Pathological proteins likely harm cognition by acting on synapses, but the details of how these proteins impair transmission largely remain a black box. In the March 4 Neuron, researchers led by Edward Stern at Bar-Ilan University in Ramat Gan, Israel, shed some light on how toxic tau might affect communication at the network level. In young tau model mice, cortical neurons fired more slowly than in controls, dampening network activity, the authors report. This seemed to be due to poor coordination of excitatory neurons, which made it harder for downstream cells to depolarize and fire. Notably, only a handful of neurons contained detectable tau tangles at this age. The data demonstrate that disrupting the timing of only a small percentage of neurons can degrade entire networks, Stern said. These young mice have spatial memory deficits, implying that the network alterations suffice to cause mild cognitive problems. The data also hint at new approaches for ameliorating cognitive symptoms in tauopathies such as frontotemporal dementia and Alzheimer’s disease, Stern noted.

“This is a fantastic paper. People have not seen this hypoexcitability at an early age before,” Kishore Kuchibhotla at New York University Langone Medical Center told Alzforum. He was not involved in the research.

Numerous studies have reported that toxic tau harms synapses, and accumulating evidence is pointing to soluble forms of the protein rather than neurofibrillary tangles (see Jul 2005 newsFeb 2007 newsMar 2013 conference news). To measure the effects of tau on overall network activity, most previous work employed acute brain slices. Researchers found hyperexcitable neurons, particularly in slices from older mice in whch many neurons had already died (see Crimins et al., 2011Crimins et al., 2012).

Stern wanted to assess network function at an earlier point in time, when, at least in mice, the cognitive deficits start to appear but neurons are still healthy. At this stage, therapeutic interventions might be most effective, Stern told Alzforum. He also wanted to study live animals, which better model what happens in the intact brain. He chose rTg4510 mice, which express human tau containing the P301L mutation linked to familial frontotemporal dementia. The mice develop spatial memory problems and sparse tau tangles in the cortex by 2.5 months, but do not lose cortical neurons until 8.5 months, when the pathology is more pronounced.

First author Noa Menkes-Caspi inserted electrodes into the frontal poles of the cortices of 5-month-old animals, and recorded electrical activity in excitatory pyramidal neurons and local field potentials as the mice sat quietly or slept. The membrane potentials of these neurons normally oscillate from hyperpolarized to depolarized at regular intervals, particularly during slow-wave sleep when network activity is highly synchronized. In the transgenics, potentials oscillated more slowly than in controls, and the difference was most pronounced during slow-wave sleep. Transgenic neurons spent more time in the hyperpolarized, “down” state, and took longer to transition to the depolarized, “up” state in which action potentials can occur. Once in the up state, transgenic neurons fired fewer action potentials, with longer delays between them, than controls did. The net effect of these changes was an overall dampening of network activity in transgenic mice.

What might cause this? The authors found that transgenic neurons experienced more “false up” transitions, in which the membrane depolarized slightly but not enough to trigger the up state—likely because transgenic neurons received insufficient synchronized synaptic inputs from upstream neurons to depolarize on schedule, Stern said. Because the mice at this young age still appear to have a normal complement of spines and synapses, the problem likely arises from faulty timing of synaptic inputs, he said. Tau in afferent neurons might interfere with transmission of information along axons, changing the timing of these inputs, Stern speculated.

This study did not address what form of tau caused the disruptions. However, only a handful of neurons in 5-month-old transgenic mice had detectable tangles. Neurons in 3-month-olds had far fewer tangles but showed similar, although less pronounced, decreases in firing. Kuchibhotla noted that the data fit with the idea that smaller species of tau, such as oligomers, may be the toxic entity.

The finding of reduced activity in these tau mice contrasts with data from amyloid models, where neurons frequently become hyperexcitable (see Sep 2007 newsNov 2009 conference newsAug 2012 news). The full picture of Alzheimer’s disease is likely more complex, as a previous study found areas of both hypo- and hyperexcitability in the brains of AD model mice (see Sep 2008 news). In Alzheimer’s brains, both amyloid and tau pathology occur together, and tau tends to track more closely with neuronal activity as measured by FDG-PET (see Feb 2015 conference news). Stern will next investigate how network activity fluctuates in a mouse model that includes both pathologies.

Stern also wants to know whether restoring normal activity can ameliorate cognitive symptoms. Several drugs that block calcium or potassium channels raise or lower neuronal firing. Stern plans to test cocktails of these drugs in mouse models to see if he can rescue neuronal function.—Madolyn Bowman Rogers


  1. It is well known that the cause of some familial cases of frontotemporal dementia (FTD) is the presence of single mutations in the MAPT gene that codes for the tau protein. Mutated (pathological) tau initiates the disease but the mechanisms for the consecutive steps are not yet well defined.

    In this work by Stern’s group, the authors clearly show that mutated tau disrupts ongoing cortical network activity that may result in cognitive deficits. Looking at neocortical activity in an FTD mouse model (expressing human tau with an FTD-mutation), they found that the membrane potential oscillations (MPOs) derived from neurons were slower than MPOs derived from wild-type mice. This result, and other conclusions of the paper, clearly demonstrates that pathological tau could reduce the activity of single neocortical cells.

    The study reports an original observation, suggesting a novel function for tau. However, to extrapolate these results to the possible dysfunctions occurring in people with FTD, we should take into account the level of the mutated tau needed to promote the dysfunction. In other words, in the model we do not have only the effect of a toxic agent but, probably, the effect of a higher concentration of that toxic agent than that occurring in FTD. We do not know the level of toxic agent that is necessary to have that toxic effect. Indeed, it is not easy to compare human tau expression in the mouse model, with that occurring in specific neurons in the brain of a patient with dementia.

    View all comments by Jesus Avila
  2. In this elegant study, Menkes-Caspi et al. provide a comprehensive insight into the dynamics of neocortical networks and activity of single neocortical pyramidal neurons in the rTg4510 mouse model of tauopathy. Using intracellular and extracellular recordings in anesthetized and freely moving animals at two stages (3- to 3.5-month-old animals and 4.5- to 6-month-old animals), they showed that spontaneous, ongoing, neocortical activity was disrupted during various arousal states and that firing rates of transgenic neurons were significantly reduced.

    The main findings include slower membrane potential oscillations during slow-wave sleep and under anaesthesia in the transgenic mice. The authors postulate (based on their intracellular recordings in anesthetized animals) that these changes may occur due to longer "down" states and state transitions of membrane potentials. The nature of these findings is not completely clear. One important point to consider may be the temperature-related effects of anaesthesia that ultimately lead to increased phosphorylation of tau via a signalling cascade that is sensitive to the body temperature drop by only a few degrees (Planel et al., 2008). It is not clear how the increase in tau phosphorylation may translate into the patterns of neuronal firing and sub-threshold membrane potential dynamics during intracellular recordings in the present study, but it may render tau mice more susceptible to the effects. Thus, the article prompts more experiments on non-anesthetized animals to clarify this. In the present study, intracellular and extracellular recordings were performed in separate experiments. While this design does not undermine the conclusions, it would be insightful to study the unitary, cellular, and network activities during simultaneous recordings.

    The study also found that neocortical activity was globally reduced to a greater degree in older animals with higher levels of neurofibrillary tangles; however, the activity was also reduced in neurons without detectable levels of pathological tau. The authors propose that the effects of pathological tau on a fraction of neurons may further amplify and spread throughout the entire cortical network. This would imply that a neuron may receive the adverse effects indirectly from the other affected neurons and in the absence of pathological tau. This mechanism may have broader implications for other neurological diseases associated with misfolding and aggregation of toxic proteins. However, it remains to be determined whether the findings can be explained by soluble tau (elevated at this stage), or even pathological tau that falls below the detection threshold of commonly used techniques, as suggested by the authors and others  (Oh et al., 2010). In short, the paper prompts more studies to improve the detection methods routinely used in the field.

    Previous studies of the rTg4510 model have also reported that structural and functional changes in layer III frontal cortical pyramidal neurons were not associated with the presence of neurofibrillary tangles, even in mice 8.5- to 9.5-months of age (Rocher et al., 2010; Crimins et al., 2011). Rigorous in vitro studies have also demonstrated that increased excitability and higher action potential firing rates (attributed to depolarized resting membrane potential) were typical for neurons of tau mice, in contrast to the present in vivo findings (Rocher et al., 2010). This discrepancy between in vivo and in vitro results may be due to the inherent differences between preparation and recording techniques (as stated by the authors) or to age-related differences. Further in vivo studies are needed. In APP/PS1 mice, for example, intracellular recordings of hippocampal CA1 neurons at late-stage disease (10 to 14 months of age) reveal elevated firing rates and frequent occurrence of action potential bursts in vivo that are confirmed by recordings in slices (Šišková et al., 2014). In the rTg4510 model it would have been interesting to compare the outcomes of slice recordings reported by Menkes-Caspi et al. with previous studies of the rTg4510 model.

    The relationship between tau and another neurotoxic protein, Aβ, has been studied extensively by others in the field. One study has observed that reducing tau levels ameliorates spontaneous epileptiform activity and also the severity of spontaneous and induced seizures in several transgenic mouse lines overexpressing human amyloid precursor protein. Following tau reduction, inhibitory currents recorded in acute hippocampal slices were increased and the balance between excitation and inhibition was restored (Roberson et al., 2011). 

    More comprehensive studies using powerful experimental approaches, such as included in the current work, will be required to finally dissect out the mechanisms underlying neuronal and brain network abnormalities induced by the interactions of several toxic proteins in individuals with Alzheimer’s disease and other neurological conditions.


    . Homeostatic responses by surviving cortical pyramidal cells in neurodegenerative tauopathy. Acta Neuropathol. 2011 Nov;122(5):551-64. PubMed.

    . Staging of Alzheimer's pathology in triple transgenic mice: a light and electron microscopic analysis. Int J Alzheimers Dis. 2010;2010 PubMed.

    . Anesthesia-induced hyperphosphorylation detaches 3-repeat tau from microtubules without affecting their stability in vivo. J Neurosci. 2008 Nov 26;28(48):12798-807. PubMed.

    . Amyloid-β/Fyn-induced synaptic, network, and cognitive impairments depend on tau levels in multiple mouse models of Alzheimer's disease. J Neurosci. 2011 Jan 12;31(2):700-11. PubMed.

    . Structural and functional changes in tau mutant mice neurons are not linked to the presence of NFTs. Exp Neurol. 2010 Jun;223(2):385-93. PubMed.

    . Dendritic structural degeneration is functionally linked to cellular hyperexcitability in a mouse model of Alzheimer's disease. Neuron. 2014 Dec 3;84(5):1023-33. Epub 2014 Nov 13 PubMed.

    View all comments by Zuzana Siskova

Make a Comment

To make a comment you must login or register.


News Citations

  1. No Toxicity in Tau’s Tangles?
  2. Tau Toxicity—Tangle-free But Tied to Inflammation
  3. In Pursuit of Toxic Tau
  4. Do "Silent" Seizures Cause Network Dysfunction in AD?
  5. Chicago: AD and Epilepsy—Joined at the Synapse?
  6. Anticonvulsants Reverse AD-like Symptoms in Transgenic Mice
  7. Hyperactive Neurons and Amyloid, Side by Side
  8. What If It’s Not Garden-Variety AD? Telling Variants Apart by Where Tau Is

Research Models Citations

  1. rTg(tauP301L)4510

Mutations Citations

  1. MAPT P301L

Paper Citations

  1. . Homeostatic responses by surviving cortical pyramidal cells in neurodegenerative tauopathy. Acta Neuropathol. 2011 Nov;122(5):551-64. PubMed.
  2. . Electrophysiological changes precede morphological changes to frontal cortical pyramidal neurons in the rTg4510 mouse model of progressive tauopathy. Acta Neuropathol. 2012 Dec;124(6):777-95. PubMed.

Further Reading

Primary Papers

  1. . Pathological tau disrupts ongoing network activity. Neuron. 2015 Mar 4;85(5):959-66. Epub 2015 Feb 19 PubMed.