{
  "id": "astrocyte",
  "title": "Astrocyte",
  "lang": "en",
  "source_lang": "en",
  "summary": "An astrocyte is a central nervous system glial cell that regulates the neural environment, supports synapses and metabolism, and responds to injury.",
  "category": "science",
  "subcategory": "biology",
  "type": "concept",
  "aliases": [
    "Astroglial cell",
    "astroglial astrocyte"
  ],
  "generated_at": "2026-11-03",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/astrocyte/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/astrocyte/",
      "json": "https://aiwiki.page/zh/astrocyte.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "An astrocyte is a major type of glial cell in the brain and spinal cord. Astrocytes maintain the chemical environment required for neuronal activity, recycle neurotransmitters, provide metabolic support, interact with blood vessels, and participate in the formation and modification of neural circuits. Unlike neurons, they generally do not communicate through rapidly propagated, all-or-none action potentials. Their influence instead depends on membrane transport, local biochemical signaling, structural interactions, and communication with neighboring cells.\n\nThe name means “star cell” and refers to the radiating processes visible with early staining methods. Mature astrocytes are considerably more elaborate than this image suggests: fine, irregular branches extend among synapses, while specialized endings contact blood vessels and the surfaces of the central nervous system. Astrocytes are diverse across anatomical regions, developmental stages, species, and physiological states. This diversity, together with their extensive interactions with neurons, immune cells, and vascular cells, makes them important both to normal nervous-system function and to the cellular responses accompanying injury and disease."
    },
    {
      "heading": "Cellular architecture and regional diversity",
      "text": "An astrocyte consists of a cell body containing the nucleus and a branching system of processes. Relatively thick branches give rise to much finer extensions, including thin leaflets that approach neuronal synapses. Many of these peripheral structures are below the resolution limit of conventional light microscopy. Their high surface-area-to-volume ratio suits them to exchange ions and small molecules with the extracellular space. Other processes end as expanded endfeet along blood vessels or contribute to the glia limitans, a thin cellular boundary beneath the membranes surrounding the brain and spinal cord.\n\nThe traditional distinction separates protoplasmic astrocytes, concentrated in gray matter, from fibrous astrocytes, common in white matter. Protoplasmic astrocytes have densely branched processes that interweave with neuronal cell bodies, dendrites, and synapses. Fibrous astrocytes generally have longer, less profusely branched processes associated with axon bundles. Specialized forms include Bergmann glia in the cerebellum and Müller glia in the retina. These retain a radial organization and support highly ordered neural circuits. The classification is useful anatomically but does not capture the full diversity revealed by molecular studies.\n\nIn much of adult rodent gray matter, neighboring protoplasmic astrocytes occupy largely separate territories, with limited overlap between their major arbors. Neuronal structures within a territory can interact with numerous processes belonging to the same astrocyte. Territorial organization varies with region, species, developmental stage, and injury, however, and should not be treated as an absolute rule. Human cortical astrocytes are generally larger and morphologically more elaborate than corresponding rodent cells; primate cortex also contains conspicuous interlaminar astrocytes with long processes spanning cortical layers.\n\nAstrocytes are often identified using glial fibrillary acidic protein, or GFAP, an intermediate-filament protein. GFAP staining reveals only part of an astrocyte’s structure and may be weak in many healthy gray-matter astrocytes. Other commonly used markers include ALDH1L1, glutamine synthetase, S100B, and the transporters encoded by SLC1A2 and SLC1A3. None is universally specific across all tissues and developmental conditions. Reliable identification therefore combines location, morphology, and several molecular features rather than relying on one marker.\n\nMolecular diversity also has functional dimensions. Astrocytes in different regions express different combinations of neurotransmitter transporters, receptors, metabolic enzymes, and extracellular proteins. Some differences follow developmental boundaries; others reflect adaptation to nearby neurons and vessels. Consequently, “astrocyte” denotes a family of related cellular identities rather than one uniform cell type repeated throughout the nervous system."
    },
    {
      "heading": "Development, maturation, and maintenance",
      "text": "Most astrocytes arise from neural stem cells and radial glial cells in the developing central nervous system. These progenitors also generate neurons and other glial lineages. In many regions, neuronal production predominates earlier, followed by increased production of glia. The timing is region- and species-dependent: substantial astrocyte production and maturation occur after birth in rodents, whereas the corresponding events extend across prenatal and postnatal development in humans.\n\nThe switch toward astrocyte production involves changing competence within progenitor cells as well as extracellular signals. Notch signaling helps maintain progenitor properties and promotes aspects of glial differentiation. Cytokine pathways involving JAK–STAT signaling and signals from bone morphogenetic proteins can activate astrocytic programs in developmentally competent cells. Transcriptional regulators, including members of the SOX and NFIA families, help establish or maintain these programs. Changes in chromatin accessibility make astrocyte-associated genes responsive at appropriate developmental stages; an extracellular signal alone is therefore not sufficient to specify cell identity in every context.\n\nAstrocyte number increases through both differentiation from progenitors and local proliferation. In the developing mouse cortex, division of locally situated astrocyte-lineage cells makes a substantial contribution to the mature population. As astrocytes mature, they elaborate branches, establish vascular contacts, increase expression of neurotransmitter transporters, and acquire more specialized membrane properties. Synapse formation, vascular development, and astrocyte maturation overlap in time, allowing reciprocal interactions among the three systems.\n\nDeveloping neurons help instruct astrocyte identity. Sonic hedgehog signaling, for example, contributes to regionally specialized astrocyte properties in several brain areas. Neuronal activity and locally produced factors also influence transporter expression and process organization. Astrocytes are therefore not simply assigned a fixed identity before entering a circuit: their mature characteristics emerge through interactions between lineage history and the environment. Some regional identities nevertheless remain stable, reflecting the spatial organization of the progenitors from which the cells originated.\n\nMost mature astrocytes in the healthy adult brain divide infrequently. Their maintenance depends largely on cellular longevity and ongoing renewal of proteins, membranes, and organelles rather than rapid population replacement. Injury can induce proliferation in subsets of astrocytes, particularly near severely damaged tissue. Although experimental manipulations can sometimes redirect astrocytes toward neuronal or progenitor-like states, such conversion is not an established routine mechanism of neuronal replacement in the intact adult brain. Apparent lineage changes must also be distinguished from experimental labeling of pre-existing neurons or neural progenitors."
    },
    {
      "heading": "Ionic balance, water movement, and extracellular homeostasis",
      "text": "Neuronal signaling repeatedly changes the composition of extracellular fluid. During an action potential, potassium leaves neurons, while sodium enters them. Synaptic activity also releases neurotransmitters and can alter extracellular acidity. Astrocytes counter these disturbances through membrane channels, transporters, pumps, and intracellular buffering. Their homeostatic activity is essential because neuronal excitability depends strongly on extracellular ion concentrations.\n\nPotassium handling is a prominent example. Mature astrocytes generally have a negative resting membrane potential and high potassium permeability, partly associated with the inwardly rectifying potassium channel Kir4.1, encoded by KCNJ10. Potassium can enter astrocytes where extracellular levels are elevated and be redistributed through the cell or an interconnected cellular network. Sodium–potassium ATPase also contributes to potassium uptake, using energy from ATP hydrolysis. The relative importance of channels, pumps, and network redistribution varies with the location and intensity of neural activity.\n\nMany astrocytes are linked by gap junctions, especially channels containing connexin 43 and connexin 30. These channels allow ions and small metabolites to pass directly between cells. Coupling can distribute ionic and metabolic loads beyond the immediate site of activity, although networks are not unrestricted, homogeneous compartments. Their connectivity depends on anatomical boundaries, developmental state, and regulation of the junctions themselves. Astrocytes can also form junctional connections with oligodendrocytes in some regions, linking their homeostatic roles to those of myelinated axons.\n\nWater movement is closely connected to ion transport. The water channel aquaporin-4 is enriched in many astrocytic endfeet facing blood vessels and fluid-containing interfaces. It increases membrane water permeability but does not actively pump water: movement follows osmotic gradients. Astrocyte swelling can reduce extracellular-space volume, thereby changing the concentrations and movement of extracellular substances. Under pathological conditions, disrupted ion gradients and water accumulation contribute to brain edema. Aquaporin-4 can have different effects during water entry into tissue and during subsequent fluid clearance, so its role cannot be reduced to uniformly causing or preventing swelling.\n\nAstrocytes also contribute to extracellular pH regulation through bicarbonate transport, proton-linked transport processes, and metabolism. Their handling of sodium, chloride, and other ions affects cell volume and transporter operation. These mechanisms are interdependent: glutamate uptake, for example, imports sodium and increases the demand on sodium–potassium ATPase. Homeostasis is thus an active, energy-consuming component of neural processing, not a passive background condition. Failure of ATP production can simultaneously compromise ion regulation, transmitter clearance, and control of extracellular volume."
    },
    {
      "heading": "Neurotransmitter clearance and synaptic organization",
      "text": "Astrocytes closely approach many synapses and regulate the persistence and spread of released neurotransmitters. Their best-characterized role is uptake of glutamate, the principal excitatory transmitter in much of the central nervous system. In adult forebrain, the transporter EAAT2, also called GLT-1 in rodents and encoded by SLC1A2, performs a large share of glutamate clearance. EAAT1, or GLAST, encoded by SLC1A3, is particularly important in several other regions and developmental settings.\n\nThese transporters couple glutamate movement to existing ion gradients. A typical transport cycle moves one glutamate anion, three sodium ions, and one proton into the cell while moving one potassium ion outward. Uptake is therefore electrogenic and depends indirectly on the ATP used to maintain ionic gradients. Efficient removal limits activation of receptors outside the releasing synapse and reduces the risk of excitotoxicity, in which excessive glutamate-receptor activation contributes to neuronal injury. During severe energy failure, collapse of ion gradients can impair uptake and, under some conditions, favor transporter reversal.\n\nWithin astrocytes, glutamine synthetase converts glutamate and ammonia into glutamine in an ATP-dependent reaction. Glutamine can be released, taken up by neurons, and converted back into glutamate by phosphate-activated glutaminase. This glutamate–glutamine cycle replenishes transmitter pools while separating parts of transmitter metabolism between cell types. The inhibitory transmitter gamma-aminobutyric acid, or GABA, is also connected to this cycle because neurons synthesize it from glutamate. Astrocytes express GABA transporters as well, although the contribution of different uptake routes varies across circuits.\n\nAstrocytes influence synapses structurally as well as chemically. During development, secreted thrombospondins promote formation of synaptic structures, while hevin and other extracellular proteins influence the assembly of particular connections. Glypicans can promote aspects of excitatory synapse function, including receptor recruitment in experimental systems. Astrocyte-derived cholesterol also supports synapse development in culture. These effects are not interchangeable: a synapse with an anatomically recognizable structure need not yet have the receptor complement required for mature transmission.\n\nAstrocytes can participate in synapse removal through phagocytosis, including pathways involving the receptors MEGF10 and MERTK. This complements, rather than duplicates, the roles of microglia in circuit remodeling. The term “tripartite synapse” emphasizes that some synapses function through interactions among a presynaptic neuron, a postsynaptic neuron, and nearby astrocytic processes. It is a functional framework, not a claim that every synapse is completely enclosed by one astrocyte or that all synaptic transmission requires an astrocytic signal."
    },
    {
      "heading": "Energy metabolism and nutritional support",
      "text": "Astrocytes participate in the distribution and transformation of energy substrates within nervous tissue. They take up glucose, perform glycolysis, and use mitochondrial oxidative metabolism. They also contain much of the brain’s mobilizable glycogen, although small neuronal glycogen pools and other exceptions prevent an absolute separation between cell types. Astrocytic glycogen provides a local reserve that can be mobilized when energy demand rises or substrate availability falls.\n\nGlutamate uptake illustrates how neurotransmission and metabolism become coupled. Sodium entering with glutamate increases the work required of sodium–potassium ATPase. This can stimulate glucose utilization and other metabolic responses. Astrocytes convert some glucose-derived pyruvate into lactate, which can cross cell membranes through monocarboxylate transporters. Neurons can take up lactate and oxidize it, allowing carbon initially metabolized in astrocytes to contribute to neuronal ATP production.\n\nThe astrocyte–neuron lactate shuttle hypothesis, formulated in 1994 by Luc Pellerin and Pierre Magistretti, proposes a specific activity-linked arrangement of this exchange. It has stimulated extensive research, but its quantitative importance varies among preparations and physiological conditions. Neurons also take up glucose directly, astrocytes consume oxidative substrates themselves, and lactate can move in either direction according to concentration and proton gradients. Evidence for lactate exchange therefore does not establish that neuronal energy supply universally depends on a one-way astrocyte-to-neuron pathway.\n\nAstrocytes also replenish metabolic intermediates through anaplerosis. Pyruvate carboxylase, strongly associated with astrocytes in the adult brain, converts pyruvate into oxaloacetate. This helps replace carbon removed from the citric acid cycle for amino-acid and neurotransmitter synthesis. The distinction matters because recycling glutamate through glutamine does not by itself replace all transmitter carbon lost through oxidation or other pathways. Astrocytic metabolism contributes both to recycling existing molecules and to maintaining the precursor pools from which new transmitter molecules are made.\n\nAdditional support involves lipid handling and antioxidant defenses. Astrocytes synthesize and distribute cholesterol in lipoprotein particles, process fatty acids, and supply precursors that help neurons maintain glutathione. Under some forms of cellular stress, neurons can transfer potentially damaging lipids to astrocytes for storage or metabolism. These exchanges are regulated and context-dependent rather than evidence that neurons are metabolically self-insufficient. Astrocytes and neurons possess overlapping metabolic capabilities, but their different transporter expression, enzyme distribution, and structural relationships create complementary roles within neural tissue."
    },
    {
      "heading": "Vascular interfaces, blood flow, and fluid exchange",
      "text": "Astrocytic endfeet form extensive coverings around blood vessels in the central nervous system. They are components of the neurovascular unit, which also includes endothelial cells, pericytes, vascular smooth-muscle cells, neurons, and extracellular matrix. This arrangement links the metabolic needs of neural tissue to its vascular supply and places astrocytes at an important interface between blood-derived substances and the brain’s extracellular environment.\n\nThe physical seal of the blood–brain barrier is formed primarily by specialized endothelial cells and their tight junctions, not by astrocytic endfeet. Brain endothelial cells also have regulated transporter systems and unusually restricted vesicular transport. Astrocytes help maintain and modulate these properties through molecular signals and interactions with the surrounding tissue. During development, endothelial barrier properties begin to appear before mature astrocytic coverage is established, demonstrating that astrocytes are not the sole inducers of the barrier.\n\nNeurovascular coupling increases local blood flow in response to neural activity. Astrocytes can detect neuronal signals and generate substances capable of influencing vessel diameter, including arachidonic-acid derivatives. Depending on receptor pathways, oxygen availability, and vascular segment, such signaling can favor dilation or constriction. Neurons, endothelial cells, pericytes, and smooth muscle also contribute directly. Experimental results do not support a single obligatory astrocytic calcium pathway for every rapid vascular response, particularly at the earliest stages of functional hyperemia.\n\nPerivascular endfeet are molecularly polarized. Aquaporin-4 is concentrated in these membranes through interactions involving the dystrophin-associated protein complex and extracellular anchoring structures. This organization helps distinguish vessel-facing membranes from other parts of the same cell. Injury, vascular disease, and aging can alter this polarization. The consequences include changes in local water handling and associations with altered movement of extracellular solutes, although the functional effects depend on the experimental and disease setting.\n\nAstrocytes are also prominent in research on the glymphatic system, a proposed organization of cerebrospinal-fluid exchange and interstitial-solute clearance involving perivascular pathways. Experiments have linked aquaporin-4 expression and localization to aspects of tracer movement in several models. Important questions remain about the relative contributions of bulk flow, diffusion, vascular pulsation, sleep state, anesthesia, and experimental tracer delivery. Astrocytic involvement in fluid regulation is well established; a universal quantitative account of directional fluid transport through brain tissue remains under investigation. These distinctions are important when relating cellular observations to claims about clearance of proteins associated with neurodegenerative disease."
    },
    {
      "heading": "Calcium signaling and communication with neural circuits",
      "text": "Although astrocytes generally lack the fast electrical excitability characteristic of neurons, they exhibit rich intracellular signaling. Changes in cytosolic calcium concentration occur in cell bodies, major branches, fine processes, and perivascular endfeet. Some signals spread over substantial portions of a cell, whereas others remain confined to tiny subcellular domains. Local signals can be frequent even when calcium measurements from the cell body appear relatively quiet.\n\nMany extracellular substances act through astrocytic G protein-coupled receptors. Receptor activation can stimulate phospholipase C, producing inositol trisphosphate, which releases calcium from the endoplasmic reticulum. Calcium can also enter across the plasma membrane or arise through other intracellular mechanisms. Receptor expression changes during development: for example, the prominent metabotropic glutamate receptor signaling seen in some immature rodent astrocytes should not be assumed to persist unchanged in mature astrocytes.\n\nAstrocytic calcium signals can reflect both local synaptic activity and widespread behavioral state. Noradrenergic input from the locus coeruleus can produce coordinated astrocytic responses during arousal and movement. Cholinergic and other neuromodulatory systems also influence astrocytes. Consequently, a signal observed near an active synapse is not necessarily generated only by that synapse; it may combine local input with brain-wide changes in vigilance or behavior.\n\nThe concept of gliotransmission proposes that astrocytes release signaling substances that directly alter neuronal activity. Candidate substances include ATP, glutamate, and D-serine. Astrocytes can release molecules through several routes, including channels, transporters, and vesicular mechanisms under particular conditions. Extracellular ATP can also be converted to adenosine, which acts on neuronal receptors. However, the existence of a release mechanism in culture or after strong experimental stimulation does not establish its prevalence during normal activity in an intact adult brain.\n\nThe physiological extent and mechanisms of calcium-dependent gliotransmission remain debated. Some experimental manipulations alter neuronal excitability or synaptic plasticity, while others find little effect after disrupting major astrocytic calcium-release pathways. Interpretations depend on targeting specificity, developmental compensation, the calcium signals actually affected, and whether stimulation reproduces physiological concentrations and timing. In particular, eliminating one endoplasmic-reticulum calcium pathway does not remove every calcium event in fine astrocytic processes.\n\nAstrocytes can influence circuits without releasing a classical transmitter. Changes in glutamate uptake, potassium clearance, extracellular volume, metabolic supply, and synaptic coverage all modify neuronal communication. These established routes are essential context when assigning a neuronal effect specifically to gliotransmitter release."
    },
    {
      "heading": "Reactive states, injury responses, and disease",
      "text": "Astrocytes change their molecular expression, morphology, metabolism, and sometimes proliferative behavior after injury or disease. These changes are collectively termed reactive astrogliosis. Responses range from altered expression of a limited set of genes to marked cellular enlargement, proliferation, and formation of a compact border around damaged tissue. Increased GFAP staining is common but is neither a complete description of reactivity nor a reliable measure of its functional consequences.\n\nReactive states depend on the initiating insult, anatomical location, distance from a lesion, and time after injury. Signals from damaged neurons, vascular cells, invading immune cells, and microglia help shape the response. Cytokines, extracellular ATP, blood-derived proteins, and changes in tissue mechanics can all contribute. The formerly influential division into “A1” and “A2” astrocytes captured responses to particular experimental conditions but does not adequately classify the diversity found across diseases. Current descriptions increasingly specify the model, molecular features, and demonstrated function rather than assigning a universal beneficial or harmful identity.\n\nAfter severe focal injury, proliferating astrocytes can form a border that separates viable neural tissue from an inflammatory lesion core. Signaling through STAT3 is important for aspects of this organization in experimental spinal-cord injury. Such borders can limit inflammatory spread and help restore tissue boundaries. At the same time, components of the surrounding extracellular matrix can restrict axon extension. The astrocytic scar is therefore not simply an obstacle to repair: it combines protective containment with a tissue environment that may constrain regeneration.\n\nIn ischemic stroke, loss of oxygen and glucose compromises ATP production, disrupting ion gradients and glutamate clearance. Astrocyte swelling and altered vascular interactions contribute to the evolving injury, while surviving astrocytes also support tissue stabilization. In epilepsy, impaired potassium buffering, altered glutamate transport, and changes in gap-junction coupling are associated with abnormal circuit excitability. These mechanisms vary among epilepsy syndromes and experimental models; they do not imply that astrocyte dysfunction is the sole cause of seizures.\n\nSome diseases provide especially direct evidence of astrocytic involvement. Alexander disease is usually caused by pathogenic variants in the GFAP gene and features astrocytic protein aggregates called Rosenthal fibers. In aquaporin-4-antibody-positive neuromyelitis optica spectrum disorder, autoantibodies target aquaporin-4, producing an astrocyte-directed inflammatory disease that can lead to secondary damage of myelin and neurons. These conditions differ fundamentally from disorders in which astrocyte alterations are mainly secondary to another primary lesion.\n\nAstrocytes also show altered states in Alzheimer disease, amyotrophic lateral sclerosis, multiple sclerosis, and many other neurological disorders. Experimental systems demonstrate that particular astrocyte states can either support or compromise neuronal survival. Human tissue studies reveal associated molecular changes but often cannot establish whether those changes initiated disease, compensated for it, or resulted from tissue damage. Distinguishing those possibilities requires temporal and cell-specific evidence rather than marker expression alone."
    },
    {
      "heading": "Discovery and experimental investigation",
      "text": "The study of astrocytes emerged from nineteenth-century investigations of the nervous system’s non-neuronal tissue. Rudolf Virchow introduced the term neuroglia in 1856, before the diversity of glial cells was understood. Improved metal-impregnation methods later made individual glial cells visible. Michael von Lenhossék introduced the term astrocyte in the 1890s, and Santiago Ramón y Cajal described varied glial forms using refined staining techniques. These studies established morphological categories long before the underlying transport mechanisms and signaling pathways could be measured.\n\nTwentieth-century electron microscopy revealed astrocytic processes around synapses and blood vessels, challenging interpretations based only on the coarse branches visible with light microscopy. Electrophysiological recordings showed that astrocytes typically have strongly negative membrane potentials and electrical properties unlike those of neurons. Work on extracellular potassium, amino-acid transport, and transmitter metabolism then established their active contribution to the neural environment. The detection of stimulus-evoked calcium signals around 1990 helped broaden research from homeostatic support toward dynamic neuron–glia communication.\n\nModern structural studies combine fluorescent labeling, high-resolution light microscopy, and serial electron microscopy. Sparse labeling can reveal the shape of an individual cell, while membrane-targeted reporters capture processes that GFAP staining misses. Electron microscopy resolves narrow extracellular spaces and fine contacts but requires difficult reconstruction to determine which processes belong to which cell. Acute slices preserve much local architecture while disrupting long-range inputs and vascular physiology; cell cultures provide experimental access but often produce properties unlike those of mature astrocytes in intact tissue.\n\nGenetic tools allow researchers to label astrocytes, delete selected genes, or express sensors and actuators. Promoters associated with GFAP or ALDH1L1 are widely used, but specificity depends on the construct, developmental stage, brain region, and delivery method. GFAP-based strategies can also target neural stem cells or their descendants. Artificial receptor activation, optogenetic manipulation, and viral gene delivery therefore require direct validation of the cells affected and of the physiological process being altered.\n\nSingle-cell and single-nucleus RNA sequencing have identified regionally specialized and disease-associated expression patterns. Single-nucleus methods are useful for frozen human tissue, but nuclear RNA does not fully represent transcripts distributed through astrocytic processes. Tissue dissociation can also induce stress responses or preferentially lose fragile cells. Spatial transcriptomics and protein measurements help connect molecular profiles to anatomy, while functional experiments test whether an expression difference changes transport, signaling, or circuit behavior.\n\nHuman induced pluripotent stem cell-derived astrocytes and multicellular brain organoids provide complementary experimental systems. Their interpretation depends on maturation, regional identity, culture environment, and comparison with primary human tissue. Across these methods, a central requirement is to connect cell identity, subcellular localization, and measured function: a gene expressed in an astrocyte does not by itself establish where its protein acts or how much that mechanism contributes in the living brain."
    }
  ],
  "links": [
    "glia",
    "brain",
    "neuron",
    "synapse",
    "protein",
    "action-potential",
    "membrane-potential",
    "adenosine-triphosphate",
    "gap-junction",
    "aquaporin",
    "ph",
    "neurotransmitter",
    "glutamate",
    "phagocytosis",
    "glucose",
    "glycolysis",
    "glycogen",
    "citric-acid-cycle",
    "blood-brain-barrier",
    "endoplasmic-reticulum",
    "microglia",
    "autoantibody",
    "rna-sequencing"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}