Using mice that develop mild to moderate tau pathology similar to that seen in early Alzheimer disease, Peter Davies, Pablo Castillo, and colleagues at the Albert Einstein College of Medicine, New York, have identified age-dependent learning problems that seem to stem from defects in synaptic function. Published in the August 26 issue of the Journal of Neuroscience, their data suggest that tau can cause memory impairment by disrupting synaptic function before extensive pathology sets in. That scenario is reminiscent of the way amyloid-β is also believed to act early in disease, and offers some hope that tau-triggered dementia could be reversible to some extent.

The work began with a simple question on Davies’s part: What do neurofibrillary tangles do to the brain? These aggregates appear in neurons in AD brain when the microtubule-binding protein tau becomes hyperphosphorylated and undergoes conformational changes to form fibrillar assemblies. The presence of tau tangles correlates with dementia, but the exact effect of tau aggregation on neuronal function is not known.

Starting with the hypothesis that neurofibrillary tangles lead to cognitive problems by impairing synaptic function, Davies and coworkers looked at both behavior and neuronal physiology in mice that overexpress the human tau gene in place of mouse tau (Andorfer et al., 2003). The mice develop an age-dependent tauopathy, with increased phosphorylation and redistribution of tau from axons to somatodendritic regions. Eventually, they show accumulation of paired helical filament structures, then neuronal loss. Compared to other models that express mutated forms of tau (see Tau Research Models), the human tau mice show a milder pathology that is more representative of tau’s role in AD, Davies told ARF, “because both the distribution and the extent of the pathology are somewhat similar to what we see in Alzheimer disease, rather than what you see in a more florid frontotemporal dementia.”

First author Manuela Polydoro looked at four-month-old versus 12-month-old tau mice, ages at which the animals show mild and moderate tau pathology, respectively. In the young mice, she noticed modest somatodendritic accumulation of phospho-tau, but very few neurons with paired helical filaments. The mice showed no behavioral changes in tests of learning and memory, including novel object recognition and the Morris water maze.

In older mice, however, the redistribution of tau was more pronounced, and there were more paired helical filament-containing cells, indicating what Davies calls a “moderate” level of tau pathology. Along with that, the older mice showed deficits on novel object recognition; in the Morris water maze, they were slower in learning the location of the platform and worse at remembering what they had learned.

Measures of synaptic function in hippocampal slices from the animals showed a similar pattern: The young mice were entirely normal, while the old mice revealed evidence of reduced probability of presynaptic glutamate neurotransmitter release (increased paired pulse ratio) and an inability to induce long-term potentiation after high-frequency stimulation. Interestingly, the LTP defect was not universal: The animals displayed normal LTP in a different induction protocol using theta burst stimulation. Further experiments tied the lack of LTP to an inability of the input cells to generate synaptic bursts specifically in response to high-frequency stimulation.

Davies emphasized that the synaptic impairment is subtle, and they do not know yet what causes it. Nonetheless, the results suggest researchers have been overestimating the amount of pathology that’s required to impair function. “It doesn’t take much pathology to see these physiologic deficits, and this kind of pathology is not uncommon in people,” he said. Davies said the level of tau pathology in the 12-month-old mice is comparable to that seen in people with mild cognitive impairment, or early Alzheimer disease.

“I’ve been looking at this pathology in human brain for decades without knowing what it meant. It was satisfying in a sense to get an answer where, yes, even this relatively mild disturbance in tau is enough to impair neuronal function,” Davies told ARF. “If this is true in humans, I think it is quite significant.” Because the neurons are not dead, he says, “There may be a reversible point that we’re identifying here where we can intervene and restore neurons.”—Pat McCaffrey


  1. Terrific paper. It helps explain what is happening with tau pathology earlier in the process.

    If the early tau pathology is reversible, with neurons still living but impaired, one wonders, What are our treatment options? I am particular curious, of course, about possible nutritional approaches.

    Emerson Lombardo, NB, Volicer L, Martin A, Wu B. Zhang XW. (2006) Memory preservation diet™ ©2005 to Reduce Risk and Slow Progression of Alzheimer's Disease (AD). In Vellas B, Grundman M, Feldman H, Fitten LJ, Winblad B, editors, Research and Practice in Alzheimer's Disease and Cognitive Decline, vol 9: 138-59.

    Emerson Lombardo NB. Martin A. Volicer L. Mandell A. Wen Zhang X. (2006) Comprehensive whole foods diet to reduce risk and slow progression of Alzheimer’s disease. J Nutri Health & Aging 10(3) 211.

    Otsuka M, Sato T, Ueki A. (2004) The effect of nutritional intervention on cognitive function in patients with AD J Nutri Health & Aging 8 (5): 428.

    View all comments by Nancy B. Emerson Lombardo
  2. This is a very carefully performed behavioral and electrophysiological study of an established tau mouse model. AD is a synapse failure.

    View all comments by Jürgen Götz

Make a Comment

To make a comment you must login or register.


Paper Citations

  1. . Hyperphosphorylation and aggregation of tau in mice expressing normal human tau isoforms. J Neurochem. 2003 Aug;86(3):582-90. PubMed.

Other Citations

  1. Tau Research Models

Further Reading

No Available Further Reading

Primary Papers

  1. . Age-dependent impairment of cognitive and synaptic function in the htau mouse model of tau pathology. J Neurosci. 2009 Aug 26;29(34):10741-9. PubMed.