A better way to study Alzheimer disease in mice

Ooh Ooh Ooh! They Finally Did It!

I am excited that my inaugural post can be about a manuscript wherein experiments that I was hoping would be performed for what felt like an eternity have finally been performed. This article reports on a highly creative way to bridge the gap in between research on Alzheimer disease populations and corresponding mouse models-which in my opinion provides useful method by which to move this type of translational research forward.

The manuscript discussed in this post is “Rigid firing sequences undermine spatial memory codes in a neurodegenerative mouse model” by Jingheng Cheng and Daoyun Ji from Baylor College of Medicine. It was published in eLife 2013:2e00647.

Special thanks to Kate Jeffery for bringing this article to my attention!

What is Being Studied and Why?

Defining appropriate outcome measures has long been a problem for those of us who study mutant and transgenic mouse models of neurodevelopmental and neurodegenerative disease. As a specific example, it has required something of a Herculean effort to develop behavioral outcomes in mice that are intuitively comparable the results of clinical/cognitive studies in human patient populations.

Up to this point, the way around this problem has been to apply creative interpretations to behavioral data to bridge research into human disease and animal models. For example, the performance of a mouse on the Morris water maze (Link) or fear conditioning (Link) paradigm is interpreted as analogous to spatial disorientation or episodic memory decline in patient populations. Although not optimal, this type of creative interpretation has to date been necessary to achieve the goals of this type of translational research-that of developing valid behavioral models for human disease in animals that are useful to evaluate candidate therapeutical interventions.

Cheng and Ji discuss the difficulty in using the water maze as the primary outcome measure. A recent report demonstrates that mouse models of Alzheimer disease show profound deficits in locating a hidden platform when placed in a pool of water at different starting positions. Interestingly, female mice seem to show more profound deficits (Link). Unfortunately, there was no clear explanation for this sex difference in behavioral outcomes.

In contrast with the behavioral data, Cheng and Ji demonstrated that female mice actually show less profound deficits as related to brain activity. Their interpretation of this discrepancy with the water maze data illustrates quite clearly one of the often unmentioned confounds in the water maze: anxiety. When mice are placed in a pool of water at 19-24ºC, stress hormone levels elevate profoundly; which has been shown repeatedly to disrupt the ability to learn. The authors suggested that it is likely that the female mice demonstrated elevated anxiety relative to males, and thus performed poorly, whereas the mice in the neural recording experiment had been handled daily for over a month, so the female mice did not show elevated anxiety.

An Innovative Approach

As a step toward overcoming this inherent weakness in the translation between data from human patients and corresponding animal models, Cheng and Ji presented on an entirely novel solution. They posited a way to simultaneously model two different aspects of Alzheimer disease: namely 1) the precipitous cognitive decline/spatial disorientation and 2) the neurodegeneration accompanied by progressive neuropathology across life (e.g., plaques and tangles) (Link). Their solution to bridge the gap between these outcome measures was to apply a method used primarily in systems neuroscience: that if recording location specific firing in the hippocampus (i.e., place cells) in mutant mice modeling Alzheimer disease.

Specifically, they proposed that it was not only important to look at the spatially selective activity of cells in the CA1 subregion of the hippocampus, but also to characterize the sequential activity across collections of cells. Cheng and Ji proposed these sequences of firing could be remembered by the mouse in two ways: 1) recalled in response to previously learned external spatial cues or 2) recalled in response to intrinsic inputs that are independent to or at least not correlated with the external spatial cues (i.e., recalled from remembered experiences rather than current sensory experience). They hypothesized that the mutant mice would remember sequences using intrinsic firing patterns and not due to direct sensory experience, the opposite from nonmutant mice. Their hypothesis was based on reports that Alzheimer disease results in false memory recall of learned information, such that patients often recall the wrong experience when provided with a recall cue to jog their memories.

Take Home Message

The take home message of this experiment is that the cells of the hippocampus in mutant mice modeling Alzheimer disease fire in sequences that at first glance appear completely normal; however, upon closer scrutiny, these sequences are not anchored to the spatial cues in the environment. This is in contrast with nonmutant mice whose sequences were strongly anchored to environmental cues. Additionally, unlike the nonmutant mice, these same sequences appeared in both novel and familiar spaces in open fields-suggesting that internal inputs may be predominant over external inputs in the mutant mice. This abnormal, erroneous sequential activity may interfere with recall of previous experience, as well as catastrophically interfere with new learning. In other words, the mutant mice are remembering the sequences and replaying them at inappropriate times when compared with nonmutant mice.


One can interpret these findings in light of cognitive decline in Alzheimer disease by the fact that patients often show evidence that older memories tend to interfere with newer memories when asked to remember specific events. However, as the authors volunteer, little can be definitively concluded about any mouse-human homology until similar studies using functional MRI (fMRI) or intracranial EEG (iEEG) would have to be performed to know for certain if Alzheimer disease patients show abnormal sequential activity patterns as those observed in the mutant mice.

An important aspect of the mouse experiment was the finding that these sequences, although not as well anchored to spatial locations as those in nonmutant mice, still seemed to be initiated or activated by space in some poorly understood manner. This is important since it allowed the researchers to propose a simple model by which the buildup of plaques and tangles observed in the mouse model resulted in erroneous sequential firing patterns. Cheng and Ji proposed that perhaps when external space is processed by a mutant mouse (and by extension Alzheimer disease patients), old activity patterns are activated in CA3, an area upstream to where they were recording in CA1. These old activity patterns from CA3 then proceed to dominate CA1 firing-thus resulting in sequences of activity that can be inappropriately repeated in response to any spatial cue, thus interfering with normal learning and recall processes.


Personally, I was excited when I read this manuscript because I had been wanting to do this exact type of experiment in my studies in mutant mouse models. I just always lacked the time, funding, equipment, and skill sets necessary to do the experiment-in other words, it was a sort of vague dream of mine and nothing more.

One of the most impressive aspects to this manuscript was the fact that the authors went to the effort to characterize the abnormal firing patterns in the mutant mice. It would have been much less work for the authors to simply state that there was abnormal firing patterns and declared it an index of hippocampus dysfunction. Instead, the authors provided a thorough characterization of these abnormalities; and, more importantly, proposed a clear, hypothesis driven mechanism that can actually be tested in the patient population using functional imaging techniques. In my opinion, Cheng and Ji have provided compelling evidence that place cell recording may provide meaningful data that can be intuitively related back to the human patient population, particularly to results of studies using fMRI, iEEG, or other neurophysiological methods.


I would love to hear your thoughts on this!

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s