Tumor Necrosis Factor Inhibition and Parkinson Disease: A Mendelian Randomization Study.

Karin Wirdefeldt,Alastair J Noyce,Xiaoying Kang,Dylan M Williams,Nancy L Pedersen,Alexander Ploner,Sara Bandres-Ciga

doi:10.1212/wnl.0000000000011630

Abstract

ObjectiveTo evaluate the effects of long-term tumor necrosis factor (TNF) inhibition on the risk and age at onset of Parkinson disease (PD), we performed a 2-sample Mendelian randomization study using genome-wide association studies (GWAS) summary statistics.MethodsGenetic variants in the vicinity of TNFRSF1A, the gene encoding TNF receptor 1 (TNFR1), were identified as predictive of pharmacologic blockade of TNFR1 signaling by anti-TNF therapy, based on genetic associations with lower circulating C-reactive protein (CRP; GWAS n = 204,402). The effects of TNF-TNFR1 inhibition were estimated for PD risk (ncases/controls = 37,688/981,372) and age at PD onset (n = 28,568) using GWAS data from the International Parkinson's Disease Genomics Consortium and 23andMe, Inc. To validate variants as proxies of long-term anti-TNF treatment, we also assessed whether variant associations reflected anticipated effects of TNFR1 inhibition on Crohn disease, ulcerative colitis, and multiple sclerosis risk (n = 38,589-45,975).ResultsTNF-TNFR1 signaling inhibition was not estimated to affect PD risk (odds ratio [OR] per 10% lower circulating CRP = 0.99; 95% confidence interval [CI] 0.91–1.08) or age at onset (0.13 years later onset; 95% CI −0.66 to 0.92). In contrast, genetically indexed TNF-TNFR1 signaling blockade predicted reduced risk of Crohn disease (OR 0.75; 95% CI 0.65–0.86) and ulcerative colitis (OR 0.84; 95% CI 0.74–0.97) and increased multiple sclerosis risk (OR 1.57; 95% CI 1.36–1.81). Findings were consistent across models using different genetic instruments and Mendelian randomization estimators.ConclusionsOur findings do not imply that TNF-TNFR1 signaling inhibition will prevent or delay PD onset.Classification of EvidenceThis study provides Class II evidence that TNF-TNFR1 signaling inhibition is not associated with the risk or age at onset of PD.

Highlights

Our findings do not imply that tumor necrosis factor (TNF)-TNF receptor 1 (TNFR1) signaling inhibition will prevent or delay Parkinson disease (PD) onset
Classification of Evidence This study provides Class II evidence that TNF-TNFR1 signaling inhibition is not associated with the risk or age at onset of PD
One single nucleotide polymorphisms (SNPs) in the region—rs767455—was retained following conservative linkage disequilibrium (LD) clumping. This exonic, synonymous SNP has a minor allele frequency of approximately 0.4 and had the strongest evidence for association with C-reactive protein (CRP) of SNPs in the region (p = 7.8 × 10−8). It is in high LD with the presumed functional variant in this region (r2 = 0.84)[32] and the variant is associated with circulating white blood cell count (WBC) (p = 4.2 × 10−3) and mean platelet volume (MPV) (p = 3.7 × 10−6), with consistent directions for all 3 inflammatory markers

Summary

Methods

Study Overview This study used a 2-sample Mendelian randomization design (figure 1).[8] We focused on single nucleotide polymorphisms (SNPs) in the vicinity of the gene TNFRSF1A, which encodes TNFR1, the principal effector of proinflammatory signaling following TNF agonism.[2] We determined the extent to which genetic variation in this gene region is indicative of long-term TNF signaling blockade, based on associations of the regional SNPs with circulating markers of systemic inflammation reported by large genome-wide association studies (GWAS) of C-reactive protein (CRP) and cell count measures.[9,10] We combined data on selected variants with corresponding association statistics from GWAS of PD traits to test the effects of TNF-TNFR1 signaling inhibition on (1) the risk of PD, measured as self-reported or clinically ascertained disease status; and (2) age at PD onset, measured as self-reported age at motor symptoms manifestation, or age at PD diagnosis when the former was unavailable.[11,12] This study provides Class II evidence that TNF-TNFR1 signaling inhibition is not associated with PD risk or age at onset, according to criteria of the American Academy of Neurology. To ensure that the CRP associations were not chance findings, we examined associations of the SNPs with 2 cell count markers of inflammation—white blood cell count (WBC) and mean platelet volume (MPV)—from independent GWAS data (table 1).[10,18]

Results

Discussion

Conclusion