The Annals of the New York Academy of Science has a paper in preprint (abstract and full PDF) on the mechanisms by which resveratrol might induce health and longevity benefits similar to those of calorie restriction in mammals. In particular, this research suggests that resveratrol doesn't act directly through sirtuins, such as SIRT1:
The natural polyphenol resveratrol stimulates sirtuins and extends lifespan. Here resveratrol inhibited expression of replicative senescence marker INK4a in human dermal fibroblasts and 47 out of 19,000 genes from microarray experiments were differentially expressed. These included genes for growth, cell division, cell signaling, apoptosis and transcription. Genes involved in Ras and ubiquitin pathways, Ras-GRF1, RAC3 and UBE2D3, were downregulated. The changes suggest resveratrol might alter sirtuin-regulated downstream pathways, rather than sirtuin activity. Serum deprivation and high confluency caused nuclear translocation of the SIRT1-regulated transcription factor FOXO3a. Our data indicate resveratrol's actions might cause FOXO recruitment to the nucleus.
we propose that resveratrol does not allosterically activate SIRT1, but rather stimulates interaction of SIRT1 with FOXO3a by inducing FOXO3a nuclear localization through the downregulation of Ras activity. Therefore, an increase in FOXO3a transcriptional activity by resveratrol might quickly decrease cellular damage, so leading to a decrease in INK4a expression, and consequently attenuate replicative senescence. Given that INK4a is induced by Ras overexpression as a tumor-suppressor mechanism the decrease in INK4a expression would be due directly to a reduction in Ras pathway activity.
Chronic resveratrol treatment did not increase the time required to reach replicative senescence, despite likely beneficial effects of resveratrol in reducing cellular damage, as judged by the reduction in INK4a it induced. On the other hand, reduced cellular p16ink4a should have a favorable effect on physiological function at the organismal level, unrelated to its effect on lifespan of fibroblasts. Effects on other cell types could, moreover, be quite different.
It's a good indicator of the complexity of metabolism: this tiny part of the whole is a dynamic dance of many different mechanisms, all feeding back into one another, their interactions poorly understood at this time. I give it another five to ten years before the scientific community has enough of calorie restriction biology nailed down to the point of being able to safely, knowingly manipulate it by means other than food intake.