Changes in neural network activity are a feature of Alzheimer's disease (AD). Epilepsy-like seizures—though they may be hard to detect—turn up in some people with AD and in transgenic mouse models of the disease. A study in the April 27 Cell blames faulty neuronal brakes that poorly suppress excitatory networks for aberrant activity such as epileptic spikes and altered oscillatory rhythms. Researchers led by Jorge Palop, Gladstone Institute of Neurological Disease in San Francisco, California, found that reduction of a specific sodium channel subunit jeopardized inhibitory interneurons in a transgenic mouse model of AD that has spontaneous silent seizures. The authors also found a dearth of the subunit, called Nav1.1, in some AD patient brains. Conversely, bumping up Nav1.1 in the mice improved their network function and rescued cognitive deficits.

"This is an important paper because it identifies one molecular explanation for the brake failure,” noted Jeffrey Noebels, Baylor College of Medicine, Houston, Texas, in an e-mail to Alzforum. Nav1.1 is encoded by the SCN1A gene, and mutations in it are responsible for epilepsies of varying severity (see Catterall et al., 2010). "Here the authors show that, through unknown mechanisms, excess Aβ peptide can result in a similar loss of sodium channel function in the same type of interneuron," wrote Noebels.

In 2007, Palop, with Lennart Mucke, also at Gladstone, and colleagues showed silent seizures in human amyloid precursor protein (APP)-expressing J20 mice. These animals lay down amyloid plaques accompanied by gliosis, exhibit signs of synaptic alteration, perform poorly in learning and memory tasks, and behave abnormally. The mice do not convulse overtly, but they freeze in their cages, and electroencephalography (EEG) recordings detected wild fluctuations in neural activity during those times (see ARF related news story on Palop et al., 2007). In the current paper, the research group asked when the aberrant neural activity happens and what causes it.

First author Laure Verret looked for clues in spikes of neural activity detectable by EEG. These are small epileptic discharges that are less intense than seizures but more frequent and easily quantified. In hAPPJ20 mice, spikes clustered during troughs in γ activity, that is, coordinated oscillations between 20 and 80 Hz. In contrast, during periods of intense γ waves, spiking stopped. Since γ activity is generated by parvalbumin (PV)-expressing inhibitory neurons, the researchers examined these cells more closely. Patch-clamp recordings showed that PV interneurons from J20 mice were more depolarized and produced action potentials of lower amplitude than those from control mice.

Since action potentials depend on voltage-gated sodium channels, the researchers looked to the core alpha subunits of four main sodium channels expressed in the central nervous system. They found that in the parietal cortex, the hAPP-expressing mice made less Nav1.1 mRNA and protein. Nav1.1 was similarly reduced in postmortem parietal cortices of 22 people who had suffered from AD, though it is unknown if any of them had seizure-like activity. The authors further found that the voltage-gated sodium channel blockers riluzole and phenytoin suppressed γ oscillations, led to more epileptiform activity, and worsened context-dependent memory in J20 mice. These blockers also bring on more seizures in human epilepsies resulting from loss-of-function mutations in Nav1.1 (see Liao et al., 2010). Riluzole is an ALS drug and phenytoin an anticonvulsant—their actions depend on blocking sodium channels in different cellular contexts. On the other hand, crossing hAPP mice to Nav1.1-overexpressing mice rescued γ activity, spiking abnormalities and cognition, and made the double-crosses long-lived.

"This paper identifies one of the possible mechanisms of this overexcitation—the reduction of neuron activity in inhibitory cells," Palop said. But the Nav1.1 channel subunit could be but one potential reason for network abnormalities, he said. "I'm sure more mechanisms will be discovered."

The larger point is that PV cells and γ activity may be therapeutic targets for AD, Palop told Alzforum. Behavioral interventions or drugs that enhance γ activity or reduce network hyperactivity could benefit cognitive function in the face of elevated Aβ, while γ-reducing drugs could actually worsen cognitive function in patients with AD.

Heikki Tanila, University of Eastern Finland, Kuopio, considered the evidence of sodium channel deficits in this model convincing. Tanila was not involved in this work. Researchers need to see if sodium channel deficits turn up in other transgenic animals, as well, he said. "Talking with colleagues, it looks like most of the APP transgenic mouse lines have epilepsy," he said, although he knows of only two that have been extensively characterized. Aside from Palop's hAPPJ20 model, Tanila's group has examined APPswe/PS1dE9 mice. These do have seizures, but unlike the hAPPJ20 line, they show increased γ activity, rather than reduced, and phenytoin treatment suppressed spikes instead of promoting them (see ARF related news story on Minkeviciene et al., 2009). "It's likely that we have more than one mechanism behind seizures in APP transgenic mice," Tanila said. He is also curious to know if reduced Nav1.1 levels are present in hAPPJ20 mice before amyloid-β starts to accumulate in the brain.

The findings not only propose future targets for AD therapeutics, but further tie AD to epilepsy, said Helen Scharfman, Nathan Kline Institute for Psychiatric Research, Orangeburg, New York, who was not part of this study. "The data would suggest there is a lot more than plaques and tangles" to AD, she added. She suggested that comparing the onset of seizure activity with that of plaque accumulation could help scientists figure out how abnormal activity and plaques affect one another. Additionally, the effects of sodium channel restoration on Aβ accumulation should be examined next, she said.

It is unclear why Nav1.1 would fall in transgenic mouse models or human patients in the first place. Here, previous work by Dora Kovacs, Massachusetts General Hospital, Boston, could be relevant. Kovacs showed that overexpression of the APP-cleaving enzyme BACE1 leads to abnormal Nav1.1 processing and compromises the channel’s insertion into neuronal membranes of adult hippocampal neurons. A lack of these channels could dampen excitability and enhance seizure activity (see ARF related news story).—Gwyneth Dickey Zakaib


  1. The incidence of epilepsy is dramatically elevated in humans with early onset Alzheimer’s disease, in particular, those bearing mutations in FAD genes (APP, PS1, PS2, trisomy 21) as well as almost every genetically engineered mouse model that overexpresses Aβ peptide in the brain. However, clinicians who work closely with dementia patients are reluctant to consider the two disorders as related, since behavioral convulsions are brief and infrequent, and the steady cognitive decline of AD is unrelenting and accompanied by brain atrophy and cell death.

    However, the presence of electrical rhythm disturbances, particularly when they occur in memory-related hippocampal networks, are not obvious; a patient may display only momentary confusion and amnesia. These seizures are also a cause of cell death and atrophy. And any neurological disorder that features neuronal degeneration is susceptible to brain rhythm disturbances at some stage, particularly once inhibitory synapses lose their ability to brake neuronal firing patterns. During this period when the brakes are failing, traffic is destabilized, and a true epileptic seizure may occur. Thus, the seizure itself is simply a warning light, telling us that underlying brain rhythms and information flow are abnormal.

    This is an important paper because it identifies one molecular explanation for the brake failure, namely, a loss of Scn1a sodium channel function in inhibitory interneurons. These channels provide the firing power for interneurons to suppress excess excitability. This defect has been previously discovered in children who inherit a defective Scn1a gene and display a spectrum of epilepsy from mild to severe. Here, the authors show that, through unknown mechanisms, excess Aβ peptide can also result in a similar loss of sodium channel function in these same interneurons, thus strengthening the functional link between the two disorders. Restoring the sodium channels to their normal levels by genetically inserting extra copies of the gene allowed recovery of more normal brain rhythms.

    There is still great uncertainty, however, as to how closely mouse models of Aβ overexpression mimic the human disorder. Many of these models lack features that may be critical to both the mechanism and therefore the correct treatment to reverse human dementia. For example, neuronal cell loss is minimal in many mouse models, allowing for more dramatic recovery with experimental therapies. And the right antiepileptic drug for Aβ toxicity has not yet been discovered.

    However, if recognized and treated correctly, reversing brain rhythm disturbances and epilepsy early in the course of cognitive decline may prove to be a powerful way to preserve memory in Alzheimer’s disease patients. I won’t be surprised if this finding stimulates research in Alzheimer’s disease to move even more aggressively to identify a drug that will strengthen the function of inhibitory interneurons.

    View all comments by Jeffrey L. Noebels
  2. I think this work is really interesting and reinforces the hypothesis that synchrony mediated by inhibition is crucial for brain computation, and these synchrony mechanisms are altered in familial or sporadic forms of AD.

    View all comments by Vincent Villette
  3. In my opinion, one of the important contributions of this study is that it proposes a novel molecular mechanism underlying AD pathogenesis in addition to Aβ synaptotoxicity on excitatory neurons. Dr. Palop’s team nicely showed that parvalbumin-positive GABAergic neurons (PV neurons) were not fully functional in hAPPJ20 AD mice. They found mRNA and protein levels of Nav1.1 voltage-gated sodium channels were significantly decreased in these neurons, especially in the parietal cortical region. Importantly, restoration of Nav1.1 levels not only rescued the inhibitory neuronal deficits, but also reduced memory deficits and premature mortality in these mice without significantly affecting Aβ peptide levels. Aβ accumulation may directly or indirectly regulate Nav1.1 transcription in PV neurons.

    Previously, we found that Nav1.1 mRNA and protein levels were also regulated by the intracellular domain of the neuronal voltage-gated sodium channel β2 subunit (Navβ2), generated by BACE1 and presenilin/γ-secretase cleavages of Navβ2 (Kim et al., 2007; Kovacs et al., 2010; Kim et al., 2011). The human APPswe/Ind transgene overproduced in hAPPJ20 mice, and endogenous Navβ2 may compete for the BACE1/γ-secretase-mediated cleavages in certain subsets of neurons, potentially contributing to the Nav1.1 decrease in these mice. Therefore, it will be interesting to explore Navβ subunit processing in PV neurons in this system. Of course, many issues need to be addressed before any of this might result in a clinical trial, including the availability of specific pharmacological modulators of Nav1.1 and/or PV neurons, additional proof-of-concept experiments with different AD models, the potential contribution of other voltage-gated ion channels to the PV neuron deficits such as Nav1.6 shown in this study, and further exploration of the underlying molecular mechanisms. However, this exciting study suggests that enhancing Nav1.1 activity/PV neuron function may reduce Alzheimer’s disease-related neuronal deficits in AD patients.


    . BACE1 regulates voltage-gated sodium channels and neuronal activity. Nat Cell Biol. 2007 Jul;9(7):755-64. PubMed.

    . Reduced sodium channel Na(v)1.1 levels in BACE1-null mice. J Biol Chem. 2011 Mar 11;286(10):8106-16. PubMed.

    . Alzheimer's secretases regulate voltage-gated sodium channels. Neurosci Lett. 2010 Dec 10;486(2):68-72. PubMed.

    . Aberrant excitatory neuronal activity and compensatory remodeling of inhibitory hippocampal circuits in mouse models of Alzheimer's disease. Neuron. 2007 Sep 6;55(5):697-711. PubMed.

    View all comments by Doo Yeon Kim

Make a Comment

To make a comment you must login or register.


News Citations

  1. Do "Silent" Seizures Cause Network Dysfunction in AD?
  2. The Many Misdeeds of Aβ—Seizures and Axonal Transport Interference
  3. BACE Regulates Sodium Channels, Neuronal Excitability

Paper Citations

  1. . NaV1.1 channels and epilepsy. J Physiol. 2010 Jun 1;588(Pt 11):1849-59. PubMed.
  2. . Aberrant excitatory neuronal activity and compensatory remodeling of inhibitory hippocampal circuits in mouse models of Alzheimer's disease. Neuron. 2007 Sep 6;55(5):697-711. PubMed.
  3. . Partial epilepsy with antecedent febrile seizures and seizure aggravation by antiepileptic drugs: associated with loss of function of Na(v) 1.1. Epilepsia. 2010 Sep;51(9):1669-78. PubMed.
  4. . Amyloid beta-induced neuronal hyperexcitability triggers progressive epilepsy. J Neurosci. 2009 Mar 18;29(11):3453-62. PubMed.

External Citations

  1. APPswe/PS1dE9 mice

Further Reading


  1. . Staged decline of neuronal function in vivo in an animal model of Alzheimer's disease. Nat Commun. 2012;3:774. PubMed.
  2. . Synaptic depression and aberrant excitatory network activity in Alzheimer's disease: two faces of the same coin?. Neuromolecular Med. 2010 Mar;12(1):48-55. PubMed.

Primary Papers

  1. . Inhibitory interneuron deficit links altered network activity and cognitive dysfunction in Alzheimer model. Cell. 2012 Apr 27;149(3):708-21. PubMed.