To investigate the effects of calorie restriction on the already extended lifespan of Ames dwarf mice, we divided 45 2-month-old Ames dwarf mice and 53 of their normal siblings into two groups, which were subjected either to calorie restriction (CR) or to continued feeding ad libitum (AL). We fed CR mice daily, reducing their food intake in successive weeks to 90%, 80% and finally 70% of that consumed daily by genotype- and sex-matched AL animals6. Because the food consumption of AL mice declines naturally with age, the amount of food given to CR animals was kept constant after the age of 2 years.

The survival curves shown in Fig. 1 indicate that calorie restriction causes a further significant increase in the longevity of Ames dwarf mice. When males and females are considered together, the difference between the CR and AL groups of Ames dwarf mice is significant (P < 0.004, log rank test). The effect of calorie restriction on lifespan in Ames dwarf mice is also significant (P < 0.05) when genders are considered separately. As expected, calorie restriction also extends the lifespan of normal mice (P < 0.002), although AL Ames dwarf mice outlive AL normal mice (P < 0.00001). Moreover, CR Ames dwarf mice outlive CR normal mice (P < 0.0001).

Figure 1
figure 1

Survival plots of Ames dwarf (DF) and normal (wild-type, WT) mice fed ad libitum (AL) or restricted to 70% of normal calorie intake (calorie restriction, CR).

The survival plots (Fig. 1) reveal a further disparity: although both dwarfism and calorie restriction extend longevity, the effect of reduced food intake is associated primarily with a change in the slope of the survival curve (that is, it reduces the rate of age-related mortality), whereas the effect of dwarfism mainly reflects a shift in the age at which the age-dependent increase in mortality risk first becomes appreciable. Calorie restriction therefore seems to decelerate ageing, whereas the Prop1df allele seems to delay it.

Our results indicate that long-lived Ames dwarf mice are not merely mimics of CR mice, and show that the pathways responsible for extending lifespan in the dwarfs and in CR animals are not identical. However, features that are shared by CR normal mice and Ames dwarf mice, and by long-lived knockout mice that lack the growth-hormone receptor7, include reduced body size and lower plasma levels of insulin, the insulin-like growth factor IGF-1, glucose and thyroid hormone. These factors may contribute to delayed ageing and increased longevity in each of these animal models.

For example, the IGF/insulin or a similar signalling pathway is involved in lifespan determination in the fruitfly Drosophila melanogaster8,9, the roundworm Caenorhabditis elegans10, and yeast11. This supports the idea that hormonal regulation of metabolic pathways in response to altered food availability may be a way of regulating lifespan that is deeply rooted in evolutionary history.