CTLs and NK cells mediate cytotoxicity

T cells and NK cells belong to the lymphoid lineage, and are more closely related to each other than they are to B cells. CTLs and NK cells both induce apoptosis in their targets by production of TNF family molecules, cytokines and cytotoxic granules, but they recognize their targets in different ways. CTLs recognize foreign antigens being presented on MHC class I, whereas NK cells respond to cells that fail to express class I. NK cells are also able to recognize stressed or antibody-coated targets directly (Fig. 10.1).

Fig. 10.1 CTLs and NK cells mediate cytotoxicity

CTLs recognize processed antigen presented on the target cell by MHC molecules, using their T cell receptor (TCR). Most CTLs are CD8⁺ and recognize antigen presented by MHC class I molecules, but a minority are CD4⁺ and recognize antigen presented on MHC class II molecules. In contrast, NK cells have receptors that recognize MHC class I on the target and inhibit cytotoxicity. NK cells also express a number of activating receptors to identify their targets positively, including NKG2D and their FC receptor (CD16).

Effector CTLs home to peripheral organs and sites of inflammation

Naive CTLs circulate in the blood and lymphatic system, but in order to kill specific target cells, they must be activated and become effector cells. This process is mediated by APCs in the lymph nodes. Naive T cells in the lymph nodes use their T cell receptor (TCR) to recognize specific antigens being presented by MHC molecules on APCs. Most CTLs are CD8⁺ and therefore recognize antigen presented on MHC class I molecules, although some CD4⁺ cells are cytotoxic and so recognize antigen presented on MHC class II molecules.

If, in addition to signaling through the TCR, the APC also delivers a costimulatory signal through CD28, the CTL becomes activated and proliferates. Activated CTLs downregulate sphingosine phosphate receptor, allowing them to exit the lymph node. They also downregulate molecules associated with homing to the lymph node, such as L-selectin and CCR7, and upregulate molecules that allow them to home to sites of inflammation, such as CD44 and LFA-1. Effector T cells may also express adhesion molecules that allow them to home to specific tissues. CD8⁺ memory T cells appear to preferentially migrate to the tissue in which they first encountered their antigen.

CTLs recognize antigen presented on MHC class I molecules

The most important role of CTLs is the elimination of virally infected cells. CTLs recognize specific antigens (e.g. viral peptides on infected cells) presented by MHC class I molecules, which are expressed by nearly all nucleated cells. Cellular molecules that have been partly degraded by the proteasome are transported to the endoplasmic reticulum, where they become associated with MHC class I molecules and are transported to the cell surface. Normal cells therefore present a sample of all the antigens they produce to CD8⁺ T cells.

Additional interactions may be required to stabilize the bond between a CTL and its target. Like CD4⁺ T cells, CD8⁺ CTLs form an immunological synapse with their target. Signaling molecules including the TCR and CD3 are found in the central zone of the supra-molecular activation cluster (cSMAC), while adhesion molecules segregate in the peripheral zone (pSMAC). In contrast to CD4⁺ T cells, the cSMAC of CTLs and NK cells is divided into signaling and secretory domains (Fig. 10.2). After signaling has occurred, the microtubule organizing center polarizes towards the synapse, directing cytotoxic granules to the secretory domain of the cSMAC. Early CTL signaling occurs within ten seconds of cell–cell contact and granule release follows some two minutes later.

Fig. 10.2 Some interactions involved in the CTL synapse

The TCR and CD8 localize to the signaling domain in the center of the synapse (cSMAC); degranulation occurs in the secretory domain. Adhesion molecules that stabilize the cell–cell interactions are located at the periphery of the supramolecular activation complex (pSMAC).

CTLs and NK cells are complementary in the defense against virally infected and cancerous cells

Natural killer (NK) cells were first identified as a subset of immune cells that were able to kill tumor cells in vitro without prior immunization of the host, even in T cell deficient mice. In 1981 Klas Kärre made the seminal observation that targets killed by NK cells tend to be spared by T cells and vice versa. From this, he reasoned that NK cells must recognize targets that are not expressing a full complement of ‘self’ molecules. Kärre’s ‘missing self’ hypothesis postulates that unlike T cells, which recognize ‘non-self’ (foreign peptide or foreign MHC), NK cells respond to the absence of MHC class I.

Several viruses (particularly herpes viruses) have evolved mechanisms to avoid recognition by CTLs. They reduce the expression of MHC class I molecules on the cell surface to reduce the likelihood that processed viral peptides can be presented to CTLs. Cancerous cells may evade the CTL response in a similar way. NK cells specifically recognize cells that have lost their MHC class I molecules. Therefore, NK cells and CTLs act in a complementary way. In effect:

Not all NK cells mediate cytotoxicity

In humans, most NK cells are CD56⁺CD3⁻ cells. However, not all cells with this phenotype are effective killers:

NK cell development

Like T and B cells, NK cells belong to the lymphoid lineage. However, the pathways and locations of NK cell development are still less well-defined than those of T and B cells. There may also be some differences between humans and mice.

In mice, the bone marrow is essential for the production of NK cells, and a complete pathway of NK development has been described at this location. Human bone marrow does contain CD34⁺ hematopoietic progenitor cells that have the potential to differentiate into NK cells, but NK-committed progenitor cells have not been identified here. On the other hand, a complete scheme of NK cell development has been described in human, but not mouse, secondary lymphoid tissue. Circulating CD34⁺ cells are thought to be recruited from the blood to the lymph nodes, where they progress through an NK-committed immature stage to give rise to CD94⁺CD56^hi mature NK cells, capable of effector functions (Fig. 10.w1). CD56^hi cells can further develop into CD56^low cells, which express KIRs and are more cytotoxic than CD56^hi cells. A thymic pathway of NK cell development has also been described in both humans and mice. However, this pathway is clearly not required for NK cell production as athymic individuals have normal NK cell numbers and function.

Fig. 10.w1 NK cell development

CD34 is a marker of hematopoietic stem cells, which is expressed by cells in early stages of NK development, but not by cells that are committed to the NK lineage. CD117 (also called SCF receptor or c-kit) is expressed by cells before they are committed to becoming NK cells, and in immature NK committed cells, but its expression is lost in more mature cells. CD122 is a marker of NK commitment in mice, but in humans it is not detectable until the mature CD56^hi stage, which is also characterized by expression of CD94. In the final stage of NK cell development, CD56^hi NK cells become CD56^low, and also start to express KIRs.

The factors required for NK cell development have been identified using in vitro cultures of human hematopoietic stem cells and by examining knockout mice. Initially, NK cell development in culture was thought to be absolutely dependent on contact with a stromal cell feeder layer, but it was then discovered that at later stages of development, exogenously added cytokines can substitute for the presence of stroma and that IL-2 or IL-15 alone can mediate the differentiation of NK-committed cells to CD56^hi mature NK cells. Although NK cells will develop in the presence of either IL-2 or IL-15, only IL-15 is made by the stroma. The phenotype of knockout mice and humans with genetic deficiencies also suggests that it is IL-15, and not IL-2, which is the critical cytokine for NK cell development. Flt3L and stem cell factor (SCF) synergistically promote NK cell development, but are not absolutely required.

For the acquisition of KIRs, which occurs at the final stage of development, as CD56^hi NK cells become CD56^low, contact with stromal cells is once again required. Defective Ly49 expression by NK cells in the Tyro/Axl/Mer triple knockout mouse suggests that this may be partially because there is a requirement for an interaction between Protein S or Gas6 on stroma and one of their receptors (Tyro, Axl or Mer) on developing NK cells. Stromal MHC class I molecules are also likely to be important for acquisition of KIR. Several of the genes that are required for immune recognition by NK cells are grouped together on the leukocyte receptor complex on the long arm of chromosome 19 (Fig. 10.w2).

Fig. 10.w2 The leukocyte receptor complex

The leukocyte receptor complex (LRC) is a 1 MB region, including the KIR, LAIR, LILR and SIGLEC, and CD66 (CEACAM) families, as well as other genes relating to NK cell function, such as the DAP adaptor proteins. The 150 kB KIR locus is shown expanded. A number of different haplotypes have been found in this region, but they can broadly be divided into ‘A’ and ‘B’ haplotypes. A haplotypes contain few KIR genes, only one of which (KIR2DS4) is likely to be activating. B haplotypes have many more KIR genes, including several activating KIR. Recombination seems to be possible around the central framework genes, as recombinant haplotypes (A at the telomeric end and B at the centromeric end, or vice versa) do exist.

Most of what is known about the transcriptional control of NK cell development comes from knockout mice, although where attempts have been made to transfect human hematopoietic stem cells with genes of interest, this has confirmed that similar transcription factors are likely to be important in human NK cell development. The master transcriptional regulator of NK cell development has recently been identified as E4bp4, which is also known as Nfil3. E4bp4 knockout mice do not have any NK cells, but are normal in all other aspects of their immune system. Downstream of E4bp4 are other factors involved in commitment to the NK lineage, most notably Id2, and factors required for their migration from the bone marrow to the periphery. T-bet, IRF2 and GATA3 are all thought to be required at this stage, as knockouts of these genes accumulate NK cells in the bone marrow but have low numbers in the periphery. Finally, there are transcription factors that are required for the acquisition of NK cell function. For example, CEBPγ knockout mice have normal numbers of phenotypically mature NK cells, but the cells are unable to kill MHC class I negative target cells, or produce IFNγ.

NK cells recognize cells that fail to express MHC class I

NK cells express inhibitory receptors that bind to MHC class I molecules. When they encounter a target cell that is not expressing MHC class I, this inhibitory signal is lost and tonic activating signals cause the NK cell to degranulate or produce cytokines in response to the target cell. Interestingly, many of the inhibitory receptors expressed by NK cells also have activating counterparts, many of which recognize the same ligands but with lower affinity. The function of these activating receptors is not known (Fig. 10.3).

Fig. 10.3 NK cell receptors with well-defined ligands

NK cell receptors with well-defined ligands.

Killer immunoglobulin-like receptors recognize MHC class I

The killer immunoglobulin-like receptors, or KIRs, are members of the immunoglobulin superfamily. They are present on the majority of CD56^low NK cells, with each individual NK cell expressing a random selection of KIRs. Almost none of the CD56^hi NK cells express KIRs.

KIRs fall into two main subsets: