Lovers of red wine the world over may once again take comfort from the amazing powers of resveratrol, a grape-derived polyphenol that has been shown to extend lifespan in yeast, worms, and fruit flies, and which activates the deacetylase SIRT1, which extends lifespan of mammalian cells (see ARF related news story). In today’s Science, Jeffrey Milbrandt and colleagues at Washington University School of Medicine, St. Louis, report that Wallerian degeneration, the active process whereby neuronal axons are destroyed, is inhibited by resveratrol and its aide d’accompli, SIRT1. The results may have implications for the protection of neurons in Alzheimer’s and other neurodegenerative diseases in which synaptic and dendritic death is often the best correlate of cognitive decline (see Hashimoto and Masliah, 2003).

First author Toshiyuki Araki and colleagues made the connection between the deacetylase and axonal degeneration by studying a mutation in mice that causes slow Wallerian degeneration. The Wlds mutation is caused by a recombination event (see ARF related news story) that results in expression of a protein chimera comprising the N-terminal of Ufd2a (ubiquitin fusion degradation protein 2a) and a full length nicotinamide mononucleotide adenylyltransferase 1 (Nmnat1). Ufd2a was strongly touted as the business end of this chimera, given that the proteasome/ubiquitin system had previously been implicated in Wallerian degeneration (see ARF related news story). But when Araki introduced some extra Nmnat1 into normal neurons and then severed their axons, he found they were protected just as well as Wlds axons. In fact, 72 hours after separation, more than 90 percent of axons are still intact in both normal neurons expressing the extra Nmnat1, and in Wlds neurons. In contrast, only about 10 percent of normal axons survive the same insult. What’s more, Araki found that knocking down Ufd2a doesn’t help. In neurons treated with small interfering Ufd2a RNAs, axons were not protected. This finding dashes the theory that the Wlds chimera acts as a dominant-negative Ufd2a.

So where do resveratrol and SIRT1 enter the picture? Well, SIRT1 is NAD-dependent, and Nmnat1 is involved in NAD biosynthesis. To prove that NAD plays a role in slowing Wallerian degeneration, Araki mutated Nmnat1 amino acids that are essential for this activity, and sure enough, these mutants failed to protect axons. In fact, Araki found that just adding NAD prior to damaging the axons was sufficient to protect them.

Armed with this knowledge, the authors sought out proteins and pathways that may be influenced by changes in NAD levels, choosing to focus on protein deacetylases and polyADP-ribose polymerases (PARPs). When they tested for slow Wallerian degeneration in neurons treated with the deacetylase inhibitor sirtinol or the PARP inhibitor 3AB, they found that only sirtinol blocked the protective effects of NAD. Taking the next logical step, they individually knocked down all seven SIRT deacetylases and found that only loss of SIRT1 abolished the protective effects of NAD. Resveratrol, which activates SIRT1, was even better at protecting axons than was the nucleotide—0.1 mM of the polyphenol offered about the same protection as 1mM NAD.

In the broadest sense, these results link neurodegeneration with diet, basic metabolism, and longevity. “It is possible that the alteration of NAD levels by manipulation of the NAD biosynthetic pathway, Sir2 protein activity, or other downstream effectors will provide new therapeutic opportunities for the treatment of diseases involving axonopathy and neurodegeneration,” the authors write.

Antonio Bedalov and Julian Simon, from the Fred Hutchinson Cancer Research Center, Seattle, are equally optimistic. “The therapeutic implication of this finding is that it may be possible to design neuroprotective drugs that boost SIRT1 activity and prevent further neurodegeneration in diseases like AD and PD,” they write in an accompanying Science perspective.

In the meantime, maybe we should keep quaffing our favorite vintage.—Tom Fagan


  1. This is a powerful and noteworthy article. The authors show with strong and well-considered data that there is a link between NAD metabolism and axonal degeneration rates. Through the use of an identified mouse mutant, Wlds, which has long-lived axons, the authors used the knowledge that a protein produced in this mouse is a fusion of parts of a ubiquitin ligase and an enzyme involved in NAD synthesis. Careful experiments dissected the roles of these disparate arms of neuronal metabolism. A noted article for contrast is that of Zhai et al., implicating the proteasome in axonal degeneration.

    The authors begin by examining the N-terminus of the Wld fusion, which comprises the first 70 amino acids of the ubiquitin ligase Ufd2a. In yeast, this protein ortholog is known to mediate cell survival under conditions of duress, so it was reasonable to hypothesize that Ufd2a activity might mediate axonal survival when toxins or neuronal transection were applied. It is arguable that different Ufd2a fusions provide different binding sites on their folded surfaces, and the authors furthermore make no attempt to discuss how a small domain of only 70 residues could possess such activity when even ubiquitin itself is larger. Nonetheless, this criticism is irrelevant, and the best available experiment as performed by the authors of inserting a GFP-Ufd2 construct into the mouse shows no effect. Furthermore, though not mentioned, the U-Box is present at the C-terminus of Ufd2, and not in the domain fused to the NAD relevant portion. It is, therefore, accepted from this evidence that the Ub ligase fusion is random and spurious, and not physiologically related to the function of the NAD-enzyme in alleviating disease phenotype. This is strongly correlated by confirmation of the lack of effect in the neurons with transfection of a dominant negative Ufd2a mutant (molecular lesion not described), and siRNA for the same. Ufd2a is not responsible for the physiological effect of the Wld fusion.

    The carboxyl terminus of the Wld fusion is Nmnat1, a nicotinamide adenylyl transferase expressed highly in neuron. Introduction of the GFP-Nmnat1 fusion isolated away from the Ub ligase produces a similar cellular phenotype as does the neuron control. Additionally, carefully constructed point mutants in the adenylase which did not synthesize NAD were unable to rescue axons that were dying back, or suppress their shrinkage post-trauma by two mechanisms. This suggested that the increased NAD synthesis was responsible for the action of the fusion in the disease process. Support for this idea came in the form of addition of NAD to cultures, which again suppressed axonal retraction.

    The report continues with a somewhat disjunct discussion of Sir2 and PARP action, two selected NAD utilizing enzymes. Inhibitors of PARP had no effect on the experimentation, and the negative results are accepted. However, inhibitors of Sir2, a neuronal histone deacetylase, blocked axonal regression. Equally, stimulators of Sir2 slowed the axonal degeneration. This is a confusing result, and resolution depends heavily on the actual biochemical activities of the sirtrinol and resveratrol used, as well as their specificity towards other NAD utilizing enzymes (not at all discussed). These results are not as strong as the first set, but imply that Sir2 may be involved in generating new transcripts or signaling which results in axonal decay.

    Next, the authors take the implied Sir2 connection further and begin a screen of bioinformatically identified proteins with a Sir2 domain. Only one of those Sir2 domain-containing proteins inhibited by siRNA methods blocked axon degeneration as well as sirtrinol. This protein is SIRT1. The report ends with the suggestion that SIRT1 may regulate the neuronal breakdown by unknown processes.

    In total, the report adequately shows that the Ubox ligase Ufd2 is not involved in the (presumably) proteasome-mediated axonal degeneration. This is not surprising, as there are many ligases in the cell, perhaps 500, and statistically the result is not unacceptable. Furthermore, the idea that NAD levels in the nucleus mediate the processes is backed by solid evidence, and in this light the paper presents its strongest foot forward. More murky are the suggestions that Sir1 is the actual mediator of these processes, and the pharmacology of the drugs used are not investigated at a deep enough level to adequately dissect from this report what is going on metabolically. Nonetheless, it is an interesting implication, and the SIRT1 data is not weak. We may assume from all the results that an unknown NAD-requiring enzyme, perhaps SIRT1 but not necessarily, is proceeding faster, as a result of higher substrate concentration beyond normal nuclear physiological levels.

    It remains to be seen if other NAD utilizing enzymes are involved in the process, which transcriptions may be affected, by deacetylation of which proteins, and how the time course of NAD addition can shed light on the specific transcriptional or post-translational events that lead to axonal protection. The paper is strong, unique, and welcome.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Who Says Chivalry is Dead?—Sir2 Fights Against Aging in Mammals
  2. Protein Chimera Found to Protect Axons from Degeneration
  3. Proteasome Implicated in Axon Degeneration

Paper Citations

  1. . Cycles of aberrant synaptic sprouting and neurodegeneration in Alzheimer's and dementia with Lewy bodies. Neurochem Res. 2003 Nov;28(11):1743-56. PubMed.

Further Reading

No Available Further Reading

Primary Papers

  1. . Neuroscience. NAD to the rescue. Science. 2004 Aug 13;305(5686):954-5. PubMed.
  2. . Increased nuclear NAD biosynthesis and SIRT1 activation prevent axonal degeneration. Science. 2004 Aug 13;305(5686):1010-3. PubMed.