Heritability of life span in mice and its implication for direct and indirect selection for longevity

We found high narrow-sense heritability of life span based on the regression of offspring on average parental (midparent) life spans. In two mouse populations prepared using the 4-way-cross design, mean ± SE heritabilities were 62 ± 11% (P ≤ 0.001) and 44 ± 15% (P ≤ 0.01). To reflect inherited rates of aging, rather than resistance to early disease, data from the first 25% to die were deleted, so that only about 40% of families were used for offspring-midparent regressions. Heritabilities still remained high, 38% and 55%, for the same two populations, respectively. Populations studied in two other experiments did not show nearly as high heritabilities; in one case probably due to environmental stress, and in the other probably because the strains used did not have sufficient additive variance in genes regulating longevity. Significant heritabilities occurred only when a wild derived inbred strain was included in the 4-way cross. The age when a female ceased to reproduce appeared to be related to the life spans of her offspring, but only weakly, not approaching significance for any individual experiment. The age when a female became infertile was related to her life span, but the relationship disappeared when short-lived mice were excluded from the analysis. Our findings indicate that, in sufficiently diverse mouse populations, selection for increased longevity should be possible and that the direct selection for parental life span will be a more efficient strategy than selection for female reproductive life span. This revised version was published online in July 2006 with corrections to the Cover Date.