Tau’s rap sheet is growing. In Alzheimer’s disease, evidence implicates the protein as the mediator of Aβ’s toxic effects. Now, two papers from Australia blame tau for the lion’s share of cell death after ischemic stroke. In the September 7 Nature Communications, researchers led by Lars Ittner and Yazi Ke at the University of New South Wales, Sydney, reported that young tau knockout mice underwent very little neurodegeneration after stroke, unlike the massive cell death seen in wild-type. Specifically, knockouts were protected from excitotoxicity by accumulation of a protective protein, SynGAP1, at their post-synapses. “This extends the strategy of lowering tau to conditions such as stroke and seizures,” Ittner suggested. He presented some of this work at the Society for Neuroscience annual conference (Dec 2013 conference news). 

  • Tau knockout mice recover quickly after ischemic stroke.
  • These mice have excess SynGAP1, which silences excitotoxic signaling.
  • Older tau knockouts lose stroke protection due to brain iron buildup.

Other researchers agreed the findings strengthened the idea that tau regulates post-synaptic signaling. Donna Wilcock at the University of Kentucky, Lexington, was impressed by the size of the protective effect. “Stroke is such an extreme, acute injury model that to see this kind of protection is very significant,” Wilcock told Alzforum.

A paper in the September 8 Molecular Psychiatry adds nuance to tau’s role, suggesting its effects depend on age. Researchers led by Ashley Bush at the University of Melbourne, Australia, and Rong Liu at Huazhong University of Science and Technology, Wuhan, China, found that while young tau knockouts recovered well from stroke, older knockouts fared as poorly as wild-type. The authors traced the problem to brain iron accumulation prior to stroke, which led to more oxidative stress and cell damage after injury. Because brain iron also rises during normal aging, Bush suggested that iron-lowering strategies could help treat other diseases of aging as well. 

Better Off Without Tau.

Wild-type mice (left) suffer extensive tissue death (white) after stroke, while tau knockouts (right) lose cells only in the immediate vicinity of the blockage (arrow). [Courtesy of Bi et al., Nature Communications.]

Researchers first linked tau to excitotoxicity in AD model mice (May 2007 newsJul 2010 newsJan 2011 news). Later, the finding was extended to epilepsy models, although it was unclear how tau influenced neuronal excitability (Jan 2013 newsAug 2013 news). 

Because much of the progressive damage in stroke arises from excitotoxicity, Ittner and colleagues wondered if tau might play a role there, too. Joint first authors Mian Bi, Amadeus Gladbach, and Janet van Eersel blocked the middle cerebral arteries of three- to six-month-old wild-type and tau knockout mice for 90 minutes, then examined their brains for progressive changes over the next 24 hours. One hour after injury, both sets of mice had similar deficits, with local cell death around the blockage site and minor movement problems. However, wild-type mice continued to worsen over the next day, as frequent electrical storms hit their brains. Cell loss spread throughout the affected hemisphere, leading to severe motor problems. Meanwhile, the tau knockouts stabilized, with no further tissue loss and no aberrant electrical activity (see image above).

Why this difference? The authors compared gene expression in wild-type and tau knockout mice after administration of pentylenetetrazol (PTZ), a drug that induces seizures. Identifying the signaling pathways that were most responsive in each line of mice, they found that only wild-type mice turned up ERK signaling. NMDA receptors activated Ras GTPase, which in turn switched on downstream effectors including ERK, leading to the expression of excitotoxic genes. In the tau knockouts, by contrast, Ras stayed quiet.

This quiescence was due to the presence of another protein, synaptic RAS GTP activating protein (SynGAP1), the authors report. SynGAP1 normally binds to PSD-95 in the post-synapse and serves to deactivate Ras. In vitro experiments revealed that tau competes with SynGAP1 for binding to PSD-95. Thus, in the knockouts, post-synaptic SynGAP1 levels soared, silencing ERK signaling (see image below). The authors noted that knocking out tau conferred a similar level of protection from stroke as directly blocking ERK signaling, strengthening the connection (Gladbach et al., 2014). 

Tau Restrains SynGAP1.

Without tau at the synapse (right), SynGAP1 accumulates, silencing excitotoxic signaling. [Courtesy of Bi et al., Nature Communications.]

The authors confirmed the mechanism by knocking down SynGAP1 in tau knockout mice. This abolished their protection against seizures induced by PTZ. The double knockouts lost as many neurons as wild-type mice did after ischemia. The opposite experiment—overexpressing SynGAP1 in cultured wild-type mouse neurons—dampened ERK signaling.

In ongoing work, Ittner will examine whether this same excitotoxic pathway activates during Alzheimer’s disease. He also wants to know whether lowering tau after stroke would help recovery. Jean-Pierre Brion at the Université Libre de Bruxelles, Belgium, noted that protection in these experiments was seen with complete absence of tau. “It is not yet obvious that acute reduction of tau, as could theoretically be done in patients, will have the same effects,” he wrote to Alzforum (see full comment below).

Ittner questions if SynGAP1 itself would make a good therapeutic target. Mutations in this gene have been linked to epilepsy and mental retardation, underlining its key role in synapse function (Hamdan et al., 2009; Carvill et al., 2013; Berryer et al., 2013). Drugs would have to activate SynGAP1 to calm excitotoxicity, and activators are notoriously hard to make, Ittner pointed out.

While the findings add to the evidence that tau can be harmful in disease states, the paper from Bush and colleagues points out a potential drawback from too little tau. The authors previously reported that tau, with amyloid precursor protein (APP), helps export iron from neurons, suggesting that absence of tau might lead to iron buildup (Sep 2010 newsFeb 2012 news). Because iron worsens stroke damage, Bush wondered how tau knockouts would fare after ischemia (Park et al., 2011; Castellanos et al., 2002; Patt et al., 1990). 

Joint first authors Qing-Zhang Tuo, Peng Lei, and Katherine Jackman also blocked the middle cerebral artery for 60 minutes in three-month-old wild-type and tau knockout mice obtained from Michael Vitek at Duke University (Dawson et al., 2001). In wild-type, as expected, iron levels rose 50 percent in the affected hemisphere several hours after the blockage, while tau fell. These animals developed extensive neuron death and motor problems. In contrast, tau knockouts had no iron increase and recovered from ischemia with little loss of neurons. Overall, the findings closely mirrored Ittner’s, though using a different tau knockout strain.

Despite tau’s proposed role in iron export from the brain prior to stroke, these young tau knockouts accumulated no more brain iron than wild-types did. Compensatory mechanisms may help to export iron and keep levels low in young knockouts, the authors hypothesized. This changed with age. Brain iron accumulates during normal aging, and the brains of year-old tau knockouts accumulated 50 percent more iron than wild-type (Jul 2010 newsBartzokis et al., 2011Jan 2012 news). After ischemia, the aged tau knockouts lost tissue at the same rate as young wild-types.

Do poor outcomes after ischemia go back to iron? To test the idea, the authors infused three-month-old wild-type mice with intravenous APP after arterial blockage to stimulate iron transport. This prevented the surge of iron in the damaged hemisphere, preserved tissue, and maintained motor abilities in the animals.

The results suggested that lowering brain iron could improve stroke recovery. Supporting this, wild-type mice treated with ceruloplasmin, a glycoprotein that oxidizes iron and speeds up its transport, bounced back quickly from ischemia, and aged tau knockouts that received ceruloplasmin fared as well as their younger counterparts, suggesting that ceruloplasmin restored the knockouts’ protection from stroke.

Recently, researchers led by Brent Stockwell at Columbia University in New York characterized an iron-dependent form of cell death called ferroptosis (see Dixon et al., 2012, and Stockwell et al., 2017, for review). Bush and colleagues treated three-month-old wild-type mice with the ferroptosis inhibitors liproxstatin-1 and ferrostatin-1 after arterial blockage. These compounds worked faster than APP or ceruloplasmin, with treated animals performing better than untreated within six hours. Liproxstatin-1 also helped aged tau knockouts recover like youngsters.

“This is the first time that ferroptosis has been revealed to be active in stroke models,” Bush told Alzforum. He is interested in extending the work to AD to see if ferroptosis plays a role in cell death there, as well.

Stockwell found the data intriguing. “The study suggests that iron has a previously unappreciated role in driving cell death in various neurological conditions. This supports the emerging hypothesis that reducing the iron burden and inhibiting ferroptosis would be beneficial in stroke and AD,” he wrote to Alzforum. Ittner noted that it is easier to sop up excess iron than to modulate tau or activate SynGAP1, making iron levels an attractive target for therapy. A Phase 2 AD study of deferiprone, an approved iron chelator, is currently enrolling in Australia. Several studies have found that iron accumulates in Alzheimer’s and Parkinson’s disease (Feb 2001 news; Jul 2001 news; Mar 2003 news). 

Bush believes the findings highlight the importance of neurodegenerative disease research in aged mice. “We think the reason stroke and AD interventions in the preclinical phase don’t translate well to the clinic is because the models are tested when they’re too young. The extra burden of iron in the older brain matters for the clinical outcome,” he explained. If his findings hold up in AD, Alzheimer’s patients might benefit from combination therapy that lowers both iron and tau, Bush suggested.—Madolyn Bowman Rogers


  1. This paper from the Ittner group is quite interesting, and primarily documents a novel "noncanonical" function of tau, i.e., a tau-dependent activation of the ERK pathway.

    The findings are in line with previous findings of this team (and others) showing that absence of tau provides some resistance to excitotoxicity, and that this mediates at least part of Aβ-mediated toxicity.

    The present study dissects in more depth the signaling mechanisms of this protective effect in tau knockout (tauKO) mice by showing that this goes though inactivation of the Ras/Raf/MEK/ERK pathway. This inactivation is associated with a post-synaptic accumulation of the synaptic protein SynGAP1, an endogenous inhibitor of Ras. 

    These data bring additional support for the "protective" role of lack of tau in several experimental settings (including in AD models), although it is not clear if a similar mechanism is involved in all of them.

    From a therapeutics point of view, it should be noted that protection provided by lack of tau is observed in models with chronic absence of tau (tauKO), and it is not yet obvious that acute reduction of tau (as could theoretically be done in patients) will have the same effects. Increasing SynGAP1, or more exactly, increasing its post-synaptic accumulation, would be a theoretical approach, but no data are provided on the effect of increasing SynGAP1 levels in in vivo models.

    The authors showed previously that Aβ-mediated toxicity in APP mice is reduced in APP/tauKO mice. It would now be quite interesting to confirm that changes are observed in the Ras/Raf/MEK/ERK pathway and in SynGAP1 localization in the latter mice, and that deficiencies in APP mice can be rescued by some degree of inactivation of ERK.

  2. Network hyperexcitability leading to rhythm disturbances in cortical and hippocampal microcircuits is a feature of many neurodegenerative disorders at some stage during their progressive course, and tau is a key switch in the appearance of abnormal information flow. Reduction of tau protein markedly and robustly interferes with the development of hyperexcitability in the context of epilepsy and stroke, yet paradoxically, tau deficiency on its own has not led to detectable neurological deficits, particularly in disease models where hyperphosphorylated tau is strongly linked to cognitive decline. Apparently, “no tau is better than bad tau.”

  3. Tau is increasingly looking more like a sensitive but nonspecific biomarker of neuronal injury than a causative agent. Is it time for a paradigm shift?

Make a Comment

To make a comment you must login or register.


News Citations

  1. Is Dendritic Tau to Blame for AD-Related Hyperexcitability?
  2. APP Mice: Losing Tau Solves Their Memory Problems
  3. Honolulu: The Missing Link? Tau Mediates Aβ Toxicity at Synapse
  4. Tau’s Synaptic Hats: Regulating Activity, Disrupting Communication
  5. One Protein Fits All? Non-AD Epilepsy Models Thrive Sans Tau
  6. In Adult Mice, Reduced Tau Quiets Agitated Neurons
  7. Iron Export? New Role Links APP, Metals, to Oxidative Stress
  8. Tau Joins APP in the Ironworks
  9. Aging Primates—Brain Iron Up, Motor Skills and Plastic Synapses Down
  10. White Matter Changes Underlie Iron’s Impact on Brain Health
  11. Defect in Iron Metabolism Leads to Neurodegeneration
  12. Evidence on Iron Accumulates
  13. Ironing out the Role of Metals in Neurodegenerative Diseases

Paper Citations

  1. . ERK inhibition with PD184161 mitigates brain damage in a mouse model of stroke. J Neural Transm (Vienna). 2014 May;121(5):543-7. Epub 2013 Dec 14 PubMed.
  2. . Mutations in SYNGAP1 in autosomal nonsyndromic mental retardation. N Engl J Med. 2009 Feb 5;360(6):599-605. PubMed.
  3. . Targeted resequencing in epileptic encephalopathies identifies de novo mutations in CHD2 and SYNGAP1. Nat Genet. 2013 Jul;45(7):825-30. Epub 2013 May 26 PubMed.
  4. . Mutations in SYNGAP1 cause intellectual disability, autism, and a specific form of epilepsy by inducing haploinsufficiency. Hum Mutat. 2013 Feb;34(2):385-94. Epub 2012 Dec 12 PubMed.
  5. . Blood-derived iron mediates free radical production and neuronal death in the hippocampal CA1 area following transient forebrain ischemia in rat. Acta Neuropathol. 2011 Apr;121(4):459-73. Epub 2010 Dec 1 PubMed.
  6. . Iron intake increases infarct volume after permanent middle cerebral artery occlusion in rats. Brain Res. 2002 Oct 11;952(1):1-6. PubMed.
  7. . Iron depletion or chelation reduces ischemia/reperfusion-induced edema in gerbil brains. J Pediatr Surg. 1990 Feb;25(2):224-7; discussion 227-8. PubMed.
  8. . Inhibition of neuronal maturation in primary hippocampal neurons from tau deficient mice. J Cell Sci. 2001 Mar;114(Pt 6):1179-87. PubMed.
  9. . Gender and iron genes may modify associations between brain iron and memory in healthy aging. Neuropsychopharmacology. 2011 Jun;36(7):1375-84. PubMed.
  10. . Ferroptosis: an iron-dependent form of nonapoptotic cell death. Cell. 2012 May 25;149(5):1060-72. PubMed.
  11. . Ferroptosis: A Regulated Cell Death Nexus Linking Metabolism, Redox Biology, and Disease. Cell. 2017 Oct 5;171(2):273-285. PubMed.

External Citations

  1. tau knockout mice
  2. Phase 2 AD study of deferiprone

Further Reading

Primary Papers

  1. . Tau exacerbates excitotoxic brain damage in an animal model of stroke. Nat Commun. 2017 Sep 7;8(1):473. PubMed.
  2. . Tau-mediated iron export prevents ferroptotic damage after ischemic stroke. Mol Psychiatry. 2017 Nov;22(11):1520-1530. Epub 2017 Sep 8 PubMed.