Genetic effects on lifespan were not constant throughout life

In a study analyzing survival data and genotypes from 6,438 mice, researchers identified 29 loci associated with lifespan and 30 loci involved in the correlation between body weight and lifespan. For some genotypes, a lifespan benefit reversed into an adverse effect with age, and the relationship between body weight and subsequent lifespan also varied by age and sex. Published in Nature, the study shows that understanding genetic effects related to longevity requires examining not only which variants are involved, but also when and in which sex they act. A locus is a genomic region associated with a trait; it does not mean that a single causal gene has been established.

Effects were distinguished by progressively raising the survival-age threshold

The researchers used data on UM-HET3 mice from the US National Institute on Aging’s Interventions Testing Program (ITP). The analysis included animals that survived to approximately 42 days of age; 559 mice lived for at least 1,100 days. The team compared lifespan and mortality risk by genotype across 72 overlapping survival groups, created by increasing the minimum survival age in 15-day increments. Here, lifespan means age at death conditional on having survived to the relevant threshold age. The analysis did not track DNA variants as they newly arose over a lifetime; it examined how the effects of inherited genotypes varied across life stages.

Of the 29 lifespan-related “Vita” loci, 12 showed a reversal in the direction of genotype effects depending on the threshold age for inclusion in a survival group. Some loci had effects only in groups including earlier or later life, while others showed relatively persistent effects. Significant interactions between locus and sex were observed at 14 loci. The researchers interpreted some of the reversals as consistent with the theory of “antagonistic pleiotropy,” in which a genetic trait that is beneficial early in life may become disadvantageous later. Not all loci can be explained by the same theory of aging.

The researchers distinguished body weight itself from its link to lifespan

Thirty “Soma” loci were involved in genotype-dependent variation in the correlation between body weight and subsequent lifespan. Among these, 19 affected the strength of the negative correlation in which greater body weight at younger ages was associated with shorter lifespan, while 11 affected the strength of the positive correlation in which greater body weight later in life was associated with longer lifespan. These effects were generally stronger in males.

The rank correlation coefficient between body weight at approximately six months of age and lifespan was about −0.28 in males and −0.11 in females. By contrast, body weight measured at 730 days of age was positively correlated with subsequent lifespan in both sexes. The overlap between loci associated with body weight itself and Soma loci was at the level expected by chance. Therefore, Soma cannot be interpreted as a confirmed list of genes that directly regulate both body weight and lifespan. The study also did not test the lifespan effects of interventions that reduce or increase body weight.

Clues for follow-up functional validation of candidate genes

In a follow-up experiment in nematodes, suppressing expression of acds-10, the counterpart of mouse Acad11, improved locomotor ability in old age. The outcome measured in this experiment was locomotion; it did not demonstrate lifespan extension in mice or humans. In a Mendelian randomization analysis using human blood gene-expression data, increased APEH expression was positively associated with male longevity. The researchers used this to help prioritize candidate genes and stated that some key assumptions required for causal inference were only partially met.

The analysis included untreated controls as well as groups given dietary additives that had shown no significant lifespan effect in individual trials. However, when the latter groups were pooled for analysis, a lifespan effect emerged. The researchers accounted for treatment status, housing facility, cohort year, and other factors in the statistical model. Unexplained factors, including effects of co-housed animals, and broad loci encompassing multiple candidate genes remain; further validation is needed to establish causal genes and mechanisms.

The source is the Nature research paper; the content above is based on the provided Europe PMC abstract and public full-text excerpt.