Skip to content
The Longevity Archive
browse 192

The biology of fasting

Sirtuins and NAD+

Also called: SIRT1, SIRT3, NAD+

Written by Reviewed Sep 2026

A family of enzymes that can only work when the cell has NAD+ to spend, which ties them to nutrient state, and a field whose two most cited supporting results did not replicate.

What it is

Sirtuins are seven mammalian enzymes, SIRT1 through SIRT7, descended from the yeast protein Sir2. Most of them remove acetyl groups from other proteins. What makes them metabolically interesting is the cofactor. Unlike the classical histone deacetylases, sirtuins are NAD+ dependent, and each reaction consumes a molecule of NAD+ outright rather than using it catalytically. Because the reaction spends NAD+ stoichiometrically, sirtuin activity is coupled to how much NAD+ the cell has, and that is a function of energy state.

Not all of them are deacetylases. SIRT4 and SIRT6 have ADP-ribosyltransferase activity, SIRT5 removes succinyl, malonyl and glutaryl groups, and SIRT6 also removes long chain fatty acyl groups.

SIRT1 is nuclear and cytoplasmic and is the closest mammalian relative of Sir2; its substrates include PGC-1 alpha, the FOXO transcription factors, p53, NF-kB p65 and histones. SIRT3 is the principal deacetylase of the mitochondrial matrix, acting on enzymes of fatty acid oxidation, the citric acid cycle, the urea cycle, ketogenesis and the antioxidant enzyme SOD2.

This page is written to be honest about a field with two serious unresolved controversies, rather than to present it as settled.

The machinery

NAD+ and NADH are the same molecule in two redox states, and it is the ratio that matters, not the total. The complication, which popular accounts flatten, is that a fast pushes that ratio in opposite directions in different compartments of the same cell.

In the mitochondrial matrix, heavy beta-oxidation generates large amounts of NADH and the matrix becomes more reduced. That is exactly why acetoacetate is reduced to beta-hydroxybutyrate during a fast: the enzyme runs in the reducing direction because NADH is abundant. In the cytosol, falling glycolytic flux means less NAD+ is being consumed and less NADH regenerated, and the balance generally moves toward a higher NAD+ to NADH ratio.

So the sentence fasting raises NAD+ and therefore activates sirtuins is directionally defensible for nuclear and cytosolic SIRT1 and considerably less clean for mitochondrial SIRT3, whose compartment is becoming more reduced at the same time.

The two best characterised outputs are PGC-1 alpha and FOXO. SIRT1 deacetylates PGC-1 alpha, and the deacetylated form is the transcriptionally active one, driving the gluconeogenic gene programme in fasted liver and mitochondrial and fatty acid oxidation genes in muscle. Note the convergence with AMPK, which phosphorylates the same coactivator; the two modifications are not redundant. For FOXO, AKT excludes it from the nucleus when insulin is high, falling insulin permits it in, and SIRT1 deacetylation shifts its output toward stress resistance and away from apoptosis in the systems examined. FOXO also drives autophagy genes and the atrophy-related ubiquitin ligases in muscle, which places it directly upstream of the recycling machinery. Beta-hydroxybutyrate independently raises Foxo3a transcription through histone deacetylase inhibition, so a fast reaches FOXO by at least two chemically unrelated routes.

What switches it on

Rising cytosolic NAD+ availability as glycolytic flux falls, plus the transcriptional and post-translational changes that accompany low insulin. AMPK also raises cellular NAD+ availability, which is one route by which the energy sensor feeds the redox sensor.

On the timing
There is no human study measuring the tissue NAD+ to NADH ratio during a fast. That is the honest statement of the timing question, and it is a genuine gap rather than a caveat. The measurement is hard for two reasons: NAD+ and NADH interconvert rapidly during sampling, and the compartment-specific ratios that matter biologically, matrix versus cytosol versus nucleus, cannot be separated in a whole tissue extract. Whole blood or plasma NAD+ does not report on liver or muscle compartmental redox state.

What has been measured

In people

What has been measured in living people is narrow: SIRT1 gene expression in blood cells on a time-restricted eating protocol, SIRT1 phosphorylation and a sirtuin-relevant substrate in muscle biopsies during the first hours of a fast, and two randomised trials of an NAD+ precursor with muscle biopsies. The NAD+ to NADH ratio itself, in human tissue, during a fast, has never been measured.

  • SIRT1 expression in blood cells. In 11 overweight adults doing a four-day randomised crossover of an 8am to 2pm eating window versus 8am to 8pm, the early window raised whole blood cell expression of SIRT1, and of the autophagy gene LC3A, in the morning before breakfast (p < 0.04), and raised MTOR expression in the evening (p = 0.007). That is gene expression in blood cells over four days, not enzyme activity in tissue (Nutrients, 2019).
  • SIRT1 in muscle during the first 8 hours of a fast, 10 healthy men, biopsies at 4 and 8 hours. SIRT1 phosphorylation was non-significantly decreased (P < 0.06), alongside decreased ACC phosphorylation. p53 acetylation, a sirtuin-relevant readout, increased over the fast in both conditions (P < 0.01). Rising p53 acetylation is the opposite direction from what a straightforward fasting-activates-sirtuins model predicts for that substrate (Pflugers Archiv, 2021).
  • NAD+ precursor trial one. Nicotinamide riboside, 6 weeks at 1000 mg daily, crossover, 13 overweight or obese adults, with muscle biopsies. Markers of increased NAD+ synthesis, nicotinic acid adenine dinucleotide and methylnicotinamide, were elevated in skeletal muscle, so the compound reached the tissue and did something there. It increased fat free mass (62.65 versus 61.32%, change 1.34%, P = 0.02), increased sleeping metabolic rate, and increased skeletal muscle acetylcarnitine (4558 versus 3025 pmol per mg dry weight, P = 0.04). Insulin sensitivity by hyperinsulinaemic euglycaemic clamp was among the outcomes and is not reported as improved in the abstract (American Journal of Clinical Nutrition, 2020).
  • NAD+ precursor trial two, and it disagrees with the first on the key point. Nicotinamide riboside, 12 weeks at 1000 mg twice daily, 40 obese insulin-resistant men, muscle biopsies before and after. Nicotinamide riboside did not affect NAD metabolite concentrations in skeletal muscle. Mitochondrial respiration by high resolution respirometry on single fibres, mitochondrial content and mitochondrial morphology were all unaffected. It decreased NAMPT protein abundance in muscle. This is the first long-term human clinical trial to report on these outcomes (Journal of Physiology, 2020).
  • Two well conducted human trials of the same compound, one finding muscle NAD-related metabolites elevated and one finding muscle NAD metabolites unchanged, and neither finding the mitochondrial functional improvements the preclinical literature predicted. That is the current state of the human evidence at this node.

In other species and in cell culture

The biochemistry is not in dispute and was established in yeast and purified enzyme systems. The fasting role of SIRT1 in liver is supported by mouse loss-of-function work. What is in dispute is the leap from sirtuins responding to nutrient state to raising sirtuin activity extending lifespan, and the invertebrate evidence originally used to make that leap did not replicate.

  • Yeast and purified enzymes: Sir2 established as an NAD+ dependent histone deacetylase, with the unique reaction product 2'-O-acetyl-ADP-ribose (Nature, 2000). This is the finding that makes the whole family metabolically interesting and it is not contested.
  • Mouse liver, hepatocytes and cell culture: SIRT1 in complex with PGC-1 alpha controlling hepatic glucose and lipid metabolism (Nature, 2005), and hepatic SIRT1 required for the fasting-dependent glucose and lipid response in mice (Proceedings of the National Academy of Sciences, 2007).
  • Replication dispute one, invertebrate lifespan. The original claims were that Sir2 overexpression extends lifespan in yeast, in C. elegans and in Drosophila, the fly claim framed as acting through a pathway related to calorie restriction (Proceedings of the National Academy of Sciences, 2004). In 2011 these were re-examined with attention to genetic background. Standardising genetic background and using appropriate transgenic controls abolished the apparent effects in both C. elegans and Drosophila. In the worm, outcrossing a high-overexpression line removed the longevity increase while leaving the overexpression intact, and the longevity co-segregated with a second-site mutation affecting sensory neurons. In the fly, a dSir2 overexpressing strain was long-lived relative to wild type as originally reported, but not relative to the appropriate transgenic controls, and a new line with stronger overexpression was not long-lived either (Nature, 2011).
  • Replication dispute two, resveratrol. The claim that resveratrol directly activates SIRT1 came from in vitro assays using a fluorophore-tagged peptide substrate. Sirtuin activation by resveratrol was shown to be substrate specific, and specifically dependent on the covalently attached fluorophore in the assay peptide. With a native substrate lacking the fluorophore, the activation was not observed (Journal of Biological Chemistry, 2005). That is an assay artefact finding, and it applies to a whole generation of reported sirtuin activating compounds screened on the same platform.
  • What survives on lifespan: SIRT6 overexpression does extend lifespan in mice, and in the original report the effect was male specific (Nature, 2012).

Why it matters

Sirtuins are the mechanism most often invoked to connect fasting to ageing, and the connection is usually presented as settled. It is not, and the reasons are specific and checkable rather than vague.

What is solid: sirtuins are NAD+ dependent, they respond to nutrient state, and SIRT1 has a demonstrated role in the hepatic fasting response in mice through PGC-1 alpha.

What is not: that raising sirtuin activity extends lifespan. The invertebrate lifespan results that were the most cited support for that claim were re-run with proper genetic controls and did not hold. The most famous pharmacological sirtuin activator turned out to depend on a fluorophore attached to the assay substrate. Neither of those retractions of confidence eliminated the field, and neither should be read as showing sirtuins do nothing. They removed the specific evidence that was doing the persuading.

And for a fasting page, the load bearing gap is the simplest one. Nobody has measured the NAD+ to NADH ratio in human tissue during a fast. The central quantity in the whole story is unmeasured in people.

A claim you will see repeated

Fasting raises NAD+, which activates sirtuins, which is how fasting slows ageing. Three problems, in order. The NAD+ part is compartment dependent and moves in opposite directions at once during a fast: the mitochondrial matrix becomes more reduced while the cytosol moves the other way. It has never been measured in human tissue during a fast in any compartment. The activation part has one human muscle measurement in the relevant window, and it went the other way: SIRT1 phosphorylation non-significantly down and p53 acetylation up. The lifespan part rests on invertebrate results that did not survive proper transgenic controls, and on a class of activator compounds identified with an assay whose readout depended on a fluorophore attached to the peptide.

Citations

  1. Human2019
    Early Time-Restricted Feeding Improves 24-Hour Glucose Levels and Affects Markers of the Circadian Clock, Aging, and Autophagy in Humans

    Nutrients

    11 overweight adults, 4-day randomised crossover of an 8am to 2pm eating window versus 8am to 8pm. The early window raised whole blood cell expression of SIRT1 and of the autophagy gene LC3A before breakfast (p < 0.04), raised MTOR expression in the evening (p = 0.007), and lowered mean 24-hour glucose by 4 mg/dL.

  2. Human2021
    Increasing whole-body energetic stress does not augment fasting-induced changes in human skeletal muscle

    Pflugers Archiv, European Journal of Physiology

    10 healthy males, vastus lateralis biopsies at baseline, 4 and 8 hours of a supervised fast. SIRT1 phosphorylation was non-significantly decreased (P < 0.06) alongside decreased ACC phosphorylation, while p53 acetylation increased over the fast in both conditions (P < 0.01).

  3. Human2020
    Nicotinamide riboside supplementation alters body composition and skeletal muscle acetylcarnitine concentrations in healthy obese humans

    American Journal of Clinical Nutrition

    13 overweight or obese adults, crossover, 6 weeks at 1000 mg nicotinamide riboside daily, with muscle biopsies. Markers of increased NAD+ synthesis, nicotinic acid adenine dinucleotide and methylnicotinamide, were elevated in skeletal muscle. Fat free mass increased (62.65 versus 61.32%, change 1.34%, P = 0.02), sleeping metabolic rate increased, and skeletal muscle acetylcarnitine increased (4558 versus 3025 pmol per mg dry weight, P = 0.04). Insulin sensitivity by clamp was among the outcomes and is not reported as improved in the abstract.

  4. Human2020
    Nicotinamide riboside does not alter mitochondrial respiration, content or morphology in skeletal muscle from obese and insulin-resistant men

    Journal of Physiology

    40 obese insulin-resistant men, 12 weeks at 1000 mg nicotinamide riboside twice daily, muscle biopsies before and after. Nicotinamide riboside did not affect NAD metabolite concentrations in skeletal muscle. Mitochondrial respiration by high resolution respirometry on single fibres, mitochondrial content and mitochondrial morphology were all unaffected. NAMPT protein abundance in muscle decreased. The first long-term human clinical trial to report these outcomes.

  5. In vitro2000
    Transcriptional silencing and longevity protein Sir2 is an NAD-dependent histone deacetylase

    Nature

    Yeast and purified enzymes. Sir2 established as an NAD+ dependent histone deacetylase, with the reaction consuming NAD+ stoichiometrically and producing nicotinamide plus 2'-O-acetyl-ADP-ribose. This coupling of enzyme activity to NAD+ availability is what makes the family metabolically interesting, and it is not in dispute.

  6. Animal2005
    Nutrient control of glucose homeostasis through a complex of PGC-1alpha and SIRT1

    Nature

    Mouse liver, hepatocytes and cell culture. SIRT1 forms a complex with PGC-1 alpha, and this controls the hepatic glucose and lipid response to fasting. Deacetylated PGC-1 alpha is the transcriptionally active form.

  7. Animal2007
    Fasting-dependent glucose and lipid metabolic response through hepatic sirtuin 1

    Proceedings of the National Academy of Sciences

    Mouse loss-of-function work confirming that hepatic SIRT1 is required for the fasting-dependent glucose and lipid metabolic response.

  8. Animal2004
    Sir2 mediates longevity in the fly through a pathway related to calorie restriction

    Proceedings of the National Academy of Sciences

    Drosophila. The original claim that Sir2 overexpression extends lifespan through a pathway related to calorie restriction. This result did not survive re-examination with appropriate transgenic controls.

  9. Animal2011
    Absence of effects of Sir2 overexpression on lifespan in C. elegans and Drosophila

    Nature

    Standardising genetic background and using appropriate transgenic controls abolished the apparent lifespan effects in both C. elegans and Drosophila. In the worm, outcrossing a high-overexpression sir-2.1 line removed the longevity increase while leaving overexpression intact, and the longevity co-segregated with a second-site mutation affecting sensory neurons. In the fly, the overexpressing strain was long-lived versus wild type but not versus appropriate transgenic controls, and a more strongly overexpressing line was not long-lived either.

  10. In vitro2005
    Substrate-specific activation of sirtuins by resveratrol

    Journal of Biological Chemistry

    In vitro enzyme assays with tagged and untagged peptide substrates. Sirtuin activation by resveratrol is substrate specific and dependent on the covalently attached fluorophore in the assay peptide; with a native substrate lacking the fluorophore, the activation was not observed.

  11. Animal2012
    The sirtuin SIRT6 regulates lifespan in male mice

    Nature

    SIRT6 overexpression extends lifespan in mice, and in the original report the effect was male specific.