Natural killer (NK) cells are a population of leukocytes that, like T‐ and B‐cells, employ receptors that can provoke their activation, the consequences of which are the secretion of cytokines, most notably IFNγ, and the delivery of signals to their target cells via Fas ligand or cytotoxic granules that are capable of kill ing the cell that provided the activation signal (Figure 1.40 and Figure 1.41; see also Videoclip 3).
However, in addition to activating NK receptors, NK cells also possess receptors that can inhibit their function. As we shall see, inhibitory NK cell receptors are critical to the correct functioning of these cells as these receptors are what prevent NK cells from indiscriminately attacking healthy host tissue. Let us dwell on this for a moment because this is quite a different set‐up to the one that prevails with T‐ and B‐cells. A T‐ or B‐lymphocyte has a single type of receptor that either recognizes antigen or it doesn’t. NK cells have two types of receptor: activating receptors that trigger cytotoxic activity upon recognition of ligands that should not be present on the target cell, and inhibitory receptors that restrain NK killing by recognizing ligands that ought to be present. Thus, NK cell killing can be triggered by two different situations: either the appearance of ligands for the activating receptors or the disappearance of ligands for the inhibitory receptors. Of course, both things can happen at once, but one is sufficient.
We have already discussed NK cell‐mediated killing in some detail in Chapter 1, here we will focus on how these cells select their targets as a consequence of alterations to the normal pattern of expression of cell surface molecules, such as classical MHC class I molecules, that can occur during viral infection. NK cells can also attack cells that have normal expression levels of classical MHC class I but have upregulated nonclassical MHC class I‐related molecules because of cell stress or DNA damage.
NK cells express diverse “hard‐wired” receptors
Unlike the antigen receptors of T‐ and B‐lymphocytes, NK receptors are “hard‐wired” and do not undergo V(D)J recombination to generate diversity. As a consequence, NK cell receptor diversity is achieved through gene duplication and divergence and, in this respect, resembles the pattern recognition receptors we discussed in Chapter 1. Thus, NK receptors are a somewhat confusing ragbag of structurally disparate molecules that share the common functional property of being able to survey cells for normal patterns of expression of MHC and MHC‐related molecules. NK cells, unlike αβ T‐cells, are not MHC‐restricted in the sense that they do not see antigen only when presented within the groove of MHC class I or MHC class II molecules. On the contrary, one of the main functions of NK cells is to patrol the body looking for cells that have lost expression of the normally ubiquitous classical MHC class I molecules; a situation that is known as “missing‐self ” recognition (Figure 4.17). Such abnormal cells are usually either malignant or infected with a microorganism that interferes with class I expression.
We saw in Chapter 1 that many pathogens activate PRRs such as Toll‐like receptors that induce transcription of interferon‐regulated factors, which subsequently direct the transcription of type I interferons (IFNα and IFNβ). PRRs, such as TLR3, TLR7–9 and the RIG‐like helicases, that reside within intracellular compartments are particularly attuned to inducing the expression of type I interferons (see Figure 1.16). Such PRRs typically detect long single or double‐stranded RNA molecules that are characteristically produced by many viruses. One of the downstream consequences of interferon secretion is the cessation of protein synthesis and consequent downregulation of, among other things, MHC class I molecules. Thus, detection of PAMPs from intracellular viruses or other intracellular pathogens can render such cells vulnerable to NK cell‐mediated attack. Which is exactly the point? Many intracellular pathogens also directly interfere with the expression or surface exposure of MHC class I molecules as a strategy to evade detection by CD8+ T‐cells that survey such molecules for the presence of nonself peptides.
Figure 4.17 Natural killer (NK) cell‐mediated killing and the “missing‐self” hypothesis. (a) Upon encounter with a normal autologous MHC class I‐expressing cell, NK inhibitory receptors are engaged and activating NK receptors remain unoccupied because no activating ligands are expressed on the target cell. The NK cell does not become activated in this situation. (b) Loss of MHC class I expression (“missing‐self”), as well as expression of one or more ligands for activating NK receptors, provokes NK‐mediated attack of the cell via NK cytotoxic granules. (c) Upon encountering a target cell expressing MHC class I, but also expressing one or more ligands for activating NK receptors (“induced‐self”), the outcome will be determined by the relative strength of the inhibitory and activating signals received by the NK cell. (d) In some cases, cells may not express MHC class I molecules or activating ligands and may be ignored by NK cells, possibly owing to expression of alternative ligands for inhibitory NK receptors.
Because of the central role that MHC class I molecules play in presenting peptides derived from intracellular pathogens to the immune system, it is relatively easy to understand why these molecules may attract the unwelcome attentions of viruses or other uninvited guests planning to gatecrash their cellular hosts. It is probably for this reason that NK cells coevolved alongside MHC‐restricted T‐cells to ensure that pathogens, or other conditions that may interfere with MHC class I expression and hence antigen presentation to αβ T‐cells, are given short shrift. Cells that end up in this unfortunate position are likely to soon find themselves looking down the barrel of an activated NK cell. Such an encounter typically results in death of the errant cell as a result of attack by cytotoxic granules containing a battery of proteases and other destructive enzymes released by the activated NK cell.
NK receptors can be activating or inhibitory
NK cells play an important role in the ongoing battle against viral infection and tumor development and carry out their task using two sets of receptors: activating receptors, which recognize molecules that are upregulated on stressed or infected cells, and inhibitory receptors that recognize MHC class I molecules or MHC‐related molecules that monitor the correct expression of classical MHC class I molecules. It is the balance between inhibitory and activating stimuli that will dictate whether NK‐mediated killing will occur (Figure 4.17).
Several structurally distinct families of NK receptors have been identified: including the C‐type lectin receptors (CTLRs) and the Ig‐like receptors. Both receptor types include inhibitory and activating receptors (Table 4.3). Those that are inhibitory contain ITIMs (immunoreceptor tyrosine‐based inhibitory motifs) within their cytoplasmic tails that exert an inhibitory function within the cell by recruiting phosphatases, such as SHP‐1, that can antagonize signal transduction events that would otherwise lead to release of NK cytotoxic granules or cytokines (Figure 4.17). Activating receptors, on the other hand, are associated with accessory proteins, such as DAP‐12, that contain positively acting ITAMs within their cytoplasmic tails that can promote events leading to NK‐mediated attack. Upon engagement with their cognate ligands (MHC class I molecules), inhibitory receptors suppress signals that would otherwise lead to NK cell activation. Cells that lack MHC class I molecules are therefore unable to engage the inhibitory receptors and are likely to suffer the consequences (Figure 4.18).
NK receptors are highly diverse and, as this is an area of active investigation, we will make some necessary generalizations.
The main class of MHC class I‐monitoring receptors in the mouse is represented by the Ly49 multigene family of receptors, which contains approximately 23 distinct genes: Ly49A to
W. These receptors are expressed as disulfide‐linked homodimers, with each monomer composed of a C‐type lectin domain connected to the cell membrane via an α‐helical stalk of ∼ 40 amino acids (Figure 4.18a). Each NK cell expresses from one to four different Ly49 genes. Individual Ly49 receptors recognize MHC class I molecules in a manner that is, in most cases, independent of bound peptide. Ly49 dimers make contact with MHC class I molecules at two distinct sites that do not significantly overlap with the TCR‐binding area on the MHC (Figure 4.18e).
Killer immunoglobulin‐like receptors
Rather remarkably, humans do not use Ly49‐based receptors to carry out the same task, but instead employ a functionally equivalent, but structurally distinct, set of receptors for this purpose, the killer immunoglobulin‐like receptors (KIRs) (Figure 4.18c,d). This is a good example of convergent evolution, where unrelated genes have evolved to fulfill the same functional role. By contrast with the mode of binding to MHC displayed by the Ly49 receptors, the KIRs make contact with MHC class I molecules in an orientation that resembles the docking mode of the TCR, where contact with bound peptide is part of the interaction. However, it is worth emphasizing that although KIRs do make contact with peptide within the MHC class I groove, these receptors do not distinguish between self and nonself peptides as TCRs do.
NK cells also use members of the CD94/NKG2 family, which belong to the CTLR class of receptor, that are present in human, rat, and mouse genomes. CD94/NKG2A heterodimers, which are inhibitory receptors, can indirectly monitor the expression of MHC class I proteins by interacting with an invariant MHC‐related molecule called HLA‐E (human) and Qa‐1 (mouse), the surface expression of which is dependent on the proper synthesis of the main MHC class I A, B, and C proteins as will be discussed in more detail below. If normal levels of HLA‐E are detected, the inhibitory receptors will suppress NK attack. CD94/NKG2 heterodimers are expressed on most NK cells as well as γδ T‐cells.
This receptor system indirectly monitors the expression of MHC class I molecules in a rather ingenious way. The MHC class I‐related molecules HLA‐E/Qa‐1 are notable for the fact that they mainly bind invariant peptides that are found in the leader sequences (amino acids 3–11) of the classical MHC class I A, B, and C molecules. In the absence of the leader sequences from these peptides, HLA‐E and Qa‐1 are not expressed on the cell surface, thereby triggering NK attack. Because many microbial agents, particularly viruses, antagonize the expression of MHC class I molecules, monitoring the expression level of such molecules is a neat way of indirectly detecting that all is not well.
Another member of this receptor family, NKG2D, does not associate with CD94 and instead forms NKG2D/NKG2D homodimers, which are activating receptors. NKG2D homodimers recognize the MHC‐related proteins, MHC class I chain‐related A chain (MICA) and the related MICB, as well as UL16‐binding proteins in human and the homologous H60/ RAE‐1/MULT‐1 proteins in mice. These ligands become upregulated in damaged or stressed cells as will be elaborated upon later.
Natural cytotoxicity receptors
Additional NK receptors that belong to the Ig‐like class are the natural cytotoxicity receptors, which include NKp30, NKp44, and NKp46, all of which are activating receptors. The ligands for these receptors remain unclear but there is some evidence that they can detect certain viral products, such as hemagglutinin of influenza virus or Sendai virus and may also be sensitive to altered patterns of heparan sulfate on the surfaces of tumors. BAT‐3 (HLA‐B associated transcript‐3), a protein that has been implicated in DNA damage response pathways, has also recently been suggested to be a ligand for NKp30.
Figure 4.18 NK receptors. (a) Schematic representation of an inhibitory Ly49 receptor dimer composed of two C‐type lectin domains (CTLDs). The cytoplasmic tails of inhibitory Ly49 receptors contain immunoreceptor tyrosine‐based inhibitory motifs (ITIMs) that can recruit phosphatases, such as SHP‐1, capable of antagonizing NK activation. Activating Ly49 receptors lack ITIMs and can associate with ITAM‐containing accessory proteins such as DAP‐12 that can promote NK cell activation. (b) C‐type lectin‐like domain of the Ly49 NK cell receptors. The three‐dimensional structure shown is the dimeric Ly49A (Protein Data Bank entry code 1Q03), the monomer A is colored blue and the monomer B is colored green. For clarity, secondary structural elements α‐helices, β‐strands, disulfide bonds and N and C termini are labeled only on one monomer. (Source: Dr. Nazzareno Dimasi. Reproduced with permission.) (c) The human KIRs (killer immunoglobulin‐like receptors) are functionally equivalent to the murine Ly49 receptors but remain structurally distinct. These receptors contain two or three Ig‐like extracellular domains and can also be inhibitory or activating depending on the presence of an ITIM motif in their cytoplasmic domains, as shown. Activating receptors can associate with the ITAM‐bearing DAP‐12 accessory complex to propagate activating signals into the NK cell that result in NK‐mediated attack. (d) Structure of the extracellular Ig‐like domains (D1 and D2) of a KIR receptor. (Source: Dr Peter Sun. Reproduced with permission.) (e) Ribbon diagram of the crystal structure of the Ly49C/H‐2Kb complex. Ly49C, the H‐2Kb heavy chain, and β ‐microglobulin (β M) are shown in red, gold, and green, respectively. The MHC‐bound peptide (gray) is drawn in ball‐and‐stick representation. (Source: Dr. Lu Deng and Professor Roy A. Mariuzza. Reproduced with permission.)
CD16 Fc receptors
Another example of an activating NK receptor is CD16, the low‐affinity Fc receptor for IgG that is responsible for anti body‐dependent cellular cytotoxicity (ADCC). In this case, the receptor ligand is IgG bound to antigen present on a target cell, which is clearly an abnormal situation.
Cell stress and DNA damage responses can activate NK cells
Cellular stress, such as heat shock, is also a matter for concern for cells of the immune system as this can also be caused by infection, or alternatively, such cells may be undergoing malig nant transformation. The HLA‐E/Qa‐1 system, which as we discussed earlier is involved in monitoring the ongoing expres sion of MHC class I proteins, is also involved in attracting the attentions of NK cells in the context of cell stress. In response to diverse forms of cellular stress, heat‐shock proteins such as HSP‐60 are induced and peptides derived from the HSP‐60 leader peptide can displace MHC class I‐derived peptides from the HLA‐E peptide‐binding cleft. Although HLA‐E/HSP‐60 peptide complexes are trafficked to the cell surface, they are no longer recognized by CD94/NKG2 heterodimers, which results in NK activation due to “missing self.”
In addition to recognizing “missing‐self,” NK cells also use their receptors to directly recognize pathogen components or nonclassical MHC class I‐like proteins, such as MICA and MICB, which are normally poorly expressed on normal healthy cells. MICA, and related ligands, have a complex pattern of expression but are often upregulated on transformed or infected cells and this may be sufficient to activate NK receptors that are capable of delivering activating signals, a phenomenon that has been termed “induced‐self” recognition (Figure 4.17). Upon ligation, the activating receptors signal the NK cell to kill the target cell and/or to secrete cytokines. The potentially anarchic situation in which the NK cells would attack all cells in the body is normally prevented because of the recognition of MHC class I by the inhibitory receptors. Thus, normal pat terns of MHC class I expression suppress NK killing, whereas the presence of abnormal patterns of self molecules induce NK activation. It is the relative intensity of these signals that determines whether an attack will occur.
Recent studies also suggest that checkpoint kinases, such as Chk1, that are involved in the DNA damage response can induce expression of a variety of activating ligands for NKG2 receptors, when a cell is damaged by γ‐irradiation, or after treatment with DNA‐damaging drugs. This suggests that cells that have suffered DNA damage may, in addition to activating their DNA repair machinery, also upregulate NK receptor ligands to alert the immune system. This makes perfect sense, as such cells are dangerous as they have the potential to escape normal growth controls and form a tumor owing to faulty or incomplete DNA repair. Indeed, tumor surveillance is thought to be one of the major roles of NK cells, a topic we will revisit again in Chapter 16.