Sleep deprivation–induced amnesia is a kind of retrograde amnesia where exhaustion causes us to forget information—a phenomenon those who have pulled an all-nighter before an important exam are all too familiar with. But the memories themselves aren’t truly lost, and a drug that’s already approved for use in people can bring them back, a mouse study published December 27 in Current Biology demonstrates—a finding which the study’s authors say may apply to other forms of memory loss.
Memories: Forgotten but not lost
Sleep deprivation is a ubiquitous problem in society today. Robert Havekes, a neuroscientist at the University of Groningen in the Netherlands, says that everybody knows things go wrong in the brain with sleep deprivation, but he wanted to investigate whether it’s possible to assist the brain in memory recall.
Prior research had suggested that the memories lost in other forms of retrograde amnesia do not vanish from the brain, but rather are suboptimally stored and, therefore difficult to access. For instance, back in 2015, molecular neuroscientist Tomás Ryan—then a postdoctoral fellow in Susumu Tonegawa’s lab at MIT—found that the protein synthesis inhibitor anisomycin could pharmacologically hamper memory consolidation by disrupting the formation of new synaptic connections in the hippocampus, which are crucial for later recalling a memory naturally. But the scientists were able to later retrieve the memories using laser stimulation of the relevant ensemble of neurons—known as a memory engram—that were activated when storing those memories. Similarly, researchers have found that they can stimulate memory recollection optogenetically in mouse models of infantile amnesia and Alzheimer’s. “I think that it’s really important that we haven’t seen a single case so far where the engram has actually been degraded,” says Ryan, who was not involved in Havekes’s study.
See “Electrically Zapping Specific Brain Regions Can Boost Memory”
To confirm that memories lost to sleep deprivation could also be recalled, Havekes and colleagues took advantage of engineered mice that are designed for labeling engrams. When these mice are fed the antibiotic doxycycline, it binds to and blocks a protein that allows neurons to express a light-sensitive protein in their membranes which responds to the lasers used for optogenetic activation. So when researchers want to label neurons involved in a specific learning event, they simply take the mice off their doxycycline diet before training, so that the neurons activated during training will express the ion channel protein in question, channelrhodopsin. “So not only we can visualize which neurons become active during a particular learning episode, those neurons also become light sensitive,” explains Havekes. “Later, by shining light on them, we can make the mice sort of think about that particular learning episode.”
Havekes had these mice perform an object location task: First, the animals were placed in an arena containing different objects to explore—this is the training phase. Then, while the mice were out of the arena (the memory consolidation phase), the researchers moved one of the objects to a novel location.
Mice are inherently curious animals, explains Ryan, so they normally investigate the moved object for longer when they’re placed back into the arena, indicating that they successfully detect the change of scenery by comparing it with their memory from training (memory retrieval).
The deprived mice do not detect the spatial change by themselves, but if we help them, they suddenly do remember.
—Robert Havekes, University of Groningen
In the experiment, some of the mice were allowed to rest during the memory consolidation phase between training and memory retrieval. But others were manually monitored by people to make sure they were sleep-deprived. “If the mice start to get sleepy, we’d gently tap on the cage, or move the cage a bit,” Havekes explains. “But we do it very gently as we do not want to stress out the animals, and we also know from previous studies that this procedure doesn’t induce any stress.”
The lack of sleep did impact their ability to remember, though: The team found that sleep-deprived mice explored all objects to an equal extent during the memory retrieval phase, which the researchers took to mean that the mice had no recollection of the objects’ previous locations. However, when these mice received optogenetic stimulation five minutes before being put in the arena, they were able to successfully detect spatial novelty. And follow up experiments demonstrated that it was the activation of the memory engram specifically—and not other neurons—that allowed the mice to recall their lost spatial memory. “So, the deprived mice do not detect the spatial change by themselves, but if we help them, they suddenly do remember,” explains Havekes.
A drug to remember?
Optogenetic stimulation worked, but people aren’t walking around with optogenetic setups surgically installed on their heads, so Havekes sought a more human-relevant way to activate those engrams. He turned to roflumilast, a clinically approved asthma drug that has been shown to boost formation of synaptic connections in the mouse brain—connections that get disrupted by sleep deprivation.
The drug reactivated the object location memories in mice as the lasers had—a hint it could have clinical relevance in treating other kinds of memory retrieval difficulties where engrams are preserved. “Roflumilast is a drug that’s approved for use in humans . . . [and] is safe,” Havekes says, “making it interesting from a translational point of view.”
However, when given the drug or lasers alone, the mice were not able to recall the memories a couple of days later. Havekes went on to show that by administering a combination of both techniques, the mice were able to retain the memories long term, as they did not require any kind of stimulation before a test two days later. That’s because the mice essentially relived the training experience when the lasers reactivated their memory engram—but this time, a drug made sure their neurons had the plasticity needed to properly store that memory. “It was sort of an artificial training,” Havekes says.
Ryan says that because this adds to the studies firmly establishing that memory engrams remain preserved, scientists can now focus their efforts on figuring out how to successfully retrieve them. “I’ll be very skeptical that these drugs will help in human cases of memory loss, but the principle is encouraging,” he says. “When I think about translational aspect to humans, [the study’s] value is more about understanding human memory in general than in treating any specific form of memory loss.”
See “How Manipulating Rodent Memories Can Elucidate Neurological Function”
He notes that human brains are much more complicated environments than mouse brains, and it is not simple to separate the target engram from other engrams. Moreover, drug delivery isn’t always very efficacious. These limitations indicate there’s a long road ahead when it comes to developing memory loss treatments for people.
Havekes agrees that there’s a lot that needs to be understood before roflumilast or any drug could be used to treat human amnesias of any kind. “You get one answer, and you get five new questions,” he says.