Treating brain-damaged mice with compounds that affect gene expression restores their ability to recall long-term memories, according to a study in this week’s Nature. Raising the mice in a stimulating environment has the same effect. The results suggest that memories, once consolidated, can remain accessible even after significant loss of brain cells. They also open up the possibility of developing drugs to treat the memory loss associated with conditions such as Alzheimer’s disease and dementia.

Previous research had produced results similar to the study’s finding that environmental enrichment can improve learning. “That’s no big deal,” says coauthor Li-Huei Tsai of MIT’s Picower Institute for Learning and Memory, “but in terms of recovery of long-term memory … we were all stunned.”

Tsai and her colleagues conducted the study on mice genetically engineered to express a protein called p25 under certain conditions. The protein triggers massive brain-cell death and has been implicated in neurodegenerative diseases. Researchers can switch the expression of p25 on and off by controlling the mice’s diet, inducing brain damage at will. Without the special diet, the mice behave like normal mice.

To test the mice’s memory, Tsai and her colleagues started by conditioning them to be afraid of a certain place. The mice were moved from their cages to a chamber providing an interesting new environment they want to explore, she explains. But in that chamber, the mice received a mild shock on their feet. “It doesn’t hurt,” Tsai says, “but they hate it.”

This fear conditioning gets coded in the hippocampus area of the brain before being transferred three to four weeks later to the cortex, where it becomes a stable, long-term memory. After that, if the mice are returned to the chamber they remember the bad experience and freeze in place instead of exploring.

After establishing this memory in the genetically engineered mice, Tsai and her colleagues induced brain damage by turning on the p25 gene. As expected, the mice lost their fear of the chamber, failing to freeze as normal mice would after the same conditioning.

But giving the brain-damaged mice a compound called a histone deacetylase (HDAC) inhibitor produced very different results. Histones are proteins that DNA strands wrap themselves around, forming a structure called chromatin. The way in which chromatin assembles itself affects gene regulation and expression. Mice given an HDAC inhibitor, which allow the DNA to unwind from the histones making the DNA accessible to transcription, were able to recover long-term memories much better than untreated mice.

This improved recall also occurred when brain-damaged mice not treated with the HDAC inhibitor were placed in an enriching environment. According to Tsai, the mice are normally kept in a small cage with not much more than food and water. Moving them to a larger cage with lots of toys, a running wheel, and other mice to interact with allows them to be “much more active, physically and mentally,” she says. Mice in this stimulating environment exhibited much better freezing behavior, showing that they were able to recover their long-term memories.

Both the HDAC inhibitor and the enriching environment probably stimulate the growth of connections between neurons, which rewire the brain in such a way as to make long-term memories more accessible, Tsai says. “You don’t necessarily see an increased number of neurons,” she explains, “but you do see an increase in the formation of dendrites and synapses.” In the case of the HDAC inhibitors, it might be that changing the structure of the chromatin causes genes that affect this synaptic growth to be expressed more.

The results are “very impressive,” says Ya-Ping Tang, a neurobiologist at the University of Chicago. It shows how epigenetics–altered gene expression that is not linked to changes in the DNA itself–is involved in learning and memory, he says.

It’s not clear why initiating brain damage in the mice doesn’t destroy the long-term memories altogether. “Our study can’t speak to that,” Tsai says, “but it shows that even with this substantial neuronal loss, it’s not enough to get rid of memory.”

Tang says that it’s too early to say if drugs based on this mechanism could be developed to help restore memory loss in humans. Tsai and her group are now trying other HDAC inhibitors in mice to see which ones function best. “It will be very interesting to see if HDAC inhibitors will help in humans,” she says. “It really provides hope for people with neuronal loss and dementia that maybe something can be done.”