Star-shaped cells called astrocytes help the brain establish long-lasting memories, Salk researchers have discovered.
The new work adds to a growing body of evidence that astrocytes, long considered to be merely supportive cells in the brain, may have more of a leading role.
The study, published in the journal Glia on July 26, 2019, could inform therapies for disorders in which long-term memory is impaired, such as traumatic brain injury or dementia.
“This is an indication that these cells are doing a lot more than just helping neurons maintain their activity,” says Professor Terrence Sejnowski, head of Salk’s Computational Neurobiology Laboratory and senior author of the new work.
“It suggests that they’re actually playing an important role in how information is transmitted and stored in the brain.”
The brain’s neurons rely on speedy electrical signals to communicate throughout the brain and release neurotransmitters, but astrocytes instead generate signals of calcium and release substances known as gliotransmitters, some of them chemically similar to neurotransmitters.
The classical view was that astrocytes‘ function was mostly to provide support to the more active neurons, helping transport nutrients, clean up molecular debris, and hold neurons in place.
Only more recently, researchers have found that they might play other, more active, roles in the brain through the release of gliotransmitters but these remain largely mysterious.
In 2014, Sejnowski, Salk postdoctoral researcher António Pinto-Duarte and their colleagues showed that disabling the release of gliotransmitters in astrocytes turned down a type of electrical rhythm known as a gamma oscillation, important for cognitive skills.
In that study, when the researchers tested the learning and memory skills of mice with disabled astrocytes, they found deficits that were restricted to their capacity to discriminate novelty.
In the new study, Sejnowski’s team looked for the first time at the longer-term memory of mice with disrupted astrocytes.
They used genetically engineered animals lacking a receptor called type 2 inositol 1,4,5-trisphosphate (IP3R2), which astrocytes rely on to release calcium for communication.
The researchers tested the mice with three different types of learning and memory challenges, including interacting with a novel object and finding the exit in a maze.
In each case, mice lacking IP3R2 showed the same ability to learn as normal mice. Moreover, when tested in the 24-48 hours after each initial learning process, the mice with disrupted astrocytes could still retain the information – finding their way through the maze, for example.
The results were in line with what had been seen in prior studies.
However, when the group waited another 2 to 4 weeks and retested the trained mice, they saw large differences; the mice missing the receptor performed much worse, making more than twice as many errors when completing the maze.
“After a few-weeks delay, normal mice actually performed better than they did right after training, because their brain had gone through a process of memory consolidation,” explains António Pinto-Duarte, who is the lead author of the new paper.
“The mice lacking the IP3R2 receptor performed much worse.”
The result is the first time that defects in astrocytes have been linked to defects in memory consolidation or remote memory.
The process of memory consolidation in the brain is known to involve several mechanisms affecting neurons.
One of those mechanisms is thought to rely in an optimal adjustment of the strength of communication between neurons through long-term potentiation, by which that strength increases, and long-term depression, by which some of these connections weaken. Sejnowski and Pinto-Duarte showed that although the mice without IP3R2 and reduced astrocyte activity had no problems with the former, they exhibited significant deficits in the latter, suggesting that astrocytes may be playing a role specifically in the long-term depression of the connections between neurons.
“The mechanism of long-term depression of neurons is not as well studied or understood,” says Sejnowski.
“And this tells us we should be looking at how astrocytes are connected to the weakening of these neural connections.”
The researchers are already planning future studies to better understand the pathways by which astrocytes affect the long-term depression of neuronal communication and memory in general.
“The long-term payout here is that if we better understand these pathways, we may be able to develop ways to manipulate memory consolidation with drugs,” says Sejnowski.
Remote memories, weeks to decades long, continuously guide our behavior, and are critically important to any organism, as the longevity of a memory is tightly connected to its significance.
However, the exact time at which each region is recruited, the duration for which it remains relevant to memory function, and the interactions between these regions, are still debated.
Astrocytes are no longer considered to merely provide homeostatic support to neurons and encapsulate synapses, as pioneering research has shown that astrocytes can also sense and modify synaptic activity as an integral part of the ‘tripartite synapse’4, 5.
Interestingly, astrocytes demonstrate extraordinary specificity in their effects on neuronal circuits6, at several levels: First, astrocytes differentially affect neurons based on their genetic identity.
For example, astrocytes in the dorsal striatum selectively respond to, and modulate, the input onto two populations of medium spiny neurons, expressing either D1 or D2 dopamine receptors7. Similarly, astrocytes selectively modulate the effects of specific inhibitory cell-types, but not others, in the same brain region8–11.
Second, astrocytes exert neurotransmitter-specific effects on neuronal circuits. For instance, astrocytic activation in the central amygdala specifically depresses excitatory inputs and enhances inhibitory inputs. Finally, astrocytes exhibit task-specific effects in-vivo, i.e. astrocytic stimulation selectively increases neuronal activity when coupled with memory acquisition, but not in the absence of learning12. An intriguing open question is whether astrocytes can differentially affect neurons based on their distant projection target.
The integration of novel chemogenetic and optogenetic tools in astrocyte research allows real-time, reversible manipulation of these cells at the population level, in combination with electrophysiological and behavioral measurements.
Such tools were used in brain slices to activate intracellular pathways in astrocytes, and show the ability of these cells to selectively modulate the activity of the neighboring neurons in the amygdala and striatum13, 14, and induce de-novo long-term potentiation in the hippocampus12, 15.
The reversibility of chemogenetic and optogenetic tools allows careful dissection of the effect of astrocytes during the different stages of memory in behaving animals16, 17. The recruitment of different intracellular signaling pathways in astrocytes using such tools is starting to shed light on their complex involvement in memory processes, with Gq activation in the CA1 during acquisition (but not during recall) resulting in enhanced recent memory12, 15, and Gs activation resulting in recent memory impairment18.
To explore the role of astrocytes in remote memory, and their ability to exert projection-specific effects, we used chemogenetics to activate the Gi pathway in these cells, and found that this astrocytic modulation in CA1 during learning resulted in a specific impairment in remote (but not recent) memory recall, accompanied by decreased activity in the anterior cingulate cortex (ACC) at the time of retrieval. In-vivo Gi activation in astrocytes disrupted synaptic transmission from CA3 to CA1 and reduced the downstream recruitment of the ACC. Finally, we show a dramatic recruitment of CA1 neurons projecting to ACC during memory acquisition, and a projection-specific inhibition of this population by Gi pathway activation in CA1 astrocytes.
More information: António Pinto‐Duarte et al, Impairments in remote memory caused by the lack of Type 2 IP 3 receptors, Glia (2019). DOI: 10.1002/glia.23679
Provided by Salk Institute