Astrocyte Ensembles are Involved in Memory Stabilization and Recall

Post by Amanda Engstrom

The takeaway

Engrams are the physical traces of memory in the neurons of the brain. This study reveals that astrocytes play a more direct role than previously thought, forming lasting ensembles that reactivate during memory recall. These astrocytic ensembles, driven by noradrenergic signaling, act as a multiday trace that helps stabilize and preserve memories over time.

What's the science?

Memory formation and stabilization involve specific neuronal ensembles, or groups of interconnected neurons that work together. Chemical and physical changes in these neurons then form memory traces – called engrams – to be activated during learning and recall. Remembering something transiently destabilizes memories, but the mechanism that subsequently re-stabilizes the memory is not completely understood and cannot be explained by neuronal engrams alone. Astrocytes are highly abundant non-neuronal cells within the central nervous system (CNS) that interact both structurally and functionally with neurons and other glial cells. They are known to be developmentally diverse and adaptive to physiological and pathological changes; however, the role of astrocytes in forming or stabilizing experience-dependent memories is not clear. This week in Nature, Dewa and colleagues investigated astrocytes that respond to memory formation and recall in order to assess their contribution to memory stabilization.

How did they do it?

The authors developed a tool to identify behaviorally relevant astrocyte ensembles in an unbiased manner by generating mice that allow them to label astrocytes that have been activated (marked by an increase in Fos gene expression) during a specific time window. For their studies, the authors “tagged” astrocytes either during a fear conditioning protocol or 24 hours later in a fear memory recall session. They then compared astrocyte activation with neuronal engram activity in the same fear conditioning paradigm. Once the authors identified the astrocyte ensembles, they performed a pharmacological screening for molecules that can activate astrocyte Fos expression and validated these target molecules through circuit analysis, imaging, and single-cell transcriptomics. Next, the authors determined the transcriptional response to memory formation and recall at different timepoints. Finally, they perturbed the astrocyte ensemble by using a peptide inhibitor to silence the target signaling cascade, as well as enhanced the ensemble signaling by overexpressing downstream signaling targets. Then they tested the efficiency of fear conditioning and recall using the same behavioral fear paradigm.  

What did they find?

The density of activated (Fos+) astrocytes did not significantly increase immediately after fear conditioning. However, 24 hours later, during recall, there was a significant increase in the number of activated astrocytes across the entire brain, especially in the amygdala, compared to fear conditioning without recall. This is unique to astrocytes, as neurons are activated both during fear conditioning and recall. There was a significant regional correlation between neuronal and astrocyte activation at recall, suggesting that astrocyte ensembles are recruited in regions of active neuronal engrams. The authors identified noradrenaline (NA) as a strong inducer of Fos expression in astrocytes, and using fiber photometry, (an imaging technique to measure cellular activity) determined that during fear recall, NA signals are stronger and last longer than during fear conditioning.

The single-cell transcriptomics on astrocytes in the amygdala revealed that after fear conditioning, astrocytes upregulate adrenergic receptor genes (Adra1a and Adrb1), which increased in expression over 24 hours, indicating a “priming” state that develops over a day and then decreases over time. Disrupting the astrocyte ensemble disrupted memory stability and reduced the mouse’s response in the fear recall test. When the ensemble was enhanced through the overexpression of Adrb1 in astrocytes, the density of activated astrocytes increased, and improved memory retention and recall. Together, these results show that fear conditioning induces an astrocyte ensemble that is primed via NA signaling and persists for roughly one day. Upon recall, the astrocyte ensemble is activated and stabilizes the memory.

What's the impact?

This study is the first to show that astrocytes form their own experience-dependent ensembles that reactivate during memory recall and help stabilize memories over multiple days. This work expands the traditional neuron-centric view of memory consolidation and argues for the critical role of astrocytes in memory stabilization and recall. 

How Psychedelics Affect Neurovascular Function

Post by Shireen Parimoo

The takeaway

Psychedelics like psilocybin exert their hallucinogenic effects by acting on 5HT2A serotonin receptors in the brain, and fMRI has been used to measure how this is reflected in neuronal activity. This research shows that psychedelics affect not only neuronal activity but also vascular activity (i.e., blood vessel activity) in the brain in distinct ways, meaning both need to be considered when using fMRI.

What's the science?

Psychedelic compounds like psilocybin have seen a surge of interest in recent years due to their potential to treat mood and substance use disorders. Many psychedelics have hallucinogenic effects through their action on serotonin receptors (5HTRs) in the brain. These serotonin receptors are primarily located in brain areas associated with higher-order cognitive functions, such as the default-mode network regions, including the retrosplenial cortex. In humans, functional magnetic resonance imaging (fMRI) studies show that psychedelics reduce functional connectivity within DMN regions but increase connectivity between the DMN and other areas of the brain. 

Functional MRI records the hemodynamic response that occurs following neural activity, but is not well-suited to differentiate between neural activity and vascular activity in the brain. Moreover, 5HTRs are located on both neurons that give rise to neural activity and astrocytes, which are involved in vascular coupling between neurons and blood vessels. Thus, it is presently unclear whether psychedelics exert their effects by modulating neural activity, vascular responses, or both. This week in Nature Neuroscience, Padawer-Curry and colleagues used optical imaging techniques to investigate the neurovascular origins of psychedelic-induced hemodynamic responses and functional connectivity in the brain.

How did they do it?

The authors performed wide-field optical imaging in mice during a whisker stimulation task and during rest. This technique can be used to measure both neural activity via calcium imaging and vascular or hemodynamic activity simultaneously. Mice were injected with either (i) 2,5-dimethoxy-4-iodoamphetamine (DOI), a hallucinogenic agonist of the 5HT2A receptor, (ii) MDL, a selective antagonist of 5HT2A receptors, (iii) DOI + MDL, (iv) saline, or (v) lisuride, a non-hallucinogenic 5HT2A receptor agonist. Notably, administering DOI and MDL together enabled them to determine whether the effects of psychedelics are specific to the 2A serotonin receptor, while administering lisuride helped determine whether these receptors were important for hallucinogenic as compared to non-hallucinogenic receptor activation.  

To distinguish between vascular and neuronal origins of the HRF, they recorded changes in hemodynamic and neural activity in response to whisker stimulation (i.e., task) and at rest. Power spectral density (PSD) was used as a measure of neural and hemodynamic activity across different frequency bands. The authors examined PSD in the infraslow (0.01-0.08 Hz), intermediate (0.08 – 0.50 Hz), and the delta frequency bands (0.50 – 4.00 Hz), which helped distinguish the neural and hemodynamic responses at different timescales. Finally, resting-state functional connectivity between regions was calculated from the neuronal and hemodynamic responses to obtain measures of intra- and inter-network connectivity. 

What did they find?

Administration of DOI altered hemodynamic and neuronal activity in a regionally specific manner. For instance, hemodynamic responses increased in regions like the somatosensory and auditory cortex. In contrast, neuronal activity in the retrosplenial cortex and somatosensory regions was reduced. The shape of the HRF following DOI was narrower, while simultaneous injection of MDL and DOI did not have this effect on the HRF, indicating that DOI disrupts neurovascular coupling through its effect on 5HT2A receptors. 

There was a divergence in the effects of DOI on neuronal compared to hemodynamic activity across frequency bands. Specifically, global neuronal activity increased in the delta band but decreased in the intermediate frequency band following DOI injection. Global hemodynamic activity, on the other hand, only increased in the delta band. At a regional level, neuronal activity in the infraslow band increased in frontal and motor regions, while hemodynamic activity increased in somatosensory regions. In the intermediate frequency band, neuronal activity showed broad reductions but an increase in frontal and cingulate regions, whereas hemodynamic activity decreased in the frontal and cingulate cortex. In the delta band, both neuronal and hemodynamic activity increased in response to DOI, but in different regions. 

Finally, DOI altered network-level organization as measured by neuronal responses. Across most of the cortex, DOI reduced intra-network connectivity and increased inter-network connectivity, while the cingulate cortex showed greater intra-network connectivity with the DMN. Conversely, hemodynamic-based network dynamics showed minimal connectivity changes in response to DOI. Interestingly, co-administration of MDL largely reversed the effects observed following DOI administration alone. This demonstrates that DOI acts specifically through the 5HT2A receptors and differentially disrupts vascular and neuronal processes in the brain. 

What's the impact?

This study found that psychedelics disrupt both neuronal and vascular responses through their action on 5HT2A receptors in the brain, identifying a potential mechanism through which psychedelics induce changes in brain network connectivity. Importantly, this indicates that hemodynamic measures of brain activity – such as those measured using fMRI in humans – may be capturing signals other than neural activity in psychedelic states.

Access the original scientific publication here.

Brain Repair Mechanisms After Stroke Differ in White and Gray Matter

Post by Anastasia Sares

The takeaway

In this study, the authors identified a gene called Lamc1, which is involved in blood vessel regeneration after a stroke, but only in the cortex (the outer layer of the gray matter) and not in the white matter, which is located deeper in the brain. This represents a potential target for medical interventions.

What's the science?

A stroke happens when a part of the brain is deprived of blood supply, either by a blockage of a blood vessel (like a heart attack, but in the brain) or by a ruptured blood vessel. When the blood supply is disrupted, neurons cannot obtain sufficient oxygen, glucose, and other essential nutrients, and they begin to die off. Stroke is one of the leading causes of death and disability worldwide, but the brain does have the ability to recover some functionality after a stroke. Understanding how this brain recovery happens and what we can do to enhance it is, therefore, a high medical priority. Strokes can happen in different brain regions depending on which blood vessels are affected.

Astrocytes (named for their sometimes star-like shape) are brain cells that are crucial for stroke recovery, modifying neural connections and interacting with systems for immune function and blood flow, especially at the border between healthy and damaged brain tissue. This week in Neuron, Gleichman and colleagues examined the activity of astrocytes in different stroke locations in mice to see if there were differences in how the brain repairs itself, identifying a key process that promotes blood vessel regeneration.

How did they do it?

The authors induced strokes in the cortex of mice, which is the brain’s outer layer of gray matter (brain tissue with more neuronal cell bodies), and others were induced in white matter (containing fat-insulated axons of neurons traveling from one region of the brain to the other).

Using a multitude of techniques (viral marking, gene sequencing, and more), the authors characterized the astrocytes and their byproducts in the dissected brains. One technique, gene set enrichment analysis (GSEA), measures the relative quantities of different genes and how they are expressed in a piece of tissue. The authors compared gene expression following strokes in the cortex and in the white matter.

Finally, the authors measured recovery in mice affected by these different strokes. To do this, they set up a test where the mice had to manipulate small objects that had peanut butter inside the grooves. To get the peanut butter, the mice needed to pick up the objects and handle them very precisely, using regions of the brain that were affected by the stroke. Their ability to perform these precise movements was a way to track the effects of injury and recovery.

What did they find?

At both different types of stroke locations, astrocytes were activated at the border between damaged and healthy tissue, initiating a cascade of processes to help the brain heal. One difference that the authors noticed was that for strokes that affected the cortex, a gene called Lamc1, which is involved in blood vessel growth, was highly expressed. However, this gene was not highly expressed for strokes in the white matter.

Recovery from white matter strokes is more difficult than for cortical strokes, so the authors decided to intervene artificially and see if activating or deactivating Lamc1 would change recovery. Knocking out Lamc1 in the cortex resulted in less blood vessel regrowth, while activating Lamc1 in white matter resulted in less long-term damage, like scarring and nerve damage. The mice also seemed to maintain more consistent performance in the object manipulation task when Lamc1 was activated after a stroke, though the difference between Lamc1 and a control condition did not quite reach significance.

What's the impact?

The authors showed that artificially triggering Lamc1 activation may promote blood vessel regeneration in white matter, where it would normally be minimal. These findings can inform future clinical research into stroke treatment and recovery.