Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Plasticity-related gene 3 (LPPR1) and age at diagnosis of Parkinson disease

Plasticity-related gene 3 (LPPR1) and age at diagnosis of Parkinson disease Objective To identify modifiers of age at diagnosis of Parkinson disease (PD). Methods Genome-wide association study (GWAS) included 1,950 individuals with PD from the Neu- roGenetics Research Consortium (NGRC) study. Replication was conducted in the Parkin- son’s, Genes and Environment study, including 209 prevalent (PAGE ) and 517 incident (PAGE ) PD cases. Cox regression was used to test association with age at diagnosis. Indi- viduals without neurologic disease were used to rule out confounding. Gene-level analysis and functional annotation were conducted using Functional Mapping and Annotation of GWAS platform (FUMA). Results The GWAS revealed 2 linked but seemingly independent association signals that mapped to LPPR1 on chromosome 9. LPPR1 was significant in gene-based analysis (p = 1E-8). The top signal (rs17763929, hazard ratio [HR] = 1.88, p = 5E-8) replicated in PAGE (HR = 1.87, p = 0.01) but not in PAGE . The second signal (rs73656147) was robust with no evidence of heterogeneity (HR = 1.95, p = 3E-6 in NGRC; HR = 2.14, p = 1E-3 in PAGE + PAGE , and HR P I = 2.00, p = 9E-9 in meta-analysis of NGRC + PAGE + PAGE ). The associations were with age P I at diagnosis, not confounded by age in patients or in the general population. The PD-associated regions included variants with Combined Annotation Dependent Depletion (CADD) scores = 10–19 (top 1%–10% most deleterious mutations in the genome), a missense with predicted destabilizing effect on LPPR1, an expression quantitative trait locus (eQTL) for GRIN3A (false discovery rate [FDR] = 4E-4), and variants that overlap with enhancers in LPPR1 and interact with promoters of LPPR1 and 9 other brain-expressed genes (Hi-C FDR < 1E-6). Conclusions Through association with age at diagnosis, we uncovered LPPR1 as a modifier gene for PD. LPPR1 expression promotes neuronal regeneration after injury in animal models. Present data provide a strong foundation for mechanistic studies to test LPPR1 as a driver of response to damage and a therapeutic target for enhancing neuroregeneration and slowing disease progression. From the Department of Neurology (Z.D.W., E.M.H.-B., H.P.), University of Alabama at Birmingham, Birmingham, AL; Department of Epidemiology and Biostatistics (H.C.), Michigan State University, East Lansing, MI; Department of Neurology (S.A.F.), Jean & Paul Amos Parkinson’s Disease and Movement Disorder Program, Emory University School of Medicine, Atlanta, GA; VA Puget Sound Health Care System and Department of Neurology (C.P.Z.), University of Washington, Seattle, WA; and Center for Genomic Medicine (H.P.), HudsonAlpha Institute for Biotechnology, Huntsville, AL. Funding information and disclosures are provided at the end of the article. Full disclosure form information provided by the authors is available with the full text of this article at Neurology.org/NG. The Article Processing Charge was funded by the authors. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal. Copyright © 2018 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary AIM = ancestry informative marker; FDR = false discovery rate; HR = hazard ratio; LD = linkage disequilibrium; MAF = minor allele frequency; MAP = moving average plot; NGRC = NeuroGenetics Research Consortium; PAGE = Parkinson’s, Genes and Environment; PC = principal component; PD = Parkinson disease. The underlying neurodegenerative process that causes Par- clinics in Portland (OR), Seattle (WA), Albany (NY), and kinson disease (PD) begins decades before the disease is di- Atlanta (GA). Controls were spouses of patients or commu- agnosed. The current view is that following an initial insult nity volunteers, self-reported as being free of neurologic dis- (e.g., toxicity, trauma, or genetic), the disease starts with an ease. The eligibility criterion for cases was diagnosis of PD by asymptomatic phase of unknown duration, followed by de- a movement disorder specialist according to the UK Brain velopment of prodromal nonmotor symptoms such as con- Bank criteria. The eligibility criteria for controls were no stipation, anosmia, and sleep disorders. Years later, cardinal neurologic disease and genetically unrelated to patients. Age motor signs appear, at which point a diagnosis of PD is made. was defined as age at study entry. Age at diagnosis was Age at onset of motor signs, and therefore the age at diagnosis extracted from medical records or ascertained by self-report. of PD, is highly variable, ranging from teen ages to the 10th Age at onset of the first motor sign was obtained using a self- decade of life. The reason for this variation is unknown, and administered questionnaire. Age at onset and age at diagnosis understanding it will likely shed light on factors that affect the were highly correlated in the NGRC (r = 0.91, p < 2E-16). All rate of disease progression. participants were whites of European descent. There is substantial evidence that genetic factors play a major PAGE is a cross-sectional study nested in the longitudinal role in age at onset of motor signs and age at diagnosis of NIH-American Association of Retired Persons Diet and 2–6 14 PD. Genome-wide studies have identified numerous loci Health Study. Participants were enrolled in 1995–1997 that associate with the risk of developing PD, but the risk (irrespective of PD) via a food frequency questionnaire 8–10 16 factors do not explain the variation in age at onset. Three mailing and in the 2004–2006 follow-up visit were asked if loci have been nominated as modifiers of age at onset in they had been diagnosed with a major chronic disease in- 11,12 familial PD. The present study was aimed at identifying cluding PD. Participants who had been diagnosed with PD genetic modifiers for common idiopathic PD. We hypothe- before enrollment (before 1998) were designated as prevalent sized that identification of the genetic basis to interindividual PD (PAGE , N = 209), participants who were diagnosed variability in age at diagnosis will provide insights into the during follow-up (1998–2006) were designated as incident intrinsic mechanisms that determine the rate of deterioration PD (PAGE , N = 517), and participants who did not have PD during preclinical disease. were designated as controls (N = 1,549). All participants in this study were non-Hispanic whites. Methods Genotyping NGRC participants were genotyped on Illumina Human- This study was a case-control GWAS, followed by replication Omni1-Quad v1-0 B array and Immunochip array. Genotypes and functional annotation. and samples were filtered by call rate, minor allele frequency Standard protocol approvals, registrations, (MAF) < 0.01, Hardy-Weinberg, and cryptic relatedness, as and patient consents described before. Imputation was performed using IMPUTE The study was approved by the institutional review boards at v2.3.0, with the 1000G Phase3 integrated variant set all participating institutions. Written informed consent was (October 2014) as reference. Imputed single nucleotide obtained from all patients and controls for participation in the polymorphisms (SNPs) with info score < 0.9 or MAF < 0.01 study. were excluded. A total of 8.5 million SNPs (900,000 geno- typed and 7.6 million imputed) were used in the analysis. Participants The study included 2 data sets. The NeuroGenetics Research PAGE participants were genotyped for rs73656147 (block 1) Consortium (NGRC) data set was used for the discovery and rs17763929 (block 2). SNPs were chosen based on sta- GWAS, gene-based test, and functional annotations. The tistical significance and availability of predesigned validated Parkinson’s, Genes and Environment (PAGE) study was TaqMan assay from Thermo Fisher (rs73656147 assay used for replication. Participants’ characteristics are shown in number = C__97534229_10; rs17763929 assay number = table 1 and figure e-1 (links.lww.com/NXG/A66). C__34297681_10). NGRC is a case-control study of genetically unrelated par- Population structure 13 18 ticipants, including 2000 PD cases and 1986 controls. Principal component (PC) analysis is used to infer population- Patients were enrolled sequentially from movement disorder specific genetic differences, which arise from ancestry differences 2 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Table 1 Data sets and participants’ characteristics Discovery (NGRC) Replication (PAGE) PD Controls PAGE PAGE Controls P I N 2,000 1,986 209 517 1,549 Male/Female 1,346/654 769/1,217 164/45 396/121 1,213/336 Age at enrollment mean ± SD 67.3 ± 10.7 70.3 ± 14.1 62.6 ± 4.9 63.2 ± 4.9 63.4 ± 4.9 Age at follow-up mean ± SD NR NR 73.9 ± 4.9 74.5 ± 4.9 74.0 ± 4.9 N with age at onset data 1,999 NR 0 0 NR Age at onset mean ± SD 58.3 ± 11.9 NR NA NA NR N with age at diagnosis data 1,950 NR 209 517 NR Age at diagnosis range 25–90 NR 42–72 53–81 NR Age at diagnosis mean ± SD 60.4 ± 11.4 NR 59.9 ± 6.6 69.4 ± 5.4 NR Abbreviations: NA = not available; NGRC = NeuroGenetics Research Consortium; NR = not relevant; PAGE = Parkinson’s, Genes and Environment. Participants were non–Hispanic whites and genetically unrelated. Data on the NGRC participants were collected at enrollment: patients already had the diagnosis of PD and controls were free of neurologic disease. NGRC participants were enrolled at 4 sites: Oregon, Washington, New York, and Georgia. Age at onset mean ± SD were as follows: Oregon = 56.6 ± 12.8, Washington = 58.7 ± 11.8, New York = 59.4 ± 11.5, and Georgia = 58.7 ± 11.1. Age at diagnosis mean ± SD were as follows: Oregon = 59.6 ± 11.7, Washington = 60.7 ± 11.6, New York = 60.9 ± 11.1, and Georgia = 60.3 ± 10.6. PAGE participants were originally enrolled in the longitudinal NIH-AARP diet study in 1995–1997. Their PD status was investigated in 2004–2006. Participants who had the diagnosis of PD before 1998 were classified as prevalent PD (PAGE ), participants who were diagnosed with PD during follow-up (between 1998 and 2006) were classified as incident PD (PAGE ), P I and participants who did not have PD were designated as controls. Because PAGE participants were of similar age at entry, the method of classifying the participants into prevalent vs incident cases inevitably assigned earlier ages at diagnosis to the prevalent group and later diagnoses to the incident group. in allele frequencies and can obscure genetic association studies used to visualize the chr9:103,865,000–104,055,000 region if not accounted for. NGRC PC analysis was conducted using (GWAS peak). Haploview v4.2 was used to generate linkage a pruned subset of 100K SNPs from the GWAS as previously disequilibrium (LD) plots of D9 and r for SNPs in the chr9: described. The top 3 PCs (effect sizes PC1 = 0.2%, PC2 = 103,865,000–104,055,000 region with GWAS p <1E-4. LD 0.06%, and PC3 = 0.06%) were included in the GWAS and between 2 SNPs was calculated using 1000G Phase3 v5 in adjusted for in all downstream analyses involving the NGRC. LDlink. Linear regression was used to estimate and test dif- The PAGE data sets used for replication did not have ancestry ferences in mean age at diagnosis (β). Conditional analysis was informative markers (AIMs); however, a subset of the partic- performed using coxph function in the survival v2.41 R pack- ipants (396 of 726 PD cases) was previously genotyped with the age. Moving average plots (MAPs) were generated using the Immunochip array. We conducted PC analysis using a pruned freqMAP v0.2 R package. set of 20K SNPs from the Immunochip array, using PLINK. Tests were conducted once using the full PAGE data set, with no Gene-based analysis was conducted using summary statistics PC adjustment, and again with a PAGE subset, adjusting for from the GWAS and LD from the 1000G Phase3 EUR to map PC1-3 (effect sizes PC1 = 0.48%, PC2 = 0.20%, and PC3 = the GWAS SNPs to 18,985 protein-coding genes (hg19 build) 0.17%). NGRC and PAGE cluster with Europeans in the and to calculate gene-based p values, using MAGMA v1.06, 1000G_Phase_3 global data set (figure e-2, links.lww.com/ as implemented in FUMA v1.3.0. Statistical significance was NXG/A67). set at Bonferroni-corrected p < 2.6E-6 (0.05/18,985). Statistics Replication Cox regression (coxph function in the survival v2.41 R Discovery package) was used to replicate the association of 2 SNPs with GWAS was conducted using PD cases only (1,950 NGRC age at diagnosis. We used the same model as the NGRC participants with known age at diagnosis). Association between (additive genetic model, treating age at diagnosis as a quanti- 8.5M SNPs and age at diagnosis was tested using Cox re- tative trait). Because of the availability of PCs only in a subset gression in ProbABEL v0.5.0., specifying an additive genetic of PAGE, analyses were conducted twice: using the full PAGE model, treating age at diagnosis as a quantitative trait, and data set without PC adjustment and using the subset that had adjusting for PC1-3. The statistical outcome of Cox regression AIMs and adjusting for PC1-3. PAGE and PAGE were was hazard ratios (HRs) and corresponding p values. Statistical I P significance was set at p < 5E-8. Manhattan plots and quantile- treated separately and were combined using meta-analysis quantile (QQ) plots were generated using FUMA v1.3.0. after testing for heterogeneity. If p of heterogeneity was <0.1, Genomic inflation factor (λ)was calculated using the the fixed-effect model was used. Meta-analysis was performed 21 22 estlambda function in GenABEL v1.8 in R. LocusZoom was using the metagen function in the meta v4.8 R package. Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 3 whether and how allele frequencies vary by age in cases or in Functional annotation controls. Allele frequencies were plotted in a moving average Functional annotation was conducted in FUMA v1.3.0, window as a function of age (figure e-3, links.lww.com/NXG/ using SNPs with GWAS p < 1E-6 and all variants in r ≥ 0.6 27 28 A68). Starting at age 45 years, allele frequencies were the same with them, and included CADD analysis, eQTL mapping, in cases and controls. In controls, allele frequencies remained 3D chromatin interaction mapping (Hi-C), annotation of the same across the age spectrum, whereas in cases, they enhancers, tissue-specific expression of genes identified via decreased sharply and significantly by age and by age at di- Hi-C and eQTL mapping, and their age-specific expression agnosis. The effect was therefore in cases and not in controls. in the brain (BrainSpan.org). The false discovery rate (FDR) Next, conditional analysis was conducted to tease age from was used to correct for multiple testing. STRUM was used to age at diagnosis (table 2). The minor alleles of rs73656147 predict the effect of a missense on the structural stability of and rs17763929 were associated with age, as was expected, a protein. given their association with age at diagnosis. However, the association with age at diagnosis persisted when adjusted for Data availability age, but the association with age was abolished when adjusted NGRC genotype and phenotype data are available at dbGaP for age at diagnosis. Hence, age at diagnosis was the driving ncbi.nlm.nih.gov/gap accession number phs000196.v3.p1. force, and association with age was a by-product of the correlation. Results To gauge robustness of the association signals with age at diagnosis and to test for heterogeneity, we stratified the data GWAS In SNP-based GWAS, the most significant signal for associ- by 8 PD-relevant variables, tested the association of each SNP ation, at p = 5E-8, mapped to LPPR1 on chromosome 9q31.1 with age at diagnosis within each stratum, and compared the (figure 1, A and B). In the gene-based test, LPPR1 achieved results across strata for evidence of heterogeneity (table e-2, p = 1E-8, surpassing the genome-wide statistical significance links.lww.com/NXG/A70). The 8 categories of stratification threshold of p < 2.6E-6 (figure 1, C and D). The p values were were family history, sex, cigarette smoking, caffeine intake, not inflated (λ = 1.007 SNP based, λ = 1.04 gene based). nonsteroidal anti-inflammatory drugs use, recruitment site, Analysis of LD in the region revealed 2 haplotype blocks with Jewish heritage, and the European country of ancestral origin. seemingly independent signals for association (figure 1, E and The association signal for rs73656147 (block 1) was robust F). There was strong LD among SNPs in each block, but weak across all strata. rs17763929 (block 2) showed evidence of LD between the blocks (r ≤ 0.2) because of a recombination heterogeneity as a function of recruitment site and the Eu- ropean country of ancestral origin. Given these results, we hot spot between them (figure 1F). The 2 blocks were in a ;200 Kb region inside LPPR1. Block 1 consisted of 51 tested the association of the 2 SNPs with PCs. rs17763929 SNPs with MAF;0.01, which yielded HR = 2.02–1.88, with was associated with PC1 (p = 7E-6) and PC3 (p = 8E-3), and p = 9E-7 to 2E-5 for association with age at diagnosis. Block 2 rs73656147 was not (p > 0.05 for PC1-3), indicating the consisted of 39 SNPs with MAF;0.02, which yielded HR = presence of population structure in block 2 but not in block 1. 1.88–1.85, with p = 5E-8 to 7E-7. We chose 1 SNP to rep- resent each block for replication: rs73656147 for block 1 Replication (MAF = 0.01, HR = 1.95, p = 3E-6) and rs17763929 for block In comparison to NGRC, which had a 65-year range for age at 2 (MAF = 0.02, HR = 1.88, p = 5E-8), both in Hardy- diagnosis, the PAGE data sets had a narrower range of less Weinberg (p > 0.3), with little correlation between them (r = than 30 years. Because PAGE participants were of similar age 0.2). Conditional analysis conducted to determine whether at study entry, the method of classifying the participants into the 2 blocks were tagging the same or different disease- prevalent PD (diagnosis before entry) vs incident PD (di- associated variants was inconclusive because although the agnosis after entry) inevitably assigned earlier ages at di- signals were weakened when adjusted for each other, neither agnosis to the prevalent group (PAGE ) and later diagnoses was abolished when conditioned on the other (table e-1, links. to the incident group (PAGE ). Mean age at diagnosis in lww.com/NXG/A69). PAGE was 59.9 ± 6.6 years, which was similar to the NGRC (60.4 ± 11.4). PAGE participants were on average 10 years There are 2 caveats in interpreting statistical evidence for older at diagnosis (69.4 ± 5.4, range 53–81 years). Given the association with age at diagnosis. First, age at diagnosis is disparity in the range and mean ages at diagnosis, we analyzed correlated with age (r = 0.74, p < 2E-16), which can result in PAGE and PAGE separately. P I spurious conclusions if the driving force responsible for the association is not identified. Second, tests of age at diagnosis Association of rs73656147 (block 1) with age at diagnosis are conducted using patients only without the benefitof replicated robustly (table 3). There was no evidence of het- controls. For example, an SNP that appears to be associated erogeneity between PAGE and PAGE in the association of I P with earlier PD diagnosis may in fact be associated with an rs73656147 with age at diagnosis, although the signal was age-related event unrelated to PD. To interpret the statistical stronger in PAGE than in PAGE , which is not surprising, P I evidence for association with age at diagnosis, we examined given that the former is enriched in cases with earlier age at 4 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Figure 1 Results of genome-wide association study for age at diagnosis of PD Genome-wide association was tested between 8.5 million SNPs and age at diagnosis in 1,950 PD cases from the NGRC, using the Cox hazard ratio regression method and adjusting for principal components (PC1-3). (A) Manhattan plot of SNP-based GWAS. Tallest peak, at p = 5E-8, was on chromosome 9q31.1. (B) QQ plot of SNP-based GWAS. The observed p values were not inflated (λ = 1.007). (C) Manhattan plot of gene-based GWAS. LPPR1 was at p = 1E-8. Statistical significance threshold was p < 2.6E-6, which is Bonferroni corrected for the 18,985 protein-coding genes tested. (D) QQ plot of gene-based GWAS. The observed p values were not inflated (λ = 1.04). (E) r (top panel) and D’ (bottom panel). Linkage disequilibrium (LD) across the SNPs that gave p < 1E-4 for association with age at diagnosis reveals 2 blocks represented by rs73656147 (left triangle) and rs17763929 (right triangle). (F) Magnified map of the associated region (chr9:103,865,000–104,055,000), showing that PD-associated SNPs map to LPPR1 and form 2 haplotype blocks separated by recombination hot spots (blue spikes). (G) Chromatin state of LPPR1 (Roadmap 111 Epigenomes), showing that active enhancers (yellow), transcription start site (red), and transcripts (green) of LPPR1 are seen only in stem cells and the brain and that the GWAS SNPs align with regulatory elements. ESC = embryonic stem cell; iPSC = induced pluripotent stem cell; TssA = active transcription start site (TSS); TssAFlnk = flanking active TSS; TxFlnk = transcription at gene 59 and 3’; Tx = strong transcription; TxWk = weak transcription; EnhG = genic enhancers; Enh = enhancers; ZNF/Rpts = zinc-finger genes and repeats; Het = heterochromatin; TssBiv = bivalent/poised TSS; BivFlnk = flanking bivalent TSS/enhancer; EnhBiv = bivalent enhancer; ReprPC = repressed polycomb; ReprPCWk = weak repressed polyComb; Quies = quiescent. Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 5 Table 2 Association of LPPR1 variants with age and age at diagnosis is driven by age at diagnosis Block 1 Block 2 rs73656147 rs17763929 Cox LR Cox LR N HR p Value β [95% CI] HR p Value β [95% CI] Ia. Association with age at diagnosis in cases 1,950 1.95 3E-6 −6.00 [−9.18 to −2.83] 1.88 5E-8 −5.65 [−8.20 to −3.11] Ib. Association with age at diagnosis in 1,950 1.95 3E-6 −5.98 [−9.16 to −2.81] 1.88 6E-8 −5.61 [−8.16 to −3.07] cases adjusted for sex II. Association with age in cases 2,000 1.48 5E-3 −4.19 [−7.1 to −1.3] 1.53 2E-4 −3.56 [−5.9 to −1.2] III. Association with age in controls 1,986 0.83 0.08 2.34 [−0.6 to 5.2] 0.84 0.07 2.37 [−0.3 to 5.1] IV. Association with age at diagnosis in 1,950 1.45 0.01 −2.30 [−3.9 to −0.7] 1.26 0.05 −2.11 [−3.4 to −0.8] cases adjusted for age V. Association with age in cases adjusted 1,950 0.92 0.56 0.78 [−0.8 to 2.3] 0.99 0.96 0.68 [−0.6 to 1.9] for age at diagnosis Abbreviations: CI = confidence interval; HR = hazard ratio; LR = linear regression; β = effect size on age at diagnosis (in years) per copy of minor allele. The associations were tested in the NGRC data set using Cox regression, and the effect sizes were estimated using linear regression (LR). HR is the age-for-age increase in the odds of event per copy of the minor allele, as estimated using Cox regression. β is the difference in years in age at diagnosis between carriers of 1 minor allele vs no minor allele, as estimated using linear regression. Age at diagnosis was the primary outcome of the study. Minor alleles of rs73656147 and rs17763929 were associated with higher HR and younger age at diagnosis (Ia). The association was not influenced by sex (Ib), which was expected because, unlike PD risk, which is significantly associated with sex (OR = 3.26, p < 2E-16), age at diagnosis is not associated with sex (HR = 0.99, p = 0.83). Minor alleles were also associated with younger ages in cases (II), but not in controls (III). Because age and age at diagnosis were correlated (r = 0.74, p < 2E-16), an association with one will show as an association with both. In conditional analysis, the association with age at diagnosis persisted when adjusted for age (IV), butthe association with age was abolished when adjusted for age at diagnosis (V), suggesting that age at diagnosis was the driving force and association with age was a by-product of the correlation. diagnosis. Nor was there evidence of heterogeneity between mapped to enhancers in the brain (table 4 and figure 1, G). PAGE and NGRC for the association of rs73656147 with age Eleven of the genes identified through Hi-C are expressed in at diagnosis. Meta-analysis yielded HR = 2.14, p = 1E-3 for the brain: LPPR1, SEC61B, MSANTD3-TMEFF1, TMEFF1, replication and HR = 2.00, p = 9E-9 for replication and dis- GALNT12, MURC, GRIN3A, NR4A3, ALG2, MRPL50, and covery. Mean difference in age at diagnosis per copy of ZNF189 (figure 2, B and C). The expression of LPPR1 in the rs73656147 minor allele was −6.0 (95% confidence interval: brain is the strongest in early prenatal stage and decreases with −9.18 to −2.83) years in the NGRC, −5.53 (−9.72 to −1.34) in developmental stage and increasing age (figure 2, C). PAGE , −0.84 (−4.22 to 2.55) in PAGE , and −4.08 (−7.45 to P I −0.70) in the meta-analysis of the 3 data sets. CADD analysis, a scoring system for deleteriousness of ge- netic variants, identified 5 SNPs in block 1 and 3 in block 2, Association of rs17763929 (block 2) with age at diagnosis with CADD = 10–19 (table 4), which places them among the showed significant heterogeneity between PAGE and PAGE top 10% (CADD > 10) to 1% (CADD > 20) of most dele- I P (table 3), as it had within the NGRC (table e-2, links.lww. terious mutations in the genome. rs41296085 (CADD = 18, com/NXG/A70). The association with rs17763929 repli- in block 1) is a missense (p.Ser12Ala) in exon 2, predicted to cated in PAGE but not in PAGE . There was significant structurally destabilize the LPPR1 protein (ΔΔG= −1.2). The P I heterogeneity between PAGE and NGRC, but not between remainder of the variants with high CADD scores are in PAGE and NGRC. Meta-analysis of PAGE and NGRC introns. eQTL analysis revealed an association between P P yielded HR = 1.88, p = 4E-9 for full PAGE data and HR = 1.95, rs117451395 (block 1) with expression levels of GRIN3A p = 3E-9 for the PAGE subsample adjusted for PC1-3. In- (FDR = 4E-4). cluding PAGE with PAGE and NGRC in a random-effects I P meta-analysis diluted the effect size to HR = 1.53, p = 0.04. Mean difference in age at diagnosis per copy of rs17763929 Discussion minor allele was −5.65 (−8.20 to −3.11) years in the NGRC, There has been intense research on PD risk factors, which so −3.62 (−7.23 to −0.02) in PAGE , and 0.62 (−1.34 to 2.58) in far has resulted in identification of numerous causative genes, PAGE . 40 susceptibility loci, several environmental factors, and a few Functional annotation genes that interact with the environmental factors to increase Hi-C analysis showed significant (FDR < 1E-6) chromatin or reduce the risk of developing PD. In contrast, we know interaction between the PD-associated LPPR1 SNPs and little about factors that affect the rate of disease progression. promoters of LPPR1 and several genes on chromosome 9 In this study, we attempted to identify genetic modifiers of age (figure 2, A). Some of the SNPs that were significant in Hi-C at diagnosis, a reflection of rate of progression, using an 6 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Table 3 Replication Age at diagnosis Block 1 rs73656147 Block 2 rs17763929 Data sets N PD cases Mean ± SD HR p Value HR p Value NGRC (discovery) 1,950 60.4 ± 11.4 1.95 3E-6 1.88 5E-8 PAGE (replication) 209 59.9 ± 6.6 2.88 7E-4 1.87 0.01 PAGE with PC1-3 113 59.9 ± 6.8 2.17 0.05 3.03 4E-3 PAGE (replication) 517 69.4 ± 5.4 1.62 0.07 1.04 0.41 PAGE with PC1-3 283 69.2 ± 5.3 1.48 0.16 1.03 0.45 Meta-analysis A Heterogeneity rs73656147 Heterogeneity rs17763929 PAGE and PAGE ns 0.08 2.14 1E-3 1.34 0.31 P I NGRC and PAGE ns ns 2.08 2E-8 1.88 4E-9 NGRC and PAGE ns 0.01 1.90 9E-7 1.42 0.23 NGRC and PAGE and PAGE ns 0.02 2.00 9E-9 1.53 0.04 P I Meta-analysis B PAGE and PAGE ns 0.02 1.73 0.07 1.67 0.34 P I NGRC and PAGE ns ns 1.97 6E-7 1.95 3E-9 NGRC and PAGE ns 0.02 1.89 2E-6 1.43 0.23 NGRC and PAGE & PAGE ns 0.02 1.91 5E-7 1.68 0.05 P I Abbreviations: HR = hazard ratio; NGRC = NeuroGenetics Research Consortium; ns = not statistically significant; PAGE = Parkinson’s, Genes and Environment; PC = principal component; PD = Parkinson disease. Two SNPs with signals for association with age at diagnosis of PD in the NGRC data set (discovery) were genotyped and tested for association with age at diagnosis of PD in the PAGE data set (replication). PAGE participants were designated as PAGE if they were diagnosed before study entry or PAGE if they were P I diagnosed during the study. Cox regression was used to test the association of SNP (additive model) with age at diagnosis (quantitative trait) and to calculate hazard ratios (HRs) and the corresponding significance (p). NGRC was adjusted for PC1-3 in GWAS and meta-analyses. Only a subset of PAGE had ancestry informative markers (AIMs) for which PC could be calculated; thus, results are shown for the full PAGE data set without PC adjustment and for the PAGE subsample with PC adjustment. p values are 2-sided for NGRC and one-sided for PAGE because of the directionality of the hypothesis being replicated. Meta-analysis A: NGRC (PC1-3 adjusted) and PAGE (all data without PC adjustment). Meta-analysis B: NGRC (PC1-3 adjusted) and PAGE (subset of data adjusted for PC1-3). rs73656147 replicated robustly with no evidence of heterogeneity across data sets. rs17763929 replicated in PAGE and showed significant heterogeneity between PAGE and PAGE or NGRC. I P Meta-analysis was conducted using the fixed-effects model if there was no evidence for heterogeneity (p ≥ 0.1) and the random effects model if there was heterogeneity (p < 0.1). unbiased genome-wide approach, followed by independent not fully capture the complexity of block 2. PAGE participants replication, and functional annotation. being significantly older than NGRC and PAGE participants 32–34 may also be a factor. LPPR1 promotes neuroregeneration, We uncovered evidence for association of genetic variants in but its expression diminishes with age to nearly undetectable neuronal plasticity-related gene 3 (LPPR1) with age at di- level by age 40 years (figure 2C). One can speculate that some agnosis of PD. Two signals of association were detected, each detrimental variants may not have an effect after a certain age representing a haplotype block of SNPs. The variants that when the gene is no longer expressed. were associated with earlier age at diagnosis had low allele frequencies (MAF = 0.01–0.02), as were the variants that Functional annotation of the PD-associated variants in LPPR1 were previously found for age at onset of familial PD. The revealed the presence of several variants with predicted del- low allele frequencies may be one reason why modifier genes eterious effects, including a missense that destabilizes the have been more difficult to detect than common variants that structure of LPPR1, a regulatory element that associates with associate with risk. expression levels of GRIN3A, and enhancers that interact with promoters of LPPR1 and several other genes in the brain. The association with block 1 replicated robustly in both PAGE Some of the candidate genes that were identified via in- and PAGE . Block 2 signal replicated in PAGE but not in teraction with LPPR1 play key roles in pathways that are I P PAGE . Block 2 has a complex LD structure, with evidence of implicated in PD, including GRIN3A (which encodes a sub- population substructure, which limits generalizability of results. unit of NMDA receptor involved in the glutamate-regulated Failure to capture a signal for block 2 in PAGE may be because ion channels in the brain), SEC61B (protein transport appa- we had genotype on only 1 SNP in block 2 for PAGE, which did ratus of the endoplasmic reticulum membrane), MURC (Rho Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 7 Figure 2 Functionally significant genes (A) 3D chromatin interaction (Hi-C) and eQTL analysis. Hi-C revealed significant interaction between GWAS variants in LPPR1 and 17 other genes on chromosome 9 (FDR < 1E-6, shown in orange). An SNP in LPPR1 was associated with the expression of GRIN3A (FDR = 4E-4, shown in green). (B) Tissue-specific expression of LPPR1, GRIN3A, and genes in Hi-C with LPPR1. Colors reflect average expression (log2 transformed) from highest (red) to lowest/absent (blue). (C) Age-specific expression of the genes in the brain. LPPR1 expression decreases with age. kinase signaling), and MRPL50 (mitochondrial ribosomal we propose that LPPR1 is involved, not necessarily in the protein). cause of PD, rather in response to damage, and influences the efficacy of regeneration and the subsequent rate of de- LPPR1 is one of the 5 members of a brain-specific gene family terioration in preclinical PD. The actual cause of injury and that modulates neuronal plasticity during development, aging, neuronal death is not stipulated in this hypothesis; it could be 32–34 and after brain injury. LPPR1 is the strongest driver of head trauma, environmental toxins or genetic, but once the axonal outgrowth in the gene family. Studies in mice have initial damage is incurred, it is the efficacy of intrinsic mech- shown that after neuronal injury, overexpression of LPPR1 anisms of repair that determine the rate of disease pro- enhances axonal growth, improves motor behavior, and pro- gression. Present findings provide a strong foundation for 33,34 motes functional recovery. Extrapolating to our findings, mechanistic studies to investigate the role of LPPR1 in PD and 8 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Table 4 Functionally significant variants Block GWAS SNP position:alleles GWASpr eQTL CADD Hi-C/EnhBrain 1 rs77351585 9:103874925:C:T 2E-06 1 — 18 Hi-C/EnhBrain 1 rs73495940 9:103875807:G:C 9E-07 Lead —— Hi-C 1 rs150164200 9:103875896:A:C 2E-06 1 — 10.4 — 1 rs117583993 9:103876647:G:A 3E-06 1 —— Hi-C/EnhBrain 1 rs148874623 9:103939117:A:C 9E-06 1 — 12.1 — 1 rs117451395 9:103941039:C:T 1E-05 1 GRIN3A — Hi-C 1 rs41296085 9:103947810:T:G 2E-05 1 — 18 (missense) Hi-C 1 rs117900237 9:103959240:G:A 2E-05 1 — 10.5 Hi-C 2 rs17763929 9:103984900:A:G 5E-08 Lead —— Hi-C 2 rs61188842 9:103988006:C:T 8E-05 0.6 —— Hi-C/EnhBrain 2 rs117058418 9:104011717:T:C 2E-07 1 — 10.4 Hi-C 2 rs117314512 9:104014244:G:A 2E-07 1 — 12.4 Hi-C 2 rs149155028 9:104032402:TTC:T 1E-05 0.7 — 18.6 Hi-C Functional annotation was conducted on SNPs with GWAS p < 1E-6 and SNPs that were in high LD with them (r > 0.6). Variants are shown if they are the lead SNPs (most significant) for the block, or an eQTL (FDR = 4E-4), or had a CADD score >10, or had both significant evidence for 3D chromatin interaction (Hi-C, FDR < 1E-6) and overlapped with an enhancer in the brain. Block 1 is a single block of SNPs in high LD. Block 2 has a complex LD structure with at least 3 subhaplotypes (figure e1-C, links.lww.com/NXG/A66). Variants are shown with their rs accession number, chromosome position and the 2 alleles (major: minor), GWAS p value for association with age at diagnosis of PD, and their correlation (r ) with the lead SNP of the block. eQTL: an SNP that is associated with gene expression, in this case, rs117451395, was associated with gene expression levels at GRIN3A (FDR = 4E-4). CADD: a predictive score for the deleteri- ousness of a variant. A CADD score of 10 usually means that the variant is among the top 10% of deleterious mutations in the genome. A CADD score of 20 puts the variant among the top 1% of deleterious mutations. Hi-C: SNPs with significant (FDR < 1E-6) evidence for interacting with the promoter region of LPPR1 or of another gene (figure 2 for the genes). Hi-C/EnhBrain: the subset of Hi-C SNPs that map to the enhancer regions of LPPR1 in the brain according to the Roadmap 111 epigenomes. a 2 One SNP was shown to represent several variants in high LD (r > 0.9) with similar MAF, GWAS p value, and Hi-C/EnhBrain evidence. This mutation yielded ΔΔG= −1.2, which predicts a destabilizing effect on the protein structure of LPPR1. determine its potential as a therapeutic target to impede data for replication (Parkinson, Genes, and Environment disease progression. (PAGE) study) was supported by the intramural research program of NIH National Institute of Environmental Health Author contributions Sciences grant Z01 ES101986. Funding agencies did not have Z.D. Wallen: statistical analysis, review, and critique of the a role in the design or execution of the study. manuscript. H. Chen: creation of the PAGE data set, acqui- sition of data, review, and critique of the manuscript. E.M. Disclosure Hill-Burns: assembly and QC of phenotype and genotype Z.D. Wallen, H. Chen, C.P. Zabetian, and E.M. Hill-Burns data, imputation, review, and critique of the manuscript. S.A. report no disclosures. S.A. Factor has received honoraria from Factor and C.P. Zabetian: creation of the NGRC data set, Neurocrine, Lundbeck, Teva, Avanir, Sunovion Pharmaceut- review, and critique of the manuscript. H. Payami: study icals, Adamas, and UCB; has received research support from concept, design and execution, creation of the NGRC data set, Ipsen, Medtronic, Teva, US WorldMeds, Sunovion Pharma- wrote the manuscript, and obtained funding. ceuticals, Solstice, Vaccinex, Voyager, the CHDI Foundation, the Michael J. Fox Foundation, and the NIH; and receives Study funding publishing royalties from Demos, Blackwell Futura, and This work was supported by the National Institute of Neuro- UpToDate. H. Payami has received research support from the logical Disorders and Stroke grant R01NS036960. Additional NIH and the University of Alabama at Birmingham. Full dis- support was provided by a Merit Review Award from the De- closure form information provided by the authors is available partment of Veterans Affairs grant 1I01BX000531; Office of with the full text of this article at Neurology.org/NG. Research & Development, Clinical Sciences Research & De- Received February 13, 2018. Accepted in final form June 11, 2018. velopment Service, Department of Veteran Affairs; the Close to the Cure Foundation; and the Sartain Lanier Family Founda- tion. Genome-wide array genotyping was conducted by the References 1. Obeso JA, Stamelou M, Goetz CG, et al. Past, present, and future of Parkinson’s Center for Inherited Disease Research, which is funded by the disease: a special essay on the 200th Anniversary of the Shaking Palsy. Mov Disord NIH grant HHSN268200782096C. Collection of samples and 2017;32:1264–1310. Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 9 2. Zareparsi S, Taylor TD, Harris EL, Payami H. Segregation analysis of Parkinson 18. Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D. Principal components analysis corrects for stratification in genome-wide association studies. disease. Am J Med Genet 1998;80:410–417. Nat Genet 2006;38:904–909. 3. Maher NE, Currie LJ, Lazzarini AM, et al. Segregation analysis of Parkinson disease 19. Aulchenko YS, Struchalin MV, van Duijn CM. ProbABEL package for genome-wide revealing evidence for a major causative gene. Am J Med Genet 2002;109:191–197. association analysis of imputed data. BMC Bioinformatics 2010;11:134. 4. McDonnell SK, Schaid DJ, Elbaz A, et al. Complex segregation analysis of Parkinson’s 20. Watanabe K, Taskesen E, van Bochoven A, Posthuma D. Functional mapping and disease: the Mayo clinic family study. Ann Neurol 2006;59:788–795. annotation of genetic associations with FUMA. Nat Commun 2017;8:1826. 5. Hamza TH, Payami H. The heritability of risk and age at onset of Parkinson’s disease 21. Aulchenko YS, Ripke S, Isaacs A, van Duijn CM. GenABEL: an R library for genome- after accounting for known genetic risk factors J Hum Genet 2010;55:241–243. wide association analysis. Bioinformatics 2007;23:1294–1296. 6. Escott-Price V, Consortium IPsDG, Nalls MA, et al. Polygenic risk of Parkinson 22. Pruim RJ, Welch RP, Sanna S, et al. LocusZoom: regional visualization of genome- disease is correlated with disease age at onset. Ann Neurol 2015;77:582–591. wide association scan results. Bioinformatics 2010;26:2336–2337. 7. Chang D, Nalls MA, Hallgrimsdottir IB, et al. A meta-analysis of genome-wide as- 23. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and sociation studies identifies 17 new Parkinson’s disease risk loci. Nat Genet 2017;49: haplotype maps. Bioinformatics 2005;21:263–265. 1511–1516. 24. Machiela MJ, Chanock SJ. LDlink: a web-based application for exploring population- 8. Nalls MA, Escott-Price V, Williams NM, et al. Genetic risk and age in Parkinson’s specific haplotype structure and linking correlated alleles of possible functional var- disease: continuum not stratum. Mov Disord 2015;30:850–854. iants. Bioinformatics 2015;31:3555–3557. 9. Lill CM, Hansen J, Olsen JH, Binder H, Ritz B, Bertram L. Impact of Parkinson’s 25. Payami H, Kay DM, Zabetian CP, Schellenberg GD, Factor SA, McCulloch CC. disease risk loci on age at onset. Mov Disord 2015;30:847–850. Visualizing disease associations: graphic analysis of frequency distributions as a func- 10. Pihlstrom L, Toft M. Cumulative genetic risk and age at onset in Parkinson’s disease. tion of age using moving average plots (MAP) with application to Alzheimer’s and Mov Disord 2015;30:1712–1713. Parkinson’s disease. Genet Epidemiol 2010;34:92–99. 11. Hill-Burns EM, Ross OA, Wissemann WT, et al. Identification of genetic modifiers of 26. de Leeuw CA, Mooij JM, Heskes T, Posthuma D. MAGMA: generalized gene-set age-at-onset for familial Parkinson’s disease. Hum Mol Genet 2016;25:3849–3862. analysis of GWAS data. PLoS Comput Biol 2015;11:e1004219. 12. Trinh J, Gustavsson EK, Vilarino-Guell C, et al. DNM3 and genetic modifiers of age of 27. Kircher M, Witten DM, Jain P, O’Roak BJ, Cooper GM, Shendure J. A general framework for estimating the relative pathogenicity of human genetic variants. Nat onset in LRRK2 Gly2019Ser parkinsonism: a genome-wide linkage and association Genet 2014;46:310–315. study. Lancet Neurol 2016;15:1248–1256. 28. Consortium G. Human genomics. The Genotype-Tissue Expression (GTEx) pilot 13. Hamza TH, Zabetian CP, Tenesa A, et al. Common genetic variation in the HLA analysis: multitissue gene regulation in humans. Science 2015;348:648–660. region is associated with late-onset sporadic Parkinson’s disease. Nat Genet 2010;42: 29. van Berkum NL, Lieberman-Aiden E, Williams L, et al. Hi-C: a method to study the 781–785. three-dimensional architecture of genomes. J Vis Exp 2010:e-1869. 14. Chen H, Huang X, Guo X, et al. Smoking duration, intensity, and risk of Parkinson 30. Consortium RE, Kundaje A, Meuleman W, et al. Integrative analysis of 111 reference disease. Neurology 2010;74:878–884. human epigenomes. Nature 2015;518:317–330. 15. Hughes AJ, Daniel SE, Kilford L, Lees AJ. Accuracy of clinical diagnosis of idiopathic 31. Quan L, Lv Q, Zhang Y. STRUM: structure-based prediction of protein stability Parkinson’s disease: a clinico-pathological study of 100 cases. J Neurol Neurosurg changes upon single-point mutation. Bioinformatics 2016;32:2936–2946. Psychiatry 1992;55:181–184. 32. Savaskan NE, Brauer AU, Nitsch R. Molecular cloning and expression regulation of PRG- 16. Schatzkin A, Subar AF, Thompson FE, et al. Design and serendipity in establishing 3, a new member of the plasticity-related gene family. Eur J Neurosci 2004;19:212–220. a large cohort with wide dietary intake distributions: the National Institutes of Health- 33. Broggini T, Schnell L, Ghoochani A, et al. Plasticity related gene 3 (PRG3) overcomes American Association of retired persons diet and Health study. Am J Epidemiol 2001; myelin-associated growth inhibition and promotes functional recovery after spinal 154:1119–1125. cord injury. Aging (Albany NY) 2016;8:2463–2487. 17. Howie BN, Donnelly P, Marchini J. A flexible and accurate genotype imputation 34. Fink KL, Lopez-Giraldez F, Kim IJ, Strittmatter SM, Cafferty WBJ. Identification of method for the next generation of genome-wide association studies. PLoS Genet intrinsic axon growth modulators for intact CNS neurons after injury. Cell Rep 2017; 2009;5:e1000529. 18:2687–2701. 10 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Neurology: Genetics Pubmed Central

Plasticity-related gene 3 (LPPR1) and age at diagnosis of Parkinson disease

Loading next page...
 
/lp/pubmed-central/plasticity-related-gene-3-lppr1-and-age-at-diagnosis-of-parkinson-UcKGxmyi7A

References (37)

Publisher
Pubmed Central
Copyright
Copyright © 2018 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology.
ISSN
2376-7839
eISSN
2376-7839
DOI
10.1212/NXG.0000000000000271
Publisher site
See Article on Publisher Site

Abstract

Objective To identify modifiers of age at diagnosis of Parkinson disease (PD). Methods Genome-wide association study (GWAS) included 1,950 individuals with PD from the Neu- roGenetics Research Consortium (NGRC) study. Replication was conducted in the Parkin- son’s, Genes and Environment study, including 209 prevalent (PAGE ) and 517 incident (PAGE ) PD cases. Cox regression was used to test association with age at diagnosis. Indi- viduals without neurologic disease were used to rule out confounding. Gene-level analysis and functional annotation were conducted using Functional Mapping and Annotation of GWAS platform (FUMA). Results The GWAS revealed 2 linked but seemingly independent association signals that mapped to LPPR1 on chromosome 9. LPPR1 was significant in gene-based analysis (p = 1E-8). The top signal (rs17763929, hazard ratio [HR] = 1.88, p = 5E-8) replicated in PAGE (HR = 1.87, p = 0.01) but not in PAGE . The second signal (rs73656147) was robust with no evidence of heterogeneity (HR = 1.95, p = 3E-6 in NGRC; HR = 2.14, p = 1E-3 in PAGE + PAGE , and HR P I = 2.00, p = 9E-9 in meta-analysis of NGRC + PAGE + PAGE ). The associations were with age P I at diagnosis, not confounded by age in patients or in the general population. The PD-associated regions included variants with Combined Annotation Dependent Depletion (CADD) scores = 10–19 (top 1%–10% most deleterious mutations in the genome), a missense with predicted destabilizing effect on LPPR1, an expression quantitative trait locus (eQTL) for GRIN3A (false discovery rate [FDR] = 4E-4), and variants that overlap with enhancers in LPPR1 and interact with promoters of LPPR1 and 9 other brain-expressed genes (Hi-C FDR < 1E-6). Conclusions Through association with age at diagnosis, we uncovered LPPR1 as a modifier gene for PD. LPPR1 expression promotes neuronal regeneration after injury in animal models. Present data provide a strong foundation for mechanistic studies to test LPPR1 as a driver of response to damage and a therapeutic target for enhancing neuroregeneration and slowing disease progression. From the Department of Neurology (Z.D.W., E.M.H.-B., H.P.), University of Alabama at Birmingham, Birmingham, AL; Department of Epidemiology and Biostatistics (H.C.), Michigan State University, East Lansing, MI; Department of Neurology (S.A.F.), Jean & Paul Amos Parkinson’s Disease and Movement Disorder Program, Emory University School of Medicine, Atlanta, GA; VA Puget Sound Health Care System and Department of Neurology (C.P.Z.), University of Washington, Seattle, WA; and Center for Genomic Medicine (H.P.), HudsonAlpha Institute for Biotechnology, Huntsville, AL. Funding information and disclosures are provided at the end of the article. Full disclosure form information provided by the authors is available with the full text of this article at Neurology.org/NG. The Article Processing Charge was funded by the authors. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal. Copyright © 2018 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary AIM = ancestry informative marker; FDR = false discovery rate; HR = hazard ratio; LD = linkage disequilibrium; MAF = minor allele frequency; MAP = moving average plot; NGRC = NeuroGenetics Research Consortium; PAGE = Parkinson’s, Genes and Environment; PC = principal component; PD = Parkinson disease. The underlying neurodegenerative process that causes Par- clinics in Portland (OR), Seattle (WA), Albany (NY), and kinson disease (PD) begins decades before the disease is di- Atlanta (GA). Controls were spouses of patients or commu- agnosed. The current view is that following an initial insult nity volunteers, self-reported as being free of neurologic dis- (e.g., toxicity, trauma, or genetic), the disease starts with an ease. The eligibility criterion for cases was diagnosis of PD by asymptomatic phase of unknown duration, followed by de- a movement disorder specialist according to the UK Brain velopment of prodromal nonmotor symptoms such as con- Bank criteria. The eligibility criteria for controls were no stipation, anosmia, and sleep disorders. Years later, cardinal neurologic disease and genetically unrelated to patients. Age motor signs appear, at which point a diagnosis of PD is made. was defined as age at study entry. Age at diagnosis was Age at onset of motor signs, and therefore the age at diagnosis extracted from medical records or ascertained by self-report. of PD, is highly variable, ranging from teen ages to the 10th Age at onset of the first motor sign was obtained using a self- decade of life. The reason for this variation is unknown, and administered questionnaire. Age at onset and age at diagnosis understanding it will likely shed light on factors that affect the were highly correlated in the NGRC (r = 0.91, p < 2E-16). All rate of disease progression. participants were whites of European descent. There is substantial evidence that genetic factors play a major PAGE is a cross-sectional study nested in the longitudinal role in age at onset of motor signs and age at diagnosis of NIH-American Association of Retired Persons Diet and 2–6 14 PD. Genome-wide studies have identified numerous loci Health Study. Participants were enrolled in 1995–1997 that associate with the risk of developing PD, but the risk (irrespective of PD) via a food frequency questionnaire 8–10 16 factors do not explain the variation in age at onset. Three mailing and in the 2004–2006 follow-up visit were asked if loci have been nominated as modifiers of age at onset in they had been diagnosed with a major chronic disease in- 11,12 familial PD. The present study was aimed at identifying cluding PD. Participants who had been diagnosed with PD genetic modifiers for common idiopathic PD. We hypothe- before enrollment (before 1998) were designated as prevalent sized that identification of the genetic basis to interindividual PD (PAGE , N = 209), participants who were diagnosed variability in age at diagnosis will provide insights into the during follow-up (1998–2006) were designated as incident intrinsic mechanisms that determine the rate of deterioration PD (PAGE , N = 517), and participants who did not have PD during preclinical disease. were designated as controls (N = 1,549). All participants in this study were non-Hispanic whites. Methods Genotyping NGRC participants were genotyped on Illumina Human- This study was a case-control GWAS, followed by replication Omni1-Quad v1-0 B array and Immunochip array. Genotypes and functional annotation. and samples were filtered by call rate, minor allele frequency Standard protocol approvals, registrations, (MAF) < 0.01, Hardy-Weinberg, and cryptic relatedness, as and patient consents described before. Imputation was performed using IMPUTE The study was approved by the institutional review boards at v2.3.0, with the 1000G Phase3 integrated variant set all participating institutions. Written informed consent was (October 2014) as reference. Imputed single nucleotide obtained from all patients and controls for participation in the polymorphisms (SNPs) with info score < 0.9 or MAF < 0.01 study. were excluded. A total of 8.5 million SNPs (900,000 geno- typed and 7.6 million imputed) were used in the analysis. Participants The study included 2 data sets. The NeuroGenetics Research PAGE participants were genotyped for rs73656147 (block 1) Consortium (NGRC) data set was used for the discovery and rs17763929 (block 2). SNPs were chosen based on sta- GWAS, gene-based test, and functional annotations. The tistical significance and availability of predesigned validated Parkinson’s, Genes and Environment (PAGE) study was TaqMan assay from Thermo Fisher (rs73656147 assay used for replication. Participants’ characteristics are shown in number = C__97534229_10; rs17763929 assay number = table 1 and figure e-1 (links.lww.com/NXG/A66). C__34297681_10). NGRC is a case-control study of genetically unrelated par- Population structure 13 18 ticipants, including 2000 PD cases and 1986 controls. Principal component (PC) analysis is used to infer population- Patients were enrolled sequentially from movement disorder specific genetic differences, which arise from ancestry differences 2 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Table 1 Data sets and participants’ characteristics Discovery (NGRC) Replication (PAGE) PD Controls PAGE PAGE Controls P I N 2,000 1,986 209 517 1,549 Male/Female 1,346/654 769/1,217 164/45 396/121 1,213/336 Age at enrollment mean ± SD 67.3 ± 10.7 70.3 ± 14.1 62.6 ± 4.9 63.2 ± 4.9 63.4 ± 4.9 Age at follow-up mean ± SD NR NR 73.9 ± 4.9 74.5 ± 4.9 74.0 ± 4.9 N with age at onset data 1,999 NR 0 0 NR Age at onset mean ± SD 58.3 ± 11.9 NR NA NA NR N with age at diagnosis data 1,950 NR 209 517 NR Age at diagnosis range 25–90 NR 42–72 53–81 NR Age at diagnosis mean ± SD 60.4 ± 11.4 NR 59.9 ± 6.6 69.4 ± 5.4 NR Abbreviations: NA = not available; NGRC = NeuroGenetics Research Consortium; NR = not relevant; PAGE = Parkinson’s, Genes and Environment. Participants were non–Hispanic whites and genetically unrelated. Data on the NGRC participants were collected at enrollment: patients already had the diagnosis of PD and controls were free of neurologic disease. NGRC participants were enrolled at 4 sites: Oregon, Washington, New York, and Georgia. Age at onset mean ± SD were as follows: Oregon = 56.6 ± 12.8, Washington = 58.7 ± 11.8, New York = 59.4 ± 11.5, and Georgia = 58.7 ± 11.1. Age at diagnosis mean ± SD were as follows: Oregon = 59.6 ± 11.7, Washington = 60.7 ± 11.6, New York = 60.9 ± 11.1, and Georgia = 60.3 ± 10.6. PAGE participants were originally enrolled in the longitudinal NIH-AARP diet study in 1995–1997. Their PD status was investigated in 2004–2006. Participants who had the diagnosis of PD before 1998 were classified as prevalent PD (PAGE ), participants who were diagnosed with PD during follow-up (between 1998 and 2006) were classified as incident PD (PAGE ), P I and participants who did not have PD were designated as controls. Because PAGE participants were of similar age at entry, the method of classifying the participants into prevalent vs incident cases inevitably assigned earlier ages at diagnosis to the prevalent group and later diagnoses to the incident group. in allele frequencies and can obscure genetic association studies used to visualize the chr9:103,865,000–104,055,000 region if not accounted for. NGRC PC analysis was conducted using (GWAS peak). Haploview v4.2 was used to generate linkage a pruned subset of 100K SNPs from the GWAS as previously disequilibrium (LD) plots of D9 and r for SNPs in the chr9: described. The top 3 PCs (effect sizes PC1 = 0.2%, PC2 = 103,865,000–104,055,000 region with GWAS p <1E-4. LD 0.06%, and PC3 = 0.06%) were included in the GWAS and between 2 SNPs was calculated using 1000G Phase3 v5 in adjusted for in all downstream analyses involving the NGRC. LDlink. Linear regression was used to estimate and test dif- The PAGE data sets used for replication did not have ancestry ferences in mean age at diagnosis (β). Conditional analysis was informative markers (AIMs); however, a subset of the partic- performed using coxph function in the survival v2.41 R pack- ipants (396 of 726 PD cases) was previously genotyped with the age. Moving average plots (MAPs) were generated using the Immunochip array. We conducted PC analysis using a pruned freqMAP v0.2 R package. set of 20K SNPs from the Immunochip array, using PLINK. Tests were conducted once using the full PAGE data set, with no Gene-based analysis was conducted using summary statistics PC adjustment, and again with a PAGE subset, adjusting for from the GWAS and LD from the 1000G Phase3 EUR to map PC1-3 (effect sizes PC1 = 0.48%, PC2 = 0.20%, and PC3 = the GWAS SNPs to 18,985 protein-coding genes (hg19 build) 0.17%). NGRC and PAGE cluster with Europeans in the and to calculate gene-based p values, using MAGMA v1.06, 1000G_Phase_3 global data set (figure e-2, links.lww.com/ as implemented in FUMA v1.3.0. Statistical significance was NXG/A67). set at Bonferroni-corrected p < 2.6E-6 (0.05/18,985). Statistics Replication Cox regression (coxph function in the survival v2.41 R Discovery package) was used to replicate the association of 2 SNPs with GWAS was conducted using PD cases only (1,950 NGRC age at diagnosis. We used the same model as the NGRC participants with known age at diagnosis). Association between (additive genetic model, treating age at diagnosis as a quanti- 8.5M SNPs and age at diagnosis was tested using Cox re- tative trait). Because of the availability of PCs only in a subset gression in ProbABEL v0.5.0., specifying an additive genetic of PAGE, analyses were conducted twice: using the full PAGE model, treating age at diagnosis as a quantitative trait, and data set without PC adjustment and using the subset that had adjusting for PC1-3. The statistical outcome of Cox regression AIMs and adjusting for PC1-3. PAGE and PAGE were was hazard ratios (HRs) and corresponding p values. Statistical I P significance was set at p < 5E-8. Manhattan plots and quantile- treated separately and were combined using meta-analysis quantile (QQ) plots were generated using FUMA v1.3.0. after testing for heterogeneity. If p of heterogeneity was <0.1, Genomic inflation factor (λ)was calculated using the the fixed-effect model was used. Meta-analysis was performed 21 22 estlambda function in GenABEL v1.8 in R. LocusZoom was using the metagen function in the meta v4.8 R package. Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 3 whether and how allele frequencies vary by age in cases or in Functional annotation controls. Allele frequencies were plotted in a moving average Functional annotation was conducted in FUMA v1.3.0, window as a function of age (figure e-3, links.lww.com/NXG/ using SNPs with GWAS p < 1E-6 and all variants in r ≥ 0.6 27 28 A68). Starting at age 45 years, allele frequencies were the same with them, and included CADD analysis, eQTL mapping, in cases and controls. In controls, allele frequencies remained 3D chromatin interaction mapping (Hi-C), annotation of the same across the age spectrum, whereas in cases, they enhancers, tissue-specific expression of genes identified via decreased sharply and significantly by age and by age at di- Hi-C and eQTL mapping, and their age-specific expression agnosis. The effect was therefore in cases and not in controls. in the brain (BrainSpan.org). The false discovery rate (FDR) Next, conditional analysis was conducted to tease age from was used to correct for multiple testing. STRUM was used to age at diagnosis (table 2). The minor alleles of rs73656147 predict the effect of a missense on the structural stability of and rs17763929 were associated with age, as was expected, a protein. given their association with age at diagnosis. However, the association with age at diagnosis persisted when adjusted for Data availability age, but the association with age was abolished when adjusted NGRC genotype and phenotype data are available at dbGaP for age at diagnosis. Hence, age at diagnosis was the driving ncbi.nlm.nih.gov/gap accession number phs000196.v3.p1. force, and association with age was a by-product of the correlation. Results To gauge robustness of the association signals with age at diagnosis and to test for heterogeneity, we stratified the data GWAS In SNP-based GWAS, the most significant signal for associ- by 8 PD-relevant variables, tested the association of each SNP ation, at p = 5E-8, mapped to LPPR1 on chromosome 9q31.1 with age at diagnosis within each stratum, and compared the (figure 1, A and B). In the gene-based test, LPPR1 achieved results across strata for evidence of heterogeneity (table e-2, p = 1E-8, surpassing the genome-wide statistical significance links.lww.com/NXG/A70). The 8 categories of stratification threshold of p < 2.6E-6 (figure 1, C and D). The p values were were family history, sex, cigarette smoking, caffeine intake, not inflated (λ = 1.007 SNP based, λ = 1.04 gene based). nonsteroidal anti-inflammatory drugs use, recruitment site, Analysis of LD in the region revealed 2 haplotype blocks with Jewish heritage, and the European country of ancestral origin. seemingly independent signals for association (figure 1, E and The association signal for rs73656147 (block 1) was robust F). There was strong LD among SNPs in each block, but weak across all strata. rs17763929 (block 2) showed evidence of LD between the blocks (r ≤ 0.2) because of a recombination heterogeneity as a function of recruitment site and the Eu- ropean country of ancestral origin. Given these results, we hot spot between them (figure 1F). The 2 blocks were in a ;200 Kb region inside LPPR1. Block 1 consisted of 51 tested the association of the 2 SNPs with PCs. rs17763929 SNPs with MAF;0.01, which yielded HR = 2.02–1.88, with was associated with PC1 (p = 7E-6) and PC3 (p = 8E-3), and p = 9E-7 to 2E-5 for association with age at diagnosis. Block 2 rs73656147 was not (p > 0.05 for PC1-3), indicating the consisted of 39 SNPs with MAF;0.02, which yielded HR = presence of population structure in block 2 but not in block 1. 1.88–1.85, with p = 5E-8 to 7E-7. We chose 1 SNP to rep- resent each block for replication: rs73656147 for block 1 Replication (MAF = 0.01, HR = 1.95, p = 3E-6) and rs17763929 for block In comparison to NGRC, which had a 65-year range for age at 2 (MAF = 0.02, HR = 1.88, p = 5E-8), both in Hardy- diagnosis, the PAGE data sets had a narrower range of less Weinberg (p > 0.3), with little correlation between them (r = than 30 years. Because PAGE participants were of similar age 0.2). Conditional analysis conducted to determine whether at study entry, the method of classifying the participants into the 2 blocks were tagging the same or different disease- prevalent PD (diagnosis before entry) vs incident PD (di- associated variants was inconclusive because although the agnosis after entry) inevitably assigned earlier ages at di- signals were weakened when adjusted for each other, neither agnosis to the prevalent group (PAGE ) and later diagnoses was abolished when conditioned on the other (table e-1, links. to the incident group (PAGE ). Mean age at diagnosis in lww.com/NXG/A69). PAGE was 59.9 ± 6.6 years, which was similar to the NGRC (60.4 ± 11.4). PAGE participants were on average 10 years There are 2 caveats in interpreting statistical evidence for older at diagnosis (69.4 ± 5.4, range 53–81 years). Given the association with age at diagnosis. First, age at diagnosis is disparity in the range and mean ages at diagnosis, we analyzed correlated with age (r = 0.74, p < 2E-16), which can result in PAGE and PAGE separately. P I spurious conclusions if the driving force responsible for the association is not identified. Second, tests of age at diagnosis Association of rs73656147 (block 1) with age at diagnosis are conducted using patients only without the benefitof replicated robustly (table 3). There was no evidence of het- controls. For example, an SNP that appears to be associated erogeneity between PAGE and PAGE in the association of I P with earlier PD diagnosis may in fact be associated with an rs73656147 with age at diagnosis, although the signal was age-related event unrelated to PD. To interpret the statistical stronger in PAGE than in PAGE , which is not surprising, P I evidence for association with age at diagnosis, we examined given that the former is enriched in cases with earlier age at 4 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Figure 1 Results of genome-wide association study for age at diagnosis of PD Genome-wide association was tested between 8.5 million SNPs and age at diagnosis in 1,950 PD cases from the NGRC, using the Cox hazard ratio regression method and adjusting for principal components (PC1-3). (A) Manhattan plot of SNP-based GWAS. Tallest peak, at p = 5E-8, was on chromosome 9q31.1. (B) QQ plot of SNP-based GWAS. The observed p values were not inflated (λ = 1.007). (C) Manhattan plot of gene-based GWAS. LPPR1 was at p = 1E-8. Statistical significance threshold was p < 2.6E-6, which is Bonferroni corrected for the 18,985 protein-coding genes tested. (D) QQ plot of gene-based GWAS. The observed p values were not inflated (λ = 1.04). (E) r (top panel) and D’ (bottom panel). Linkage disequilibrium (LD) across the SNPs that gave p < 1E-4 for association with age at diagnosis reveals 2 blocks represented by rs73656147 (left triangle) and rs17763929 (right triangle). (F) Magnified map of the associated region (chr9:103,865,000–104,055,000), showing that PD-associated SNPs map to LPPR1 and form 2 haplotype blocks separated by recombination hot spots (blue spikes). (G) Chromatin state of LPPR1 (Roadmap 111 Epigenomes), showing that active enhancers (yellow), transcription start site (red), and transcripts (green) of LPPR1 are seen only in stem cells and the brain and that the GWAS SNPs align with regulatory elements. ESC = embryonic stem cell; iPSC = induced pluripotent stem cell; TssA = active transcription start site (TSS); TssAFlnk = flanking active TSS; TxFlnk = transcription at gene 59 and 3’; Tx = strong transcription; TxWk = weak transcription; EnhG = genic enhancers; Enh = enhancers; ZNF/Rpts = zinc-finger genes and repeats; Het = heterochromatin; TssBiv = bivalent/poised TSS; BivFlnk = flanking bivalent TSS/enhancer; EnhBiv = bivalent enhancer; ReprPC = repressed polycomb; ReprPCWk = weak repressed polyComb; Quies = quiescent. Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 5 Table 2 Association of LPPR1 variants with age and age at diagnosis is driven by age at diagnosis Block 1 Block 2 rs73656147 rs17763929 Cox LR Cox LR N HR p Value β [95% CI] HR p Value β [95% CI] Ia. Association with age at diagnosis in cases 1,950 1.95 3E-6 −6.00 [−9.18 to −2.83] 1.88 5E-8 −5.65 [−8.20 to −3.11] Ib. Association with age at diagnosis in 1,950 1.95 3E-6 −5.98 [−9.16 to −2.81] 1.88 6E-8 −5.61 [−8.16 to −3.07] cases adjusted for sex II. Association with age in cases 2,000 1.48 5E-3 −4.19 [−7.1 to −1.3] 1.53 2E-4 −3.56 [−5.9 to −1.2] III. Association with age in controls 1,986 0.83 0.08 2.34 [−0.6 to 5.2] 0.84 0.07 2.37 [−0.3 to 5.1] IV. Association with age at diagnosis in 1,950 1.45 0.01 −2.30 [−3.9 to −0.7] 1.26 0.05 −2.11 [−3.4 to −0.8] cases adjusted for age V. Association with age in cases adjusted 1,950 0.92 0.56 0.78 [−0.8 to 2.3] 0.99 0.96 0.68 [−0.6 to 1.9] for age at diagnosis Abbreviations: CI = confidence interval; HR = hazard ratio; LR = linear regression; β = effect size on age at diagnosis (in years) per copy of minor allele. The associations were tested in the NGRC data set using Cox regression, and the effect sizes were estimated using linear regression (LR). HR is the age-for-age increase in the odds of event per copy of the minor allele, as estimated using Cox regression. β is the difference in years in age at diagnosis between carriers of 1 minor allele vs no minor allele, as estimated using linear regression. Age at diagnosis was the primary outcome of the study. Minor alleles of rs73656147 and rs17763929 were associated with higher HR and younger age at diagnosis (Ia). The association was not influenced by sex (Ib), which was expected because, unlike PD risk, which is significantly associated with sex (OR = 3.26, p < 2E-16), age at diagnosis is not associated with sex (HR = 0.99, p = 0.83). Minor alleles were also associated with younger ages in cases (II), but not in controls (III). Because age and age at diagnosis were correlated (r = 0.74, p < 2E-16), an association with one will show as an association with both. In conditional analysis, the association with age at diagnosis persisted when adjusted for age (IV), butthe association with age was abolished when adjusted for age at diagnosis (V), suggesting that age at diagnosis was the driving force and association with age was a by-product of the correlation. diagnosis. Nor was there evidence of heterogeneity between mapped to enhancers in the brain (table 4 and figure 1, G). PAGE and NGRC for the association of rs73656147 with age Eleven of the genes identified through Hi-C are expressed in at diagnosis. Meta-analysis yielded HR = 2.14, p = 1E-3 for the brain: LPPR1, SEC61B, MSANTD3-TMEFF1, TMEFF1, replication and HR = 2.00, p = 9E-9 for replication and dis- GALNT12, MURC, GRIN3A, NR4A3, ALG2, MRPL50, and covery. Mean difference in age at diagnosis per copy of ZNF189 (figure 2, B and C). The expression of LPPR1 in the rs73656147 minor allele was −6.0 (95% confidence interval: brain is the strongest in early prenatal stage and decreases with −9.18 to −2.83) years in the NGRC, −5.53 (−9.72 to −1.34) in developmental stage and increasing age (figure 2, C). PAGE , −0.84 (−4.22 to 2.55) in PAGE , and −4.08 (−7.45 to P I −0.70) in the meta-analysis of the 3 data sets. CADD analysis, a scoring system for deleteriousness of ge- netic variants, identified 5 SNPs in block 1 and 3 in block 2, Association of rs17763929 (block 2) with age at diagnosis with CADD = 10–19 (table 4), which places them among the showed significant heterogeneity between PAGE and PAGE top 10% (CADD > 10) to 1% (CADD > 20) of most dele- I P (table 3), as it had within the NGRC (table e-2, links.lww. terious mutations in the genome. rs41296085 (CADD = 18, com/NXG/A70). The association with rs17763929 repli- in block 1) is a missense (p.Ser12Ala) in exon 2, predicted to cated in PAGE but not in PAGE . There was significant structurally destabilize the LPPR1 protein (ΔΔG= −1.2). The P I heterogeneity between PAGE and NGRC, but not between remainder of the variants with high CADD scores are in PAGE and NGRC. Meta-analysis of PAGE and NGRC introns. eQTL analysis revealed an association between P P yielded HR = 1.88, p = 4E-9 for full PAGE data and HR = 1.95, rs117451395 (block 1) with expression levels of GRIN3A p = 3E-9 for the PAGE subsample adjusted for PC1-3. In- (FDR = 4E-4). cluding PAGE with PAGE and NGRC in a random-effects I P meta-analysis diluted the effect size to HR = 1.53, p = 0.04. Mean difference in age at diagnosis per copy of rs17763929 Discussion minor allele was −5.65 (−8.20 to −3.11) years in the NGRC, There has been intense research on PD risk factors, which so −3.62 (−7.23 to −0.02) in PAGE , and 0.62 (−1.34 to 2.58) in far has resulted in identification of numerous causative genes, PAGE . 40 susceptibility loci, several environmental factors, and a few Functional annotation genes that interact with the environmental factors to increase Hi-C analysis showed significant (FDR < 1E-6) chromatin or reduce the risk of developing PD. In contrast, we know interaction between the PD-associated LPPR1 SNPs and little about factors that affect the rate of disease progression. promoters of LPPR1 and several genes on chromosome 9 In this study, we attempted to identify genetic modifiers of age (figure 2, A). Some of the SNPs that were significant in Hi-C at diagnosis, a reflection of rate of progression, using an 6 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Table 3 Replication Age at diagnosis Block 1 rs73656147 Block 2 rs17763929 Data sets N PD cases Mean ± SD HR p Value HR p Value NGRC (discovery) 1,950 60.4 ± 11.4 1.95 3E-6 1.88 5E-8 PAGE (replication) 209 59.9 ± 6.6 2.88 7E-4 1.87 0.01 PAGE with PC1-3 113 59.9 ± 6.8 2.17 0.05 3.03 4E-3 PAGE (replication) 517 69.4 ± 5.4 1.62 0.07 1.04 0.41 PAGE with PC1-3 283 69.2 ± 5.3 1.48 0.16 1.03 0.45 Meta-analysis A Heterogeneity rs73656147 Heterogeneity rs17763929 PAGE and PAGE ns 0.08 2.14 1E-3 1.34 0.31 P I NGRC and PAGE ns ns 2.08 2E-8 1.88 4E-9 NGRC and PAGE ns 0.01 1.90 9E-7 1.42 0.23 NGRC and PAGE and PAGE ns 0.02 2.00 9E-9 1.53 0.04 P I Meta-analysis B PAGE and PAGE ns 0.02 1.73 0.07 1.67 0.34 P I NGRC and PAGE ns ns 1.97 6E-7 1.95 3E-9 NGRC and PAGE ns 0.02 1.89 2E-6 1.43 0.23 NGRC and PAGE & PAGE ns 0.02 1.91 5E-7 1.68 0.05 P I Abbreviations: HR = hazard ratio; NGRC = NeuroGenetics Research Consortium; ns = not statistically significant; PAGE = Parkinson’s, Genes and Environment; PC = principal component; PD = Parkinson disease. Two SNPs with signals for association with age at diagnosis of PD in the NGRC data set (discovery) were genotyped and tested for association with age at diagnosis of PD in the PAGE data set (replication). PAGE participants were designated as PAGE if they were diagnosed before study entry or PAGE if they were P I diagnosed during the study. Cox regression was used to test the association of SNP (additive model) with age at diagnosis (quantitative trait) and to calculate hazard ratios (HRs) and the corresponding significance (p). NGRC was adjusted for PC1-3 in GWAS and meta-analyses. Only a subset of PAGE had ancestry informative markers (AIMs) for which PC could be calculated; thus, results are shown for the full PAGE data set without PC adjustment and for the PAGE subsample with PC adjustment. p values are 2-sided for NGRC and one-sided for PAGE because of the directionality of the hypothesis being replicated. Meta-analysis A: NGRC (PC1-3 adjusted) and PAGE (all data without PC adjustment). Meta-analysis B: NGRC (PC1-3 adjusted) and PAGE (subset of data adjusted for PC1-3). rs73656147 replicated robustly with no evidence of heterogeneity across data sets. rs17763929 replicated in PAGE and showed significant heterogeneity between PAGE and PAGE or NGRC. I P Meta-analysis was conducted using the fixed-effects model if there was no evidence for heterogeneity (p ≥ 0.1) and the random effects model if there was heterogeneity (p < 0.1). unbiased genome-wide approach, followed by independent not fully capture the complexity of block 2. PAGE participants replication, and functional annotation. being significantly older than NGRC and PAGE participants 32–34 may also be a factor. LPPR1 promotes neuroregeneration, We uncovered evidence for association of genetic variants in but its expression diminishes with age to nearly undetectable neuronal plasticity-related gene 3 (LPPR1) with age at di- level by age 40 years (figure 2C). One can speculate that some agnosis of PD. Two signals of association were detected, each detrimental variants may not have an effect after a certain age representing a haplotype block of SNPs. The variants that when the gene is no longer expressed. were associated with earlier age at diagnosis had low allele frequencies (MAF = 0.01–0.02), as were the variants that Functional annotation of the PD-associated variants in LPPR1 were previously found for age at onset of familial PD. The revealed the presence of several variants with predicted del- low allele frequencies may be one reason why modifier genes eterious effects, including a missense that destabilizes the have been more difficult to detect than common variants that structure of LPPR1, a regulatory element that associates with associate with risk. expression levels of GRIN3A, and enhancers that interact with promoters of LPPR1 and several other genes in the brain. The association with block 1 replicated robustly in both PAGE Some of the candidate genes that were identified via in- and PAGE . Block 2 signal replicated in PAGE but not in teraction with LPPR1 play key roles in pathways that are I P PAGE . Block 2 has a complex LD structure, with evidence of implicated in PD, including GRIN3A (which encodes a sub- population substructure, which limits generalizability of results. unit of NMDA receptor involved in the glutamate-regulated Failure to capture a signal for block 2 in PAGE may be because ion channels in the brain), SEC61B (protein transport appa- we had genotype on only 1 SNP in block 2 for PAGE, which did ratus of the endoplasmic reticulum membrane), MURC (Rho Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 7 Figure 2 Functionally significant genes (A) 3D chromatin interaction (Hi-C) and eQTL analysis. Hi-C revealed significant interaction between GWAS variants in LPPR1 and 17 other genes on chromosome 9 (FDR < 1E-6, shown in orange). An SNP in LPPR1 was associated with the expression of GRIN3A (FDR = 4E-4, shown in green). (B) Tissue-specific expression of LPPR1, GRIN3A, and genes in Hi-C with LPPR1. Colors reflect average expression (log2 transformed) from highest (red) to lowest/absent (blue). (C) Age-specific expression of the genes in the brain. LPPR1 expression decreases with age. kinase signaling), and MRPL50 (mitochondrial ribosomal we propose that LPPR1 is involved, not necessarily in the protein). cause of PD, rather in response to damage, and influences the efficacy of regeneration and the subsequent rate of de- LPPR1 is one of the 5 members of a brain-specific gene family terioration in preclinical PD. The actual cause of injury and that modulates neuronal plasticity during development, aging, neuronal death is not stipulated in this hypothesis; it could be 32–34 and after brain injury. LPPR1 is the strongest driver of head trauma, environmental toxins or genetic, but once the axonal outgrowth in the gene family. Studies in mice have initial damage is incurred, it is the efficacy of intrinsic mech- shown that after neuronal injury, overexpression of LPPR1 anisms of repair that determine the rate of disease pro- enhances axonal growth, improves motor behavior, and pro- gression. Present findings provide a strong foundation for 33,34 motes functional recovery. Extrapolating to our findings, mechanistic studies to investigate the role of LPPR1 in PD and 8 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG Table 4 Functionally significant variants Block GWAS SNP position:alleles GWASpr eQTL CADD Hi-C/EnhBrain 1 rs77351585 9:103874925:C:T 2E-06 1 — 18 Hi-C/EnhBrain 1 rs73495940 9:103875807:G:C 9E-07 Lead —— Hi-C 1 rs150164200 9:103875896:A:C 2E-06 1 — 10.4 — 1 rs117583993 9:103876647:G:A 3E-06 1 —— Hi-C/EnhBrain 1 rs148874623 9:103939117:A:C 9E-06 1 — 12.1 — 1 rs117451395 9:103941039:C:T 1E-05 1 GRIN3A — Hi-C 1 rs41296085 9:103947810:T:G 2E-05 1 — 18 (missense) Hi-C 1 rs117900237 9:103959240:G:A 2E-05 1 — 10.5 Hi-C 2 rs17763929 9:103984900:A:G 5E-08 Lead —— Hi-C 2 rs61188842 9:103988006:C:T 8E-05 0.6 —— Hi-C/EnhBrain 2 rs117058418 9:104011717:T:C 2E-07 1 — 10.4 Hi-C 2 rs117314512 9:104014244:G:A 2E-07 1 — 12.4 Hi-C 2 rs149155028 9:104032402:TTC:T 1E-05 0.7 — 18.6 Hi-C Functional annotation was conducted on SNPs with GWAS p < 1E-6 and SNPs that were in high LD with them (r > 0.6). Variants are shown if they are the lead SNPs (most significant) for the block, or an eQTL (FDR = 4E-4), or had a CADD score >10, or had both significant evidence for 3D chromatin interaction (Hi-C, FDR < 1E-6) and overlapped with an enhancer in the brain. Block 1 is a single block of SNPs in high LD. Block 2 has a complex LD structure with at least 3 subhaplotypes (figure e1-C, links.lww.com/NXG/A66). Variants are shown with their rs accession number, chromosome position and the 2 alleles (major: minor), GWAS p value for association with age at diagnosis of PD, and their correlation (r ) with the lead SNP of the block. eQTL: an SNP that is associated with gene expression, in this case, rs117451395, was associated with gene expression levels at GRIN3A (FDR = 4E-4). CADD: a predictive score for the deleteri- ousness of a variant. A CADD score of 10 usually means that the variant is among the top 10% of deleterious mutations in the genome. A CADD score of 20 puts the variant among the top 1% of deleterious mutations. Hi-C: SNPs with significant (FDR < 1E-6) evidence for interacting with the promoter region of LPPR1 or of another gene (figure 2 for the genes). Hi-C/EnhBrain: the subset of Hi-C SNPs that map to the enhancer regions of LPPR1 in the brain according to the Roadmap 111 epigenomes. a 2 One SNP was shown to represent several variants in high LD (r > 0.9) with similar MAF, GWAS p value, and Hi-C/EnhBrain evidence. This mutation yielded ΔΔG= −1.2, which predicts a destabilizing effect on the protein structure of LPPR1. determine its potential as a therapeutic target to impede data for replication (Parkinson, Genes, and Environment disease progression. (PAGE) study) was supported by the intramural research program of NIH National Institute of Environmental Health Author contributions Sciences grant Z01 ES101986. Funding agencies did not have Z.D. Wallen: statistical analysis, review, and critique of the a role in the design or execution of the study. manuscript. H. Chen: creation of the PAGE data set, acqui- sition of data, review, and critique of the manuscript. E.M. Disclosure Hill-Burns: assembly and QC of phenotype and genotype Z.D. Wallen, H. Chen, C.P. Zabetian, and E.M. Hill-Burns data, imputation, review, and critique of the manuscript. S.A. report no disclosures. S.A. Factor has received honoraria from Factor and C.P. Zabetian: creation of the NGRC data set, Neurocrine, Lundbeck, Teva, Avanir, Sunovion Pharmaceut- review, and critique of the manuscript. H. Payami: study icals, Adamas, and UCB; has received research support from concept, design and execution, creation of the NGRC data set, Ipsen, Medtronic, Teva, US WorldMeds, Sunovion Pharma- wrote the manuscript, and obtained funding. ceuticals, Solstice, Vaccinex, Voyager, the CHDI Foundation, the Michael J. Fox Foundation, and the NIH; and receives Study funding publishing royalties from Demos, Blackwell Futura, and This work was supported by the National Institute of Neuro- UpToDate. H. Payami has received research support from the logical Disorders and Stroke grant R01NS036960. Additional NIH and the University of Alabama at Birmingham. Full dis- support was provided by a Merit Review Award from the De- closure form information provided by the authors is available partment of Veterans Affairs grant 1I01BX000531; Office of with the full text of this article at Neurology.org/NG. Research & Development, Clinical Sciences Research & De- Received February 13, 2018. Accepted in final form June 11, 2018. velopment Service, Department of Veteran Affairs; the Close to the Cure Foundation; and the Sartain Lanier Family Founda- tion. Genome-wide array genotyping was conducted by the References 1. Obeso JA, Stamelou M, Goetz CG, et al. Past, present, and future of Parkinson’s Center for Inherited Disease Research, which is funded by the disease: a special essay on the 200th Anniversary of the Shaking Palsy. Mov Disord NIH grant HHSN268200782096C. Collection of samples and 2017;32:1264–1310. Neurology.org/NG Neurology: Genetics | Volume 4, Number 5 | October 2018 9 2. Zareparsi S, Taylor TD, Harris EL, Payami H. Segregation analysis of Parkinson 18. Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D. Principal components analysis corrects for stratification in genome-wide association studies. disease. Am J Med Genet 1998;80:410–417. Nat Genet 2006;38:904–909. 3. Maher NE, Currie LJ, Lazzarini AM, et al. Segregation analysis of Parkinson disease 19. Aulchenko YS, Struchalin MV, van Duijn CM. ProbABEL package for genome-wide revealing evidence for a major causative gene. Am J Med Genet 2002;109:191–197. association analysis of imputed data. BMC Bioinformatics 2010;11:134. 4. McDonnell SK, Schaid DJ, Elbaz A, et al. Complex segregation analysis of Parkinson’s 20. Watanabe K, Taskesen E, van Bochoven A, Posthuma D. Functional mapping and disease: the Mayo clinic family study. Ann Neurol 2006;59:788–795. annotation of genetic associations with FUMA. Nat Commun 2017;8:1826. 5. Hamza TH, Payami H. The heritability of risk and age at onset of Parkinson’s disease 21. Aulchenko YS, Ripke S, Isaacs A, van Duijn CM. GenABEL: an R library for genome- after accounting for known genetic risk factors J Hum Genet 2010;55:241–243. wide association analysis. Bioinformatics 2007;23:1294–1296. 6. Escott-Price V, Consortium IPsDG, Nalls MA, et al. Polygenic risk of Parkinson 22. Pruim RJ, Welch RP, Sanna S, et al. LocusZoom: regional visualization of genome- disease is correlated with disease age at onset. Ann Neurol 2015;77:582–591. wide association scan results. Bioinformatics 2010;26:2336–2337. 7. Chang D, Nalls MA, Hallgrimsdottir IB, et al. A meta-analysis of genome-wide as- 23. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and sociation studies identifies 17 new Parkinson’s disease risk loci. Nat Genet 2017;49: haplotype maps. Bioinformatics 2005;21:263–265. 1511–1516. 24. Machiela MJ, Chanock SJ. LDlink: a web-based application for exploring population- 8. Nalls MA, Escott-Price V, Williams NM, et al. Genetic risk and age in Parkinson’s specific haplotype structure and linking correlated alleles of possible functional var- disease: continuum not stratum. Mov Disord 2015;30:850–854. iants. Bioinformatics 2015;31:3555–3557. 9. Lill CM, Hansen J, Olsen JH, Binder H, Ritz B, Bertram L. Impact of Parkinson’s 25. Payami H, Kay DM, Zabetian CP, Schellenberg GD, Factor SA, McCulloch CC. disease risk loci on age at onset. Mov Disord 2015;30:847–850. Visualizing disease associations: graphic analysis of frequency distributions as a func- 10. Pihlstrom L, Toft M. Cumulative genetic risk and age at onset in Parkinson’s disease. tion of age using moving average plots (MAP) with application to Alzheimer’s and Mov Disord 2015;30:1712–1713. Parkinson’s disease. Genet Epidemiol 2010;34:92–99. 11. Hill-Burns EM, Ross OA, Wissemann WT, et al. Identification of genetic modifiers of 26. de Leeuw CA, Mooij JM, Heskes T, Posthuma D. MAGMA: generalized gene-set age-at-onset for familial Parkinson’s disease. Hum Mol Genet 2016;25:3849–3862. analysis of GWAS data. PLoS Comput Biol 2015;11:e1004219. 12. Trinh J, Gustavsson EK, Vilarino-Guell C, et al. DNM3 and genetic modifiers of age of 27. Kircher M, Witten DM, Jain P, O’Roak BJ, Cooper GM, Shendure J. A general framework for estimating the relative pathogenicity of human genetic variants. Nat onset in LRRK2 Gly2019Ser parkinsonism: a genome-wide linkage and association Genet 2014;46:310–315. study. Lancet Neurol 2016;15:1248–1256. 28. Consortium G. Human genomics. The Genotype-Tissue Expression (GTEx) pilot 13. Hamza TH, Zabetian CP, Tenesa A, et al. Common genetic variation in the HLA analysis: multitissue gene regulation in humans. Science 2015;348:648–660. region is associated with late-onset sporadic Parkinson’s disease. Nat Genet 2010;42: 29. van Berkum NL, Lieberman-Aiden E, Williams L, et al. Hi-C: a method to study the 781–785. three-dimensional architecture of genomes. J Vis Exp 2010:e-1869. 14. Chen H, Huang X, Guo X, et al. Smoking duration, intensity, and risk of Parkinson 30. Consortium RE, Kundaje A, Meuleman W, et al. Integrative analysis of 111 reference disease. Neurology 2010;74:878–884. human epigenomes. Nature 2015;518:317–330. 15. Hughes AJ, Daniel SE, Kilford L, Lees AJ. Accuracy of clinical diagnosis of idiopathic 31. Quan L, Lv Q, Zhang Y. STRUM: structure-based prediction of protein stability Parkinson’s disease: a clinico-pathological study of 100 cases. J Neurol Neurosurg changes upon single-point mutation. Bioinformatics 2016;32:2936–2946. Psychiatry 1992;55:181–184. 32. Savaskan NE, Brauer AU, Nitsch R. Molecular cloning and expression regulation of PRG- 16. Schatzkin A, Subar AF, Thompson FE, et al. Design and serendipity in establishing 3, a new member of the plasticity-related gene family. Eur J Neurosci 2004;19:212–220. a large cohort with wide dietary intake distributions: the National Institutes of Health- 33. Broggini T, Schnell L, Ghoochani A, et al. Plasticity related gene 3 (PRG3) overcomes American Association of retired persons diet and Health study. Am J Epidemiol 2001; myelin-associated growth inhibition and promotes functional recovery after spinal 154:1119–1125. cord injury. Aging (Albany NY) 2016;8:2463–2487. 17. Howie BN, Donnelly P, Marchini J. A flexible and accurate genotype imputation 34. Fink KL, Lopez-Giraldez F, Kim IJ, Strittmatter SM, Cafferty WBJ. Identification of method for the next generation of genome-wide association studies. PLoS Genet intrinsic axon growth modulators for intact CNS neurons after injury. Cell Rep 2017; 2009;5:e1000529. 18:2687–2701. 10 Neurology: Genetics | Volume 4, Number 5 | October 2018 Neurology.org/NG

Journal

Neurology: GeneticsPubmed Central

Published: Oct 5, 2018

There are no references for this article.