Gene expression profiling of 20,687 unique transcripts using Affymetrix Mouse Genome 430 2.0 arrays was performed in young (5-months old) control and old control, CR and resveratrol fed animals (30-months old, all groups). (A). A panel of genes corresponding to significant (P≤0.01) changes in gene expression in the comparison between young and old control groups was examined in both CR (left) and resveratrol fed mice (right). Numbers on X and Y axes represent fold changes in the young vs. old comparison (X axis), and the old CR vs old control (Y axis, left graphs) or old resveratrol vs. old control group (Y axis, right graphs). Each dot corresponds to an specific gene. Red dots correspond to genes that are significantly altered in expression at P≤0.01 in both aging and CR, or aging and resveratrol comparisons. Opposite fold changes in this analysis represent prevention of aging changes. (B). A panel of genes significantly changed by CR and resveratrol (but not by age) as compared to old control mice (P≤0.01) is plotted. Similar fold changes in this analysis represent resveratrol mimicry of CR. Pie charts represent the proportion of genes changed by CR only, resveratrol only or both treatments. Gene expression changes plotted in graphs correspond to the genes changed significantly expression by both treatments (red).

(A) Cardiac function as determined by transthoracic echocardiography (n = 7, YC, n = 10 OC, n = 8 CR, n = 9 Resv) (B) Serum glucose and plasma IGF-1 in control, CR and resveratrol treated mice (C) Glucose uptake and insulin signaling in control, CR and resveratrol fed animals (n = 16). Rate of 2-deoxyglucose uptake and Akt threonine³⁰⁸ (T³⁰⁸) phosphorylation were determined in isolated paired soleus and extensor digitorum longus (EDL) muscles with or without 0.36 nM insulin. Data were analyzed by two-way ANOVA, and the source of significant variance was tested using Student-Newman-Keuls post-hoc test (n = 15–16 muscles per group). There were no differences between groups at baseline so values shown reflect insulin-stimulated values only. GLUT4 protein abundance was determined in soleus and EDL muscles by western blotting (n = 16 muscles per group). Data were analyzed by one-way ANOVA on ranks, and the source of significant variance was tested using Dunn's post-hoc test.

(A) SIRT1 levels in liver, skeletal muscle and brain (n = at least four animals per group) were determined by Western blot analysis. A loading correction factor based on HSP70 band intensity data was used to normalize the SIRT1 band intensity data. (B) mRNA abundance for known Pgc-1α transcriptional targets is shown for heart, muscle and brain. Data on Y axis represent percentage changes relative to young controls. Results represent n = 5, values in bar graphs are means and SE. * p<0.01.

A class matrix, which describes the functional categories and specifies what genes are members of what classes, was based on Gene Ontology (GO), and included classes with at least 10 genes, for a total of 571 classes. The SAFE analysis was then run using the default settings for the local statistic (Student's t) and GO terms that differed at P≤0.05 were considered significantly different. Only classes that show a significant effect for at least one treatment in one tissue are shown. Significance values for all functional classes are shown in Supplemental Table S1. Classes highlighted in blue were changed by both CR and resveratrol in all tissues examined.

Means identified with † were significantly different from age-matched controls at p<0.05; means identified with * were significantly different from age-matched controls at p<0.01.