In line with the changes in acetylation levels of PGC-1α, a key transcriptional regulator of mitochondrial biogenesis, we could observe either an elevated expression or a strong tendency towards an increase (P < 0.1) of nuclear genes encoding transcriptional regulators of oxidative metabolism (Sirt1, PGC-1α, mitochondrial transcription factor A (Tfam)) and mitochondrial proteins (Mitofusin 2 (Mfn2), Cytochrome C (Cyt C), Medium Chain Acyl-coA Dehydrogenase (MCAD), Carnitine palmitoyltransferase-1b (CPT-1b), Citrate Synthase (CS) or ATP synthase lipid binding protein (ATP5g1)) in quadriceps muscles from NR-fed mice (Fig.6B). Conversely, in brain, where NAD+ and sirtuin activity levels were not affected by NR feeding, the expression of these genes was not altered (Fig.6B). Consistently also with enhanced mitochondrial biogenesis, mitochondrial DNA content was increased in muscle, but not in brain from NR-fed mice (Fig.6C). Finally, mitochondrial protein content also confirmed that mitochondrial function was only enhanced in tissues in which NAD+ content was increased (Fig.6D and Fig.S5D). This way, while muscle, liver and BAT showed a prominent increase in mitochondrial proteins (Complex V - ATP synthase subunit α and porin), such change was not observed in brain or WAT. Altogether, these results suggest that NR feeding increases mitochondrial biogenesis in a tissue-specific manner, consistent with the tissue-specific nature of the increases in NAD+ and sirtuin activity observed in NR-fed mice. The higher number of mitochondria, together with the different morphological mitochondrial profiles found in NR-fed mice (Fig.4C) would contribute to explain the higher oxidative profile, energy expenditure and protection against metabolic damage observed upon NR feeding.
Discussion
While increased NAD+ levels in response to NR supplementation were already reported in yeast (Belenky et al., 2007; Bieganowski and Brenner, 2004) and cultured human cells (Yang et al., 2007b), we extend these observations to a wide range of mammalian cell lines and demonstrate that NR supplementation also enhances NAD+ bioavailability in mammalian tissues. Also, our work provides evidence that the increase in NAD+ after NR administration stimulates the activity of mammalian sirtuins. This further supports the role of sirtuins as a family of proteins whose basal activity can be largely modulated by NAD+ availability (Houtkooper et al., 2010).
The fact that the activity of both SIRT1 and SIRT3 are positively regulated by NR both in vitro and in vivo favours the hypothesis that the increase in NAD+ promoted by NR affects, at least, the mitochondrial and the nuclear compartments. This is a key difference compared to other strategies that boost intracellular NAD+ levels. For example, Parp-1 knock-out mice show a two-fold increase in intracellular NAD+, yet SIRT3 activity is not affected (Bai et al., 2011b). Given that PARP-1 is a nuclear enzyme, it is tempting to hypothesize that, in the Parp-1 knockout mice, the increase in NAD+ is mainly confined to the nucleus, while the NR-induced increase in NAD+ levels, despite being more discrete in amplitude, reaches different subcellular compartments.
The inability of NR to activate GPR109A illustrates how NR provides an alternative way to increase NAD+ and ameliorate metabolic homeostasis in the absence of the undesirable effects seen with NA. Since NR does not activate GPR109A, we speculate that NR reduces cholesterol levels through the activation of SIRT1 in liver, which influences the activity of a number of transcription factors and cofactors linked to cholesterol homeostasis (Kemper et al., 2009; Li et al., 2007; Shin et al., 2003; Walker et al., 2010). Interestingly, the mechanism of action of NA on VLDL/LDL and hepatic metabolism is not clear, as GPR109A is not expressed in the liver (Tunaru et al., 2003). Our data raise the possibility that NA, like NR, achieves such effects by elevating NAD+ levels, activating SIRT1 and its downstream targets. However, future experiments will be required to verify this hypothesis. The activation of SIRT1 and SIRT3, impacting on targets such as PGC-1α, FOXOs and SOD2, also provides a likely explanation for the mitochondrial fitness and metabolic flexibility of NR-supplemented mice. In this sense, it is remarkable that while SIRT1 transgenic mice have some shared phenotypes with NR-treatment mice, such as protection against metabolic damage, there are also a number of discrepancies. For example, SIRT1 transgenics show similar body weight gain upon HFD to wild-type littermates (Pfluger et al., 2008). The differences between NR supplemented mice and SIRT1 trangenics can be explained by numerous reasons. First, NR affects different sirtuin activities; second, sirtuin expression does not necessarily correspond with sirtuin activity; third, NR actions are tissue-specific. Importantly, we should also consider that NR might trigger other actions unrelated to sirtuin activity that might also contribute to the beneficial effects of NR (Fig.7).