Telomere length and Alzheimer's disease endophenotypes: a Mendelian Randomization study

Abstract

Background: Telomere length (TL) is a putative biomarker of biological aging and aging-related outcomes. Observational studies are limited to conclude whether TL is causally associated with Alzheimer’s Disease (AD) or a marker of an underlying pathological process. Mendelian randomization (MR) was developed for assessing causality using genetic variants in epidemiological research. Objective: Our main objective was to test the potential causal role of TL in cognitive performance, brain vulnerability and cerebrospinal fluid (CSF) AD biomarkers through a MR analysis. Methods: Our analysis was conducted in the context of the ALFA (ALzheimer and FAmilies) study. We created episodic memory, executive function, and global cognitive performance composites. We calculated an AD signature as a composite measure reflecting cortical thickness of specific AD vulnerable brain regions. We measured CSF levels of core AD and neurodegeneration biomarkers, that is, amyloid-β (Aβ) 42, Aβ40, p-tau, t-tau, and neurofilament light (NfL), using NeuroToolKit and Elecsys® immunoassays. Genome-wide genotyping and imputation were performed. MR analyses using 7 Single Nucleotide Polymorphisms (SNPs) (rs10936599, rs7675998, rs2736100, rs9420907, rs755017, rs11125529, rs8105767) associated with TL and identified in Genome-wide Association studies (GWAS) were used as instrumental variables to determine the predicted effect of TL on AD endophenotypes: cognitive performance, AD brain signature, and CSF biomarkers. Causal effects of TL were estimated using the inverse-variance weighted method. Sensitivity analyses using the MR-Egger method were conducted to correct for directional pleiotropic effects. Stratified analyses by APOE-ɛ4 status were performed. Results: MR analysis revealed a significant association between SNPs predicting shorter TL and worse performance on executive function. Effect modification by APOE-ɛ4 genotype was also suggested: associations between SNPs predicting shorter TL and worse executive function were only observed among APOE- ɛ4 non-carriers. Conclusion: We showed an association between SNPs predicting shorter TL and cognitive-related endophenotypes of AD. Our results suggest a potential causal role of telomeres in AD vulnerability, which may compromise cognitive resilience, specifically in individuals not at increased genetic risk of AD. Further observational and genetic analyses in larger datasets are warranted to better understand these associations.