Dalhousie University
   
 



Dr. David Hoskin
Natural Killer Cell Recognition of Target Cells

References:
KL McQueen and P Parham. Curr Opinion Immunol 14:615, 2002
L Lanier. Curr Opinion Immunol 15:308, 2003.
PJ Leibson. Immunity 6:655, 1997.

  • Natural killer (NK) cells were discovered by Swedish researchers looking for tumor-specific CTL in human peripheral blood. Surprisingly, normal controls had peripheral blood lymphocytes which were capable of lysing certain tumor cells!

NK activity

  1. is not restricted to target cells expressing only self MHC, i.e., NK cells kill allogeneic targets or targets with no class I MHC.
  2. does not require prior immunization.
  3. can occur in the presence (ADCC) or absence (natural cytotoxicity) of specific antibody.

NK cells express the following cell-surface markers:

  • LGL-1 (also have large granular lymphocyte morphology)
  • CD2
  • CD16 (FcgammaRIIIA)
  • CD56 and CD2 (humans)
  • asialoGM1 (mice)
The operational definition of an NK cell is the ability to lyse NK-sensitive target cells such as YAC-1 lymphoma cells (in mice) or K562 erythroleukemia cells (in humans). This operational definition has led to confusion and problems in the past prior to the identification of NK-specific cell-surface markers! Freshly isolated tumor cells are frequently resistant to NK-mediated killing, unless the NK cells are first exposed to IL-2 in which case they become LAK cells. IL-12 and interferons (alpha, ß, and gamma) will also activate NK cells.

  • NK cells are believed to be derived from a common T cell/NK cell progenitor.

    Stem cell
        • pre-B cell
        • NK/T progenitor
        • pre-T cell
        • pre-NK cell
  • This scheme is supported by studies which indicate that mature T cells are able to revert to an NK-like cell under certain conditions, eg., stimulation with anti-CD3 antibody plus IL-2. Also immature fetal thymocytes can be induced to differentiate into NK cells in vitro in the presence of PMA, IL-4 and IL-2. An extrathymic environment seems important for NK development. NK cells arise in the BM, most likely from prothymocytes that do not go to thymus.

Functions of NK cells:

  1. Lyse malignant tumor cells (transformed white blood cells are especially susceptible).
  2. Lyse virally-infected cells, limiting viral spread until CTL are activated.
  3. Kill certain bacteria (either directly or indirectly by the secretion of cytokines, i.e., IFNgamma, the release of which leads to macrophage activation).
  4. Inhibit hematopoietic colony formation, apparently by TNFalpha and IFN-gamma release.
  5. Inhibit antibody production (but do not actually kill B cells), probably by cytokine release, i.e., IFN-gamma would shut down the Th2 response.
  6. Promote Th1 responses and CTL development through IFN-gamma release following exposure to IL-12 or stimulation by target cells.
  • The primary function of NK cells may be to prevent tumor metastases, most of which occur through the blood. This hypothesis is supported by a correlation between NK deficiency and high metastatic frequency of certain tumors in beige mice (lack functional NK cells due to a defect in formation, fusion, or trafficking of granules). The human equivalent of the beige mutation is Chediak-Higashi Syndrome. Both beige mice and individuals with Chediak-Higashi Syndrome are prone to develop leukemias and lymphomas, suggesting an important role for NK cells in immune surveillance of these types of cancer.

NK cell cytolytic process:

  1. NK conjugation to target cells and strengthening/stabilization of conjugates involves adhesion molecules such as LFA-1 and CD2.
  2. Triggering of NK cytolytic machinery involves protein kinase C- and protein tyrosine kinase-dependent (phosphatidylinositol 3-kinase-independent) signaling mediated through NK activating receptors; protein tyrosine kinase- and phosphatidylinositol 3-kinase-dependent (protein kinase C-independent) signaling mediated through CD16.
  3. Delivery of lethal hit involves granule exocytosis and release of perforin, granzymes and other cytocidal factors. Activated NK cells express Fas ligand (can therefore kill Fas-bearing targets) as well as Fas, providing a potential mechanism to downregulate NK responses.
  4. NK recycling and target cytolysis.

How do NK cells recognize tumor targets?

  • NK recognition of tumor target cells is currently believed to be the result of a delicate balance between positive signals that initiate NK effector function and negative signals that prevent cytolysis. The triggering structures recognized on tumor cells by NK cells are believed to be glycoproteins. There is ample evidence that transformed cells exhibit aberrant glycosylation of surface proteins. The negative signals come from NK cell recognition of MHC class I molecules. This recognition event prevents NK cell-mediated cytolysis. Conversely, cells which down-regulate or lack MHC class I molecules (e.g., some virally infected cells and metastatic tumor cells, possibly selected for by CTL responses) are susceptible to NK cell-mediated killing.
  • NK receptors redistribute to the interface between the NK cell and the target cell when an appropriate MHC class I ligand is present for an inhibitory receptor. Protein kinases as well as phosphatases localize to the intercellular interface that is also rich in lipid rafts.

NK cell activating receptors:

  • There are 4 general categories of NK cell activating receptors which may all be expressed by a single NK cell. This allows the NK cell to be activated by many different target cell types. Unlike the T cell receptor, NK cell activating receptors are coded for by non-rearranging genes. There does not appear to be any dominant stimulatory receptor on NK cells. Instead, activation seems to be an additive process.
  1. Molecules that are adhesion or costimulatory molecules on other hematopoietic cells may alone or in combination with other molecules trigger NK cell activation.

    LFA-1
    • binds to ICAM-1/ICAM-2/ICAM-3 on target cells.
    • NK cell recognition of certain targets is blocked by anti-LFA-1 mAb; LAD patients (lack ß2 integrins) are NK-defective.
    • signalling is through associated protein tyrosine kinases.

    CD2
    • binds to LFA-3 (CD58).
    • may represent an alternative form of NK cell-target cell conjugate stabilization for certain targets, i.e., instead of LFA-1/ICAM interactions.
    • involved in NK signalling through associated protein tyrosine kinases.

  2. Activating isoforms of human killer cell inhibitory receptors (KIRs) and murine Ly-49 are expressed by some NK cells. These molecules differ from their inhibitory counterparts by lacking inhibitory motifs (ITIMs) in their shorter cytoplasmic domains and possessing a charged transmembrane region that associates with additional signal-transducing elements such as disulfide-linked dimers of DAP12 (an ITAM-bearing 13 kDa protein similar to the CD3 zeta chain of the T cell receptor/CD3 complex). Like the CD3 zeta chain, once DAP12 is phosphorylated by protein tyrosine kinases it recruits ZAP70 and syk protein tyrosine kinase.
  3. Molecules which are dedicated recognition/activation structures.

    NKR-P1
    • involved in signal transduction in rat NK cells since antibodies to NKR-P1 stimulate BLT esterase release and reverse ADCC against NKR FcR+ targets.
    • 60 kD disulfide-linked homodimer that is a C-type lectin with specificity for carbohydrate ligands. Thus, NKR-P1 is a type II integral membrane protein like Ly-49.
    • binds to carbohydrate ligands on NK-sensitive targets and triggers NK activation. Note that many tumor cells exhibit aberrant glycosylation patterns.
    • NK 1.1 is one of several analogous structures expressed by mouse NK cells.
    • genes coding for NKR-P1 in mice (at least 3) exhibit allelic polymorphism.
    • not involved in conjugate stabilization. This process requires adhesion molecules such as LFA-1 and CD2.

    NKG2D
    • a member of the C-type lectin family expressed by most human NK cells and some T cells.
    • - does not associate with CD94 or use the DAP12 adaptor molecule for signaling
    • expressed as a homodimer in association with a DAP10 signaling subunit (contains an SH2 domain-binding site which is capable of recruiting the p85 subunit of phosphatidylinositol 3-kinase and Grb2 upon phosphorylation).

    • ligands are MICA and MICB, which are distant relatives of class I MHC, and certain cytomegalovirus proteins. MICA and MICB have conserved alpha1 and alpha2 domains but do not bind peptide or associate with ß2 microglobulin. MICA and MICB expression is induced by cellular stress. Normal tissues, with the exception of intestinal epithelium, express minimal MICA and MICB. However, MICA and MICB expression is induced by viral infection and malignant transformation.

    • the murine form of NKG2D recognizes Rae-1 and H60 molecules since there is no mouse equivalent to MICA and MICB.

    Natural cytotoxic receptors (NCR)
    • designated NKp30, NKp44, NKp46, and NKp80.
    • all are members of the immunoglobulin superfamily, and with the exception of NKp80 (also expressed on a T cell subset) are selectively expressed on human NK cells.
    • ligands for NKp44 and NKp46 are the influenza and sendai virus hemagglutinins.

    • do not themselves have signal transducing capability, i.e., the cytoplasmic domain lacks ITAMs but binds to various adaptor signaling molecules via a charged residue in the transmembrane domain (NKp30 associates with the CD3 zeta chain; NKp44 associates with DAP12; NKp46 is coupled to the CD3 zeta chain and FcRIgamma chain, both of which are able to recruit ZAP-70 and syk protein tyrosine kinases).

  4. CD16 remains the best characterized NK activating receptor.
    • important for NK cell-mediated ADCC and cytokine production.
    • associated with a disulfide-linked homodimer or heterodimer consisting of the zeta chain and/or gamma chain of CD3 (role in signal transduction).
    • ITAMs (immunoreceptor tyrosine-based activation motifs) in the zeta and gamma subunits are responsible for generating second messengers involved in signal transduction. ITAMs associated with the gamma subunit appear to be critical since gamma-deficient mice (but not zeta-deficient mice) are not able to mediate ADCC.
    • Src-family and Syk-family protein tyrosine kinases are involved in CD16 signalling. This process involves tyrosine phosphorylation of ITAMs and phospholipase C-gamma. Once activated, phospholipase C-gamma cleaves membrane phosphoinositides to generate inositol trisphosphate and diacyl glycerol.
    • inositol trisphosphate induces release of calcium from intracellular stores. Elevated intracellular calcium is required for granule exocytosis and the assembly of functional NFAT involved in cytokine gene transcription.
    • diacyl glycerol activates protein kinase C (not required for ADCC but needed for cytokine gene transcription).
    • phosphatidylinositol 3-kinase is also activated and is critical for ADCC (but is not needed for natural cytotoxicity).

MHC class I-specific inhibitory receptors:

  • Four families of candidate molecules have been described; Ly-49 (mice) and LIRs (Leukocyte Immunoglobulin-like Receptors; in humans), KIRs (p58 and p70 Killer-cell Immunoglobulin-like Receptors; in humans), and CD94/NKG2 (in mice and humans). See figures for structure.
  • All use a common inhibitory mechanism involving ITIMs (which are tyrosine-phosphorylated upon ligation and crosslinking of the receptor by MHC I on the target cell) and recruitment of a tyrosine phosphatases (SHP-1 and SHP-2) to the phosphorylated ITIMs resulting in the inhibition of proximal protein tyrosine kinases (PTK) involved in NK activation through the activating receptors.
  • An individual NK cell may therefore express several different inhibitory receptors which function independently of each other. Note that triggering receptors coupled to DAP12 also exist and form pairs with inhibitory receptors.

What is Ly-49?

  • Type II membrane glycoprotein with extensive homology to C-type lectin superfamily.
  • Disulfide linked homodimer of 44 kDa subunits coded for by a family of genes located on chromosome 6.
  • At least 10 possible variants with specificity for different allelic forms of class I MHC.
  • Diversity provided by alternate mRNA splicing and allelic polymorphism.
  • Evidence of a requirement for carbohydrate for binding.
  • Best studied family member is Ly-49A (on 15-20% of NK cells) which binds to alpha1/alpha2 domains of H-2Dd and delivers a dominant negative signal to the NK cell. Ly-49C, on the other hand, binds H-2Kb. Each NK cell can express several different Ly-49 receptors.
  • Some Ly-49 isoforms (Ly-49D, Ly-49H) are coupled to DAP12 and, therefore, have activating function.
What is CD94/NKG2?
  • Both subunits are glycoproteins and members of the C-type lectin superfamily with homology to NKR-P1 and Ly-49.
  • CD94 covalently assembles with NKG2 family members (NKG2A, NKG2C, NKG2E) to form a heterodimer.
  • Ligand in humans is HLA-E, a non-classical class I MHC molecule; ligand in mice is a specific peptide presented by Qa-1, another non-classical class I MHC molecule.
  • NKG2A bears ITIMs and has inhibitory activity. In contrast, NKG2C and NKG2E lack ITIMs and are coupled to DAP12 via charged transmembrane residues, i.e., are stimulatory rather than inhibitory.

What are KIRs (Killer-cell Immunoglobulin-like Receptors)?

  • Type I membrane glycoproteins belonging to the immunoglobulin superfamily.
  • No homology to mouse Ly-49 gene products.
  • Consists of a single chain molecule of 58 kDa (2 Ig-like loops) or 70 kDa (3 Ig-like loops).
  • Coded for by a family of genes located on chromosome 19 which produce at least 12 different proteins, each of which presumably has allelic specificity for HLA-A, -B, or -C.
  • Carbohydrates are not required for receptor function.
  • Like Ly-49, KIRs bind to alpha1 and alpha2 domains of class I MHC.
  • Inhibitory KIRs have long cytoplasmic tails containing 2 ITIM motifs whereas activating KIRs have short cytoplasmic tails, lack ITIMs, and possess charged transmembrane amino acid residues which allow association with DAP12.

What are LIRs (Leukocyte Immunoglobulin-like Receptors)?

  • Proteins with 2 or 4 immunoglobulin-like domains.
  • Some contain cytoplasmic ITIMs whereas others lack ITIMs but contain charged transmembrane residues that may be associated with activating function.
  • Are expressed on NK cells, T cells, B cells, and macrophages.

Why should humans and mice have evolved very different molecules to perform the same function?

  • A human Ly-49-like gene (Ly-49L) has recently been identified that is inactive as the result of a point mutation. However, the corresponding gene in baboons and cows is active. In mice, gp49A and gp49B are close relatives to human KIRs, and are expressed on activated NK cells. Gp49B contains an ITIM and functions as an inhibitory receptor. True KIRs are found in species other than humans, suggesting that this gene family is older than primates. Members of each family may be specific for certain MHC class I loci or they may be functionally redundant. In any case, an individual NK cell may express several different inhibitory receptors.

 

 

 
   
 
 
Dalhousie University Faculty of Medicine Department of Microbiology