Dynamics of cognitive variability with age and its genetic underpinning in NIHR BioResource Genes and Cognition cohort participants

The high global projected number of dementia cases by 2050 (approximately 139 million) necessitates understanding the preclinical and prodromal stages of neurodegenerative diseases. A significant challenge is the translational failure of many promising drugs, possibly because clinical trials are initiated too late in the disease progression. This study aimed to address this by establishing a large cohort to allow early intervention studies. The NIHR BioResource Genes and Cognition (G&C) cohort provides a platform for longitudinal studies to identify individuals at various stages of cognitive decline, allowing for early intervention trials. The study aimed to characterize cognitive trajectories across a wide age range, investigate the genetic underpinnings of cognitive performance, and assess the influence of established risk factors like APOE genotype and Alzheimer's disease polygenic risk scores (PRS). The potential for participant recall stratified by genotype and cognitive phenotype for future studies is highlighted.

Existing literature highlights the substantial societal burden of dementia and the frequent failure of promising drug candidates in large-scale clinical trials. Several hypotheses for this translational failure include insufficient understanding of the disease's pathophysiology, inadequate animal models, and the timing of intervention. The latter suggests that effective treatments may not show benefit in clinical trials if introduced late in the disease course. Previous research has explored the relationship between APOE genotype and cognition in smaller studies with mixed results, and similarly, investigations of AD PRS and cognition in healthy individuals have yielded inconsistent outcomes. The need for large-scale, well-characterized cohorts capable of supporting longitudinal studies of preclinical and prodromal stages of neurodegenerative diseases to maximize the efficacy of early interventions is well-established. This study built upon these existing needs and gaps in knowledge.

The G&C study enrolled 21,051 participants (aged 17–85 years) from the NIHR BioResource. Participants completed an online cognitive test battery comprising 11 tests assessing various cognitive domains. Two data-driven measures of general cognitive ability (G4 and G6) were derived through principal component analysis. Participants also provided information on demographics, lifestyle factors, and self-reported health conditions. Genome-wide genotype array data were available for a subset of participants. Statistical analyses included examination of cognitive trajectories across age and gender, assessment of associations between cognitive performance and APOE genotype, AD PRS, education, socioeconomic status, and self-reported health conditions. Genome-wide association studies (GWAS) were conducted for G4 and G6 to identify associated genetic loci. Functional annotation, pathway analysis, colocalization analysis, summary-based Mendelian randomization (SMR), and statistical fine mapping were used to explore the functional relevance of identified loci.

The study found that cognitive performance declined with age in most domains, with the exception of vocabulary (VY), which increased with age. Males generally exhibited higher scores (indicating poorer performance) on specific tests (SD, TMN, TMA, PR), while females scored lower on others. However, the overall pattern of cognitive change across genders was remarkably similar, with gender accounting for only a small percentage of the variation in cognitive phenotypes. Consistent with previous studies, lower educational attainment and higher levels of deprivation correlated with lower cognitive performance. Unexpectedly, APOE genotype had a negligible impact on cognitive performance, even after accounting for other covariates. This was also true for AD PRS; AD-PRS-high group did show some subtle increases in response time and other cognitive tests in later life but no significant changes in adjusted models. Heritability estimates ranged from 0.06 to 0.28 for the cognitive phenotypes. GWAS identified distinct genome-wide significant loci for G4 and G6. For G4 (summarizing short-term memory, fluid, and crystallized intelligence), functional annotation implicated microglial-mediated immunological processes, highlighting the involvement of TUFM, SULT1A1, and SULT1A2. For G6 (summarizing reaction time, attention, processing speed, and executive functioning), GBE1 was strongly associated, suggesting a role for glycogen metabolism in cognitive ability. Genetic correlations between G4 and G6 and childhood/adulthood intelligence were high, but the genetic correlation between summary measures of cognitive abilities (G4 and G6) and AD was weak.

The large-scale design of this study provided a unique opportunity to investigate the impact of age, sex, genetic factors and environmental factors on cognitive performance. The finding of negligible effects of APOE genotype and AD PRS on cognition in healthy individuals is noteworthy. The identified loci associated with G4 and G6 suggest distinct biological pathways underlying different aspects of cognition, distinct from those implicated in dementia risk. The implication of microglial function in G4 and glycogen metabolism in G6 provide novel targets for interventions aimed at slowing age-related cognitive decline. The findings highlight the potential to identify specific genetic subgroups of individuals with different response profiles to treatments, to use in clinical trials.

This study leveraged the NIHR BioResource G&C cohort to comprehensively characterize cognitive trajectories and their genetic underpinnings in a large population of healthy adults. The minimal impact of known dementia risk factors on general cognitive performance, in contrast to the identified novel loci, underscores the importance of focusing on distinct molecular pathways to combat age-related cognitive decline. The resource is readily available for future studies. Future studies should examine longitudinal trajectories and include more diverse ethnic groups. Further research is needed to investigate the functional roles of the identified genes in cognition.

The cross-sectional nature of the data limits causal inference about age-related cognitive decline. The predominantly white European ancestry of the cohort limits generalizability to other populations. The device dependence of the cognitive tests, though accounted for in the analyses, could still introduce some bias. The cognitive tests used might not represent the full spectrum of cognitive domains. Lastly, while the study characterized genetic and cognitive factors, it did not include neurodegenerative disease-specific biomarkers, meaning that recalled participants for specific trials may not be representative for all measures.