Report

Calorie Restriction Promotes Mitochondrial Biogenesis by Inducing the Expression of eNOS

See allHide authors and affiliations

Science  14 Oct 2005:
Vol. 310, Issue 5746, pp. 314-317
DOI: 10.1126/science.1117728

Abstract

Calorie restriction extends life span in organisms ranging from yeast to mammals. Here, we report that calorie restriction for either 3 or 12 months induced endothelial nitric oxide synthase (eNOS) expression and 3′,5′-cyclic guanosine monophosphate formation in various tissues of male mice. This was accompanied by mitochondrial biogenesis, with increased oxygen consumption and adenosine triphosphate production, and an enhanced expression of sirtuin 1. These effects were strongly attenuated in eNOS null-mutant mice. Thus, nitric oxide plays a fundamental role in the processes induced by calorie restriction and may be involved in the extension of life span in mammals.

Calorie restriction (CR) extends life span in numerous organisms from yeast (1) to rodents and possibly primates (2). In mammals, CR delays the onset of age-associated diseases including cancer, atherosclerosis, and diabetes (3). A decrease in oxidative damage has been proposed as a mechanism (4); however, a lack of correlation between reactive oxygen species (ROS) production and life span was recently reported in Drosophila (5). Furthermore, increasing evidence suggests that SIRT1, the mammalian ortholog of the SIR2 gene that mediates the life-extending effect of CR in yeast (1, 6), is a key regulator of cell defenses and survival in mammals in response to stress (7).

Eight-week-old male wild-type mice were fed either ad libitum (AL) or with a CR diet (food provided on alternate days) for 3 or 12 months (8). Mice maintained on a CR feeding schedule consume 30 to 40% fewer calories over time compared with animals fed AL, have a lower body weight (fig. S1), and are known to have an extended life span (9). At 3 months of treatment, the amounts of mitochondrial DNA (mtDNA, a marker of mitochondrial content), the expression of peroxisome proliferator-activated receptor–γ coactivator 1α (PGC-1α), nuclear respiratory factor–1 (NRF-1), and mitochondrial transcription factor A (Tfam) [master regulators of mitochondrial biogenesis (10)], and expression of cytochrome c oxidase (COX-IV) and cytochrome c (Cyt c) (two mitochondrial proteins involved in cell respiration) were higher in white adipose tissue (WAT) and many other tissues of CR mice when compared with AL mice (Figs. 1A and 2A). This was consistent with increased mitochondrial biogenesis and mitochondrial gene expression (1113).

Fig. 1.

CR induces mitochondrial biogenesis in WAT of wild-type (wt) but not eNOS–/– mice through eNOS expression and cGMP formation. (A and D) PGC-1α, NRF-1, Tfam, Mfn1, and Mfn2 mRNA were analyzed by means of quantitative reverse transcription polymerase chain reaction (RT-PCR); COX IV, Cyt c, and eNOS proteins were detected by immunoblot analysis. (Insets) WAT mtDNA (gel shows a representative experiment with two mice per group). The relative values were obtained by densitometric analysis, with those measured in the AL mice taken as 1.0. (B and E) O2 consumption and (C and F) cGMP concentrations in WAT. Each experiment (n = 10) was repeated at least three times. Triple asterisks indicate P < 0.001, and single asterisk, P < 0.05, compared with AL-fed mice. Error bars indicate SEM.

Fig. 2.

CR induces mitochondrial biogenesis in different tissues of wt mice through eNOS expression and cGMP formation. (Large bar graphs) PGC-1α, NRF-1, Tfam, Mfn1, and Mfn2 mRNA were analyzed by means of quantitative RT-PCR with gene-specific oligonucleotide probes. COX IV, Cyt c, and eNOS proteins were detected by immunoblot analysis. (Top images) MtDNA (gel shows one representative experiment from one mouse per group). The relative values were obtained by densitometric analysis, with those measured in the AL mice taken as 1.0. (Top small bar graphs) O2 consumption and (bottom small bar graphs) cGMP concentrations. Each experiment (n = 10 animals) was repeated at least three times. Triple asterisks, P < 0.001, and single asterisk, P < 0.05, compared with AL-fed mice. Error bars indicate SEM.

To confirm that CR increases mitochondrial function, we measured oxygen consumption and expression of mitofusin (Mfn) 1 and 2 [the mitochondrial transmembrane guanosine triphosphatases crucial to the mitochondrial fusion process and metabolism (14, 15)]. These parameters were higher in several tissues, particularly in WAT, of CR than in AL animals (Figs. 1, A and B, and 2). This suggests that CR induces mitochondrial biogenesis with increased respiration and expression of genes crucial for the dynamic fusion processes required for oxidative function. We then investigated whether the increase in respiration was associated with an increase in adenosine triphosphate (ATP) synthesis and found that CR increased ATP concentrations in WAT (0.025 ± 0.001 nmol/mg tissue in CR mice compared with 0.018 ± 0.002 nmol/mg tissue in AL mice, P < 0.001, n = 4 animals) and in other tissues (table S1). Similar results were obtained in mice treated for 12 months. Thus, the molecular changes induced by CR occur early and are long-lasting, consistent with the early onset and persistent effect of CR on life span (9).

Nitric oxide (NO) generated by eNOS increases mitochondrial biogenesis and enhances respiration and ATP content in various mammalian cells by acting through its second messenger, 3′,5′-cyclic guanosine monophosphate (cGMP) (11, 16). We investigated whether eNOS and cGMP play a role in the mitochondrial biogenesis induced by CR. The expression of eNOS, unlike neuronal and inducible NOS, was higher in CR than in AL mice (Figs. 1A and 2) and was accompanied by higher concentrations of cGMP (Figs. 1C and 2) in WAT and in several other tissues. The increased serum concentrations of nitrite and nitrate (an index of NO production) and plasma cGMP in obese subjects exposed to CR in controlled weight loss trials (17, 18) are consistent with our findings.

To verify the role of eNOS in the mitochondrial biogenesis induced by CR, we fed 8-week-old male eNOS null-mutant (eNOS–/–) mice either an AL or a CR diet for 3 months (Fig. 1, D to F, and fig. S2, A to C). In particular, mtDNA content and PGC-1α, NRF-1, Tfam, Mfn1, and Mfn2 mRNA amounts, although different from those in wild-type animals, were not significantly greater in CR eNOS–/– mice compared to in AL eNOS–/– animals. Moreover, COX IV and Cyt c expression did not increase significantly in CR animals except in WAT and brain, where these parameters increased to a much lesser extent than those in wild-type animals (Fig. 1D and fig. S2A). Thus, CR was unable to induce significant mitochondrial biogenesis in a number of tissues of eNOS–/– mice, including WAT. To confirm this, we measured oxygen consumption (Fig. 1E and fig. S2B) and cGMP (Fig. 1F and fig. S2C) and ATP concentrations (table S1) in WAT and other tissues (table S1) of both CR and AL eNOS–/– animals. These parameters also did not increase significantly as a result of CR in knock-out compared to in wild-type mice. AL eNOS–/– mice displayed greater feed efficiency (body weight gain per food intake) than their wild-type counterparts (11), suggesting that both energy expenditure and oxidative metabolism are partly NO-dependent. The CR wild-type mice showed lower feed efficiency values than AL wild-type animals (0.295 ± 0.023 compared with 0.488 ± 0.028, respectively; P < 0.001, n = 10 animals), whereas there was no difference between CR eNOS–/– mice and AL eNOS–/– animals (0.67 ± 0.025 and 0.654 ± 0.019, respectively; n = 10 animals). Thus, the CR-induced increase in oxidative metabolism appears to be blunted in the absence of eNOS expression in mammals.

Given the role of yeast SIR2 protein in life span extension by CR (1, 6), we studied the expression of SIRT1 and found it to be higher in many tissues (fig. S3) of CR wild-type animals than of AL wild-type mice, including WAT (Fig. 3A) (19), where SIRT1 triggers lipolysis and loss of fat (20). SIRT1 mRNA and protein were ∼threefold higher in cultured white adipocytes exposed either to NO donors, such as (Z)-1-[2-(2-aminoethyl)-N-(2-ammonioethyl)amino] diazen-1-ium-1,2 diolate (DETA-NO) and S-nitrosoacetyl penicillamine (SNAP), or to a cGMP analog (8 Br-cGMP) than in untreated cells (Fig. 3B) and ∼80% lower in WAT of eNOS–/– mice when compared with wild-type animals (Fig. 3, C and D). Thus, the expression of SIRT1 in WAT during CR might be partly mediated by NO acting via cGMP.

Fig. 3.

SIRT1 expression is regulated by NO in WAT and white adipocytes. (A) SIRT1 protein levels in WAT of either AL or CR wt and eNOS–/– mice. A representative experiment is shown, with means ± SEM of densitometric measurements performed in 10 animals per group. The histogram values were obtained by densitometric analysis, with values measured in the AL mice taken as 1.0. (B) SIRT1 protein concentrations in white adipocytes cultured for 3 days with or without SNAP (100 μM), DETA-NO (50 μM), and 8 Br-cGMP (3 mM). A representative experiment is shown, with means ± SEM of densitometric measurements performed in five separate experiments. The protein concentrations were obtained by densitometric analysis, with values measured in the untreated cells (c) taken as 1.0. (C and D) SIRT1 protein and mRNA expression, respectively, in WAT of male eNOS–/– mice compared with wt mice. Each experiment (n = 10 animals) was repeated at least three times. The relative values of mRNA were obtained by densitometric analysis, with values measured in wt mice taken as 1.0. Triple asterisks, P < 0.001 and single asterisk, P < 0.05 compared with AL-fed wild-type mice or untreated cells.

To investigate whether the CR-induced SIRT1 expression was dependent on eNOS-derived NO, we performed immunoblot analysis in WAT of eNOS–/– mice fed either an AL or a CR diet. In eNOS–/– mice fed a CR diet, SIRT1 expression was also increased (∼30%) in WAT compared with that of eNOS–/– mice fed an AL diet (Fig. 3A), although the change was much smaller than that in wild-type animals (∼120%, P < 0.001). Similar results were obtained in the other tissues tested (fig. S3).

Thus, CR induces an increase in eNOS expression, which in turn is involved in both mitochondrial biogenesis and SIRT1 expression in a variety of tissues. The enhanced expression of SIRT1 by CR is consistent with a potential increase in life span. This transcription factor may be an evolutionarily ancient biological stress response that slows aging, promoting the mobilization of fat into the blood from WAT stores (20), the down-regulation of adipogenesis (20), and the long-term survival of irreplaceable cells (7, 19). The increase in mitochondrial activity, i.e., in oxidative metabolism, that we see in CR animals is intriguing in view of the widely accepted hypothesis that CR increases longevity by slowing metabolism and reducing mitochondrial ROS and accompanying cellular damage (4). In fact, metabolic rate normalized to body weight does not decline in CR mice, and the lifetime metabolic output of these animals is therefore larger than that of their AL cohorts (21). Respiration actually increases during CR in yeast (22) and the nematode worm Caenorhabditis elegans (23). The effects of CR on life span may be independent of excessive ROS production.

The effects of CR in mammals are complex, affecting many organs and physiological pathways. Nevertheless, the significantly reduced effects observed in eNOS–/– animals point to a role for NO in the response to CR. eNOS–/– mice are characterized by a reduced life span (24) due to age-related diseases (25). One possibility is that in wild-type CR animals NO, acting via mitochondrial biogenesis and expression of SIRT1, increases β-oxidation and lipolysis. This would result in a reduction in the accumulation of fat, which is known to have an impact on life span (26, 27).

Supporting Online Material

www.sciencemag.org/cgi/content/full/310/5746/314/DC1

Materials and Methods

Figs. S1 to S3

Table S1

References

References and Notes

View Abstract

Navigate This Article