Researchers have discovered that star-shaped brain cells called astrocytes help determine whether long-term memory is retained, rather than simply supporting the neurons that store it. In mice, removing one protein from these cells did not disrupt memories formed one day earlier, while memories from two weeks ago quietly faded away.
This finding separates two processes that the brain was long thought to handle together: memory formation and its preservation over time. It also gives scientists a new target for studying memory loss seen in aging and disease—a target located outside neurons.
The research team is led by Uh-Hyun Ko from the Institute for Basic Science (IBS) in Daejeon, South Korea. The study results were published in the journal Nature Communications.
For decades, memory research focused on neurons—cells that transmit and process signals. Astrocytes were viewed as a passive scaffold. To test whether astrocytes perform any other functions, the research team bred mice lacking the protein ankyrin-2, or Ank2, specifically in these cells. Ank2 is found in astrocytes at concentrations as high as in neurons, yet no one had questioned its role there.
Initially, nothing suspicious was observed. The modified mice moved and interacted with their cage mates normally. However, one day after a mild shock in an experimental chamber, they froze upon returning to it—a typical sign that a new memory had formed.
Two weeks later, the situation changed. Mice lacking Ank2 in astrocytes showed far fewer freezing behaviors when returned to the same chamber compared to normal animals, as if the memory had weakened. Recent memory was intact. The issue lay in the brain’s ability to preserve memories over time.
Behind the fading memories was a physical change. When researchers examined astrocytes lacking Ank2, the cells had retracted inward. Their branches were shorter and fewer, and the overall volume was reduced, making them less able to reach neuronal connection points.
This retraction was most evident in one specific set of connections. After learning, astrocytes typically touch engram neurons—cells that store a particular memory. These new contacts help sustain the memory. In mice with Ank2 deficiency, far fewer such contacts formed.
The neural circuits themselves also struggled to maintain their strength. When neurons are repeatedly activated simultaneously, their connections strengthen and persist, a stable change considered the cellular basis of memory. In genetically modified mice, this strengthening initially appeared but then vanished instead of lasting for several hours.
The idea that astrocytes actively participate in memory storage is gaining traction. A recent study described how groups of these cells form a stable, multi-day trace that helps support memory, suggesting they carry their own record of it. The new work adds a missing piece by naming the protein underpinning the entire process.
Ank2 functions as an anchor. Inside an astrocyte, it holds another protein in place that releases calcium when the cell receives a specific growth signal. This signal comes from BDNF, a molecule the brain produces after learning to promote neural connection formation.
Without Ank2, this chain broke down. Calcium activity inside astrocytes decreased, and the cells stopped growing the thin extensions they typically develop. Their connection to nearby junctions weakened. The growth signal arrived, but the astrocyte could no longer respond to it.
The team observed the effect by adding the growth signal directly: an injection of BDNF into the hippocampus—the brain region most linked to this type of memory—about 12 hours after learning typically boosts memory persistence. In mice with astrocytic Ank2 deficiency, the same injection had no effect.
None of this depended on the basic electrical processes in astrocytes, which remained normal. The issue was solely in the mechanism allowing these cells to respond and rebuild. It is the same delicate molecular control that, according to other studies, determines how long a memory lasts.
Removing a protein reveals what a cell cannot do without it. To find out what astrocytes can do when the process is reversed, the team created a light-controlled tool they named Opto-T1.
When blue light hits astrocytes, it activates the growth signaling pathway. Twelve hours after a mild training session, researchers shone light on the animals in their cages. The astrocytes expanded, and two weeks later, the mice recalled the event far better than unexposed animals. Recent memory remained unaffected. Increasing astrocyte activity alone was sufficient to preserve the memory.
The study results rewrite the understanding of a cell long thought to perform only a maintenance role. “Our data show that astrocytes are not passive support cells but active regulators determining how long memories are retained,” said Ko.
The issue extends beyond the lab. The Ank2 gene has already been implicated in autism, intellectual disability, and epilepsy. Weakened astrocytes may also contribute to the memory loss that occurs with age and disease.
Other studies examining neuron-astrocyte interactions for long-term memory retention follow the same path, targeting the partner cell long considered background.
What was once vague has now become concrete. Memory is not simply stored by neurons and left alone; astrocytes continue working long after learning to keep it stable, and Ank2 is the protein that enables them to do so.
This opens the possibility of investigating memory impairments outside neurons and provides a method for intervention using light. The next question is whether the same cells can be made to protect memory not only in mice but also in humans.
About the Author
