Genetic meta-analysis of levodopa induced dyskinesia in Parkinson’s disease

Importance Forty percent of Parkinson’s disease patients develop levodopa-induced-dyskinesia (LiD) within 4 years of starting levodopa. The genetic basis of LiD remains poorly understood, and there have been few well powered studies.

Objective To discover common genetic variants in the PD population that increase the probability of developing LiD.

Design, setting and Participants We performed survival analyses to study the development of LiD in 5 separate longitudinal cohorts. We performed a meta-analysis to combine the results of genetic association from each study based on a fixed effects model weighting the effect sizes by the inverse of their standard error. The selection criteria was specific to each cohort. We studied individuals that were genotyped from each cohort and that passed our analysis specific inclusion criteria.

Main Outcomes and Measures We measured the time for PD patients on levodopa treatment to develop LiD as defined by reaching a score higher or equal than 2 from the MDS-UPDRS part IV, item 1, which is equivalent to a range of 26%-50% of the waking time with dyskinesia. We carried out a genome-wide analysis of the hazard ratio and the association of genome-wide SNPs with the probability of developing LiD using cox proportional hazard models (CPH).

Results This study included 2,784 PD patients of European ancestry, of whom 14.6% developed LiD. Consistent with previous studies, we found female gender (HR = 1.35, SE = 0.11, P = 0.007) and younger age at onset (HR = 1.8, SE = 0.14, P = 2 × 10⁻⁵) to increase the probability of developing LiD. We identified three loci significantly associated with time-to-LiD onset. rs72673189 on chromosome 1 (HR = 2.77, SE = 0.18, P = 1.53 × 10⁻⁸) located in the LRP8 locus, rs189093213 on chromosome 4 (HR = 3.06,, SE = 0.19, P = 2.81 × 10⁻⁹) in the non-coding RNA LINC02353 locus, and rs180924818 on chromosome 16 (HR = 3.13, SE = 0.20, P = 6.27 × 10⁻⁹) in the XYLT1 locus. Subsequent colocalization analyses on chromosome 1 identified DNAJB4 as a candidate gene associated with LiD through a change in gene expression. We computed a PRS based on our GWAS meta-analysis and found high accuracy to stratify between PD-LID and PD (AUC 83.9). We also performed a stepwise regression analysis for baseline features selection associated with LiD status. We found baseline anxiety status to be significantly associated with LiD (OR = 1.14, SE = 0.03, P = 7.4 × 10⁻⁵). Finally, we performed a candidate variant analysis and found that genetic variability in ANKK1 (rs1800497, Beta = 0.24, SE = 0.09, P = 8.89 × 10⁻³) and BDNF (rs6265, Beta = 0.19, SE = 0.10, P = 4.95 × 10⁻²) loci were significantly associated with time to LiD in our large meta-analysis.

Conclusion In this association study, we have found three novel genetic variants associated with LiD, as well as confirming reports that variability in ANKK1 and BDNF loci were significantly associated with LiD probability. A PRS nominated from our time-to-LiD meta-analysis significantly differentiated between PD-LiD and PD. In addition, we have found female gender, young PD onset and anxiety to be significantly associated with LiD.