Array-based expression analysis of mouse liver genes: Effect of age and ofthe longevity mutant Prop1(df)

Ames dwarf mice, homozygous fur the df allele at the Prop1 locus, live 40% to 70% longer than nonmutant siblings and represent the first single-gene m utant that extends life span in a mammal. To gain insight into the basis fo r the longevity of the Ames dwarf mouse, we measured liver mRNA levels for 265 genes in a group of 11 df/df mice, (three to four mice per age group), at ages 5, 13, and 22 months, and in 13 age- and sex-matched control mice. The analysis showed seven genes where the effects of age reach p < .01 in n ormal mice and six others with possible age effects in dwarf mice, but none of these met Bonferroni-adjusted significance thresholds. Thirteen genes s howed possible effects of the df/df genotype at p <.01. One of these, insul in-like growth factor I (IGF-I), was statistically significant even after a djustment fur multiple comparisons; and genes for two IGF-binding proteins, a cyclin, a heat shock protein, p38 mitogen-activated protein kinase, and an inducible cytochrome P450 were among those implicated by the survey. In young control mice, half of the expressed genes showed SDs that were more t han 58% of the mean, and a simulation study showed that genes with this deg ree of interanimal variation would often produce false-positive findings wh en conclusions were based on ratio calculations alone (i.e., without formal significance testing). Many genes in our data set showed apparent young-to -old or normal-to-dwarf ratios above 2, but the large majority of these pro ved to be genes where high interanimal variation could create high ratios b y chance alone, and only a few of the genes with large ratios achieved p < .05. The proportion of genes showing relatively large changes between 5 and 13 months, or from 13 to 22 months of age, was not diminished by the df/df genotype, providing no support for the idea that the dwarf mutation leads to global delay or deceleration of the pare of age-dependent changes in gen e expression. These survey data provide the foundation for replication stud ies that should provide convincing proof for age- and genotype-specific eff ects on gene expression and thus reveal key similarities among the growing number of mouse models of decelerated aging.