What Are Astrocytes?

For most of the twentieth century, neuroscience treated the brain as a network of neurons. The story was simple: neurons fire, synapses connect, circuits compute. Everything else—the cells that outnumber neurons, that fill the space between them, that wrap around blood vessels and synapses alike—was backdrop. Glue, even. The word itself, glia, comes from the Greek for glue.

Astrocytes are the most abundant of these glial cells, named for their star-like shape when stained and viewed under a microscope. They were first described in the late nineteenth century by pathologist Rudolf Virchow and later characterized in exquisite detail by Santiago Ramón y Cajal, who nonetheless considered them structural scaffolding. For decades, textbooks echoed this view: astrocytes provide support, maintain homeostasis, clean up after neurons. Passive. Ancillary. Necessary, perhaps, but not interesting.

That view has collapsed. Over the past three decades, a wave of research has revealed that astrocytes are active participants in nearly every aspect of brain function. They listen to synaptic transmission, modulate it, and in some cases initiate it. They regulate blood flow in response to neural activity, recycle neurotransmitters, maintain the blood-brain barrier, coordinate metabolic support, and communicate with one another via waves of calcium signaling that sweep across brain tissue like slow, silent thought. When the brain is injured, they transform. When it is diseased, they are implicated. And when neuroscience finally began to pay attention, the entire conceptual architecture of how the brain works had to be rewritten.

This article is about what astrocytes are, what they do, and why they matter. It is also about a larger shift in how we understand the nervous system: not as a machine made of wires, but as an ecosystem in which neurons are only one kind of participant.

Astrocytes are a type of glial cell found throughout the central nervous system. They are star-shaped, with a small cell body and dozens of fine, branching processes that extend outward in all directions. These processes are not passive structures. Each one makes contact with thousands of synapses, wrapping around the junctions where neurons communicate. A single astrocyte in the human cortex can contact more than a million synapses. In the rodent brain, that number is closer to one hundred thousand, but the principle is the same: astrocytes are positioned to sense and influence an enormous volume of neural activity.

Astrocytes also send specialized processes called endfeet to blood vessels, where they form part of the blood-brain barrier. This barrier is not a wall but a highly selective interface, and astrocytes are the gatekeepers. They regulate what enters the brain from the bloodstream and what leaves. They also couple neural activity to blood flow, a phenomenon called neurovascular coupling, which underlies functional brain imaging techniques like fMRI.

Morphologically, astrocytes come in several varieties. Protoplasmic astrocytes are found in gray matter and have thick, highly branched processes. Fibrous astrocytes populate white matter and have longer, thinner processes. There are also specialized astrocytes in the cerebellum (Bergmann glia), the retina (Müller cells), and other regions. Each type is adapted to the local architecture and functional demands of its environment.

Astrocytes do not generate action potentials. They do not spike. This is one reason they were overlooked for so long: the tools of twentieth-century neuroscience were built to detect electrical activity in neurons. But astrocytes are far from electrically silent. They express a wide array of ion channels, receptors, and transporters. They respond to neurotransmitters. And they communicate using calcium signals—slow, wave-like changes in intracellular calcium concentration that propagate from cell to cell, sometimes across millimeters of tissue.

In short, astrocytes are cells that occupy a unique position in the brain's architecture: intimately connected to both neurons and blood vessels, capable of sensing and modulating activity across spatial and temporal scales that neurons alone cannot access.

The term astrocyte—from the Greek astron, meaning star, and kytos, meaning cell—was coined in the early twentieth century to describe the stellate appearance of these cells when visualized with metallic stains. Ramón y Cajal's drawings of astrocytes are still reproduced in textbooks, beautiful and precise, their processes radiating outward like the arms of a starfish frozen in ink.

But the name also reflects a conceptual error. Stars are distant, decorative, inert. For most of the twentieth century, that is how astrocytes were treated: as pretty background objects in a drama centered on neurons. The oversight was not malicious. It was methodological. Electrophysiology, the dominant technique in neuroscience, measures voltage changes. Astrocytes do not spike, so they did not register. Behavioral neuroscience traced circuits from sensation to action, and astrocytes did not fit neatly into those diagrams. Molecular biology focused on genes and proteins that were conserved across species, and astrocytes—especially human astrocytes—turned out to be far more complex and diverse than those in the model organisms most commonly studied.

It was not until the 1990s, with the advent of calcium imaging and the discovery that astrocytes could release gliotransmitters, that the field began to reckon with what had been missed. The star-shaped cells were not glue. They were participants.

The traditional view of astrocytes was that they provided structural and metabolic support to neurons. They were thought to maintain ion balance, buffer extracellular potassium, supply neurons with lactate, and clean up excess neurotransmitters. All of this is true. But it is also radically incomplete.

The first crack in the passive model came in 1994, when researchers showed that astrocytes in culture could respond to neurotransmitters with increases in intracellular calcium, and that these calcium signals could propagate from cell to cell in waves. This was surprising, but it was in a dish. The real shift came when similar calcium signals were observed in vivo, in living brain tissue, and when it became clear that these signals were not random but correlated with neural activity and behavior.

Then came the discovery of the tripartite synapse. The classical synapse is a two-part structure: a presynaptic neuron releases neurotransmitter, a postsynaptic neuron receives it. But in most synapses, an astrocytic process is also present, enwrapping the synaptic cleft and expressing receptors for the same neurotransmitters that neurons use. When the synapse fires, the astrocyte listens. And in response, the astrocyte can release its own signaling molecules—glutamate, ATP, D-serine, and others—that modulate synaptic strength, neuronal excitability, and even the probability of neurotransmitter release.

This was not support. This was regulation. Astrocytes were not stagehands; they were part of the performance.

The implications were profound. If astrocytes modulate synaptic transmission, then they are involved in learning, memory, and plasticity. If they regulate blood flow, then they are central to brain energetics and the signals measured by neuroimaging. If they respond to and release neurotransmitters, then they are part of the signaling networks that underlie cognition and behavior. The brain, it turned out, was not a network of neurons. It was a network of neurons and glia, interacting in ways that were only beginning to be understood.

One of the most fundamental roles of astrocytes is the recycling of neurotransmitters, particularly glutamate, the brain's primary excitatory neurotransmitter. When a neuron releases glutamate into the synaptic cleft, it must be cleared quickly. If it lingers, it can overstimulate receptors and cause excitotoxicity, a process implicated in stroke, epilepsy, and neurodegenerative disease.

Astrocytes express high-affinity glutamate transporters, particularly GLT-1 (also called EAAT2), which remove glutamate from the extracellular space. Once inside the astrocyte, glutamate is converted to glutamine by the enzyme glutamine synthetase. Glutamine is then released back into the extracellular space, taken up by neurons, and converted back into glutamate or GABA. This is the glutamate-glutamine cycle, and it is essential for sustained synaptic transmission.

Without astrocytes, this cycle breaks down. Glutamate accumulates, neurons become hyperexcitable, and the system destabilizes. In animal models, selective loss of astrocytic glutamate transporters leads to seizures and neuronal death. In humans, mutations in these transporters are associated with epilepsy and other neurological disorders.

But astrocytes do more than clean up. They also sense the glutamate they take up. Glutamate uptake triggers changes in astrocytic metabolism, calcium signaling, and gene expression. The astrocyte is not a passive sponge; it is a sensor and responder, adjusting its activity based on the intensity and pattern of synaptic transmission around it.

The blood-brain barrier is one of the brain's defining features. It protects the neural environment from pathogens, toxins, and fluctuations in blood composition. But it is not a static wall. It is a dynamic interface, and astrocytes are central to its formation and regulation.

Astrocytic endfeet cover more than ninety percent of the brain's vasculature. These endfeet express aquaporin-4, a water channel that regulates fluid balance, and a host of transporters and receptors that sense and respond to signals from both blood and brain. During development, astrocytes induce the tight junctions between endothelial cells that make the barrier impermeable. In adulthood, they maintain it.

Astrocytes also mediate neurovascular coupling, the process by which active neurons receive increased blood flow. When neurons fire, they release neurotransmitters and ions into the extracellular space. Astrocytes detect these signals and respond by releasing vasoactive molecules—such as prostaglandins, nitric oxide, and arachidonic acid metabolites—that cause nearby blood vessels to dilate or constrict. This ensures that metabolically active regions receive the oxygen and glucose they need.

This coupling is what makes functional brain imaging possible. When you see a region of the brain light up in an fMRI scan, you are seeing increased blood flow, which is driven in large part by astrocytes. The neuron fires, the astrocyte listens, the blood vessel responds. It is a three-way conversation, and astrocytes are the translators.

Disruptions in neurovascular coupling are implicated in a range of conditions, from migraine to Alzheimer's disease. In Alzheimer's, astrocytes become reactive and dysfunctional, and neurovascular coupling is impaired, contributing to the cognitive decline that defines the disease.

The concept of the tripartite synapse, introduced in the late 1990s, fundamentally altered how neuroscientists think about synaptic transmission. In this model, the synapse is not a two-part structure but a three-part one: presynaptic neuron, postsynaptic neuron, and astrocyte.

Astrocytes ensheath synapses with fine processes that express receptors for neurotransmitters including glutamate, GABA, ATP, and acetylcholine. When these receptors are activated, they trigger calcium signals within the astrocyte. These calcium signals can, in turn, cause the astrocyte to release its own signaling molecules—termed gliotransmitters—including glutamate, D-serine, ATP, and adenosine.

Gliotransmitters can modulate synaptic transmission in multiple ways. Glutamate released by astrocytes can activate extrasynaptic NMDA receptors on neurons, influencing synaptic plasticity. D-serine, a co-agonist of the NMDA receptor, is synthesized and released by astrocytes and is required for certain forms of long-term potentiation, a cellular mechanism of learning and memory. ATP can be broken down to adenosine, which acts as a neuromodulator and is involved in sleep regulation.

The synapse is not a wire. It is a conversation, and astrocytes are part of the dialogue.

The mechanisms of gliotransmitter release are still debated. Some evidence suggests vesicular release, similar to neurons. Other evidence points to non-vesicular pathways, including volume-regulated anion channels and hemichannels. What is clear is that astrocytes can influence the strength, timing, and plasticity of synaptic transmission, and that this influence is bidirectional: neurons affect astrocytes, and astrocytes affect neurons.

This has profound implications for understanding learning, memory, and cognition. It means that plasticity is not solely a property of neurons but emerges from the interaction of neurons and glia. It means that drugs targeting synaptic function may have unintended effects on astrocytes, and vice versa. And it means that the computational capacity of the brain may be far greater than models based on neurons alone would suggest.

Neurons are metabolically expensive. They consume vast amounts of energy, primarily in the form of ATP, to maintain ion gradients, synthesize neurotransmitters, and support synaptic transmission. But neurons are not self-sufficient. They rely on astrocytes for metabolic support.

Astrocytes store glycogen, the brain's only significant energy reserve. When neurons are active, astrocytes break down glycogen into glucose, which is then metabolized via glycolysis to produce lactate. This lactate is released into the extracellular space and taken up by neurons, where it is converted to pyruvate and used to fuel oxidative phosphorylation in mitochondria. This is the astrocyte-neuron lactate shuttle, proposed by Pierre Magistretti and Luc Pellerin in the 1990s.

The shuttle is not just a backup system. There is evidence that lactate is a preferred fuel for neurons during periods of high activity, and that it may also act as a signaling molecule, influencing synaptic plasticity and memory consolidation. Blocking lactate transport impairs long-term memory in animal models, suggesting that astrocytic metabolism is not merely supportive but integral to cognitive function.

Astrocytes also regulate the availability of other metabolic substrates, including amino acids and lipids. They synthesize cholesterol, which neurons require for synapse formation and maintenance. They provide precursors for neurotransmitter synthesis. And they buffer the extracellular environment, maintaining the pH and ion concentrations that neurons need to function.

In disease, this metabolic partnership breaks down. In Alzheimer's disease, astrocytes become less efficient at providing metabolic support, and neurons suffer. In epilepsy, dysregulated astrocytic metabolism contributes to hyperexcitability. In stroke, the loss of astrocytic function exacerbates neuronal injury. The brain is not a collection of independent cells. It is a metabolic ecosystem, and astrocytes are the ecosystem engineers.

Astrocytes do not fire action potentials, but they are not silent. They communicate using calcium. Changes in intracellular calcium concentration serve as the primary signaling mechanism in astrocytes, and these signals can be remarkably complex: localized to individual processes, propagating as waves across the cell body, or spreading from cell to cell across networks of astrocytes.

Calcium signals in astrocytes are triggered by a variety of stimuli, including neurotransmitters, neuromodulators, and mechanical stress. When an astrocyte detects glutamate or ATP, for example, it activates G-protein-coupled receptors that trigger the release of calcium from internal stores. This calcium can then propagate through the cell via calcium-induced calcium release, and it can spread to neighboring astrocytes through gap junctions formed by connexin proteins.

These calcium waves can travel across millimeters of tissue, coordinating the activity of large populations of astrocytes. In some cases, they are correlated with neural activity and behavior. In the visual cortex, for example, astrocytic calcium signals are modulated by visual stimuli. In the hippocampus, they are associated with spatial navigation and memory encoding.

Recent advances in imaging have revealed that astrocytic calcium signals are far more diverse than previously thought. They occur at multiple spatial scales, from microdomains within individual processes to global signals that encompass the entire cell. They occur at multiple temporal scales, from rapid transients lasting milliseconds to slow oscillations lasting minutes. And they are heterogeneous: different astrocytes in the same region can exhibit different patterns of calcium activity, suggesting functional specialization.

What these calcium signals mean—how they encode information, how they influence neural circuits, and how they contribute to behavior—is still being worked out. But it is clear that astrocytes are not passive. They are active, dynamic, and capable of integrating and transmitting information across spatial and temporal scales that neurons cannot access alone.

Not all astrocytes are the same. They vary in morphology, gene expression, and function depending on where they are in the brain. This regional diversity is increasingly recognized as a key feature of astrocyte biology, and it has important implications for understanding both normal brain function and disease.

In the cortex, protoplasmic astrocytes have bushy, highly branched processes that contact thousands of synapses. In white matter, fibrous astrocytes have longer, thinner processes that run parallel to axon tracts. In the cerebellum, Bergmann glia extend radial processes that guide the migration of neurons during development and ensheath synapses in the molecular layer. In the retina, Müller cells span the entire thickness of the tissue, providing structural and metabolic support to photoreceptors and other retinal neurons.

Recent transcriptomic studies have revealed even finer distinctions. Astrocytes in different cortical layers express different sets of genes. Astrocytes in the striatum differ from those in the hippocampus. And astrocytes in the human brain are larger, more complex, and more diverse than those in rodents, with some human astrocytes contacting up to two million synapses.

This diversity suggests that astrocytes are adapted to the specific functional demands of their local environment. In regions with high metabolic demand, astrocytes may prioritize energy supply. In regions with dense synaptic connectivity, they may prioritize neurotransmitter recycling and synaptic modulation. In regions involved in rhythmic activity, they may coordinate network oscillations.

Understanding this diversity is critical for developing therapies that target astrocytes. A drug that affects astrocytes in the cortex may have different effects in the striatum or cerebellum. A mutation that disrupts astrocyte function in one brain region may have little effect in another. The brain is not uniform, and neither are its astrocytes.

When the brain is injured or diseased, astrocytes transform. They become reactive, a state characterized by changes in morphology, gene expression, and function. Reactive astrocytes hypertrophy, their processes thicken, and they upregulate intermediate filament proteins such as GFAP (glial fibrillary acidic protein), which is often used as a marker of reactivity.

Reactive astrogliosis is not a single, uniform response. It exists on a spectrum, from mild and reversible changes to severe and permanent alterations. In some cases, reactive astrocytes are protective: they form a glial scar that walls off damaged tissue, limiting the spread of inflammation and preventing further injury. They upregulate antioxidant defenses, secrete neurotrophic factors, and help restore the blood-brain barrier.

But in other cases, reactive astrocytes are harmful. They can release pro-inflammatory cytokines, produce reactive oxygen species, and lose their ability to support neurons metabolically. They can downregulate glutamate transporters, leading to excitotoxicity. And in chronic conditions, they can contribute to a hostile environment that prevents repair and regeneration.

Recent work has identified distinct subtypes of reactive astrocytes. In response to acute injury such as stroke or trauma, astrocytes may adopt an A1 phenotype, characterized by the expression of complement proteins and other pro-inflammatory molecules. In response to ischemia or other insults, they may adopt an A2 phenotype, characterized by the expression of neurotrophic factors. These subtypes are not absolute, and individual astrocytes may express features of both, but the framework highlights the complexity and context-dependence of astrocyte reactivity.

Understanding reactive astrogliosis is essential for developing therapies for brain injury and disease. Blocking all astrocyte reactivity is likely to be harmful, as it would eliminate protective responses. The challenge is to selectively modulate astrocyte function, enhancing beneficial responses and suppressing harmful ones.

Astrocytes are implicated in nearly every neurological and psychiatric disorder. Their dysfunction contributes to disease pathology, and in some cases, they may be primary drivers of disease.

In Alzheimer's disease, astrocytes become reactive early in the disease process, even before significant neuronal loss. They cluster around amyloid plaques and express inflammatory markers. They lose their ability to clear glutamate and support neuronal metabolism. And they may contribute to the spread of tau pathology, the other hallmark of Alzheimer's. Recent evidence suggests that astrocytes can take up and propagate tau, transmitting it from cell to cell in a prion-like manner.

In multiple sclerosis, astrocytes are involved in both the inflammatory and neurodegenerative phases of the disease. They contribute to the formation of glial scars around demyelinated lesions, which can be both protective and inhibitory to repair. They also interact with immune cells, modulating the inflammatory response.

In epilepsy, astrocytes play a central role. Loss of astrocytic glutamate transporters leads to increased extracellular glutamate and neuronal hyperexcitability. Dysregulation of astrocytic potassium buffering can also contribute to seizures. And in some genetic forms of epilepsy, mutations in astrocyte-specific genes are the primary cause.

In amyotrophic lateral sclerosis (ALS), astrocytes become toxic to motor neurons. Mutations in genes such as SOD1, which are expressed in both neurons and astrocytes, cause astrocytes to release factors that kill motor neurons. Replacing mutant astrocytes with healthy ones in animal models can slow disease progression, suggesting that astrocytes are a viable therapeutic target.

In chronic pain, astrocytes in the spinal cord become reactive and release pro-inflammatory cytokines and chemokines that sensitize pain pathways. Blocking astrocyte activation can reduce pain in animal models, and there is growing interest in targeting astrocytes for pain management in humans.

In migraine, astrocytes are thought to contribute to cortical spreading depression, the wave of neuronal and glial depolarization that underlies migraine aura. They also regulate the release of vasoactive substances and inflammatory mediators that contribute to headache pain.

• Alzheimer's disease: reactive astrocytes, impaired glutamate clearance, tau propagation
• Multiple sclerosis: glial scar formation, immune modulation, demyelination
• Epilepsy: loss of glutamate transport, potassium buffering deficits, genetic mutations
• ALS: astrocyte toxicity to motor neurons, SOD1 mutations, therapeutic target
• Chronic pain: spinal astrocyte activation, cytokine release, pain sensitization
• Migraine: cortical spreading depression, vasoactive signaling, inflammation

Across all these conditions, a common theme emerges: astrocytes are not bystanders. They are active participants in disease pathology, and targeting them may offer new therapeutic opportunities.

Despite the explosion of research on astrocytes over the past three decades, misconceptions persist, both in the scientific literature and in popular accounts of neuroscience.

The first misconception is that astrocytes are support cells. This framing, inherited from the early twentieth century, is still common in textbooks and introductory courses. It is not wrong, exactly—astrocytes do provide support—but it is incomplete and misleading. Astrocytes are regulatory cells. They modulate synaptic transmission, control blood flow, coordinate metabolism, and respond to injury. Calling them support cells is like calling the conductor of an orchestra a stagehand.

The second misconception is that astrocytes are uniform. In fact, astrocytes are highly diverse, varying by brain region, developmental stage, and functional state. Treating all astrocytes as interchangeable obscures this diversity and limits our understanding of how they contribute to brain function and disease.

The third misconception is that astrocytes are slow. Because they do not fire action potentials, astrocytes are often assumed to operate on timescales of seconds to minutes, far slower than neurons. But recent imaging studies have revealed that astrocytic calcium signals can occur on millisecond timescales, and that astrocytes can respond to and modulate synaptic transmission in real time.

The fourth misconception is that astrocytes are less important in humans than in rodents. In fact, the opposite is true. Human astrocytes are larger, more complex, and more diverse than rodent astrocytes. They contact more synapses, express more genes, and exhibit more sophisticated calcium signaling. Some researchers have speculated that the expansion and elaboration of astrocytes may be one of the key evolutionary changes that enabled the cognitive capacities unique to humans.

The fifth misconception is that astrocytes are a recent discovery. Astrocytes were first described in the nineteenth century, and their basic functions—such as glutamate uptake and potassium buffering—have been known for decades. What is recent is the recognition that these functions are not passive but active, and that astrocytes are integral to nearly every aspect of brain function.

The story of astrocytes is the story of a paradigm shift. For most of the twentieth century, neuroscience was built on a simple model: the brain is a network of neurons, and understanding the brain means understanding how neurons fire, connect, and compute. This model was productive. It led to profound insights into sensory processing, motor control, learning, and memory. But it was incomplete.

Astrocytes reveal that the brain is not a wiring diagram. It is an ecosystem. Neurons do not operate in isolation. They are embedded in a dense, dynamic matrix of glial cells that sense, modulate, and support their activity. Synaptic transmission is not a two-way conversation but a three-way one, with astrocytes as active participants. Plasticity is not solely a property of neurons but emerges from the interaction of neurons and glia. And the computational capacity of the brain may be far greater than models based on neurons alone would suggest.

This has profound implications for how we think about nervous system intelligence. Intelligence is not just about the speed or precision of neural firing. It is about the integration of information across multiple timescales and spatial scales. It is about the coordination of metabolism, blood flow, and immune function with neural activity. It is about the ability to adapt, repair, and reorganize in response to experience and injury. All of these capacities depend on astrocytes.

Astrocytes also challenge the reductionist impulse that has dominated neuroscience. You cannot understand the brain by studying neurons alone, just as you cannot understand an ecosystem by studying only one species. The brain is a community of cells, and the interactions between those cells are as important as the cells themselves.

This shift in perspective is not just conceptual. It has practical implications. Drugs that target neurons may have unintended effects on astrocytes, and vice versa. Therapies that modulate astrocyte function may offer new treatments for diseases that have been resistant to neuron-centric approaches. And understanding the role of astrocytes in learning, memory, and cognition may open new avenues for enhancing nervous system intelligence.

The brain is not a machine. It is a living system, and astrocytes are central to its life. They are the cells that listen, respond, coordinate, and adapt. They are the cells that make the brain more than the sum of its neurons. And they are the cells that remind us that intelligence is not a property of individual components but an emergent property of the whole.