Genes and age reveal new evidence for cognitive variation

A recent study published in Nature Medicine examines the effects of specific genes and age on cognitive ability. The researchers discuss the potential utility of their findings for creating cognitively and genotypically stratified cohorts for future epidemiological and intervention studies.

Current estimates suggest that up to 140 million people could develop dementia by 2050, despite the development of new treatments.

Many new drugs approved for the treatment of neurodegenerative diseases are initially tested in people with advanced and irreversible stages of the disease, which often results in limited effectiveness of these therapies. Therefore, improving the current understanding of preclinical and early stages of neurodegeneration can help assess the effectiveness of new treatments to prevent further neurodegeneration and restore patients’ quality of life.

This motivated the current study, which included people who could be followed over long periods to find out the development of dementia and possibly the effect of drugs on it.

All study participants were from the National Institute for Health and Care Research (NIHR) in England, which was originally set up as a database of volunteers for experimental medicine and clinical trials.

Both genotypes and phenotypes were available for all study participants, with the majority of them healthy at baseline. For this purpose, the Genes and Cognition (G&C) cohort, comprising over 21,000 participants within the NIHR BioResource, was identified for targeted calling.

The current study examined changes in cognitive performance (phenotype) with age, associated genotypes, and demographic and socioeconomic information. The study included eleven cognitive tests across multiple domains, as well as two new measures of cognitive ability, designated G6 and G4.

G4 is a summary measure including short-term memory, fluid and crystallized intelligence, while G6 is a measure summarizing reaction time, attention, information processing speed and executive functions. The genetic background for both measures was used to identify new genetic loci influencing cognitive status across the lifespan.

The results of the study showed that all 13 parameters were positively correlated with each other, except for vocabulary (VY), which showed both positive and negative correlations.

The study results were adjusted for the type of device used, which would otherwise have affected test scores. However, future studies should also take into account that device type varies by age, socioeconomic status, and educational status, which contributes to different phenotypes.

Cognitive performance declined with age in all tests except VY, which increased with age. This finding contradicts earlier studies reporting a decline in VY in people over 60 years of age.

Gender explained 0.1–1.33% of the variation in cognitive performance, indicating that both genders experience similar types and degrees of cognitive decline over time. G4 and G6 explained the majority of the variation in each test.

The two groups with the least education performed the worst, with education vs. cognitive ability being linear. The presence of deprivation was negatively related to cognitive performance on almost all tests.

Apolipoprotein E (APOE) genotype, data for which were available for nearly 10,000 participants, did not correlate with phenotype in any of the tests. The Alzheimer's disease polygenic risk score (AD-PRS) approach showed no significant effect on cognitive performance.

Genotype-phenotype correlations were stronger than phenotypic correlations. Moreover, the heritability of the phenotype ranged from 0.06 to 0.28, which was similar to previous studies.

Functional mapping of G4-associated genes identified genes involved in microglia-mediated immunological pathways in cognitive impairment in older adults. For G6, glycogen branching enzyme 1 (GBE1), which is involved in glycogen metabolism, was associated with cognitive performance, suggesting its role in general cognitive ability.

Genome-wide association studies (GWAS) identified several new loci, one of which explained 185 times more variation in G4 compared to APOE. A strong genetic correlation was also found between IQ and G4 and G6.

The fluid and crystallized intelligence domain may be a better marker of future educational success, as G4 had a more than two-fold genetic correlation with educational achievement compared to G6. Importantly, G4 and G6 did not show strong correlations with Alzheimer's disease (AD), indicating that normal cognition and AD have different genetic factors.

Conclusions The current study used multiple tools to distinguish genetic mechanisms of normal cognition from those of neurodegeneration. Recognizing these distinct pathways is necessary to identify molecular targets to prevent or alleviate age-related cognitive decline.

All study participants were white Europeans, limiting the generalizability of the results. Furthermore, the current study did not assess all cognitive domains.

Future studies are needed to perform functional mapping of G4-related genes. However, this is an extremely challenging task because animal cognition does not reflect changes in normal human cognition with age.

We are currently repeating cognitive profiling of all participants to determine cognitive trajectories over time, expanding it to include more diverse ethnic groups, and conducting long-read genome sequencing to enrich the potential challenge for both academic and industrial researchers.