Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Apr 1.
Published in final edited form as: Mov Disord. 2019 Jan 24;34(4):460–468. doi: 10.1002/mds.27614

The endocytic membrane trafficking pathway plays a major role in the risk of Parkinson disease

S Bandres-Ciga 1,#, S Saez-Atienzar 2,#, L Bonet-Ponce 3,#, K Billingsley 1, D Vitale 4, C Blauwendraat 1, JR Gibbs 4, L Pihlstrom 5, Z Gan-Or 6,7; The International Parkinson’s Disease Genomics Consortium (IPDGC), MR Cookson 3, MA Nalls 1,8, AB Singleton 1
PMCID: PMC6467788  NIHMSID: NIHMS1005117  PMID: 30675927

Abstract

Background

Parkinson’s disease (PD) is a complex polygenic disorder. In recent years, several genes from the endocytic membrane-trafficking pathway (EMTP) have been suggested to contribute to disease etiology. However, a systematic analysis of pathway-specific genetic risk factors is yet to be performed.

Objectives

To comprehensively study the role of the EMTP in the risk of PD.

Methods

Linkage disequilibrium score regression was used to estimate PD heritability explained by 252 genes involved in the EMTP including GWAS data from 18,869 cases and 22,452 controls. We used pathway-specific SNPs to construct a polygenic risk score (PRS) reflecting the cumulative risk of common variants. To prioritize genes for follow-up functional studies, summary-data based Mendelian randomization analyses were applied to explore possible functional genomic associations with expression or methylation quantitative trait loci.

Results

The heritability estimate due to EMTP was 3.58 % (se = 1.17). Excluding previously nominated PD EMTP genes, the missing heritability was 2.21 % (se = 0.42). Random heritability simulations were estimated to be 1.44 % (SD= 0.54), indicating that the unbiased total heritability explained by the EMTP was 2.14 %. PRS based on EMTP showed a 1.25 times increase of PD risk per standard deviation of genetic risk. Finally, Mendelian randomization identified 11 EMTP genes showing functional consequence associated to PD risk.

Conclusions

We provide compelling genetic evidence that the EMTP plays a relevant role in disease etiology. Further research on this pathway is warranted as critical effort should be made to identify potential avenues within this biological process suitable for therapeutic interventions.

Keywords: endocytosis, genetic risk, Parkinson disease, polygenic risk score, heritability

Introduction

Parkinson disease (PD) is a progressive neurodegenerative disorder clinically characterized by the manifestation of motor disturbances such as bradykinesia, rigidity, tremor, and a wide range of non-motor symptoms. Disease is caused by the death of the dopaminergic neurons in the Substantia Nigra (SN) with the consequent deficit of striatal dopamine1. The etiology of PD is complex, including genetic, epigenetic and environmental factors. Familial PD accounts for less than 10% of the total cases, mostly due to monogenic forms with autosomal dominant (including SNCA, LRRK2 and VPS35) or recessive pattern of inheritance, such as PARKIN, PINK1 and DJ‐1, among others2,3. However, the vast majority of the disease is considered to be sporadic, likely caused by the cumulative effect of multiple common or rare variants of small or moderate effect sizes in addition to other unknown environmental and stochastic factors.

Over the last two decades, extensive work in genetics has uncovered up to 92 common risk loci associated with idiopathic PD4. As expected from complex diseases, the PD related signals identified to date only account for a small proportion of the total phenotypic variation, defined as heritability5. It is estimated that only about 10–15% of the overall heritable risk of PD is explained by current GWAS4. The most plausible explanation for the “missing heritability” might be small effect sizes of the yet to be discovered risk loci which do not reach GWAS significance, as well as structural genomic and rare variation poorly covered by current technology6. Undoubtedly, the estimated heritability in PD suggests that a significant amount of risk has not been uncovered yet.

Aligned with the genetic discoveries, a big effort has been made to identify how those genes interact within biological pathways to elucidate the cellular and molecular processes that could explain the neurodegenerative mechanism7. Recent findings point out that dysfunction in the endosomal membrane-trafficking pathway (EMTP) could be a driving force in the pathogenesis of PD8. The EMTP is a complex multi-step pathway that involves different process and organelles. First, cells uptake macromolecules and other cargo from the extracellular space to a group of organelles called early endosomes. From this point, internalized molecules can recycle back to the plasma membrane directly or through recycling endosomes. Alternatively, endocytosed cargo can be retained in early endosomes that will eventually mature into late endosomes and fuse with lysosomes to ensure cargo degradation. In parallel, proteins can also traffic between endosomes and the Golgi apparatus (for a review9). A substantial amount of genetic variation in PD and parkinsonism previously nominated genes has been associated with vesicle trafficking via endosomal gene alterations10, including monogenic genes (VPS3511, DNAJC618,13) and GWAS nominated genes (RAB7L112,13, SH3GL214, GAK15,16 and CHMP2B15,17). In addition, other well-known PD genes such as LRRK212,13 and PLA2G618 seem to be linked to the aforementioned pathway through their biological interaction with EMPT genes. LRRK2 binds to RAB7L112,13 and PLA2G6 interacts with VPS35 and VPS26, although a clear role of these genes in the EMTP is yet to be elucidated. The genes involved in the EMTP play different roles within the pathway. For instance, DNAJC6, SYNJ1, GAK and SH3GL2 are linked to clathrin-coated vesicles whereas VPS35, DNAJC13 or RAB7L1 participate in recycling components from the endosomes to the Golgi. Hence, it seems plausible that other genes in that pathway could be involved in PD risk and therefore account for the non-reported heritability.

Based on these premises, we aimed to benefit from a cellular perspective to systematically explore the genetic contribution of 252 endosomal membrane-trafficking related genes on the susceptibility for PD. We set out to study to what extent common variation in these genes might play a role on the genetic architecture of PD by performing genome-wide heritability estimations and polygenic risk versus disease status and age at onset in the International Parkinson’s Disease Genomics Consortium (IPDGC) GWAS dataset. We also applied Mendelian randomization to attempt to investigate possible functional genomic associations between nominated variants of interest within this pathway versus expression and methylation quantitative trait loci. Finally, we used biological pathway analysis and expression data from murine SN to point out the biological contribution of the genetic findings.

Methods

1. Subjects

The IPDGC GWAS dataset consisted of 41,321 individuals (18,869 cases and 22,452 controls) of European ancestry. Clinical and demographic characteristics of the cohorts under study are shown in Supplementary Table 1. Additional details of these cohorts can be found in Nalls et al., 2018.

No individuals with known pathogenic mutations were excluded from following analyses. When the case age at onset or control age at last exam was not provided, individuals were likewise included.

2. Genes selection and pathway analysis.

The EMTP is a general and extense pathway that includes several sub-pathways. EMTP related genes under study were selected by using The Molecular Signatures Database (MSigDB) of annotated gene set for “endocytosis”19,20 based on the Kyoto Encyclopedia of Genes and Genomes information (KEGG-endocytosis). However, the gene list provided by KEGG was not completely accurate and failed to represent some relevant genes within the EMTP. In order to validate whether each gene flagged as “endocytosis” was involved in the EMTP, a detailed literature search was performed. In addition, since the current research in the EMTP is very dynamic and there are constantly new discovered proteins involved in some aspects of the pathway, we included relevant genes that have been proved to participate in the EMPT but were not nominated by KEGG. In addition, a thorough review of the literature was performed using the terms “endocytosis”, “endosome to Golgi retrograde transport”, “endosomal maturation”, “endosomal recycling”. We have provided a table (Supplementary Table 2) that includes each gene involved in our analysis with the literature source that refers to.A total of 262 genes were flagged as ‘endocytic membrane-trafficking’ (Supplementary Table 2) of which 252 were autosomal and 10 spanned in chromosome X. The latter were excluded for the following analyses due to lack of data. The PANTHER web-based Classification System was used to obtain the Top 10 significant enriched Reactome Pathways with an adjusted False Discovery Rate (FDR) p-value < 0.05. Additionally, we obtained expression data from single cell RNA-seq (scRNA-seq) performed in 243 putative dopaminergic neurons at postnatal day 28–56 from Slc6a3-Cre/tdTomato mice (GSE76381)21. We applied an arbitrary selected threshold of expression. Only genes with total expression values across the 243 cells higher than 100 were nominated as expressed. Human genes were converted to their ortholog genes using Mouse Genome Informatics (MGI) (http://www.informatics.jax.org) (Supplementary Table 3). For Venn diagram visualization the online-based website http://bioinformatics.psb.ugent.be/webtools/Venn/ and Inkscape software version 0.92.2 (https://inkscape.org) were used.

3. Quality control

Genome-wide association data

Quality control (QC) procedures in the IPDGC GWAS dataset have been previously described elsewhere4. In brief, samples with call rates < 95% and whose genetically determined sex from X chromosome heterogeneity did not match that from clinical data were excluded. Individuals exhibiting heterozygosity greater than 6 standard deviations from the population mean and samples of non-European ancestry were not included. Individuals related at more than the level of first cousin estimated from linkage pruned genetic data on common variants were excluded (one proband from any family cluster was extracted randomly). Regarding variant QC, variants exhibiting Hardy-Weinberg Equilibrium (HWE) p-value < 1E-5 and missingness rates > 5 % as well as palindromic SNPs were excluded from further analyses. Data was imputed using the Haplotype Reference Consortium (http://www.haplotype-reference-consortium.org), under default settings with phasing using the EAGLE option https://www.nature.com/articles/ng.3643 https://www.ncbi.nlm.nih.gov/pubmed/27571263. VCFs were filtered for imputation quality > 0.30 for inclusion in analyses.

4. Heritability analysis

In order to estimate the phenotypic variance of PD explained by genes in the extended EMTP pathway, we estimated heritability using Linkage disequilibrium score regression (LDSC) 22,23. We used LDSC to estimate heritability in this study because it requires only summary-level data and is more computationally efficient for larger data sets. We extracted EMTP regions of interest from a meta-analysis of GWAS data and then ran LDSC under default settings. Reference LD scores were computed with the European ancestry subset of the 1000 Genomes data for SNPs within 500 kb of the SNP to be scored.

Firstly, we estimated heritability explained by variants in the whole set of EMTP genes and missing heritability by excluding previously PD associated EMTP related genes. Subsequently, to evaluate significant enrichment of heritability within our pathway of interest compared to any similar set of genes, we re-sampled 50 random sets of 252 genes in the same sample series (after excluding EMTP genes and regions +/− 250kb from PD GWAS loci). During resampling, we implemented the exact sample LDSC analyses and extracted the heritability per sampling iteration. Next, we calculated the mean and standard deviations of the heritability across sampling iterations. Finally the mean heritability of our re-sampled series was compared to the test set to determine if our previous estimates were significantly divergent from random sets of genes of similar size using a t-test.

In an attempt to explore the genes that contribute the most to the estimated heritability, we performed the Sequence Kernel Association Test (SKAT) in imputed data adjusting by age, gender and data subset membership without applying any frequency filtering. Significance was evaluated using 10000 permutations at alpha= 0.001.

5. Polygenic risk profiles versus disease status and age at onset

For the risk profiling analysis, we mirrored the workflow used in the most recent PD meta-analysis to date4, with the exception that prior to running this analysis pipeline we extracted only genotypes from the pathway of interest. In brief, PRS profiling is calculated traditionally based on weighted allele dose manner as per previous work25–32. We used the R package PRSice233. Permutation testing and p-value aware linkage disequilibrium (LD) pruning was used to identify best P thresholds in external GWAS data to construct the PRS, allowing us to utilize variants below the common GWAS significance threshold of 5E-08. External summary statistics utilized in this phase of analysis included data from leave-one-out meta-analyses (LOOMAs) that exclude the study in which the PRS was being tested, avoiding overfitting/circularity to some degree. LD clumping was implemented under default settings (window size = 250kb, r2 > 0.1) and for each dataset 10,000 permutations were used to generate empirical P estimates for each GWAS derived P threshold ranging from 5E-08 to 0.5, at a minimum increment of 5E-08. Each permutation test in each dataset provided a Nagelkerke’s pseudo r2 after adjustment for an estimated prevalence of 0.005 (aged population estimate as per Gasser et al.) and study-specific eigenvector 1–5, age and sex as covariates. GWAS derived P threshold with the highest pseudo r2 was selected for further analysis. Summary statistics were meta-analyzed using random effects (REML) per study-specific dataset by using the R package PRSice-233.

For the age at onset risk profiling, we performed similar analyses with the exception that we used age at onset as a continuous variable.

6. Mendelian randomization analysis to explore pathway specific quantitative trait loci

We utilized two-sample Mendelian randomization to investigate possible functional genomic associations between nominated variants of interest in the EMTP pathway versus expression and methylation patterns. QTL association summary statistics across well-curated methylation and expression datasets from the SMR (http://cnsgenomics.com/software/smr)34 were compared to PD outcome summary statistics4 after extracting the pathway-specific variants considered as the instrumental variables. These include estimates for methylation and cis-expression from the Genotype-Tissue Expression (GTEx) Consortium v6; (whole blood and 10 brain regions)35, the CommonMind Consortium (CMC; dorsolateral prefrontal cortex)36, the Religious Orders Study and Memory and Aging Project (ROSMAP)37, and the Brain eQTL Almanac project (Braineac; 10 brain regions)38 as reported elsewhere. Furthermore, we studied expression patterns in blood from the largest eQTL meta-analysis so far39.

Wald ratios were generated for each of the instrumental variable SNPs tagging a cis-QTL (probes within each gene and meeting a QTL p-value of at least 5E-8 in the original QTL study) and for a methylation or expression probe with a nearby gene. Linkage pruning and clumping were carried out using default SMR protocols. With each data type, p-values per instrument substrate (gene level expression summaries for eQTLs and cis-gene probe level for mQTLs) were adjusted by false discovery rate.

Results

1. Pathway analysis and expression of EMTP related genes

EMTP variants from imputed data were extracted from 252 genes selected as described in the Methods section. Distribution of variants within the EMTP in the IPDGC cohort are highlighted in Supplementary Table 4. Gene Ontology analysis was performed on these genes showing an enrichment for trafficking-related biological processes such as lysosomal biogenesis, clathrin-mediated endocytosis, retrograde transport at the trans-Golgi network or vesicle-mediated transport among others (Figure 1A). Remarkably, 151 genes among the analyzed genes were found expressed in dopaminergic neurons in adult mouse brains (Figure 1B).

Figure 1. Biological implication of EMPT genes in DA neurons.

Figure 1.

Open in a new tab

A. Top 10 significant Enriched Reactome Pathways with an adjusted False Discovery Rate p-value < 0.05. Pathway enrichment analysis was performed using Panther. Only the TOP 10 enriched pathways were selected for visualization. B. Overlap between genes expressed in DA neurons and EMTP gene set. The EMTP human gene were transform in their mouse homolog genes. From the 252 total human genes, a total of 244 were found in mouse (Table S3). The DA neuron gene expression were obtain from GSE76381) (La Manno et al. 2016) after an arbitrary selected threshold of expression.

2. Heritability analysis

We measured heritability explained by the 252 genes under study. EMTP genetic contribution to PD heritability was estimated to be 3.58 % (se = 1.17). After excluding previously nominated PD and parkinsonism genes related to the EMTP, including DNAJC13, VPS35, DNAJC6, SYNJ1, SYT4, SYT11, GAK, RAB7L1, CHMP2B, and SH3GL2, missing heritability explained by EMTP related genes was estimated to be 2.21 % (se = 0.42). Reported heritability estimates might be underestimated since there are possibly rare variants within these genes poorly covered by current genotyping technologies or lost after imputation. These estimates were significantly different from 50 randomly generated sets of 252 genes (t-test, p < 2.2E-16) giving a heritability estimate of 1.44 % (SD= 0.54) and showing that the heritability enrichment in the EMTP pathway is not likely by chance (Supplementary Figure 1). After subtracting the randomly generated heritability estimates (1.44 %) that might have biased our initial EMTP estimates, the EMTP genetic contribution to PD heritability was estimated to be 2.14 %. When considering 22.9 % as the latest estimate of total heritability for PD using the same methodological approach4, the EMTP explains roughly 9.34 % of the overall heritability of PD.

Additionally, we explored what genes contributed to a greater extent to the PD heritability estimates within the EMTP. Nineteen genes surpassed multiple test correction after the SKAT analysis with false discovery rate adjusted p-values < 0.05 (Supplementary Table 5). Only GAK has been previously associated with PD.

3. Polygenic risk profiles versus disease status and age at onset

EMTP risk profiling versus disease status was found to be significantly associated with PD (random-effects p-value = 2.55E-12, beta = 0.227, se = 0.032) and an odds ratio of 1.20 per standard deviation increase in the PRS from the population mean (see Figure 2 for forest plot showing effect estimates within gene sets). Heterogeneity was estimated at an I2 of 59.32% across all arrays. In all datasets, the directionality of effect was similar. The data obtained from the different datasets strongly enhances the general conclusion for our study. EMTP risk profiling versus age at onset was not found to be significantly associated (random-effects p-value = 0.182).

Figure 2. Polygenic Risk Score across EMTP.

Figure 2.

Open in a new tab

Forest plots of PRS estimates across cohorts. For each effect estimate, the size of the square is proportional to the size of the study, with 95% confidence intervals as horizontal error bars. The random-effects estimates are represented by the red diamond, with fixed-effects in blue, and the centerline of each diamond representing the summary PRS estimate. The number of samples per cohort are reported as follows: SHULMAN: 789 cases and 195 controls, MCGILL: 583 cases and 906 controls, SPAIN3: 2120 cases and 1333 controls, TUBI: 741 cases and 944 controls, VANCE: 621 cases and 303 controls, OSLO: 476 cases and 462 controls, HBS:, PDBP: 543 cases and 284 controls, FINLAND: 386 cases and 493 controls, PPMI: 363 cases and 165 controls.

4. Mendelian randomization analysis to explore pathway specific quantitative trait loci (QTL)

We also tested a number of QTL resources across varied tissues of interest to make functional inferences relating to biological features that might be involved in the genetic etiology of the disease. After adjustment for false discovery rate correction, 11 EMTP related genes showed functional consequence via two-sample Mendelian randomization in expression and methylation datasets (see Table 1). Increased blood expression of VAMP4, ARL8B and GAK was found to be inversely associated to PD risk, while a positive risk association was found for RABGEF1, VAMP8, CLTCL1 and ITSN1. Increased brain expression of SH3GL2 and GAK was causally linked to PD risk while ARL8B expression exhibited a negative association. mQTL Mendelian randomization analyses revealed 8 CpG sites linked to PD risk after multiple test correction. Increased methylation of GAK and HSPA1B was found to be positively associated with disease while PLEKHM1 and VPS39 showed a negative directionality versus PD risk.

Table 1.

Significant functional associations for EMTP related genes via two-sample Mendelian randomization

Gene Probe beta se P-adjusted Data source Analyte Top QTL SNP Chr Bp
VAMP4 ILMN_1761363 −0.121 0.022 5.21E-06 Expression (Westra et al., 2013) Blood rs10913529 1 171679465
VAMP4 ILMN_1804676 −0.156 0.029 5.21E-06 Expression (Westra et al., 2013) Blood rs11580522 1 171720978
ARL8B ILMN_1752837 −0.174 0.045 4.80E-03 Expression (Westra et al., 2013) Blood rs4499578 3 5167618
GAK ILMN_1813775 −0.090 0.025 1.12E-02 Expression (Westra et al., 2013) Blood rs6964 4 843184
RABGEF1 ILMN_2230577 0.266 0.080 1.94E-02 Expression (Westra et al., 2013) Blood rs4718424 7 66376095
VAMP8 ILMN_2190084 0.041 0.012 2.20E-02 Expression (Westra et al., 2013) Blood rs1058588 2 85808871
ITSN1 ILMN_1718769 0.077 0.025 3.66E-02 Expression (Westra et al., 2013) Blood rs2251854 21 35056247
CLTCL1 ILMN_1694584 0.066 0.022 3.88E-02 Expression (Westra et al., 2013) Blood rs2854643 22 19189706
SH3GL2 ENSG00000107295 0.252 0.045 2.45E-06 Expression (Qi et al., 2018) Brain rs10756899 9 17684784
GAK ENSG00000178950 0.508 0.103 3.47E-05 Expression (Qi et al., 2018) Brain rs11248057 4 906131
ARL8B ENSG00000134108 −0.033 0.010 1.90E-02 Expression (Qi et al., 2018) Brain rs6787725 3 5201054
GAK cg14517359 0.158 0.030 1.51E-05 Methylation (Qi et al., 2018) Brain rs3775127 4 886061
HSPA1B cg00970279 0.048 0.010 4.99E-05 Methylation (Qi et al., 2018) Brain rs506770 6 31785228
HSPA1B cg16237409 0.051 0.011 4.99E-05 Methylation (Qi et al., 2018) Brain rs506770 6 31785228
PLEKHM1 cg06925179 −0.094 0.026 4.92E-03 Methylation (Qi et al., 2018) Brain rs4523963 17 43570226
GAK cg23992470 0.093 0.027 1.20E-02 Methylation (Qi et al., 2018) Brain rs114066161 4 825635
PLEKHM1 cg04703951 −0.144 0.043 1.20E-02 Methylation (Qi et al., 2018) Brain rs4523963 17 43570226
VPS39 cg04726019 −0.092 0.028 1.28E-02 Methylation (Qi et al., 2018) Brain rs57393446 15 42494536
GAK cg26978381 0.119 0.037 1.38E-02 Methylation (Qi et al., 2018) Brain rs2306253 4 884475
Open in a new tab

Chr; chromosome; Bp, base pair; se, standard error; P-adjusted, adjusted false discovery rate p value after performing Mendelian randomization between QTL resources (exposures) versus Nalls et al., 2018 (outcome)

Discussion

The present study provides compelling evidence for EMTP variation in PD risk. Using a set of pathway-specific genes/variants, we demonstrate that the phenotypic variance attributed to this complex process contributes to the heritability of PD outside of what is explained by GWAS.

Genetic variance in PD has been estimated to be ~ 22.9 % 4 and to date only a small proportion of the total heritability has been explained4. Our data shows that about 2.14 % of the common variant heritability in PD can be explained by genetic variants in the EMTP, explaining roughly 9.34 % of the overall estimates in PD. Here, we have only included the transcribed genes assuming the limitation that the heritability residing in regulatory regions adjacent to the upstream or downstream of these genes is not included in this estimate. The reported estimates suggest that there are additional PD genes within the EMTP yet to be discovered. Our results are in concordance with the largest and most recent meta-analysis in PD which tested up to 10,651 biological pathways. Among these, only 27 were enriched for PD associations after multiple testing correction4, and notably endosomal trafficking was included with significant p-value.

PRS has previously demonstrated how cumulative small effect variants can contribute to PD risk and age at onset30. PRS has emerged as a probabilistic tool to infer how independent loci with small effects can be collapsed into a single numerical value useful in predicting the risk of a disease. It has been successfully used to dissect the genetic complexity of brain disorder and quantify risk in many common complex conditions40,41 for early detection, prevention and therapeutics42,43. Here we assessed the overall cumulative contribution of EMTP specific loci/genes in PD risk through PRS calculations. Our data show the PRS based on 252 EMTP gene regions is associated with a 1.2 times increased of PD risk per standard deviation, confirming a cumulative pathogenic effect of EMPT variants in PD.

When assessing whether the EMTP had any effect on the AAO we did not find any significant association. PD genetic studies using AAO as a predictor report greater genetic burden with earlier AAO44,29. Linking specific phenotypic aspects of PD to pathways constitutes a critical effort but remains challenging since larger longitudinal cohorts of well clinically characterized patients are necessary. We believe that the PRS approach we report here should be further explored in other clinical outcomes related to PD.

Growing evidence has suggested that abnormalities in endosomes or dysregulation in their trafficking plays an important role in several neurological diseases such as Alzheimer’s disease or Lewy body dementia45,46. This pathway is emerging as key to understanding the mechanisms underlying both protein degradation and neurodegeneration. We performed SKAT analysis in order to nominate the genes that contribute the most to the risk of PD within the EMTP. Nineteen genes were significant after FDR correction, including PLEKHM1, IQSEC3, GAK, VPS39, AMPH, SMURF1, CHMP1A, AP2B1, VPS28, VAMP8, AP3B2, RABGEF1, PLEKHG7, TBC1D15, AP4E1, AP2M1, STX8, ARFGAP2, and AP3M1. Interestingly, nine of them are involved in cargo degradation and eight of them play a role in cargo uptake which strongly suggests that these 2 sub-pathways are specially relevant for PD risk. Only GAK has already been associated with PD risk through GWAS30,31,4, and for the remaining eighteen genes not identified in previous studies, only five including IQSEC3, VPS39, CHMP1A, AP3B2 and AP4E1, have been associated with neurological disorders. For instance, IQSEC3 protein expression is downregulated in the brain of patients that suffer from autism and schizophrenia47, and loss-of-function mutations in CHMP1A cause pontocerebellar hypoplasia and microcephaly48. Autosomal recessive mutations in AP3B2 cause early-onset epileptic encephalopathy49 and a mutation in AP4E1 has been found in a familiar case of spastic paraplegia50. Furthermore, QTL analysis using Mendelian randomization indicates that five genes among the ones nominated by SKAT showed functional relevance associated with PD (PLEKHM1, VPS39, GAK, VAMP8, and RABGEF1).

The present study shows that EMTP genes contribute to PD risk. However, we are aware of some limitations. First our work is restricted by the sensitivity of array-based genotyping methods used to calculate heritability. Likewise, heritability estimates might be biased, appearing lower than what should be expected. The remaining variance in PD etiology not ascribed to the EMTP identified heritable factors, suggests a contribution of joint effects of rare and structural variants poorly covered by current genotyping technology. Therefore, further comprehensive studies are needed to identify novel variants within the EMTP that are relevant for disease. Additionally, there may be some selection bias in the PRS, although the meta-analysis suggest consistent effect estimates across studies.

The endosomal trafficking membrane system is a complex and dynamic process where internalized proteins, receptors and ligands interact and move between organelles in order to fulfill the cell’s needs. Here we point to PD, at least to some extent, as an endosomal membrane-trafficking disease. This study demonstrates the contribution of this unknown pathway to the oligogenic nature of PD and highlights the importance of pathway-based comprehensive approaches as a way to prioritize functional studies. Following the most recent meta-analysis4, pathway analysis established that the nominated candidate genes were linked to autophagy, mitochondrial biology, immune response, and lysosomal function. We envisage that our methodology could be applied to remaining relevant pathways in PD as a way to better understand genetic risk factors. Unraveling the genetics underlying PD will provide with therapeutic options for drug discovery and ultimately could lead to the development of effective interventions.

Supplementary Material

Supp TableS1
Supp TableS2
Supp TableS3
Supp TableS4
Supp TableS5
Supp figS1

Figure Supplementary 1. Percentage of heritability per simulation based on 50 randomly generated sets of 252 genes. Heritability estimates of random sampling sets, expressed as mean +/− SD.

ACKNOWLEDGMENTS:

We would like to thank all of the subjects who donated their time and biological samples to be a part of this study. This work was supported in part by the Intramural Research Programs of the National Institute of Neurological Disorders and Stroke (NINDS), the National Institute on Aging (NIA), and the National Institute of Environmental Health Sciences both part of the National Institutes of Health, Department of Health and Human Services; project numbers 1ZIA-NS003154, Z01-AG000949-02 and Z01-ES101986. In addition this work was supported by the Department of Defense (award W81XWH-09–2-0128), and The Michael J Fox Foundation for Parkinson’s Research.

Funding agencies: This research was supported in part by the Intramural Research Program of the National Institutes of Health (National Institute on Aging, National Institute of Neurological Disorders and Stroke; project numbers: project numbers 1ZIA-NS003154, Z01-AG000949–02 and Z01-ES101986. In addition this work was supported by the Department of Defense (award W81XWH-09–2-0128), and The Michael J Fox Foundation for Parkinson’s Research

Footnotes

Financial Disclosures:

Mike A. Nalls’ participation is supported by a consulting contract between Data Tecnica International and the National Institute on Aging, NIH, Bethesda, MD, USA, as a possible conflict of interest Dr. Nalls also consults for SK Therapeutics Inc, Lysosomal Therapeutics Inc, the Michael J. Fox Foundation and Vivid Genomics among others. No other disclosures were reported.

REFERENCES

  • 1.Surmeier DJ, Obeso JA & Halliday GM Selective neuronal vulnerability in Parkinson disease. Nat. Rev. Neurosci 18, 101–113 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Hernandez DG, Reed X & Singleton AB Genetics in Parkinson disease: Mendelian versus non-Mendelian inheritance. J. Neurochem 139 Suppl 1, 59–74 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Billingsley KJ, Bandres-Ciga S, Saez-Atienzar S & Singleton AB Genetic risk factors in Parkinson’s disease. Cell Tissue Res 373, 9–20 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Nalls MA et al. Parkinson’s disease genetics: identifying novel risk loci, providing causal insights and improving estimates of heritable risk. bioRxiv 388165 (2018). 10.1101/388165 [DOI] [Google Scholar]
  • 5.Yang J, Lee SH, Goddard ME & Visscher PM GCTA: a tool for genome-wide complex trait analysis. Am. J. Hum. Genet 88, 76–82 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Maher B Personal genomes: The case of the missing heritability. Nature 456, 18–21 (2008). [DOI] [PubMed] [Google Scholar]
  • 7.Fujita KA et al. Integrating pathways of Parkinson’s disease in a molecular interaction map. Mol. Neurobiol 49, 88–102 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Abeliovich A & Gitler AD Defects in trafficking bridge Parkinson’s disease pathology and genetics. Nature 539, 207–216 (2016). [DOI] [PubMed] [Google Scholar]
  • 9.Schreij AMA, Fon EA & McPherson PS Endocytic membrane trafficking and neurodegenerative disease. Cell. Mol. Life Sci 73, 1529–1545 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Perrett RM, Alexopoulou Z & Tofaris GK The endosomal pathway in Parkinson’s disease. Mol. Cell. Neurosci 66, 21–28 (2015). [DOI] [PubMed] [Google Scholar]
  • 11.McGough IJ et al. Retromer binding to FAM21 and the WASH complex is perturbed by the Parkinson disease-linked VPS35(D620N) mutation. Curr. Biol 24, 1670–1676 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Beilina A et al. Unbiased screen for interactors of leucine-rich repeat kinase 2 supports a common pathway for sporadic and familial Parkinson disease. Proc. Natl. Acad. Sci. U. S. A 111, 2626–2631 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.MacLeod DA et al. RAB7L1 interacts with LRRK2 to modify intraneuronal protein sorting and Parkinson’s disease risk. Neuron 77, 425–439 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Ambroso MR, Hegde BG & Langen R Endophilin A1 induces different membrane shapes using a conformational switch that is regulated by phosphorylation. Proc. Natl. Acad. Sci. U. S. A 111, 6982–6987 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Song L et al. Auxilin Underlies Progressive Locomotor Deficits and Dopaminergic Neuron Loss in a Drosophila Model of Parkinson’s Disease. Cell Rep 18, 1132–1143 (2017). [DOI] [PubMed] [Google Scholar]
  • 16.Yim Y-I et al. Endocytosis and clathrin-uncoating defects at synapses of auxilin knockout mice. Proc. Natl. Acad. Sci. U. S. A 107, 4412–4417 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Stuchell-Brereton MD et al. ESCRT-III recognition by VPS4 ATPases. Nature 449, 740–744 (2007). [DOI] [PubMed] [Google Scholar]
  • 18.Lin G et al. Phospholipase PLA2G6, a Parkinsonism-Associated Gene, Affects Vps26 and Vps35, Retromer Function, and Ceramide Levels, Similar to α-Synuclein Gain. Cell Metab (2018). 10.1016/j.cmet.2018.05.019 [DOI] [PubMed]
  • 19.Subramanian A et al. Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles. Proceedings of the National Academy of Sciences 102, 15545–15550 (2005). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Liberzon A et al. The Molecular Signatures Database Hallmark Gene Set Collection. Cell Systems 1, 417–425 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.La Manno G et al. Molecular Diversity of Midbrain Development in Mouse, Human, and Stem Cells. Cell 167, 566–580.e19 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Zheng J et al. LD Hub: a centralized database and web interface to perform LD score regression that maximizes the potential of summary level GWAS data for SNP heritability and genetic correlation analysis. Bioinformatics 33, 272–279 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Bulik-Sullivan BK et al. LD Score regression distinguishes confounding from polygenicity in genome-wide association studies. Nat. Genet 47, 291–295 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Finucane HK et al. Partitioning heritability by functional annotation using genome-wide association summary statistics. Nat. Genet 47, 1228–1235 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.International Parkinson Disease Genomics Consortium et al. Imputation of sequence variants for identification of genetic risks for Parkinson’s disease: a meta-analysis of genome-wide association studies. Lancet 377, 641–649 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.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 7, e1002142 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Hernandez DG et al. Genome wide assessment of young onset Parkinson’s disease from Finland. PLoS One 7, e41859 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Ripatti S et al. A multilocus genetic risk score for coronary heart disease: case-control and prospective cohort analyses. Lancet 376, 1393–1400 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Nalls MA et al. Genetic risk and age in Parkinson’s disease: Continuum not stratum. Mov. Disord 30, 850–854 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Nalls MA et al. Large-scale meta-analysis of genome-wide association data identifies six new risk loci for Parkinson’s disease. Nat. Genet 46, 989–993 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Chang D et al. A meta-analysis of genome-wide association studies identifies 17 new Parkinson’s disease risk loci. Nat. Genet 49, 1511–1516 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Bandrés-Ciga S et al. Genome-wide assessment of Parkinson’s disease in a Southern Spanish population. Neurobiol. Aging 45, 213e3–213.e9 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Euesden J, Lewis CM & O’Reilly PF PRSice: Polygenic Risk Score software. Bioinformatics 31, 1466–1468 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Zhu Z et al. Integration of summary data from GWAS and eQTL studies predicts complex trait gene targets. Nat. Genet 48, 481–487 (2016). [DOI] [PubMed] [Google Scholar]
  • 35.GTEx Consortium. Human genomics. The Genotype-Tissue Expression (GTEx) pilot analysis: multitissue gene regulation in humans. Science 348, 648–660 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Fromer M et al. Gene expression elucidates functional impact of polygenic risk for schizophrenia. Nat. Neurosci 19, 1442–1453 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Ng B et al. An xQTL map integrates the genetic architecture of the human brain’s transcriptome and epigenome. Nat. Neurosci 20, 1418–1426 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Trabzuni D et al. Quality control parameters on a large dataset of regionally dissected human control brains for whole genome expression studies. J. Neurochem 119, 275–282 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Westra H-J et al. Systematic identification of trans eQTLs as putative drivers of known disease associations. Nat. Genet 45, 1238–1243 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Wang T et al. Polygenic risk for five psychiatric disorders and cross-disorder and disorder-specific neural connectivity in two independent populations. Neuroimage Clin 14, 441–449 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Reus LM et al. Association of polygenic risk for major psychiatric illness with subcortical volumes and white matter integrity in UK Biobank. Sci. Rep 7, 42140 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Natarajan P et al. Polygenic Risk Score Identifies Subgroup With Higher Burden of Atherosclerosis and Greater Relative Benefit From Statin Therapy in the Primary Prevention Setting. Circulation 135, 2091–2101 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Maas P et al. Breast Cancer Risk From Modifiable and Nonmodifiable Risk Factors Among White Women in the United States. JAMA Oncol 2, 1295–1302 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Lill CM et al. Impact of Parkinson’s disease risk loci on age at onset. Mov. Disord 30, 847–850 (2015). [DOI] [PubMed] [Google Scholar]
  • 45.Gao S, Casey AE, Sargeant TJ & Mäkinen V-P Genetic variation within endolysosomal system is associated with late-onset Alzheimer’s disease. Brain 141, 2711–2720 (2018). [DOI] [PubMed] [Google Scholar]
  • 46.Neefjes J & van der Kant R Stuck in traffic: an emerging theme in diseases of the nervous system. Trends Neurosci 37, 66–76 (2014). [DOI] [PubMed] [Google Scholar]
  • 47.Ellis SE, Panitch R, West AB & Arking DE Transcriptome analysis of cortical tissue reveals shared sets of downregulated genes in autism and schizophrenia. Transl. Psychiatry 6, e817 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Mochida GH et al. CHMP1A encodes an essential regulator of BMI1-INK4A in cerebellar development. Nat. Genet 44, 1260–1264 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Assoum M et al. Autosomal-Recessive Mutations in AP3B2, Adaptor-Related Protein Complex 3 Beta 2 Subunit, Cause an Early-Onset Epileptic Encephalopathy with Optic Atrophy. Am. J. Hum. Genet 99, 1368–1376 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Kong X-F et al. A novel homozygous p.R1105X mutation of the AP4E1 gene in twins with hereditary spastic paraplegia and mycobacterial disease. PLoS One 8, e58286 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Robak LA et al. Excessive burden of lysosomal storage disorder gene variants in Parkinson’s disease. Brain 140, 3191–3203 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp TableS1
NIHMS1005117-supplement-Supp_TableS1.docx (15.7KB, docx)
Supp TableS2
NIHMS1005117-supplement-Supp_TableS2.doc (265KB, doc)
Supp TableS3
NIHMS1005117-supplement-Supp_TableS3.docx (28.3KB, docx)
Supp TableS4
NIHMS1005117-supplement-Supp_TableS4.docx (39.7KB, docx)
Supp TableS5
NIHMS1005117-supplement-Supp_TableS5.docx (16.9KB, docx)
Supp figS1

Figure Supplementary 1. Percentage of heritability per simulation based on 50 randomly generated sets of 252 genes. Heritability estimates of random sampling sets, expressed as mean +/− SD.

NIHMS1005117-supplement-Supp_figS1.tiff (381.2KB, tiff)

RESOURCES