{
  "id": "toll-like-receptor",
  "title": "Toll-like Receptor",
  "lang": "en",
  "source_lang": "en",
  "summary": "Toll-like receptors are innate immune sensors that recognize microbial and damage-associated molecules and initiate inflammatory and antimicrobial responses.",
  "category": "science",
  "subcategory": "biology",
  "type": "concept",
  "aliases": [
    "Toll-like receptors",
    "TLR",
    "TLRs"
  ],
  "generated_at": "2026-11-03",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/toll-like-receptor/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/toll-like-receptor/",
      "json": "https://aiwiki.page/zh/toll-like-receptor.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "A Toll-like receptor (TLR) is a member of a family of membrane-spanning proteins that detect molecular features associated with microorganisms and cellular injury. TLRs belong to the broader class of pattern-recognition receptors and are central to innate immunity. Their activation connects the recognition of a potential threat to changes in gene expression, production of cytokines, and mobilization of antimicrobial defenses. Unlike the antigen receptors of adaptive immune cells, their recognition properties are encoded in inherited genes rather than generated by somatic rearrangement.\n\nIn humans, ten family members, TLR1–TLR10, are recognized. Some operate mainly at the cell surface, where they detect microbial membrane components or proteins; others operate in intracellular compartments, where they recognize nucleic acids and their breakdown products. TLRs are particularly important in macrophages, dendritic cells, and other immune cells, but they also function in epithelial and stromal tissues. Their discovery established a molecular explanation for how early recognition of infection helps initiate adaptive immunity. Their activities also contribute to inflammatory disease, inherited susceptibility to infection, and the effects of several vaccine adjuvants and immune-modifying medicines."
    },
    {
      "heading": "Discovery and evolutionary relationships",
      "text": "The name derives from Toll, a gene first characterized through studies of embryonic development in the fruit fly Drosophila melanogaster. Work on maternal-effect mutations established that the Toll pathway helps organize the embryo’s dorsal–ventral axis. The same signaling machinery later proved important in immunity. In 1996, Bruno Lemaitre and colleagues showed that disruption of the Drosophila Toll pathway impaired antifungal defense, connecting a developmental receptor system with resistance to infection.\n\nA crucial distinction separates this insect pathway from many mammalian TLR systems. Drosophila Toll is activated by the host protein Spätzle, which is proteolytically processed through extracellular recognition cascades. It does not itself serve as a direct receptor for the major microbial structures that initiate those cascades. Many mammalian TLRs, by contrast, participate directly in complexes that bind microbial molecules. Conservation therefore concerns the receptor architecture and signaling logic, not an identical recognition mechanism in every organism.\n\nIn 1997, Ruslan Medzhitov, Paula Preston-Hurlburt, and Charles Janeway described a human Toll-related receptor, subsequently designated TLR4, and showed that an activated form could induce inflammatory signaling and expression of a costimulatory molecule. This provided an experimental bridge between innate recognition and the activation of lymphocytes. In 1998, Alexander Poltorak and colleagues identified mutations in Tlr4 that explained the defective response of certain mouse strains to bacterial endotoxin. These findings established TLR4 as a necessary component of the endotoxin-recognition system.\n\nThe broader conceptual framework had been advanced by Janeway in 1989: inherited receptors should recognize conserved microbial features and instruct adaptive responses. TLR research supplied molecular examples of that proposal. The 2011 Nobel Prize in Physiology or Medicine recognized Bruce Beutler and Jules Hoffmann for discoveries concerning the activation of innate immunity; Ralph Steinman received the other half for discovery of the dendritic cell and its role in adaptive immunity.\n\nTLR repertoires have diversified during evolution. Humans have ten family members, whereas laboratory mice have twelve functional receptors: TLR1–TLR9 and TLR11–TLR13. Mouse Tlr10 is disrupted, and human TLR11 is a pseudogene. Fish, birds, and other vertebrates possess additional combinations, including receptors absent from mammals. Gene duplication, loss, and changes in recognition specificity make receptor numbering an imperfect guide to equivalent function across species. Comparative studies therefore distinguish shared ancestry from experimentally demonstrated ligand recognition."
    },
    {
      "heading": "Molecular architecture and receptor activation",
      "text": "TLRs are type I transmembrane glycoproteins. Each has an amino-terminal ligand-recognition region, a single membrane-spanning segment, and a carboxy-terminal signaling domain exposed to the cytoplasm. At the plasma membrane, the recognition region faces the extracellular space; in an endosome, it faces the compartment’s lumen. This arrangement allows the receptor to sample material outside the cytoplasm while communicating with intracellular signaling proteins.\n\nThe recognition region contains multiple leucine-rich repeats, sequence motifs that assemble into a curved, solenoid-like structure. Their repeating structural framework supports a large surface on which ligand-binding sites can evolve. Individual TLRs differ in the number and organization of repeats, the shape of their binding surfaces, and the accessory proteins needed for recognition. Glycosylation and terminal structural caps contribute to appropriate folding, stability, and receptor transport.\n\nThe intracellular region is a Toll/interleukin-1 receptor domain, usually abbreviated TIR domain. Related domains occur in interleukin-1 receptor family members and in cytoplasmic adaptor proteins. TIR domains organize signaling through selective protein interactions rather than through intrinsic catalytic activity. A TLR is therefore not a receptor kinase: its cytoplasmic domain recruits proteins that subsequently assemble kinase-containing signaling complexes.\n\nActivation generally involves formation or stabilization of a productive receptor dimer. The active complex brings the intracellular TIR domains into a geometry that permits adaptor recruitment. Some TLRs form homodimers, whereas TLR2 commonly forms heterodimers with TLR1 or TLR6. Ligands may bridge two receptor molecules, fit into complementary pockets, or induce rearrangements in pre-existing dimers. Dimerization alone is not always equivalent to activation; orientation and the accessibility of intracellular interaction surfaces are also important.\n\nStructural studies have made these principles concrete. In the TLR1–TLR2 complex, a triacylated bacterial lipopeptide places two lipid chains in a pocket in TLR2 and a third in a channel in TLR1. In the TLR4 complex, endotoxin binds chiefly to the accessory protein MD-2, and the ligand-loaded complex forms an activating assembly containing two TLR4 molecules and two MD-2 molecules. These examples show that “TLR recognition” may describe recognition by a multi-component receptor complex rather than direct binding to the TLR polypeptide alone.\n\nX-ray crystallography, biochemical reconstitution, mutational analysis, and cryo-electron microscopy have collectively established much of this architecture. An isolated ligand-binding structure nevertheless captures only part of the mechanism: membrane organization, trafficking, proteolysis, and assembly of intracellular signaling complexes also determine whether binding produces a cellular response."
    },
    {
      "heading": "Recognition at the cell surface",
      "text": "The best-characterized surface TLR systems detect bacterial lipids, lipid-linked peptides, and flagellar proteins. These molecules often have important structural or physiological roles in microorganisms, making them useful indicators of microbial presence. They are not unique markers of pathogenicity: harmless and commensal bacteria can carry the same structures. Whether detection produces a protective response or damaging inflammation depends on tissue location, microbial access, signal strength, and accompanying recognition pathways.\n\nTLR4 is the principal signaling receptor for lipopolysaccharide, or LPS, a major constituent of the outer membrane of Gram-negative bacteria. Its most potent stimulatory region is lipid A. Recognition normally requires an accessory system: LPS-binding protein helps extract and transfer LPS, CD14 facilitates delivery, and MD-2 provides the principal hydrophobic binding pocket associated with TLR4. A typical strongly stimulatory lipid A contains six acyl chains, but variations in acylation and phosphorylation markedly change biological activity. The same lipid A variant may also behave differently in human and mouse receptor systems.\n\nTLR2 works mainly with TLR1 or TLR6 to recognize microbial lipoproteins and lipopeptides. TLR2–TLR1 is classically associated with triacylated ligands, whereas TLR2–TLR6 recognizes many diacylated ligands. Accessory molecules, including CD14 and CD36 in appropriate settings, can influence ligand delivery and cellular responsiveness. These recognition systems contribute to responses against numerous Gram-positive and Gram-negative bacteria and some other microorganisms, rather than mapping neatly onto one taxonomic group.\n\nTLR5 recognizes flagellin, the principal protein subunit of many bacterial flagella. Its recognition surface includes regions constrained by the requirements of filament assembly. Such recognition illustrates how an immune sensor can target a microbial structure whose function limits the organism’s freedom to change it. Some bacteria nevertheless possess flagellins that stimulate TLR5 weakly, or regulate flagellin expression during infection.\n\nTLR10 is less completely characterized than these receptors. It can associate with TLR2, and experimental work has identified immunomodulatory, including inhibitory, effects in several systems. Its physiological ligands and functions are not as firmly established as those of TLR4 or TLR5, so it cannot be assigned a single universally accepted microbial target.\n\nInterpretation of ligand specificity requires attention to purity. Early claims that TLR2 directly recognizes many chemically unrelated bacterial products were sometimes revised when highly purified preparations became available. Contaminating lipoproteins can account for apparent activity attributed to other cell-wall constituents, including some peptidoglycan preparations. Receptor-deficient cells, defined synthetic ligands, and biochemical binding assays help distinguish direct recognition from contamination or indirect activation."
    },
    {
      "heading": "Endosomal recognition of nucleic acids",
      "text": "TLR3, TLR7, TLR8, and TLR9 are the major human TLRs specialized for recognizing nucleic acids in endosomal compartments. Their localization helps them encounter material taken up from microorganisms, infected cells, and extracellular debris. It also reduces routine exposure to the host’s own nucleic acids. This compartmental arrangement is a safeguard rather than an absolute distinction between self and non-self: host-derived material can stimulate these receptors if it reaches the appropriate compartment in a stimulatory form.\n\nTLR3 recognizes double-stranded RNA. Such RNA can occur during viral replication, but the receptor can also respond to RNA released from damaged cells. A sufficiently long RNA duplex provides a scaffold for receptor dimerization. The synthetic double-stranded RNA analogue polyinosinic:polycytidylic acid, usually called poly(I:C), is widely used experimentally. Its effects are not specific to TLR3, however, because delivery method and molecular properties can also permit detection by cytoplasmic RNA sensors.\n\nTLR7 and TLR8 detect products associated with single-stranded RNA degradation. Their specificity is more precise than the shorthand description “single-stranded RNA receptors” suggests. Human TLR7 recognizes guanosine together with particular short RNA fragments, whereas human TLR8 recognizes uridine together with short oligoribonucleotides. Endosomal ribonucleases help generate these stimulatory products. Both receptors also respond to certain synthetic small molecules, although ligand selectivity and cellular activity differ between them.\n\nTLR9 recognizes DNA containing stimulatory sequence contexts, classically described in terms of unmethylated CpG motifs. CpG frequency, sequence context, chemical modification, and DNA delivery all influence activity. Bacterial DNA often provides favorable stimulatory patterns, but TLR9 does not identify microbial origin with certainty. Mammalian DNA can activate it when concentrated, complexed with carrier molecules, or delivered into the relevant intracellular compartment.\n\nTrafficking is essential to all these systems. The membrane protein UNC93B1 escorts several nucleic-acid-sensing TLRs from the endoplasmic reticulum toward their functional destinations. Endosomal acidity and proteases support ligand processing and receptor maturation. Proteolytic cleavage is an important regulatory step for several family members; for TLR9, cleavage of the ectodomain is closely associated with signaling competence.\n\nSpecies differences are especially consequential in this group. Human and mouse TLR8 have substantially different response properties, so a compound active through human TLR8 need not reproduce the same effects in ordinary mouse experiments. Mouse TLR13, which humans lack, recognizes a conserved sequence in bacterial ribosomal RNA. Endosomal recognition therefore combines conserved principles of compartmentalization with lineage-specific receptor repertoires and ligand preferences."
    },
    {
      "heading": "Intracellular signaling pathways",
      "text": "TLR signaling is organized around two principal adaptor systems: the MyD88 pathway and the TRIF pathway. All human TLRs except TLR3 can signal through MyD88; TLR3 uses TRIF, and TLR4 can engage both systems. This division is useful but does not imply that every receptor produces the same output through a shared adaptor. Receptor location, adaptor availability, signaling kinetics, and cell identity strongly shape the resulting response.\n\nMyD88 contains a TIR domain that connects to receptor-associated complexes and a death domain that recruits members of the interleukin-1 receptor-associated kinase family. Assembly of MyD88 with IRAK4 and IRAK1 or IRAK2 generates an oligomeric signaling platform called the Myddosome. These proteins recruit and regulate additional components, including TRAF6. Ubiquitin-based scaffolds then help organize downstream kinase complexes. Ubiquitination in this context is not simply a signal for protein destruction; particular ubiquitin linkages create binding platforms that coordinate signal transmission.\n\nA major downstream branch activates the IκB kinase complex. This promotes removal of inhibitory IκB proteins and allows NF-κB transcription factors to enter the nucleus. Other branches activate mitogen-activated protein kinases, including p38, JNK, and ERK, which regulate AP-1 and additional transcriptional programs. Together these pathways induce inflammatory mediators, chemokines, antimicrobial factors, and molecules that support immune-cell activation.\n\nTRIF signaling is central to the production of type I interferons after TLR3 or TLR4 activation. TRIF recruits signaling machinery that includes TBK1 and IKKε, leading to activation of interferon regulatory factor 3, or IRF3. It also connects to inflammatory pathways, so its function is not restricted to interferon production. At TLR4, the adaptor TRAM helps connect the receptor to TRIF, whereas TIRAP helps recruit MyD88 at appropriate membrane sites.\n\nThe spatial organization of TLR4 illustrates how one receptor can generate distinct signals. Its plasma-membrane activity strongly supports TIRAP–MyD88 signaling, while internalization enables TRAM–TRIF signaling from endosomal compartments. Ligand delivery and receptor trafficking therefore influence both the magnitude and the character of the response.\n\nA different specialization occurs in plasmacytoid dendritic cells. Their TLR7 and TLR9 pathways use MyD88 to drive strong IRF7-dependent type I interferon production. Consequently, equating MyD88 exclusively with inflammation and TRIF exclusively with antiviral responses is misleading. Shared adaptors operate within different cellular environments, and those environments determine which transcription factors are available, how long signaling persists, and which genes can be activated."
    },
    {
      "heading": "Regulation, tolerance, and integration with other sensors",
      "text": "TLR responses must be constrained because microbial products are common at body surfaces and inflammatory mediators can injure host tissues. Regulation occurs before ligand binding, during receptor trafficking, within signaling complexes, and at the level of transcription and RNA stability. Receptor abundance is only one component. The accessibility of a receptor to its ligand, the availability of accessory proteins, and the state of downstream signaling machinery can be equally decisive.\n\nSeveral intracellular proteins limit signaling. IRAK-M restrains MyD88-associated responses in myeloid cells, while the ubiquitin-editing protein A20 helps terminate signaling pathways that activate NF-κB. SOCS proteins and other inhibitory regulators also influence responses in context-dependent ways. These mechanisms combine with receptor internalization, degradation, and changes in adaptor recruitment to prevent an initially useful signal from persisting indefinitely.\n\nRepeated exposure to LPS can produce endotoxin tolerance, in which cells respond less strongly to subsequent stimulation. This is not simply loss of receptor expression or generalized cellular inactivity. Different groups of genes are affected differently: many inflammatory cytokine genes become difficult to reactivate, while some antimicrobial and regulatory functions remain responsive. Changes in chromatin, transcription-factor recruitment, and metabolic state help establish this altered response. Endotoxin tolerance is a form of innate response reprogramming, distinct from antigen-specific immune tolerance.\n\nTLRs also work alongside other pattern-recognition systems. Cytoplasmic RIG-I-like receptors detect viral RNA, the cGAS–STING pathway detects DNA in the cytoplasm, and NOD-like receptors detect bacterial products or cellular disturbances. Responses to intact microorganisms therefore cannot usually be assigned to a single receptor. Signals may reinforce one another, induce different effector programs, or provide checks that distinguish tissue damage from an active infection.\n\nA particularly important relationship concerns the inflammasome. In many experimental systems, a TLR stimulus supplies a priming signal that increases expression of pro-interleukin-1β and the sensor NLRP3. A second stimulus, associated with events such as ion imbalance or membrane injury, promotes inflammasome assembly and proteolytic maturation of interleukin-1β. The two-signal model is useful but has cell-type- and species-specific exceptions; a TLR is not itself an inflammasome.\n\nAt mucosal surfaces, tissue organization further limits inappropriate activation. Epithelial polarity, mucus, antimicrobial products, and separation of microbes from receptor-accessible compartments all matter. A microbial molecule encountered in an intact intestinal lumen can have a different consequence from the same molecule entering damaged tissue. TLR activity therefore reflects both molecular identity and anatomical context."
    },
    {
      "heading": "Functions in host defense and adaptive immunity",
      "text": "TLR activation coordinates several early responses to infection. Macrophages and neutrophils increase production of inflammatory mediators and modify antimicrobial activities. Chemokines recruit additional leukocytes, while cytokines alter local blood vessels and influence systemic responses. Depending on the cell and stimulus, TLR pathways can support antimicrobial peptide production, enhanced killing mechanisms, or changes in cellular metabolism. These effects complement, rather than replace, recognition through phagocytic receptors and other innate sensors.\n\nA major consequence is activation of the dendritic cell. After appropriate stimulation, dendritic cells increase expression of antigen-presentation and costimulatory molecules and can migrate toward lymphoid tissues. They process microbial proteins and display derived peptides through major histocompatibility complex molecules. A T cell recognizing a peptide must also receive suitable accessory signals to develop an effective response. TLR-induced dendritic-cell maturation helps provide those signals.\n\nThe cytokine environment influences the character of adaptive immunity. For example, interleukin-12 produced by activated antigen-presenting cells promotes type 1 immune responses, important against many intracellular pathogens. Other combinations of signals support different T-cell programs. No single TLR specifies a fixed adaptive outcome under all conditions: ligand dose, delivery route, accompanying receptors, and the responding dendritic-cell subset all affect the result.\n\nTLRs can also act directly on B cells. In humans, TLR7 and TLR9 are important examples of receptors that can cooperate with B-cell antigen recognition. An antigen receptor can bind a molecular complex, internalize it, and deliver associated nucleic acid to an endosomal TLR. This provides a mechanism by which recognition of an antigen and recognition of microbial material reinforce one another. In suitable contexts, these signals support proliferation and antibody production; in other contexts, the same linkage contributes to autoreactivity.\n\nAntiviral defense illustrates both the importance and the redundancy of TLRs. TLR3, TLR7, TLR8, and TLR9 detect different forms of viral material and can initiate interferon responses. Yet many infected cells rely heavily on cytoplasmic sensors. The relative contribution of a TLR therefore depends on the virus, the tissue, and whether the responding cell is infected or instead has taken up material from another cell.\n\nHuman genetic deficiencies show that apparently broad sensing pathways may have unusually specific indispensable roles. Other receptors often compensate for their absence, but compensation can fail in particular tissues or infections. Protective immunity is thus produced by overlapping recognition systems whose redundancy is substantial but incomplete."
    },
    {
      "heading": "Inherited defects and inflammatory disease",
      "text": "Rare inherited defects have provided direct evidence for the roles of TLR pathways in human infection. Deficiencies of MyD88 or IRAK4 cause susceptibility to invasive pyogenic bacterial infections, particularly in childhood. Documented organisms include Streptococcus pneumoniae and Staphylococcus aureus. Inflammatory signs can be less pronounced than expected for the severity of infection because the affected pathways contribute to fever and inflammatory cytokine production. These disorders affect signaling shared by multiple TLRs and interleukin-1 receptor family members, rather than a single microbial receptor.\n\nDefects in the TLR3 pathway illustrate a more anatomically restricted vulnerability. Pathogenic variants in TLR3 and several associated signaling components have been identified in some patients with herpes simplex encephalitis. Experimental studies support a role for impaired local antiviral defense in cells of the central nervous system. The clinical pattern is not equivalent to generalized susceptibility to every virus, and incomplete penetrance means that a disease-associated variant does not inevitably produce encephalitis.\n\nTLR7 loss-of-function variants have been identified in a subset of males with severe COVID-19, consistent with its role in antiviral interferon production. Conversely, rare gain-of-function variants in TLR7 can cause lupus. These contrasting genetic observations demonstrate how insufficient and excessive activity in the same recognition pathway can produce different kinds of disease.\n\nIn systemic lupus erythematosus, complexes containing host nucleic acids can enter endosomal compartments through antibody-dependent uptake or B-cell antigen receptors. This allows self material to engage nucleic-acid-sensing TLRs. TLR7 is strongly implicated in this process. TLR9 has more complex effects, particularly in animal models, where it can support anti-DNA antibody production while also restraining some TLR7-associated disease features. It is therefore inaccurate to treat every nucleic-acid-sensing TLR as having an identical role in autoimmunity.\n\nTLR4 has been extensively studied in sepsis, where microbial recognition can contribute to dysregulated systemic inflammation and organ dysfunction. Sepsis nevertheless involves many interacting host pathways, pathogen differences, vascular changes, and metabolic disturbances. The importance of endotoxin sensing does not make TLR4 the sole cause, and blocking one receptor has not produced a generally effective treatment for this heterogeneous syndrome.\n\nSterile injury and chronic inflammatory conditions have also been associated with TLR activity. Claims that particular host-derived molecules directly bind and activate TLRs require careful scrutiny, because microbial contamination, carrier molecules, and indirect release of other mediators can produce similar experimental results. Human genetic evidence, reproducible biochemical mechanisms, and controlled functional studies provide stronger causal support than increased receptor expression alone."
    },
    {
      "heading": "Experimental tools, vaccine adjuvants, and therapeutic targeting",
      "text": "TLR research uses complementary experimental approaches. Cells engineered to express individual receptors can identify candidate agonists, while knockout models test whether a receptor is necessary for a response. Reporter assays measure downstream transcriptional activity, often through NF-κB or interferon-responsive elements. Cytokine measurements, imaging, biochemical interaction studies, and gene-expression profiling reveal additional consequences. No single assay establishes the entire recognition mechanism: downstream activation can arise indirectly, and receptor overexpression can alter normal signaling thresholds.\n\nDefined ligands are especially useful. Synthetic lipopeptides probe TLR2-containing complexes; ultrapure LPS probes TLR4; purified flagellin probes TLR5; and synthetic CpG oligonucleotides probe TLR9. Poly(I:C) is useful for studying double-stranded RNA recognition but can also activate cytoplasmic sensors. Imidazoquinoline compounds can stimulate TLR7 or TLR8, depending on their structure. Appropriate receptor-deficient controls and attention to species differences are essential for interpreting these reagents.\n\nAn important clinical application is the vaccine adjuvant, a component that enhances immune responses to an antigen. Monophosphoryl lipid A, a modified lipid A preparation with reduced toxicity relative to strongly inflammatory LPS, stimulates TLR4. It is included in licensed adjuvant systems such as AS04 and AS01. AS04 combines it with an aluminum salt, whereas AS01 incorporates it with the saponin QS-21 in a liposomal formulation. These combinations show that an adjuvant’s action depends on formulation and interactions among components, not merely the identity of one receptor agonist.\n\nCpG 1018, a synthetic TLR9 agonist, is used in a licensed adult hepatitis B vaccine. By providing a defined innate stimulus alongside antigen, such adjuvants help activate antigen-presenting cells and support antibody responses. Their effects also depend on antigen composition, dose, route of administration, and the populations studied.\n\nTLR agonists are also used outside vaccination. Imiquimod, a topical TLR7 agonist, is licensed for specified dermatological indications, including certain external genital warts, actinic keratoses, and superficial basal cell carcinomas. Its effects involve local immune activation rather than direct, selective destruction of a microbial target. Other TLR agonists have been investigated in cancer immunotherapy, including approaches intended to stimulate immune activity within tumors.\n\nAttempts to inhibit TLR signaling have faced a different problem: suppressing inflammation can also weaken protective recognition. Eritoran, an antagonist of the MD-2–TLR4 system, did not reduce mortality in a large phase III severe-sepsis trial. This result illustrates the difficulty of translating a well-defined molecular mechanism into treatment for a heterogeneous clinical condition, especially when disease timing, pathogen identity, and parallel inflammatory pathways differ among patients."
    }
  ],
  "links": [
    "protein",
    "pattern-recognition-receptor",
    "innate-immunity",
    "gene-expression",
    "cytokine",
    "cell-membrane",
    "adaptive-immunity",
    "gene",
    "evolution",
    "species",
    "glycosylation",
    "x-ray-crystallography",
    "cryo-electron-microscopy",
    "bacteria",
    "flagellum",
    "peptidoglycan",
    "rna",
    "dna",
    "endoplasmic-reticulum",
    "ribosomal-rna",
    "cell-nucleus",
    "interferon",
    "transcription-factor",
    "chromatin",
    "immune-tolerance",
    "neutrophil",
    "leukocyte",
    "dendritic-cell",
    "major-histocompatibility-complex",
    "t-cell",
    "b-cell",
    "antibody",
    "systemic-lupus-erythematosus",
    "sepsis",
    "vaccine-adjuvant",
    "cancer"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}