Analysis of lifespan across diversity outbred mouse studies identifies multiple longevity-associated loci

Abstract

Lifespan is an integrative phenotype whose genetic architecture is likely to highlight multiple processes with high impact on health and aging. Here, we conducted a genetic mega-analysis of longevity in Diversity Outbred (DO) mice that included 2,444 animals from 3 independently conducted lifespan studies. We identified 8 loci that contributed significantly to lifespan independently of diet and drug treatment in at least one study. One of these loci also influenced lifespan in a sex-dependent manner, and we detected an additional locus with a diet-specific effect on lifespan. Collectively, these loci explained over half of the estimated heritable variation in lifespan across these studies and provided insight into the genetic architecture of lifespan in DO mice.

Keywords: QTL mapping; aging; diversity outbred mice; genetics; lifespan; quantitative trait analysis.

Fig. 1.

Survival in DO mice across 4 independent studies. a) Kaplan–Meier survival curves for ad libitum females in each of the 4 studies. Dashed vertical lines indicate median survival in days and bands around each curve indicate 95% confidence intervals. Vertical bars on each curve indicate censorship events. b) Kaplan–Meier survival curves by intervention in the Harrison study, which involved female mice on ad libitum (AL), 30% calorie restricted (30% CR), or rapamycin-treated (Rapamycyin) diets. c) Kaplan–Meier survival curves by intervention in the Svenson study, which involved female mice on AL, 30% calorie restricted (30% CR), or rapamycin-treated (Rapamycyin) diets. d) Kaplan–Meier survival curves by dietary intervention in the DRiDO study Di Francesco et al. (2024), which involved female mice on AL, 1 day intermittent fasting (1D IF), 2 day intermittent fasting (2D IF), 20% calorie restricted (20% CR), or 40% calorie restricted (40% CR) diets. e) Kaplan–Meier survival curves by sex in the Shock study, which included male and female mice on an AL diet. f) Baseline hazard and mortality doubling times and 95% confidence intervals estimated by a Gompertz log-linear model.

Fig. 2.

Genetic analysis of lifespan in DO mice. a) QTL profile of the additive genome-wide scan on lifespan in the Harrison study. The solid horizontal line indicates genome-wide significance, while the dashed horizontal line indicates a significance threshold of LOD 6. The narrow-sense heritability (h²) of lifespan is reported, with standard error about the h² estimate reported in parentheses (top right). b) QTL profile of lifespan in the Shock study. c) QTL profile of lifespan in the DRiDO study. d) Allelic effects (BLUPs) and corresponding standard errors of the chromosome 18 locus detected in the Harrison (left) and DRiDO (middle) studies. The allelic effects at this locus are also shown in the Shock study (right), although it did not reach statistical significance in that dataset. e) Comparisons of allelic effects (BLUPs) and standard errors of the chromosome 18 locus in each pair of studies. f) Allelic effects (BLUPs) and corresponding standard errors of the chromosome 12 locus detected in the DRiDO study. g) Allelic effects (BLUPs) and corresponding standard errors of the chromosome 16 locus detected in the DRiDO study.

Fig. 3.

Mega-analysis of lifespan in DO mice. a) QTL profile of the additive genome-wide scan on lifespan using combined data from the Harrison, Shock, and DRiDO studies. The solid horizontal line indicates genome-wide significance, while the dashed horizontal line indicates a significance threshold of LOD ≥ 6. The narrow-sense heritability (h²) of lifespan is reported, with standard error about the h² estimate reported in parentheses (top right). b) QTL profile of the forward regression scan on the combined data, in which statistically significant loci were included as additive covariates in the model. c) Manhattan plot of the additive genome-wide scan on combined lifespan data using GxEMM. Markers surpassing the horizontal line are considered statistically significant. d–i) Allelic effects (BLUPs) and corresponding standard errors of d) the chromosome 18 locus, e) the chromosome 12 locus, f) the chromosome 16 locus previously detected in the Shock study, g) the second chromosome 16 locus, h) the chromosome 2 locus i) the chromosome 4 locus detected in the forward regression. j) Mean residual lifespan and standard error of DO mice as a function of the number of CAST alleles at the chromosome 18 locus. The dotted black line denotes the expected residual lifespan of individuals carrying a single CAST allele at this locus. k) Mean residual lifespan and standard error of DO mice as a function of the number of CAST alleles at the chromosome 12 locus. The dotted black line denotes the expected residual lifespan of individuals carrying a single CAST allele at this locus.

Fig. 4.

Variant association mapping of loci detected at statistical significance in one or more studies. a) Variant association mapping of the QTL on chromosome 18 depicting the LOD scores for each variant within the 3 LOD support interval around the locus in the mega-analysis (top). The area encompassed by the plot corresponds to the 3 LOD support interval around the locus in the mega-analysis. The most likely candidate SNPs, defined as those with an LOD score within 1 LOD of the maximum variant association, are highlighted in pink. Genes within the genomic interval are depicted (bottom). b) Variant association mapping of the chromosome 12 locus in the mega-analysis. Here, the bottom panel is zoomed on the coordinates of the most likely causal variants for clarity (dashed lines). c) Variant association mapping of the chromosome 16.1 locus in the mega-analysis.

Fig. 5.

Genome-wide scans for loci with diet- or sex-specific effects on lifespan. a) Sex-specific effects of the chromosome 16 locus within the Shock study. Mean lifespan after correcting for sex and generation wave is shown by the black line. Error bars denote standard error. b) Whole genome scan for loci influencing lifespan via interaction with diet using GxEMM. Red points denote loci with significant effects at P < 1 × 10⁻⁴. c) Whole genome scan for loci influencing lifespan via interaction with sex using GxEMM. d) Allelic effects of the chromosome 5 locus on lifespan as a function of dietary intervention or drug treatment, with standard errors. e) Map of the 2LOD support interval about the peak marker at chromosome 5 depicting all protein coding genes within the interval.