Multi‐cohort genome‐wide association study on flortaucipir PET identifies novel risk loci associated with tau deposition and its role in AD pathology
Tamil Iniyan Gunasekaran, Annie J Lee, Logan C Dumitrescu, Elizabeth C. Mormino, Reisa A. Sperling, Andrew J. Saykin, Timothy J. Hohman, Badri N Vardarajan,- Psychiatry and Mental health
- Cellular and Molecular Neuroscience
- Geriatrics and Gerontology
- Neurology (clinical)
- Developmental Neuroscience
- Health Policy
- Epidemiology
Abstract
Background
Amyloid‐β plaques and tau neurofibrillary tangles are the pathological hallmarks of Alzheimer’s disease (AD). Previously, genetic studies on amyloid‐β mediated AD pathogenesis dominated the field, and genetic studies on tau pathology are limited by comparison. Prior studies suggest that the spread of tau neurofibrillary tangles within the medial temporal lobe occurs following abnormal deposition of amyloid‐β, leading to cognitive impairment. Although amyloid‐β and tau both play a role in AD pathogenesis, exact genetic pathways associated with them are still unknown.
Method
We performed genome‐wide association study (GWAS) with brain Tau standard uptake value ratios (SUVRs) measured from positron emission tomography (PET) images from the A4 ((N = 311 with preclinical AD;female = 62%;tau mean‐SUVR = 1.076±0.06) and ADNI studies (N = 375‐280 cognitively normal, 76 mild cognitively impaired and 19 AD;female = 53%;tau mean‐SUVR = 1.12±0.13). Models were adjusted for age, sex and APOE‐ε4 dosage. Additionally, we evaluated the association of amyloid‐β and tau SUVRs with AD polygenic risk scores (PRS). PRS was calculated using genome‐wide variants, excluding 1MB region flanking APOE, and effect sizes from Kunkle et al GWAS study of clinical AD.
Result
Analysis of tau‐SUVR in 686 participants, identified five genome‐wide significant loci: rs78636169 (P = 1.37×10−9) in JARID2, rs114272033 (P = 7.87×10−9) in ISCA1P2, rs7292124 (P = 1.73×10−8) in RP1‐272J12.1, rs114742337 (P = 2.85×10−8) in RP4‐771M4.2, and rs138338441 (P = 3.91×10−8) in AC092684.1. Rs13412014 in AC007364.1 and rs9393067 in RP11‐314C16.1 were also associated with amyloid‐SUVR in a previous large GWAS. Two loci previously associated with clinical AD in a recent large GWA study, INPP5D (rs10933431) and SEC61G (rs76928645), were also nominally associated (P<0.05) with tau deposition in our study. Genes associated with Tau‐SUVR are enriched in the Schaffer axonal collaterals (P<1×10−05) that form synapses in hippocampus. Genome‐wide PRS of AD (excluding APOE region) was strongly associated with amyloid‐SUVR (P = 3.21×10−11;R2 = 0.077) but only weakly associated with tau‐SUVR (P = 1.38×10−04;R2 = 0.027). We are in the process of replicating our results in an independent cohort of 700 participants with measurements of tau deposition.
Conclusion
We identified five novel genetic loci associated with tau pathology. Several genes involved in tau deposition were also associated with amyloid load in the brain. Taken together, our findings will clarify the genetic pathways that effect both amyloid and tau pathology in AD.