Search

Summary: Background: Neuroimaging-based brain-age delta has been shown to be a mediator linking cardiovascular risk factors to cognitive function. We aimed to assess the mediating role of brain-age delta in the association between modifiable risk factors of dementia and longitudinal cognitive decline in middle-aged and older individuals who are asymptomatic, stratified by Alzheimer's disease pathology. We also explored whether the mediation effect is specific to cognitive domain. Methods: In this cohort study, we included participants from the ALFA+ cohort aged between 45 years and 65 years who were cognitively unimpaired and who had available structural MRI, cerebrospinal fluid β-amyloid (Aβ)42 and Aβ40 measurements obtained within 1 year of each other, modifiable risk factors assessment, and cognitive evaluation over 3 years. Participants were recruited from the Barcelonaβeta Brain Research Center (Barcelona, Spain). Included individuals underwent a first assessment between Oct 25, 2016, and Jan 28, 2020, and a follow-up cognitive assessment 3·28 (SD 0·27) years later. We computed brain-age delta and composites of different cognitive function domains (preclinical Alzheimer's cognitive composite [PACC], attention, executive function, episodic memory, visual processing, and language). We used partial least squares path modelling to explore mediation effects in the associations between modifiable risk factors (including cardiovascular, mental health, mood, metabolic or endocrine history, and alcohol use) and changes in cognitive composites. To assess the role of Alzheimer's disease pathology, we computed separate models for Aβ-negative and Aβ-positive individuals. Findings: Of the 419 participants enrolled in ALFA+, 302 met our inclusion criteria, of which 108 participants were classified as Aβ-positive and 194 as Aβ-negative. In Aβ-positive individuals, brain-age delta partially mediated (percent mediation proportion 15·73% [95% CI 14·22–16·66]) the association between modifiable risk factors and decline in overall cognition (across cognitive domains). Brain-age delta fully mediated (mediation proportion 28·03% [26·25–29·21]) the effect of modifiable risk factors on the PACC, wherein increased values for risk factors correlated with an older brain-age delta, and, consequently, an older brain-age delta was linked to greater PACC decline. This effect appears to be primarily driven by memory decline. Mediation was not significant in Aβ-negative individuals (3·52% [0·072–4·17]) on PACC, although path coefficients were not significantly different from those in the Aβ-positive group. Interpretation: Our findings suggest that brain-age delta captures the association between modifiable risk factors and longitudinal cognitive decline in middle-aged and older people. In asymptomatic middle-aged and older individuals who are Aβ-positive, the pathology might be the strongest driver of cognitive decline, whereas the effect of risk factors is smaller. Our results highlight the potential of brain-age delta as an objective outcome measure for preventive lifestyle interventions targeting cognitive decline. Funding: La Caixa Foundation, the TriBEKa Imaging Platform, and the Universities and Research Secretariat of the Catalan Government. Translation: For the Spanish translation of the abstract see Supplementary Materials section.

Background: While numerous studies have identified blood proteins that modulate brain aging in mice, the direct translation of these findings to human health remains a substantial challenge. Bridging this gap is critical for developing interventions that can effectively target human brain aging and associated diseases., Methods: We first identified 12 proteins with aging or rejuvenating properties in murine brains through a systematic review. Using protein quantitative trait loci data for these proteins, we developed polygenic scores to predict plasma protein levels, which we then validated in two independent human cohorts. We employed association models to explore the association between these genetically predicted protein levels and cognitive performance, focusing specifically on their interaction with key genetic markers such as sex, APOE- ε4 and Aβ42 status., Results: Predicted plasma levels of Tissue Inhibitor of Metalloproteinases 2 (TIMP2) were significantly associated with improved global cognition and memory performance in humans, also when the models were stratified by sex, APOE- ε4, and Aβ42 status., Conclusions: This finding aligns with TIMP2's brain-rejuvenating role in murine models, suggesting it as a promising therapeutic target for brain aging and age-related brain diseases in humans., Competing Interests: Conflicts of interest CC has received research support from: GSK and EISAI. CC is a member of the scientific advisory board of Circular Genomics and owns stocks. CC is a member of the scientific advisory board of ADmit. MS-C has given lectures in symposia sponsored by Almirall, Eli Lilly, Novo Nordisk, Roche Diagnostics, and Roche Farma; received consultancy fees (paid to the institution) from Roche Diagnostics; and served on advisory boards of Roche Diagnostics and Grifols. He was granted a project and is a site investigator of a clinical trial (funded to the institution) by Roche Diagnostics. In-kind support for research (to the institution) was received from ADx Neurosciences, Alamar Biosciences, Avid Radiopharmaceuticals, Eli Lilly, Fujirebio, Janssen Research & Development, and Roche Diagnostics. The remaining co-authors have no conflicts to disclose.

Imaging genetic studies aim to test how genetic information influences brain structure and function by combining neuroimaging-based brain features and genetic data from the same individual. Most studies focus on individual correlation and association tests between genetic variants and a single measurement of the brain. Despite the great success of univariate approaches, given the capacity of neuroimaging methods to provide a multiplicity of cerebral phenotypes, the development and application of multivariate methods become crucial. In this article, we review novel methods and strategies focused on the analysis of multiple phenotypes and genetic data. We also discuss relevant aspects of multi-trait modelling in the context of neuroimaging data., (© 2021 The Authors.)