Data-Driven Genome-Wide Association Study of Pain Trajectories in Persons with Parkinson’s Disease
Shiying Liu1, Doug Gunzler2, Steven Gunzler3, Dana Crawford1, Farren Briggs1
1Department of Population and Quantitative Health Sciences, 2Center for Health Care Research and Policy, Case Western Reserve University, 3Neurological Institute, University Hospitals Cleveland Medical Center and Case Western Reserve University School of Medicine
Objective:

To conduct the first genome-wide association study (GWAS) of pain in persons with Parkinson’s disease (PWP).

Background:

Pain is a common and complex non-motor symptom experienced by ~60% of PWP. Little is known about the genetic and non-genetic determinants of pain in PWP.

Design/Methods:

Using data spanning ~4.5 years from 8,612 PWP who are early in their disease course (<3 years), and available through the Fox Insight Data Exploration Network, we empirically discerned 5 subgroups of PWP with similar longitudinal patterns of pain using latent class growth analysis. 4,411 PWP were also genotyped across two customized Illumina platforms that had significant overlap. Quality control criteria (minor allele frequency <1%, Hardy Weinberg Equilibrium p<5x10-5) were imposed prior imputation using the TOPMed reference panel. After post-imputation quality control (same as above + imputation R2>70%), there were 7.7x106 genetic variants. We retained 4,159 PWP who self-identified as White and who genetically clustered with other European samples. For this analysis, we focused on a GWAS of the most severe pain subgroup (N=97) to the least impaired subgroup (N=934). Gene-based and pathway enrichment analyses were also conducted.

Results:

For the GWAS of extreme pain phenotypes, top genic SNP association were MAPK8-rs72794357 (OR=4.6; p=1x10-6) and VLDLR-rs4741753 (OR=2.2; p=1.5x10-6) and top gene-based results included CTNNB1 (p=3.5x10-5) and KLK7 (p=7.7x10-5). These results are intriguing since MAPK8 is involved in opioid receptor pathways, VLDLR is a ApoE receptor involved in brain development through Reelin signaling, CTNNB1 promotes neurogenesis and de novo mutations have been associated with polyneuropathy in lower limbs, and KLK7 is an astrocyte‐derived amyloid-β degrading enzyme. Overall there was an enrichment of genes associated with oxytocin signaling (p=0.14), opioid signaling (p=0.045), and Wnt signaling pathway (p=0.021).

Conclusions:

The findings are consistent with known neuropathic mechanisms, thus therapeutic targets for these mechanisms may be relevant for PWP. 

10.1212/WNL.0000000000203039