Dementia-Linked Tau Protein May Be Necessary For Long-Term Memory

Getting your Trinity Audio player ready...

In this preprint study submitted to Nature, researchers investigated the role of tau protein, best known for mis-folding and aggregating into fibrous tangles in neurodegenerative diseases. They report a previously unknown role of correctly folded tau protein in long-term memory. Their findings also suggest that in the earlier stages of dementia, some memories may not be permanently lost, but rather are inaccessible as a result of tau pathology.

A Quick Introduction to Tau

The accumulation of tau protein is a well established characteristic of a range of neurodegenerative diseases, including Alzheimer’s disease. Individual tau proteins are not a threat by themselves, but can undergo a modification called hyperphosphorylation that causes them to mis-fold and join together into filaments that then aggregate in to neurofibrillary tangles. The exact relationship between neurofibrillary tangles and diseases like Alzheimer’s is debated, but it is generally understood that they contribute to the progression of dementias to some extent.

Due to the focus on its role in dementia, many people may be unaware of the role of normal, correctly folded tau protein. Tau helps to stabilise microtubules – long tubular structures that form part of the cell’s ‘skeleton’ and act as a kind of ‘rail’ along which proteins can be transported within the cell. This is particularly important for neurons, since proteins produced in the body of the cell (where the DNA resides) must be transported all the way along the nerve fibre. One might therefore be forgiven for thinking that tau protein was quite important, and that attempts to eliminate tau protein before it formed into tangles could have some quite negative effects on the brain. However, studies have suggested that tau-deficient mice are protected against Alzheimer’s without suffering any significant drawbacks in terms of important neurological functions like memory. This suggested that the function of tau might be mostly redundant, and that other microtuble-stabilising proteins were sufficient to maintain microtubles in the event that tau was depleted. This would mean that targeting tau before it formed into tangles, which become significantly harder to remove, could be a viable strategy to prevent some neurodegenerative diseases.

The issue with previous studies is that they had primarily been looking at memory retrieval over timespans of hours to days. In the present study, researchers hypothesised that formation of longer-term memories, which are stored differently, might be more dependent on tau protein.

The Long And The Short Of It

First, the researchers took mice that had been genetically engineered to completely lack tau protein, and compared their recent and remote (long-term) memory with that of control mice. To do this, they used various memory tests where mice had a learning phase (such as learning to navigate a maze) and their recall ability was tested shortly after (within 24 hours) or after a much longer time gap (weeks to months). In agreement with previous studies, mice lacking tau fared no worse than control mice when recalling recent information, but in the remote recall experiments they fared significantly worse, suggesting that tau deficiency did in fact harm long-term memory.

By genetically engineering mice so that tau protein production could be ‘switched back on’ for different periods of the learning process, researchers were able to demonstrate that tau was only important during the ‘encoding window’ – that is to say, the moment in which the memory is created. So long as tau was produced during this window, it could be suppressed during the weeks that followed and long-term memory would still be intact. They were also able to identify a specific site within the tau protein (named T205) that needed to be phosphorylated in order for memory to be preserved. Phosphorylation is a common way in which proteins can be modified, causing their shape and function to change. The researchers showed that when tau was genetically engineered so that T205 could not be phosphorylated, mice that produced this modified tau had similar long-term memory deficits to mice lacking tau entirely.

A Question Of Precision

Next, researchers set out to establish how tau was enabling the establishment of long-term memories. Memories are encoded by a specific set of neurons called an engram. When those neurons – and only those neurons – are activated at once, the information is recalled. However, if too many neurons outside of the engram also fire, recall fails, so precision is key. Using some clever staining and fluorescent tagging techniques, researchers were able to tag engram neurons during memory encoding so that they could then observe how precisely these neurons were being activated later on. They found that if mice lacked tau or lacked T205 during the encoding window, many cells outside of the engram would activate during long-term recall. As previously, restoring normal tau to the mice only during the encoding window was enough to make recall normal. This suggested that tau was necessary during the encoding for ensuring precise reactivation of the correct neurons later down the line.

Researchers then asked a very interesting question: if the memory deficits in tau-deficient mice were caused by the activation of extra neurons outside of the engram, did this mean that the original memory was still intact and retrievable? To test this, researchers introduced targeted light-sensitive ion channels into engram cells during memory formation. When exposed to light, these channels would open, allowing electrically charged ions to flood into the cells and activating them. This meant that by delivering light to these cells via a wirelessly controlled brain implant, researchers could specifically activate only the engram cells. They found that when they did this, long-term recall in mice lacking tau was restored. This showed that without tau, mice were still technically able to form long-term memories, as the engrams for them were still there. However, if tau is not present during encoding, they struggle to activate those engrams with the precision necessary for recall to occur.

The Implications

With the disclaimer that this study was conducted in mice and is the first of its kind (and so may need to be replicated), this research has two important implications. The first is that normal, correctly folded tau may be far more important for memory than was previously thought, which means that therapies aimed at reducing tau levels in the brain so that it doesn’t form neurofibrillary tangles will need to be approached with caution.

The more positive implication of this study is that memories can be inaccessible, but not lost, as a result of tau deficiency. While humans with tau pathology are not tau-deficient in the way that these mice were, they do have some deficiency in functional, correctly folded tau, as this normal tau is being phosyphorylated and diverted into neurofibrillary tangles at an accelerated rate. This means that there is a possibility that, in the earlier stages of neurodegenerative disease before large numbers of neurons have died, some of the memory problems may be a consequence of the mechanism described here. The physical foundations for new memories are potentially still being formed and could theoretically be retrieved.