | 41.2% | 1.01 (0.87 - 1.19) | 0.86 | 1.07 (0.85 - 1.34) | 0.57 | 0.95 (0.71 - 1.26) | 0.71 |
| rs156429 | G | 40.2% | 1.00 (0.85 - 1.17) | 0.96 | 0.92 (0.74 - 1.16) | 0.49 | 1.13 (0.84 - 1.51) | 0.42 |
| rs7077361 | C | 13.3% | 0.76 (0.60 - 0.96) | 0.019 | 0.76 (0.59 - 0.98) | 0.033 | 0.51 (0.22 - 1.21) | 0.13 |
| rs1491942 | C | 21.6% | 1.16 (0.96 - 1.40) | 0.11 | 1.08 (0.86 - 1.35) | 0.51 | 2.19 (1.27 - 3.78) | 0.0047 |
| rs10847864 | T | 34.5% | 1.37 (1.16 - 1.62) | 0.0002 | 1.65 (1.31 - 2.07) | 1.5 x 10-5 | 1.19 (0.84 - 1.68) | 0.34 |
| rs2942168 | A | 21.6% | 0.66 (0.55 - 0.80) | 1.6 x 10-5 | 0.58 (0.46 - 0.73) | 2.9 x 10-6 | 0.72 (0.44 - 1.17) | 0.19 |
| rs12456492 | G | 31.8% | 1.07 (0.90 - 1.26) | 0.46 | 1.02 (0.82 - 1.27) | 0.84 | 1.28 (0.88 - 1.85) | 0.19 |
ORs, 95% CIs, and p-values result from logistic regression models adjusted for age and gender. ORs correspond to an additional minor allele (additive models), presence of the minor allele (dominant models), and presence of two copies of the minor allele (recessive models). N/A is given for variants with no homozygotes of the minor allele in either patients with PD or controls. MAF=minor allele frequency; OR=odds ratio; CI=confidence interval.
Table 3D.
Single variant associations with PD in the Polish series (345 patients with PD, 234 controls)
| Additive model | Dominant model | Recessive model | ||||||
|---|---|---|---|---|---|---|---|---|
|
|
||||||||
| Variant | Minor allele | MAF | OR (95% CI) | P-value | OR (95% CI) | P-value | OR (95% CI) | P-value |
| rs2230288 | T | 0.9% | 1.61 (0.39 - 6.68) | 0.51 | 1.61 (0.39 - 6.68) | 0.51 | N/A | N/A |
| rs34372695 | T | 1.0% | 1.28 (0.36 - 4.61) | 0.71 | 1.28 (0.36 - 4.61) | 0.71 | N/A | N/A |
| rs708723 | C | 39.2% | 0.68 (0.52 - 0.89) | 0.0055 | 0.66 (0.45 - 0.96) | 0.029 | 0.55 (0.33 - 0.91) | 0.019 |
| rs10928513 | T | 31.4% | 0.94 (0.71 - 1.25) | 0.69 | 0.96 (0.68 - 1.38) | 0.84 | 0.83 (0.43 - 1.60) | 0.58 |
| rs2102808 | T | 14.4% | 1.31 (0.89 - 1.92) | 0.17 | 1.24 (0.83 - 1.86) | 0.29 | N/A | N/A |
| rs11711441 | A | 11.4% | 0.76 (0.51 - 1.14) | 0.18 | 0.77 (0.50 - 1.17) | 0.22 | 0.43 (0.06 - 2.90) | 0.39 |
| rs6599388 | T | 33.9% | 1.07 (0.82 - 1.39) | 0.61 | 0.94 (0.66 - 1.34) | 0.72 | 1.57 (0.90 - 2.76) | 0.11 |
| rs11724635 | C | 43.8% | 0.85 (0.66 - 1.09) | 0.20 | 1.06 (0.72 - 1.54) | 0.77 | 0.55 (0.35 - 0.86) | 0.0084 |
| rs6812193 | T | 35.4% | 0.81 (0.63 - 1.05) | 0.11 | 0.73 (0.51 - 1.05) | 0.093 | 0.81 (0.48 - 1.36) | 0.42 |
| rs356219 | G | 39.2% | 1.47 (1.14 - 1.91) | 0.0034 | 1.48 (1.03 - 2.14) | 0.034 | 2.08 (1.24 - 3.49) | 0.0055 |
| rs3129882 | G | 48.5% | 0.85 (0.66 - 1.10) | 0.23 | 1.25 (0.83 - 1.88) | 0.28 | 0.51 (0.34 - 0.78) | 0.0019 |
| rs156429 | G | 38.1% | 1.01 (0.79 - 1.31) | 0.91 | 1.01 (0.71 - 1.46) | 0.94 | 1.03 (0.63 - 1.67) | 0.91 |
| rs7077361 | C | 11.2% | 0.73 (0.50 - 1.07) | 0.11 | 0.69 (0.45 - 1.07) | 0.10 | 0.69 (0.21 - 2.33) | 0.55 |
| rs1491942 | C | 21.2% | 1.12 (0.82 - 1.52) | 0.48 | 1.18 (0.82 - 1.70) | 0.37 | 0.94 (0.40 - 2.24) | 0.89 |
| rs10847864 | T | 34.2% | 1.10 (0.85 - 1.43) | 0.47 | 1.21 (0.84 - 1.73) | 0.31 | 1.00 (0.59 - 1.69) | 0.99 |
| rs2942168 | A | 14.3% | 1.03 (0.73 - 1.46) | 0.85 | 1.24 (0.82 - 1.88) | 0.30 | 0.35 (0.12 - 0.98) | 0.045 |
| rs12456492 | G | 36.9% | 1.12 (0.86 - 1.46) | 0.40 | 1.31 (0.90 - 1.90) | 0.16 | 0.92 (0.55 - 1.53) | 0.74 |
ORs, 95% CIs, and p-values result from logistic regression models adjusted for age and gender. ORs correspond to an additional minor allele (additive models), presence of the minor allele (dominant models), and presence of two copies of the minor allele (recessive models). N/A is given for variants with no homozygotes of the minor allele in either patients with PD or controls. MAF=minor allele frequency; OR=odds ratio; CI=confidence interval.
Other variants showing significant evidence of an association with PD in the combined series prior to correction for multiple testing were GBA rs2230288, PARK16 rs708723, ACMSD rs10928513, STK39 rs2102808, MCCC1/LAMP3 rs11711441, GAK rs6599388, and STBD1/SCARB2 rs6812193. Of these associations, the strongest were observed for STK39 rs2102808 under an additive model (OR: 1.24, P=0.008) and STBD1/SCARB2 rs6812193 under an additive model (OR: 0.85, P=0.007), both of which were consistent across the three individual series. There was no evidence of an association with PD in the combined series for SYT11 rs34372695, GPNMB rs156429, ITGA8 rs7077361, or RIT2 rs12456492 (all P≥0.10).
No role for LRRK2 coding variants in the GWAS association signal
Associations of LRRK2 p.N551K, p.R1398H, p.K1423K, and p.M1646T with PD are displayed in Table 4. There was no statistically significant evidence of an association between any of these LRRK2 variants and PD in the combined series (all P≥0.38), though the direction of effects (protective for p.N551K, p.R1398H, and p.K1423K; risk for p.M1646T) is similar to what has been previously observed in our larger series and by others [13,18]. The 3-variant p.N551K-R1398H-K1423K haplotype was also not significantly associated with PD (OR: 0.97, 95% CI: 0.71 - 1.34, P=0.71). There was very low correlation of LRRK2 rs1491942 with LRRK2 p.N551K, p.R1398H, p.K1423K, and p.M1646T (all r2<0.01), and additionally the association of LRRK2 rs1491942 with PD was consistent with adjusting for each of the four aforementioned LRRK2 variants (Table 5), both of which indicate that the association signal for the GWAS-nominated rs1491942 is independent of the LRRK2 coding variants.
Table 4.
Associations of LRRK2 p.M1646T, p.N551K, p.R1398H, and p.K1423K with PD
| US series (674 patients, 724 controls) | Irish series (362 patients, 370 controls) | Polish series (345 patients, 234 controls) | Combined series (1,381 patients, 1,329 controls) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
|
||||||||||||
| Variant | MAF | OR (95 CI) | P-value | MAF | OR (95 CI) | P-value | MAF | OR (95 CI) | P-value | MAF | OR (95 CI) | P-value |
| p.M1646T (rs35303786) | 1.4% | 1.54 (0.78, 3.02) | 0.21 | 2.3% | 0.82 (0.39, 1.71) | 0.60 | 1.0% | 0.75 (0.22, 2.58) | 0.64 | 1.5% | 1.16 (0.74, 1.81) | 0.53 |
| p.N551K (rs7308720) | 7.0% | 1.03 (0.75, 1.41) | 0.87 | 5.6% | 0.55 (0.33, 0.92) | 0.022 | 6.0% | 1.10 (0.63, 1.94) | 0.73 | 6.4% | 0.90 (0.71, 1.14) | 0.38 |
| p.R1398H (rs7133914) | 6.8% | 1.07 (0.77, 1.47) | 0.69 | 6.5% | 0.61 (0.38, 0.99) | 0.044 | 6.0% | 1.10 (0.63, 1.94) | 0.73 | 6.5% | 0.92 (0.73, 1.17) | 0.50 |
| p.K1423K (rs11175964) | 6.6% | 1.07 (0.77, 1.49) | 0.68 | 6.3% | 0.57 (0.35, 0.92) | 0.022 | 6.0% | 1.10 (0.63, 1.94) | 0.73 | 6.4% | 0.92 (0.72, 1.16) | 0.47 |
ORs, 95% CIs, and p-values result from logistic regression models adjusted for age, gender, and series (combined series only). LRRK2 p.M1646T, p.N551K, p.R1398H, and p.K1423K were considered under a dominant model in all analysis.
Table 5.
Associations of individual LRRK2 variants with PD when adjusting for other LRRK2 variants
| Association/Model adjustment | OR (95% CI) | P-value |
|---|---|---|
| Association of LRRK2 rs1491942 with PD under an additive model adjusting for: | ||
| LRRK2 p.N551K | 1.15 (1.00, 1.32) | 0.042 |
| LRRK2 p.R1398H | 1.17 (1.02, 1.33) | 0.026 |
| LRRK2 p.K1423K | 1.17 (1.02, 1.34) | 0.024 |
| LRRK2 p.M1646T | 1.16 (1.01, 1.33) | 0.032 |
| LRRK2 p.N551K, p.R1398H, p.K1423K, and p.M1646T | 1.17 (1.02, 1.34) | 0.025 |
| Association of LRRK2 rs1491942 with PD under a recessive model adjusting for: | ||
| LRRK2 p.N551K | 1.80 (1.22, 2.66) | 0.0026 |
| LRRK2 p.R1398H | 1.81 (1.23, 2.68) | 0.0023 |
| LRRK2 p.K1423K | 1.82 (1.23, 2.70) | 0.0024 |
| LRRK2 p.M1646T | 1.80 (1.22, 2.66) | 0.0027 |
| LRRK2 p.N551K, p.R1398H, p.K1423K, and p.M1646T | 1.80 (1.21, 2.68) | 0.0029 |
ORs, 95% CIs, and p-values result from logistic regression models adjusted for age, gender, series, and the given LRRK2 variants. LRRK2 p.N551K, p.R1398H, p.K1423K, and p.M1646T were considered under a dominant model in all analysis.
Interaction between LRRK2 and PARK16
To evaluate the interaction between PARK16 rs708723 and LRRK2 rs1491942, we performed logistic regression analyses. The results are displayed in Table 6, where we considered PARK16 rs708723 under additive and dominant models and LRRK2 rs1491942 under additive and recessive models owing to the significant associations with PD that were observed in these scenarios. There was no statistically significant evidence of an interaction between these two variants (all interaction P≥0.36); the protective effect of PARK16 rs708723 on risk of PD was observed across LRRK2 rs1491942 genotypes, and the risk effect of LRRK2 rs1491942 was seen across genotypes of PARK16 rs708723. To mirror the presentation of results by MacLeod et al. [15], the association between LRRK2 rs1491942 under an additive model and PD risk was similar for individuals with (OR: 1.19, P=0.038, N=1,806) and without (OR: 1.11, P=0.40, N=871) the PARK16 protective allele, and similarly the association between LRRK2 rs1491942 under a recessive model and risk of PD was comparable for individuals with (OR: 2.05, P=0.0031) and without (OR: 1.51, P=0.22) the protective PARK16 allele.
Table 6.
Interaction between PARK16 rs708723 and LRRK2 rs1491942
| Test of association | |||||
|---|---|---|---|---|---|
|
|
|||||
| LRRK2 rs1491942 model/genotype | PARK16 rs708723 model/genotype | Sample genotype count and frequency | OR (95% CI) | P-value | Test of interaction |
| Additive model | Additive model | ||||
| GG | TT | 520 (19.4%) | 1.00 (reference) | N/A | OR: 1.09 |
| GG | CT | 815 (30.4%) | 0.85 (0.68, 1.06) | 0.15 | 95% CI: 0.90 - 1.31 |
| GG | CC | 322 (12.0%) | 0.73 (0.55, 0.97) | 0.031 | P=0.36 |
| CG | TT | 309 (11.5%) | 1.02 (0.76, 1.36) | 0.90 | |
| CG | CT | 430 (16.1%) | 0.85 (0.66, 1.11) | 0.24 | |
| CG | CC | 162 (6.1%) | 0.85 (0.60, 1.23) | 0.39 | |
| CC | TT | 42 (1.6%) | 1.52 (0.78, 2.96) | 0.22 | |
| CC | CT | 49 (1.8%) | 1.73 (0.92, 3.23) | 0.088 | |
| CC | CC | 28 (1.0%) | 1.69 (0.76, 3.79) | 0.20 | |
| Recessive model | Additive model | ||||
| GG or CG | TT | 829 (30.9%) | 1.00 (reference) | N/A | OR: 1.22 |
| GG or CG | CT | 1245 (46.5%) | 0.84 (0.70, 1.01) | 0.064 | 95% CI: 0.73 - 2.05 |
| GG or CG | CC | 484 (18.1%) | 0.76 (0.61, 0.96) | 0.022 | P=0.44 |
| CC | TT | 42 (1.6%) | 1.51 (0.78, 2.91) | 0.22 | |
| CC | CT | 49 (1.8%) | 1.71 (0.92, 3.18) | 0.087 | |
| CC | CC | 28 (1.0%) | 1.68 (0.76, 3.74) | 0.20 | |
| Additive model | Dominant model | ||||
| GG | TT | 520 (19.4%) | 1.00 (reference) | N/A | OR: 1.09 |
| GG | CT or CC | 1137 (42.5%) | 0.81 (0.66, 1.00) | 0.055 | 95% CI: 0.82 - 1.45 |
| CG | TT | 309 (11.5%) | 1.02 (0.76, 1.36) | 0.90 | P=0.57 |
| CG | CT or CC | 592 (22.1%) | 0.85 (0.67, 1.09) | 0.20 | |
| CC | TT | 42 (1.6%) | 1.52 (0.78, 2.96) | 0.22 | |
| CC | CT or CC | 77 (2.9%) | 1.71 (1.03, 2.85) | 0.038 | |
| Recessive model | Dominant model | ||||
| GG or CG | TT | 829 (31.0%) | 1.00 (reference) | N/A | OR: 1.38 |
| GG or CG | CT or CC | 1729 (64.6%) | 0.82 (0.69, 0.97) | 0.023 | 95% CI: 0.61 - 3.13 |
| CC | TT | 42 (1.6%) | 1.51 (0.78, 2.91) | 0.23 | P=0.44 |
| CC | CT or CC | 77 (2.9%) | 1.70 (1.03, 2.80) | 0.036 | |
ORs and p-values result from logistic regression models. For tests of association, LRRK2 rs1491942 and PARK16 rs708723 were combined into one variable, and the model was adjusted for age, gender, and series. For tests of interaction, models included LRRK2 rs1491942, PARK16 rs708723, the interaction between these two variants, age, gender, and series. OR=odds ratio. CI=confidence interval.
Interactions of PARK16 rs708723 with LRRK2 p.N551K, p.R1398H, and p.K1423K are examined in Table 7. There were no significant interactions of PARK16 rs708723 with LRRK2 p.N551K, p.R1398H, or p.K1423K in relation to risk of PD, though non-significant trends toward interaction were observed. This was most evident for LRRK2 p.N551K, where the protective effect of the minor allele for PARK16 rs708723 was strongest in individuals with a copy of the minor allele for p.N551K, and vice versa (Interaction OR: 0.61, P=0.057). Similar non-significant trends were observed for p.R1398H and p.K1423K (Table 7).
Table 7.
Interactions of PARK16 rs708723 with LRRK2 p.N551K, p.R1398H, and p.K1423K
| Test of association | |||||
|---|---|---|---|---|---|
|
|
|||||
| LRRK2 variant/genotype | PARK16 rs708723 model/genotype | Sample genotype count and frequency | OR (95% CI) | P-value | Test of interaction |
| LRRK2 p.N551K | Additive model | ||||
| CC | TT | 757 (28.6%) | 1.00 (reference) | N/A | OR: 0.77 |
| CC | CT | 1116 (42.1%) | 0.92 (0.76, 1.11) | 0.40 | 95% CI: 0.55 - 1.09 |
| CC | CC | 447 (16.9%) | 0.84 (0.66, 1.07) | 0.16 | P=0.14 |
| CG or GG | TT | 107 (4.0%) | 1.28 (0.84, 1.96) | 0.25 | |
| CG or GG | CT | 163 (6.2%) | 0.70 (0.49, 0.99) | 0.044 | |
| CG or GG | CC | 59 (2.2%) | 0.71 (0.41, 1.23) | 0.22 | |
| Dominant model | |||||
| CC | TT | 757 (28.6%) | 1.00 (reference) | N/A | OR: 0.61 |
| CC | CT or CC | 1563 (59.0%) | 0.90 (0.75, 1.07) | 0.24 | 95% CI: 0.37 - 1.02 |
| CG or GG | TT | 107 (4.0%) | 1.28 (0.84, 1.96) | 0.25 | P=0.057 |
| CG or GG | CT or CC | 222 (8.4%) | 0.70 (0.52, 0.96) | 0.025 | |
| LRRK2 p.R1398H | Additive model | ||||
| GG | TT | 751 (28.3%) | 1.00 (reference) | N/A | OR: 0.87 |
| GG | CT | 1116 (42.1%) | 0.91 (0.75, 1.09) | 0.30 | 95% CI: 0.61 - 1.22 |
| GG | CC | 446 (16.8%) | 0.81 (0.64, 1.03) | 0.083 | P=0.41 |
| GA or AA | TT | 109 (4.1%) | 1.19 (0.78, 1.80) | 0.42 | |
| GA or AA | CT | 171 (6.5%) | 0.72 (0.51, 1.02) | 0.062 | |
| GA or AA | CC | 57 (2.2%) | 0.79 (0.46, 1.38) | 0.41 | |
| Dominant model | |||||
| GG | TT | 751 (28.3%) | 1.00 (reference) | N/A | OR: 0.71 |
| GG | CT or CC | 1562 (58.9%) | 0.88 (0.73, 1.05) | 0.15 | 95% CI: 0.43 - 1.18 |
| GA or AA | TT | 109 (4.1%) | 1.19 (0.78, 1.80) | 0.42 | P=0.19 |
| GA or AA | CT or CC | 228 (8.6%) | 0.74 (0.55, 1.00) | 0.052 | |
| LRRK2 p.K1423K | Additive model | ||||
| GG | TT | 749 (28.5%) | 1.00 (reference) | N/A | OR: 0.85 |
| GG | CT | 1109 (42.2%) | 0.92 (0.76, 1.11) | 0.39 | 95% CI: 0.60 - 1.20 |
| GG | CC | 448 (17.0%) | 0.81 (0.64, 1.03) | 0.091 | P=0.36 |
| GA or AA | TT | 104 (4.0%) | 1.21 (0.79, 1.85) | 0.39 | |
| GA or AA | CT | 166 (6.3%) | 0.72 (0.51, 1.02) | 0.062 | |
| GA or AA | CC | 54 (2.1%) | 0.79 (0.45, 1.39) | 0.41 | |
| Dominant model | |||||
| GG | TT | 749 (28.5%) | 1.00 (reference) | N/A | OR: 0.69 |
| GG | CT or CC | 1557 (59.2%) | 0.89 (0.74, 1.06) | 0.19 | 95% CI: 0.41 - 1.15 |
| GA or AA | TT | 104 (4.0%) | 1.21 (0.79, 1.85) | 0.39 | P=0.15 |
| GA or AA | CT or CC | 220 (8.4%) | 0.74 (0.54, 1.00) | 0.051 | |
ORs and p-values result from logistic regression models. For tests of association, the two given variants were combined into one variable, and the model was adjusted for age, gender, and series. For tests of interaction, models included each of the two variants, their interaction, age, gender, and series. OR=odds ratio. CI=confidence interval.
Discussion
The results of this study provide evidence that confirm associations with PD for a number of variants that have previously been nominated as risk-modifying susceptibility factors for PD in GWAS. The strongest associations with PD were observed for variants in BST1, SNCA, HLA, CCDC62/HIP1R, MAPT, and LRRK2 all of which were significant after correction for multiple testing with the exception of the LRRK2 variant which was almost significant. Additionally, prior to multiple testing correction there was significant (P≤0.05) evidence of associations with PD for variants in the GBA, PARK16, ACMSD, STK39, MCCC1/LAMP3, GAK, and STBD1/SCARB2 genes. Although there is some degree of overlap between these findings and those reported in the aforementioned larger consortium replication effort that utilized more than 17,000 patients and controls from 19 different countries16, it is of interest that the many of the previously replicated significant associations (ACMSD, STK39, MCCC1/LAMP3, BST1, SNCA, LRRK2, CCDC62/HIP1R, MAPT) were still observed when considering only subjects from the US, Ireland, and Poland. Additionally, the significant associations with PD that we identified for variants in GBA, PARK16, and STBD1/SCARB2 have been previously unreported in our series. We also observed a very low degree of correlation between LRRK2 rs1491942 and LRRK2 p.N551K, p.R1398H, p.K1423K, and p.M1646T, indicating that the associations of these variants with PD are independent of one another. Finally, no significant interaction was noted between PARK16 rs708723 and LRRK2 rs1491942, though we did observe a non-significant interaction between PARK16 rs708723 and LRRK2 p.N551K which requires further study.
Of the 13 variants showing significant evidence of an association with PD prior to multiple testing correction, the magnitude of the effect was similar to previous studies for all variants with the exception of HLA-DRA rs3129882, where we observed the risk of PD was lower for individuals with two copies of the minor allele as previously reported [19]. This highlights one limitation of GWAS, where due to the extremely large amount of variants that are included, the only statistical model that is usually considered is an additive model, and this may not be the most appropriate model for a given variant. Indeed, the strongest associations that we observed in this study for GAK rs6599388, BST1 rs11724635, HLA rs3129882, and LRRK2 rs1491942 were under a recessive model.
Variants in SYT11, GPNMB, ITGA8, and RIT2 did not show evidence of an association with PD even before correction for multiple testing. Given the much smaller sample size of this study compared to the aforementioned GWAS, these results should be interpreted carefully, and the possibility of Type II error (i.e. false-negative association) is important to consider. It should be noted that for ITGA8 rs7077361, the estimated OR of 0.87 that was observed in this study is nearly identical to what has been observed in previous studies, and 95% confidence limits for GPNMB rs156429 and RIT2 rs12456492 are generally consistent with the findings of previous studies. SYT11 rs34372695 is rare with a MAF of 1.8% in the combined series, and this may in part explain the lack of a significant association between this variant and PD in the current study, where 95% confidence limits are also consistent with previous findings.
Common variation in the SNCA, MAPT and LRRK2 genes are well-established to affect susceptibility to disease and were known prior to GWA approaches. Consistent replication of the other loci now nominated from GWAS is required to determine true associations and effect sizes. It is also crucial to discern ethnic-specific effects which have been previously observed for a number of loci including LRRK2 [13,20]. Studies have started to publish replication results for some of these loci [21-23]. Sharma and colleagues in the Genetic Epidemiology of Parkinson’s disease consortium attempted to replicate 11 variants from the initial IPDGC study in a large patient (n=8750)-control (n=8955) series of Caucasian and Asian descent [23]. This study replicated nine of the loci in either series with no evidence observed for the HLA and ACMSD association with disease.
Pihlstrom et al. examined 18 loci in a Scandinavian series of 1345 unrelated PD patients and 1225 control subjects collected in Norway and Sweden [22]. Only four loci showed significant association following statistical correction, although P-values <0.05 were observed for eleven loci. Lack of association was observed for BST1, PARK16 and LRRK2 candidates, which may reflect some ethnic specificity given previous associations in Asian series. Interestingly, Liu and colleagues recently examined the association of 17 loci in 1,737 subjects (989 patients and 748 controls) of Chinese descent [21]. It was observed that nine of the selected variants were monomorphic and four other had very low minor allele frequencies. The only variants that showed association were in the SNCA and BST1 loci. This study highlights the ethnic-specific frequencies of nominated variants and the need for further studies in under-represented populations.
Our recent studies of LRRK2 coding variation in PD susceptibility identified a common protective haplotype (p.N551K-p.R1398H-p.K1423K) and risk factor (p.M1646T) in Caucasian populations [13]. In the present study we assessed whether the coding susceptibility variants accounted for a proportion of the GWAS LRRK2 signal (rs1491942). Our results suggest that the LRRK2 signal is independent of the coding variation and is therefore driven by non-coding variation most likely in regulatory regions affecting gene/transcript expression. In addition, the study by MacLeod and colleagues showed the PARK16 candidate protein RAB7L1 may act in a pathway with LRRK2 [15]. They observed that overexpression of RAB7L1 rescued a neurodegeneration phenotype in LRRK2 mutant neurons. The study also suggested a genetic interaction between the PARK16 and LRRK2 loci. Our study examined the potential interaction and although we did not show an interaction with our GWAS SNPs at these loci, we did observe a trend with the LRRK2 common protective variants and the PARK16 GWAS SNP. These findings suggest that further gene-gene interaction studies are warranted and it is crucial to determine if the RAB7L1 gene is accounting for the PARK16 association signal.
The characterization of population-based genetic susceptibility factors for PD will be an important step forward in our understanding of the disease. Sequencing studies will help pinpoint those functional variants affecting disease risk, which can be used as diagnostic and prognostic markers. In addition, examining how these variants interact will be important to generate accurate predictions of risk and direct therapeutic interventions strategies.
Acknowledgements
We would like to thank all those who have contributed to our research, particularly the patients and families who donated DNA samples for this work. This work is supported in part by a gift from Carl Edward Bolch, Jr. and Susan Bass Bolch, Michael J. Fox Foundation, NINDS R01 NS078086, and a Morris K. Udall Parkinson’s Disease Research Center of Excellence (NINDS P50 #NS072187).
References
- 1.Fung HC, Scholz S, Matarin M, Simon-Sanchez J, Hernandez D, Britton A, Gibbs JR, Langefeld C, Stiegert ML, Schymick J, Okun MS, Mandel RJ, Fernandez HH, Foote KD, Rodriguez RL, Peckham E, De Vrieze FW, Gwinn-Hardy K, Hardy JA, Singleton A. Genome-wide genotyping in Parkinson’s disease and neurologically normal controls: first stage analysis and public release of data. Lancet Neurol. 2006;5:911–916. doi: 10.1016/S1474-4422(06)70578-6. [DOI] [PubMed] [Google Scholar]
- 2.Maraganore DM, de Andrade M, Lesnick TG, Strain KJ, Farrer MJ, Rocca WA, Pant PV, Frazer KA, Cox DR, Ballinger DG. High-resolution whole-genome association study of Parkinson disease. Am J Hum Genet. 2005;77:685–693. doi: 10.1086/496902. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Pankratz N, Wilk JB, Latourelle JC, DeStefano AL, Halter C, Pugh EW, Doheny KF, Gusella JF, Nichols WC, Foroud T, Myers RH PSG-PROGENI and GenePD Investigators, Coordinators and Molecular Genetic Laboratories. Genomewide association study for susceptibility genes contributing to familial Parkinson disease. Hum Genet. 2009;124:593–605. doi: 10.1007/s00439-008-0582-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Satake W, Nakabayashi Y, Mizuta I, Hirota Y, Ito C, Kubo M, Kawaguchi T, Tsunoda T, Watanabe M, Takeda A, Tomiyama H, Nakashima K, Hasegawa K, Obata F, Yoshikawa T, Kawakami H, Sakoda S, Yamamoto M, Hattori N, Murata M, Nakamura Y, Toda T. Genome-wide association study identifies common variants at four loci as genetic risk factors for Parkinson’s disease. Nat Genet. 2009;41:1303–1307. doi: 10.1038/ng.485. [DOI] [PubMed] [Google Scholar]
- 5.Simon-Sanchez J, Schulte C, Bras JM, Sharma M, Gibbs JR, Berg D, Paisan-Ruiz C, Lichtner P, Scholz SW, Hernandez DG, Kruger R, Federoff M, Klein C, Goate A, Perlmutter J, Bonin M, Nalls MA, Illig T, Gieger C, Houlden H, Steffens M, Okun MS, Racette BA, Cookson MR, Foote KD, Fernandez HH, Traynor BJ, Schreiber S, Arepalli S, Zonozi R, Gwinn K, van der Brug M, Lopez G, Chanock SJ, Schatzkin A, Park Y, Hollenbeck A, Gao J, Huang X, Wood NW, Lorenz D, Deuschl G, Chen H, Riess O, Hardy JA, Singleton AB, Gasser T. Genome-wide association study reveals genetic risk underlying Parkinson’s disease. Nat Genet. 2009;41:1308–1312. doi: 10.1038/ng.487. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Hamza TH, Zabetian CP, Tenesa A, Laederach A, Montimurro J, Yearout D, Kay DM, Doheny KF, Paschall J, Pugh E, Kusel VI, Collura R, Roberts J, Griffith A, Samii A, Scott WK, Nutt J, Factor SA, Payami H. Common genetic variation in the HLA region is associated with late-onset sporadic Parkinson’s disease. Nat Genet. 2010;42:781–785. doi: 10.1038/ng.642. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Lesage S, Brice A. Role of mendelian genes in “sporadic” Parkinson’s disease. Parkinsonism Relat Disord. 2012;18(Suppl 1):S66–70. doi: 10.1016/S1353-8020(11)70022-0. [DOI] [PubMed] [Google Scholar]
- 8.International Parkinson Disease Genomics Consortium. Nalls MA, Plagnol V, Hernandez DG, Sharma M, Sheerin UM, Saad M, Simón-Sánchez J, Schulte C, Lesage S, Sveinbjörnsdóttir S, Stefánsson K, Martinez M, Hardy J, Heutink P, Brice A, Gasser T, Singleton AB, Wood NW. Imputation of sequence variants for identification of genetic risks for Parkinson’s disease: a meta-analysis of genome-wide association studies. Lancet. 2011;377:641–649. doi: 10.1016/S0140-6736(10)62345-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.International Parkinson’s Disease Genomics Consortium (IPDGC); Wellcome Trust Case Control Consortium 2 (WTCCC2) A two-stage meta-analysis identifies several new loci for Parkinson’s disease. PLoS Genet. 2011;7:e1002142. doi: 10.1371/journal.pgen.1002142. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Do CB, Tung JY, Dorfman E, Kiefer AK, Drabant EM, Francke U, Mountain JL, Goldman SM, Tanner CM, Langston JW, Wojcicki A, Eriksson N. Web-based genome-wide association study identifies two novel loci and a substantial genetic component for Parkinson’s disease. PLoS Genet. 2011;7:e1002141. doi: 10.1371/journal.pgen.1002141. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Lill CM, Roehr JT, McQueen MB, Kavvoura FK, Bagade S, Schjeide BM, Schjeide LM, Meissner E, Zauft U, Allen NC, Liu T, Schilling M, Anderson KJ, Beecham G, Berg D, Biernacka JM, Brice A, DeStefano AL, Do CB, Eriksson N, Factor SA, Farrer MJ, Foroud T, Gasser T, Hamza T, Hardy JA, Heutink P, Hill-Burns EM, Klein C, Latourelle JC, Maraganore DM, Martin ER, Martinez M, Myers RH, Nalls MA, Pankratz N, Payami H, Satake W, Scott WK, Sharma M, Singleton AB, Stefansson K, Toda T, Tung JY, Vance J, Wood NW, Zabetian CP 23andMe Genetic Epidemiology of Parkinson’s Disease Consortium; International Parkinson’s Disease Genomics Consortium; Parkinson’s Disease GWAS Consortium; Wellcome Trust Case Control Consortium 2) Young P, Tanzi RE, Khoury MJ, Zipp F, Lehrach H, Ioannidis JP, Bertram L. Comprehensive research synopsis and systematic meta-analyses in Parkinson’s disease genetics: The PDGene database. PLoS Genet. 2012;8:e1002548. doi: 10.1371/journal.pgen.1002548. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Pankratz N, Beecham GW, DeStefano AL, Dawson TM, Doheny KF, Factor SA, Hamza TH, Hung AY, Hyman BT, Ivinson AJ, Krainc D, Latourelle JC, Clark LN, Marder K, Martin ER, Mayeux R, Ross OA, Scherzer CR, Simon DK, Tanner C, Vance JM, Wszolek ZK, Zabetian CP, Myers RH, Payami H, Scott WK, Foroud T PD GWAS Consortium. Meta-analysis of Parkinson disease: Identification of a novel locus, RIT2. Ann Neurol. 2012;71:370–84. doi: 10.1002/ana.22687. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Ross OA, Soto-Ortolaza AI, Heckman MG, Aasly JO, Abahuni N, Annesi G, Bacon JA, Bardien S, Bozi M, Brice A, Brighina L, Van Broeckhoven C, Carr J, Chartier-Harlin MC, Dardiotis E, Dickson DW, Diehl NN, Elbaz A, Ferrarese C, Ferraris A, Fiske B, Gibson JM, Gibson R, Hadjigeorgiou GM, Hattori N, Ioannidis JP, Jasinska-Myga B, Jeon BS, Kim YJ, Klein C, Kruger R, Kyratzi E, Lesage S, Lin CH, Lynch T, Maraganore DM, Mellick GD, Mutez E, Nilsson C, Opala G, Park SS, Puschmann A, Quattrone A, Sharma M, Silburn PA, Sohn YH, Stefanis L, Tadic V, Theuns J, Tomiyama H, Uitti RJ, Valente EM, van de Loo S, Vassilatis DK, Vilarino-Guell C, White LR, Wirdefeldt K, Wszolek ZK, Wu RM, Farrer MJ. Association of LRRK2 exonic variants with susceptibility to Parkinson’s disease: a case-control study. Lancet Neurol. 2011;10:898–908. doi: 10.1016/S1474-4422(11)70175-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Elbaz A, Ross OA, Ioannidis JP, Soto-Ortolaza AI, Moisan F, Aasly J, Annesi G, Bozi M, Brighina L, Chartier-Harlin MC, Destée A, Ferrarese C, Ferraris A, Gibson JM, Gispert S, Hadjigeorgiou GM, Jasinska-Myga B, Klein C, Krüger R, Lambert JC, Lohmann K, van de Loo S, Loriot MA, Lynch T, Mellick GD, Mutez E, Nilsson C, Opala G, Puschmann A, Quattrone A, Sharma M, Silburn PA, Stefanis L, Uitti RJ, Valente EM, Vilariño-Güell C, Wirdefeldt K, Wszolek ZK, Xiromerisiou G, Maraganore DM, Farrer MJ Genetic Epidemiology of Parkinson’s Disease (GEO-PD) Consortium. Independent and joint effects of the MAPT and SNCA genes in Parkinson disease. Ann Neurol. 2011;69:778–792. doi: 10.1002/ana.22321. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.MacLeod DA, Rhinn H, Kuwahara T, Zolin A, Di Paolo G, McCabe BD, Marder KS, Honig LS, Clark LN, Small SA, Abeliovich A. RAB7L1 interacts with LRRK2 to modify intraneuronal protein sorting and Parkinson’s disease risk. Neuron. 2013;77:425–439. doi: 10.1016/j.neuron.2012.11.033. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Schaid DJ, Rowland CM, Tines DE, Jacobson RM, Poland GA. Score tests for association between traits and haplotypes when linkage phase is ambiguous. Am J Hum Genet. 2002;70:425–434. doi: 10.1086/338688. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Dudoit S, van der Laan MJ, Pollard KS. Multiple testing. Part I. Single-step procedures for control of general type I error rates. Stat Appl Genet Mol Biol. 2004;3:Article13. doi: 10.2202/1544-6115.1040. [DOI] [PubMed] [Google Scholar]
- 18.Tan EK, Peng R, Teo YY, Tan LC, Angeles D, Ho P, Chen ML, Lin CH, Mao XY, Chang XL, Prakash KM, Liu JJ, Au WL, Le WD, Jankovic J, Burgunder JM, Zhao Y, Wu RM. Multiple LRRK2 variants modulate risk of Parkinson disease: a Chinese multicenter study. Hum Mutat. 2010;31:561–568. doi: 10.1002/humu.21225. [DOI] [PubMed] [Google Scholar]
- 19.Puschmann A, Verbeeck C, Heckman MG, Soto-Ortolaza AI, Lynch T, Jasinska-Myga B, Opala G, Krygowska-Wajs A, Barcikowska M, Uitti RJ, Wszolek ZK, Ross OA. Human leukocyte antigen variation and Parkinson’s disease. Parkinsonism Relat Disord. 2011;17:376–378. doi: 10.1016/j.parkreldis.2011.03.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Peeraully T, Tan EK. Genetic variants in sporadic Parkinson’s disease: East vs West. Parkinsonism Relat Disord. 2012;18(Suppl 1):S63–65. doi: 10.1016/S1353-8020(11)70021-9. [DOI] [PubMed] [Google Scholar]
- 21.Liu J, Xiao Q, Wang Y, Xu ZM, Yang Q, Wang G, Tan YY, Ma JF, Zhang J, Huang W, Chen SD. Analysis of genome-wide association study-linked loci in Parkinson’s disease of Mainland China. Mov Disord. 2013;28:1892–5. doi: 10.1002/mds.25599. [DOI] [PubMed] [Google Scholar]
- 22.Pihlstrom L, Axelsson G, Bjornara KA, Dizdar N, Fardell C, Forsgren L, Holmberg B, Larsen JP, Linder J, Nissbrandt H, Tysnes OB, Ohman E, Dietrichs E, Toft M. Supportive evidence for 11 loci from genome-wide association studies in Parkinson’s disease. Neurobiol Aging. 2013;34:1708, e1707–1713. doi: 10.1016/j.neurobiolaging.2012.10.019. [DOI] [PubMed] [Google Scholar]
- 23.Sharma M, Ioannidis JP, Aasly JO, Annesi G, Brice A, Van Broeckhoven C, Bertram L, Bozi M, Crosiers D, Clarke C, Facheris M, Farrer M, Garraux G, Gispert S, Auburger G, Vilariño-Güell C, Hadjigeorgiou GM, Hicks AA, Hattori N, Jeon B, Lesage S, Lill CM, Lin JJ, Lynch T, Lichtner P, Lang AE, Mok V, Jasinska-Myga B, Mellick GD, Morrison KE, Opala G, Pramstaller PP, Pichler I, Park SS, Quattrone A, Rogaeva E, Ross OA, Stefanis L, Stockton JD, Satake W, Silburn PA, Theuns J, Tan EK, Toda T, Tomiyama H, Uitti RJ, Wirdefeldt K, Wszolek Z, Xiromerisiou G, Yueh KC, Zhao Y, Gasser T, Maraganore D, Krüger R GEO-PD Consortium. Large-scale replication and heterogeneity in Parkinson disease genetic loci. Neurology. 2012;79:659–667. doi: 10.1212/WNL.0b013e318264e353. [DOI] [PMC free article] [PubMed] [Google Scholar]