The twin lesions of Alzheimer disease—amyloid plaques and tau tangles—are mentioned in the same breath but have never achieved equal status among researchers. Most of the efforts to find treatments for AD focus on reducing the accumulation of amyloid-β (Aβ) peptides, while tau’s reputation as a “downstream player” has relegated it to the sidelines of drug discovery.

A new study puts the microtubule-associated protein tau front and center in Aβ toxicity. In the May 4 issue of Science, Erik Roberson, Lennart Mucke, and colleagues at the Gladstone Institute and the University of California in San Francisco show that transgenic mice with half of the normal amount of tau protein are resistant to cognitive deficits induced by Aβ overproduction. The scientists show that tau reduction protects neurons not only against Aβ, but against other excitotoxic insults as well. The work points to a new physiological role for tau in excitotoxicity, and suggests that reducing levels of normal, endogenous tau could help the brain resist toxic effects of Aβ in AD.

Until now, therapeutic strategies around tau have primarily targeted kinase inhibitors, aiming to reverse tau hyperphosphorylation, aggregation, and tangle formation. One obstacle in these efforts has been that researchers do not really know which forms of tau are “good” or “bad” for neurons, leading to questions about which kinases or tau forms are the right target.

Roberson and colleagues decided to skirt this issue by lowering tau concentration wholesale in a mouse model of AD amyloidosis. To do that, they crossed tau knockout mice with human amyloid precursor protein (hAPP)-overexpressing mice. With tau lowered, whether by half in heterozygotes, or completely in homozygotes, the mice fared dramatically better by several measures. In the Morris water maze, the hAPP mice with lower tau performed equally to wild-type, with no discernable deficits in learning or memory. The hyperactivity and early death seen in hAPP mice was gone, too.

Somehow, tau reduction prevented the major Aβ-dependent adverse effects in hAPP mice. How did this happen? Reducing tau did not change hAPP expression, Aβ levels, Aβ variant ratios, or plaque load. The researchers found no change in levels of Aβ*56, an oligomeric form of the peptide that has been shown to cause memory deficits in rats (see ARF related news story). They concluded that tau reduction must somehow act downstream to protect neurons.

Nor did tau reduction protect neurons by preventing the production of modified forms of tau protein. In the hAPP mouse, the researchers found no Aβ-induced phosphorylation of tau, a pathological modification tied to cytotoxicity in other disease models. A previous study showing that tau reduction abolished acute Aβ toxicity in cells in culture implicated a 17 kDa tau fragment (Park and Ferreira, 2005), but Roberson and colleagues did not find a similar fragment in their mice. Thus, the protection against Aβ-dependent cognitive impairment in this study did not involve the loss of a large pool of modified tau. The results point to a physiological form of tau as the culprit in sensitizing neurons to the effects of Aβ.

This idea was borne out when the researchers looked at how tau depletion affected excitotoxicity. Previous observations had established that some Aβ-expressing strains of mice become hypersensitive to excitotoxic insults. Consistent with this, the researchers found that the hAPP mice showed a heightened response to the seizure-inducing effects of GABA receptor antagonist pentylenetetrazole (PTZ). A dose that normal mice tolerated actually killed 20 percent of hAPP mice. Knocking out tau abolished this sensitivity. In hAPP/tau-/- mice, no mice died after PTZ treatment, and their seizures were less severe than in hAPP/tau+/+ mice. Importantly, even mice without hAPP showed the tau-dependent difference in sensitivity, and tau depletion protected normal mice against kainite-induced seizures.

“Our results suggest that tau is not a downstream effector specifically of Aβ. By lowering tau, we changed the physiological function of the neuron so as to make it more resistant to being put into overdrive by either kainite or GABA antagonists or Aβ,” Roberson told ARF.

In terms of translating these observations to treatment strategies, Roberson stressed the need to be careful about extrapolating from mice to humans. Lacking neuronal loss, the hAPP mouse model does not fully recapitulate AD. On the other hand, he said, the absence of neuronal loss makes the model useful to study reversible neuronal dysfunction, which Mucke and colleagues feels is likely very important in AD (see review by Palop et al., 2006). The idea is that the well-known but mysterious clinical observation of wide fluctuations in the day-to-day functioning of a given AD patient indicates some component of reversible cognitive impairment, even late in disease. This neuronal dysfunction, caused by a continuous onslaught of Aβ, might be separable from irreversible neuronal loss, they speculate, and is likely to be a network function, not an expression of neurodegeneration.

“We know that in people, the range of abilities is wide at any given amount of neuronal loss, so the ability to modulate such dysfunction on top of cell loss is very important,” Roberson explained.

“Our results indicate that tau may play a pretty important role in cognitive impairment independent of neuron loss. This makes it a good treatment target, which could give us a chance to maximize neuronal function even in the face of neuron loss,” he said.

Roberson acknowledges it could be pharmacologically tricky to find a way to reduce endogenous tau. As a start, he points to work by Michael Hutton and colleagues, who screened a library of small molecules and found some that reduce tau protein levels in cultured cells (Dickey et al., 2006).

Genetic hints that endogenous tau levels could affect AD risk exist, as well. Work from John Hardy’s team connected a haplotype associated with increased tau expression with a higher risk of AD in people (see ARF related news story). That study does not definitively isolate expression level as a cause, because the tau isoform profile is affected as well. So far, there is no genetic evidence that decreased tau expression is protective. Roberson says he is interested in looking at a broader study of people who carry the high-risk ApoE4 allele, but have low tau expression, to see if they might be at lower risk for AD.—Pat McCaffrey


  1. The paper supports the notion that tauopathy plays a major role in AD pathogenesis.

    View all comments by Takaomi Saido
  2. This is a remarkable report. It addresses key aspects of the underlying mechanisms of AD and has significant implications for AD drug discovery. Guided in large part by the amyloid cascade hypothesis, most potential treatments for AD target one or more aspects of Aβ amyloidosis, by reversing amyloid plaques, reducing levels or Aβ, inhibiting Aβ fibrillization, or promoting clearance of toxic Aβ fibrils or oligomeric species of Aβ. The microtubule-associated protein tau also is implicated in mechanisms of AD, but tau has often been considered to be downstream of the toxic effects of Aβ. However, the data presented here by Roberson et al. show that reducing endogenous levels of tau prevented behavioral deficits in transgenic mice expressing human amyloid precursor protein without altering their high levels of brain Aβ. Significantly, these data suggest that by reducing levels of brain tau, it may be possible to block Aβ-mediated neuronal dysfunction and neurodegeneration. The authors suggest this strategy may represent a novel approach to developing better therapies for AD and related disorders known as tauopathies.

    Since most insights into tauopathies have come from studies of AD, there is extensive literature on tau pathologies in AD. But a substantial amount of data also has come from studies of related tauopathies, wherein tau pathology is the critical underlying abnormality that links all of these disorders to a shared mechanism of neurodegeneration. For example, when tau becomes hyperphosphorylated, it forms amyloid filaments that aggregate to form neurofibrillary tangles (NFTs). As a result, this leaves less tau available to stabilize microtubules (MTs), and when MTs are destabilized, this compromises intraneuronal transport leading to neurodegeneration. Thus, several types of tau-focused interventions are being developed, including some that are directed at abrogating tau fibrillization or hyperphosphorylation, and others that are designed to stabilize MTs by compensating for the sequestration of tau in NFTs (Schenk et al.; Skovronsky et al.). However, while it may be desirable to suppress mutant forms of tau in hereditary tauopathies or in tauopathies with an abnormal ratio of 3- versus 4-repeat tau isoforms, reducing tau levels to the extent that MT stability is compromised is likely to have long-term deleterious effects. For example, the report by Ikegami et al. on tau knockout mice demonstrated that these tau-deficient mice develop cognitive and motor abnormalities with age. This signifies that reducing tau levels may have negative consequences. However, there are few studies examining the behavioral and CNS consequences of reducing tau levels over the lifespan, and hopefully the studies by Roberson et al. make it clear that far more research is needed on this topic if we are to exploit their findings for therapeutic benefit of patients with AD or a related tauopathy.


    . Muscle weakness, hyperactivity, and impairment in fear conditioning in tau-deficient mice. Neurosci Lett. 2000 Feb 4;279(3):129-32. PubMed.

    . Cytoskeletal modulators and pleiotropic strategies for Alzheimer drug discovery. Alzheimers Dement. 2006 Oct;2(4):275-81. PubMed.

    . Neurodegenerative diseases: new concepts of pathogenesis and their therapeutic implications. Annu Rev Pathol. 2006;1:151-70. PubMed.

    View all comments by John Trojanowski

Make a Comment

To make a comment you must login or register.


News Citations

  1. Aβ Star is Born? Memory Loss in APP Mice Blamed on Oligomer
  2. Tau Shows Subtle Hints of Genetic Association

Paper Citations

  1. . The generation of a 17 kDa neurotoxic fragment: an alternative mechanism by which tau mediates beta-amyloid-induced neurodegeneration. J Neurosci. 2005 Jun 1;25(22):5365-75. PubMed.
  2. . A network dysfunction perspective on neurodegenerative diseases. Nature. 2006 Oct 19;443(7113):768-73. PubMed.
  3. . Pharmacologic reductions of total tau levels; implications for the role of microtubule dynamics in regulating tau expression. Mol Neurodegener. 2006;1:6. PubMed.

Further Reading

Primary Papers

  1. . Reducing endogenous tau ameliorates amyloid beta-induced deficits in an Alzheimer's disease mouse model. Science. 2007 May 4;316(5825):750-4. PubMed.