Mess with a good guy, and you may discover his shadier side. Still, it came as a shock when scientists blocked the activity of sirtuins—molecular heroes known for promoting longevity and protecting against age-related maladies—and found that this inhibition restored cognition in a mouse model for Alzheimer disease. In this week’s Journal of Neuroscience, researchers led by Frank LaFerla at the University of California, Irvine, report that the sirtuin inhibitor nicotinamide prevents memory defects and selectively reduces a phospho-tau species associated with microtubule stability in 3xTg-AD mice. The findings suggest that oral nicotinamide, the biologically active form of niacin (vitamin B3), could be used to treat AD—a prospect that will be tested in an early-stage clinical trial.

Sirtuins are class III histone deacetylases (HDAC) whose activation slows age-associated disease in mice through transcriptional pathways closely resembling those induced by caloric restriction (see ARF related news story). This led many, including first author Kim Green, to presume that blocking the activity of sirtuins would make things worse for mice that harbor both hallmark AD pathologies and show memory impairments at four months of age.

Instead, Green and colleagues found striking cognitive improvement in eight-month-old 3xTg-AD mice that were given nicotinamide-spiked drinking water (200 mg/kg/d), as opposed to vehicle-treated, for four months. In tests of long-term, short-term, and hippocampal-dependent spatial memory, nicotinamide rescued 3xTg-AD mice to wild-type levels, and even boosted short-term memory in non-transgenic animals. Nicotinamide-treated 3xTg-AD mice fared better than vehicle-treated counterparts in a fear conditioning test requiring the amygdala—a brain area with prominent pathology in the triple transgenics. However, both groups did equally well on an object recognition task involving the cortex, a region largely spared of Aβ pathology. Consistent with their equally good performance on the cortex-dependent task, nicotinamide-treated 3xTg-AD mice had roughly similar brain Aβ levels as did control mice—measured by ELISA on detergent-soluble and -insoluble brain homogenates, and by immunohistochemical staining of hippocampus, amygdala, and cortex.

Given nicotinamide’s lack of effect on Aβ levels, the researchers figured the compound must be improving cognition through another mechanism. Analyzing protein extracts from whole brain samples of treated and control 3xTg-AD mice, they found a 20 percent reduction in steady-state levels of human tau in the nicotinamide-treated animals. They saw no differences at several tau sites typically phosphorylated in 3xTg-AD mice by eight months (Thr212/Ser214, Ser199/202, or Thr181), but a whopping 60 percent reduction in Thr231-phospho-tau in the nicotinamide group compared with vehicle. “It's incredibly dramatic,” Green told ARF. “This thing is just wiped from the brain very specifically.”

To address whether phosphorylation at Thr231 affects stability and accumulation of tau, the researchers transfected fibroblast cells with wild-type human tau or a phospho-mimic human tau (T231E) containing a glutamic acid substitution at the Thr231 site. Compared to cells overexpressing wild-type tau, the T231E-expressing cells had reduced steady-state levels of tau (yet similar myc-actin levels). In a filter retardation assay, accumulation of insoluble tau was seen for wild-type but not T231E mutant proteins. Phosphorylation at Thr231 “seems to target tau for degradation,” Green said. “We think that somehow the nicotinamide is facilitating that degradation in the AD brain.” Interestingly, another group recently showed that inhibition of another histone deacetylase, HDAC6, specifically reduces tau phosphorylation at the very same residue (Thr231) (Ding et al., 2008).

In the current study, nicotinamide did not affect levels of cyclin-dependent kinase 5 (Cdk5) or GSK3β, two kinases that mediate tau hyperphosphorylation in 3xTg-AD mice and human brain. The researchers did, however, find that nicotinamide treatment increased levels of Cdk5’s coactivator, p25, whose upregulation may be either pathological or physiological (see ARF related news story). Relative to vehicle-treated counterparts, nicotinamide-treated 3xTg-AD mice also had more than doubled levels of di-acetyl-α-tubulin, a protein associated with increased microtubule stability, and a greater than fourfold increase in microtubule-associated protein 2c (MAP2c), which shares 80 percent homology with tau in its microtubule-binding domain.

To address, albeit indirectly, whether nicotinamide’s effects are in fact being mediated by sirtuins, the scientists crossed homozygous 3xTg-AD mice with heterozygous SIRT1 knockouts (homozygous null mice are embryonic lethal). Total human tau levels in the brain did not differ between mice with one or two copies of SIRT1, but Thr231-phospho-tau levels were about 50 percent lower in the single-copy mice, suggesting that SIRT1 deletion and nicotinamide have similar effects on tau pathology. However, SIRT1 knockdown did not lead to increased levels of α-tubulin or p25, suggesting that these proteins are regulated by other nicotinamide targets.

Considering recent reports of SIRT2 inhibitors that can prevent α-synuclein toxicity in cells and flies (see Outeiro et al., 2007 and ARF related news story) and suppress pathogenesis in a fly model of Huntington disease (Pallos et al., 2008), the new data might not seem so surprising. Though activating sirtuins is beneficial earlier in life, Green said, when it comes to age-related neurodegeneration, inhibiting brain sirtuins may activate “a protective mechanism against a whole bunch of different cytosolic aggregating proteins.”

Nailing down the role of sirtuins in AD could still prove tricky, though. In an ICAD 2008 poster (see preliminary abstract), researchers led by Frederic Calon at Laval University Medical Center, Quebec, Canada, report decreased SIRT1 protein levels in parietal cortex of AD patients but not of patients with mild cognitive impairment (MCI). They also found no change in SIRT1 levels in the cortex of 3xTg-AD mice at 12, 16, and 20 months of age. These data suggest that loss of SIRT1 occurs at relatively late stages of AD, Calon told ARF. The full paper is currently in press in the Journal of Neuropathology and Experimental Neurology.

Based on the nicotinamide data in mice, LaFerla’s group is recruiting mild to moderate AD patients for a six-month clinical trial. Of the 50 patients, half will receive placebo pills, the other half 1,500 mg of nicotinamide twice a day. This dose is the human equivalent of the 200 mg/kg per day received by the 3xTg-AD mice, Green said. A standard multivitamin contains just tens of milligrams of niacin, which seemed to protect against AD and age-related cognitive decline in a prospective study (Morris et al., 2004). The primary outcome measure in the new trial will be the Alzheimer Disease Assessment Scale-Cognitive (ADAS-Cog), which will be administered to participants at six-week intervals. In addition, the researchers will perform spinal taps at the beginning and end of the study to measure CSF phospho-tau levels. They will also monitor liver function to see if the high doses of nicotinamide are safe.—Esther Landhuis


  1. One must be careful when calling nicotinamide an "inhibitor" in this experiment. While it is true that our lab showed that nicotinamide is a direct inhibitor of SIRT1 enzyme, it is also a precursor of NAD+, and NAD+ is a co-substrate (i.e., activator) of SIRT1.

    In vivo, there is an abundant enzyme called Nampt in cells and serum that initiates the conversion of nicotinamide to NAD+. Therefore we should entertain the possibility that nicotinamide is activating SIRT1 in vivo, not inhibiting it. This would fit with other papers showing that SIRT1 is neuroprotective.

  2. The experimental dose used in the study was 200 mg/kg/day. This would translate to a daily dose of nearly 14,000 mg for a 70 kg (154 lb.) person. Yet in the proposed clinical trial the experimental group will be receiving a daily dose of 3,000 mg. How does one explain the lower dose being used in the clinical trial?

  3. I am responding to William Polsky's comment on computation of the human dose of nicotinamide.

    Following the publication of a study on the use of resveratrol in mice to improve their health and maximum lifespan, the press reported that a human would have to consume an enormous amount of wine or supplements to gain similar benefits. This statement shows a lack of understanding of the appropriate criteria for dosage translations between species.

    There are a number of acceptable ways to compute the human equivalent dose from animal studies. The key is to consider energy-expenditure differences between species. Energy expenditure is a measure of metabolic rate. The method favored by the FDA (see uses the body surface area (BSA) normalization method. Basal metabolic rate is directly related to surface area. As the FDA notes, the BSA method correlates well across several mammalian species with several parameters of biology, including oxygen utilization, caloric expenditure, basal metabolism, blood volume, circulating plasma proteins, and renal function. However, there are important differences, such as different sensitivities, that make the BSA method a guide rather than a rule.

    A recent article in the FASEB Journal criticized the media for its misunderstanding (or ignorance of) what a human equivalent dose would be for the amount of resveratrol used in the Sinclair mouse study to which the comment refers (1). Immediately after that paper was published, the popular press—along with a contingent of the scientific community—voiced concerns regarding the relevance to humans of the resveratrol dose used by the researchers. Almost without exception, the press scaled the amount of resveratrol given to the mice—22.4 mg per kg of body weight—to humans on a straight weight basis. According to their reports, a person weighing 175 lb. (about 80 kg) would have to ingest 22.4 x 80 = 1,792 mg/day. Furthermore, the media typically wrote that to get that much resveratrol from red wine (using an estimate of 2 mg of resveratrol per bottle), a person would have to drink 896 bottles per day.

    Pharmacology 101 teaches us, however, that ratios involving body weight, energy expenditure, and body surface area are far more realistic than weight ratios alone in scaling dosages from one species to another. This has been known for over a century, and the relevant scaling factors are familiar to most scientists. The media concluded that the human equivalent dose of the Sinclair study was ridiculously large and impractical.

    This does an injustice to the researchers, not to mention impede implementation. It's frustrating considering that resveratrol has been found to be safe in extremely large doses.

    Returning to the article in the FASEB Journal, the authors assert that the mouse dose in the Sinclair study should be multiplied by the appropriate mouse/human scaling factor of 3/37, which gives a value of 1.82 mg per kg per day. Using the 80-kg person as an example again, the human dosage would therefore be 1.82 x 80 = 146 mg/day, an amount easily achieved with supplements, but not so easily with wine (73 bottles!). But the mice were not fed wine.

    We do not know for certain if resveratrol can do for humans what it does for mice and other creatures, but the upside potential is great, and there does not appear to be a downside as yet.

    Applying this line of reasoning to nicotinamide yields about 1,298 mg/day for a 80-kg person (200 * 0.081 * 80).


    . Dose translation from animal to human studies revisited. FASEB J. 2008 Mar;22(3):659-61. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. SIRT1, Resveratrol and More: Moving Closer to Anti-aging Elixir?
  2. SfN: P25 at Synapses—A Bite Peps Up, A Binge Crashes the System
  3. Parkinson Disease—Potential Targets, Therapies

Paper Citations

  1. . Histone deacetylase 6 interacts with the microtubule-associated protein tau. J Neurochem. 2008 Sep;106(5):2119-30. PubMed.
  2. . Sirtuin 2 inhibitors rescue alpha-synuclein-mediated toxicity in models of Parkinson's disease. Science. 2007 Jul 27;317(5837):516-9. PubMed.
  3. . Inhibition of specific HDACs and sirtuins suppresses pathogenesis in a Drosophila model of Huntington's disease. Hum Mol Genet. 2008 Dec 1;17(23):3767-75. PubMed.
  4. . Dietary niacin and the risk of incident Alzheimer's disease and of cognitive decline. J Neurol Neurosurg Psychiatry. 2004 Aug;75(8):1093-9. PubMed.

Other Citations

  1. 3xTg-AD

External Citations

  1. clinical trial
  2. preliminary abstract

Further Reading


  1. . Sirtuin 2 inhibitors rescue alpha-synuclein-mediated toxicity in models of Parkinson's disease. Science. 2007 Jul 27;317(5837):516-9. PubMed.
  2. . Dietary niacin and the risk of incident Alzheimer's disease and of cognitive decline. J Neurol Neurosurg Psychiatry. 2004 Aug;75(8):1093-9. PubMed.

Primary Papers

  1. . Nicotinamide restores cognition in Alzheimer's disease transgenic mice via a mechanism involving sirtuin inhibition and selective reduction of Thr231-phosphotau. J Neurosci. 2008 Nov 5;28(45):11500-10. PubMed.