The Genetic Architecture of Parkinson Disease in Spain: Characterizing Population-Specific Risk, Differential Haplotype Structures, and Providing Etiologic Insight

Abstract

Background: The Iberian Peninsula stands out as having variable levels of population admixture and isolation, making Spain an interesting setting for studying the genetic architecture of neurodegenerative diseases.

Objectives: To perform the largest PD genome-wide association study restricted to a single country.

Methods: We performed a GWAS for both risk of PD and age at onset in 7,849 Spanish individuals. Further analyses included population-specific risk haplotype assessments, polygenic risk scoring through machine learning, Mendelian randomization of expression, and methylation data to gain insight into disease-associated loci, heritability estimates, genetic correlations, and burden analyses.

Results: We identified a novel population-specific genome-wide association study signal at PARK2 associated with age at onset, which was likely dependent on the c.155delA mutation. We replicated four genome-wide independent signals associated with PD risk, including SNCA, LRRK2, KANSL1/MAPT, and HLA-DQB1. A significant trend for smaller risk haplotypes at known loci was found compared to similar studies of non-Spanish origin. Seventeen PD-related genes showed functional consequence by two-sample Mendelian randomization in expression and methylation data sets. Long runs of homozygosity at 28 known genes/loci were found to be enriched in cases versus controls.

Conclusions: Our data demonstrate the utility of the Spanish risk haplotype substructure for future fine-mapping efforts, showing how leveraging unique and diverse population histories can benefit genetic studies of complex diseases. The present study points to PARK2 as a major hallmark of PD etiology in Spain. © 2019 International Parkinson and Movement Disorder Society.

Keywords: Parkinson's disease; Spanish population; age at onset; polygenic risk score; risk haplotype.

FIG. 1.

Manhattan plot showing results of PD GWA testing. Based on unrelated individuals (4,639 cases and 2,949 controls) using 9,945,565 SNPs. Four genome-wide significant loci were identified: SNCA, LRRK2, HLA-DQB1, and MAPT.

FIG. 2.

Manhattan plot showing results of PD GWA with AAO testing. Based on 3,997 unrelated cases with available age at onset information using 9,945,565 SNPs. One genome-wide significant loci was identified: PARK2.

FIG. 3.

Polygenic risk score versus disease status and AAO. (A) Polygenic risk score versus disease status. R² estimates at various P-value thresholds. (B) ORs by quantile of PD polygenic risk score. (C) Polygenic risk score versus AAo. R² estimates at various P-value thresholds. (D) ORs by quantile of AAO polygenic risk score.