Figure 1. T cell populations. The same T cell precursor can give rise to several T cell subsets with different functions. Conventional αβ T cells develop into CD4 and CD8 populations in the absence of foreign antigens. When they become activated by antigen they can develop further into different effector subsets (e.g. T h1, T h2 and perhaps T c1, T c2 cells, etc.) or into memory T cells. Some can also develop into regulatory T cells (not shown). Non‐conventional T cell populations include γδ T cells and iNKT cells. (There are also non‐conventional αβ cells that express neither CD4 nor CD8, and different types of NKT cells; not shown.)

Figure 2. Interactions of T cell receptors with peptide–MHC complexes. αβ TCRs recognize peptides bound to MHC molecules. These peptides are generated from proteins within cells by mechanisms collectively known as antigen processing and bind to MHC molecules before they are exported to the plasma membrane. The TCR α and β chains each contain three hypervariable CDRs (CDR1–3). The CDR3 regions interact primarily with the peptide while the others interact predominantly with the MHC molecule itself.

Figure 3. Interactions of CD4 and CD8 with MHC molecules. CD4 and CD8 molecules are first expressed on thymocytes during T cell development, and which one of these a T cell eventually expresses determines whether it will interact with MHC class I or II. CD4 and CD8 bind to invariant parts of MHC class II and class I respectively. The TCR is associated with CD3, a multimolecular complex (not shown). CD4 and CD8 are constitutively associated with the tyrosine kinase, Lck, which phosphorylates specific regions ITAM motifs of CD3 (not shown) in the early stages of T cell activation.

Figure 4. Genetic organization of human and mouse MHC. The MHC consists of a large stretch of DNA, in humans called HLA and in mice H‐2. Several distinct structural genes encode the α chain of classical MHC class I molecules (which associate with β 2‐microglobulin, encoded outside of the MHC), and the α and β chains of class II molecules. Many other molecules are also encoded in the MHC region. Some of these are involved in peptide delivery to, or loading of, classical MHC class I or class II molecules; these, respectively, include TAP, and in humans HLA‐DM and ‐DO (which are also dimers; A, B); the mouse equivalents are H‐2M and H‐2O. There are also non‐classical MHC molecules, such as MIC(‐A and ‐B), and HLA‐E, with specialized functions.

Figure 5. Interactions of peptides with MHC molecules. The peptide‐binding grooves of MHC molecules consist of two stretches of α‐helix which overlie a β‐pleated sheet base. Most of the differences (polymorphisms) between MHC molecules are found in the groove and hence determine precisely which peptides can be bound by different MHC alleles. Class I molecules have grooves closed at both ends, restricting the size of peptides that can bind. They also contain pockets into which side chains from the peptide amino acids can bind non‐covalently as anchor residues to hold the peptide in place. Class II molecules have more open ends, allowing peptides of greater and more varied lengths to bind.

Figure 6. Origins of peptides bound to MHC molecules. Under steady‐state conditions, peptides bound to MHC class I molecules are mostly derived from abnormally‐folded cytosolic proteins (defective ribosomal products; DRIPs). In contrast, peptides bound to class II molecules may originate from extracellular proteins that have been endocytosed or from plasma or endosome membranes. During infection peptides from viruses or microbes that have gained access to the cytoplasm, or which have been endocytosed, may become loaded onto either MHC class I or class II molecules, respectively.

Figure 8. Cellular localization of proteins recognized by CD8 cytotoxic T cells. Different strains of influenza virus express different forms of a relatively small number of proteins. One of these is the nucleoprotein (NuP) that is never expressed on the surface of infected cells. Mice were infected with one strain of the virus and their CTLs were tested for the ability to kill infected target cells in culture. The target cells were infected with strains of virus that shared different variant proteins with the original virus used to infect the mice. In some cases it was found that in order to render the target cells sensitive to killing, the only protein that needed to be identical in the two viruses was NuP. At the time this was a real puzzle. However, we now understand that the NuP protein is degraded intracellularly and peptides from it are about to MHC class I molecules and transported to the cell surface where these complexes can be recognized by CTL.

Figure 7. MHC restriction of CD8 cytotoxic T cells. Mice infected with many viruses generate CD8 + CTLs. These can be isolated and are able to kill cells in culture that are infected with the same virus. Doherty and Zinkernagel, using a particular virus, LCMV, found that only cells of the original mouse strain could be killed (e.g. A), while cells from other strains (e.g. B) could not even if they were infected with the same virus. Eventually they discovered that, if infected cells were to be killed, they needed to express the same MHC class I molecules as the mouse that was initially infected. This became known as MHC restriction. This principle applies to both CD8 and CD4 T cells: the former are restricted to MHC class I molecules, the latter to MHC class II.

Figure 9. The peptide–major histocompatibility complex class I pathway. (1) Cytosolic proteins, often attached to ubiquitin, (2) are targeted to proteasomes (modified as immunoproteasome; no shown) where they degraded. (3) Peptides of appropriate sizes are transported into the RER by ATP‐dependent TAP transporters. (4) These are associated, via tapasin, to newly assembled MHC class I molecules, with associated β 2‐microglobulin, to form the peptide‐loading complex. Suitable peptides can bind to the MHC molecules, (5) which are then released and (6) transported through the Golgi apparatus to the plasma membrane. (The peptide‐loading complex also contains chaperone proteins; not shown.)

Figure 10. The peptide–major histocompatibility complex class II processing. Proteins in the endocytic pathway are broken down to peptides by lysosomal proteases. Meanwhile, (1) MHC class II α, β and invariant chains are synthesized and (2) assembled in the RER. (3) The invariant chain both blocks the peptide‐binding groove in the MHC molecule and targets the complex, through the Golgi, to the endocytic pathway. (4) Here, the invariant chain is digested in a step‐wise manner to leave a small fragment, CLIP, which blocks the groove. (5) CLIP then either dissociates spontaneously or dissociation is facilitated by a MHC class II‐related molecule, human HLA‐DM (or mouse H‐2M), which additionally promotes peptide exchange until a high‐affinity peptide binds into the groove. (6) The peptide–MHC complex can then be transported to the plasma membrane.

Figure 11. Cross‐presentation. In some DCs, peptides from endocytosed proteins can be displayed on MHC class I molecules. This is known as cross‐presentation. The mechanisms are still not fully understood. (1) Proteins or (2) peptides may exit the endocytic pathway into the cytosol. (3) Alternatively, endosomes may acquire MHC class I, possibly from the RER, permitting direct peptide binding. (4) Cross‐presentation enables peptides derived from microbes in peripheral tissues to be presented on MHC class I molecules to CD8 T cells, without the need for the DCs to be actively infected. Such DCs can also present exogenous peptides on MHC class II, additionally permitting CD4 T cell activation through the normal route.

Figure 12. The lipid–CD1 pathway. CD1 molecules are structurally similar to classical MHC class I molecules, but possess a hydrophobic groove, permitting lipid‐containing molecules to bind. Unlike classical MHC class I molecules, CD1 molecules may associate with the invariant chain and are targeted to the endosomal pathway. They may acquire lipid‐containing molecules at different stages of their life history, including the RER (1) and the endosome (2). Different CD1 molecules traffic through different endosomes. The natural ligands of CD1 molecules may include glycolipids derived from mycobacterial cell walls and phospholipids. Glycolipids bound to CD1 molecules can be presented to non‐conventional T cells such as iNKT cells in the case of human CD1d.

Figure 13. Costimulation of CD4 T cells. Naive CD4 T cells may become unresponsive (anergic), die (apoptose) or differentiate into regulatory T cells (T reg) if they recognize a peptide–MHC complex on an APC in the absence of further stimuli (Signal 1 alone). This can lead to antigen‐specific tolerance. If however the T cell additionally receives positive signals through CD28, from B7 family members CD80 and/or CD86, it will become full activated (Signal 1 plus Signal 2). (In contrast, negative signals delivered by molecules such as CTLA‐4 or PD1 can inhibit T cell activation; not shown.) Activated (mature) DCs are the usual source of such costimulatory signals for initial, naïve CD4 T cell activation. Other signals regulate the type of effector response that develops, such as into T h1 versus T h2 cells (Signal 3). The T cells may also be instructed as to where they should migrate (Signal 4: not shown)

Figure 14. The immunological synapse. When a CD4 T cell interacts with a DC the cytoskeleton of both cells reorganizes to form a tight cell–cell interaction zone. This is the immunological synapse in which different sets of molecules are arranged in concentric rings. Molecules involved in T cell activation – the TCR and MHC, and costimulatory molecules – are concentrated in the central areas and are surrounded by adhesion molecules. Some large molecules, such as CD45 which has inhibitory functions, and CD43, which may serve as a barrier molecule preventing close contact between cells, are excluded from the synapse. (Such synapses also form when other cells of adaptive immunity interact, e.g. CD4 T cells with B cells and CD8 T cells (CTLs) with target cells; not shown.

Figure 15. Intracellular signalling pathways for T cell activation. A highly over‐simplified scheme of some major intracellular signalling pathways involved in T cells. Signal 1. Ligation of peptide–MHC complexes (and co‐recognition by CD4 or CD8; not shown) initiates signalling via the CD3 complex which activates multiple downstream pathways. Several of these lead to the activation and nuclear translocation of transcription factors AP‐1, NF‐κB and NF‐AT, the latter two via pathways involving DAG and IP 3 respectively. Rac is involved in activation of the actin cytoskeleton (which also involves WASP). Signal 2. Additional signalling from CD28 increases activation of key transcription factors and the PI3 kinase pathway (involving mTOR) which regulates T cell growth, proliferation and survival. See the text for further details.

Figure 16. Dendritic cell populations. Classical DCs and plasmacytoid DCs arise from different precursor cells. Some classical DC precursors enter peripheral tissues and then migrate to secondary lymphoid tissues; a similar pathway is followed by specialized Langerhans cells of skin. Other classical precursors enter lymph nodes directly from the blood and comprise a separate lymph node‐resident DC population. DCs can also develop from monocytes that are recruited to sites of inflammation, and may also migrate to secondary lymphoid tissues. Plasmacytoid DCs are a distinct population that can migrate directly from the blood into secondary lymphoid tissues (or to peripheral tissues and tumours in disease settings; not shown). These cells must not be confused with follicular DCs (FDCs), which are entirely unrelated in origin or in function (not shown).

Figure 17. Induction of T h0 CD4 T cells. DCs acquire microbial antigens and present them, as peptide–MHC II complexes, to CD4 T cells. In response to signalling through TLRs, for example, they also up‐regulate CD80 and CD86 (B7) which provide costimulation for naïve T cell activation. These signals stimulate the T cell to express the high‐affinity IL‐2 receptor (CD25) and also to secrete IL‐2. The IL‐2 acts in an autocrine manner to induce anti‐apoptotic mechanisms and to stimulate proliferation. The result is increased survival and clonal expansion of the activated T cell, which can secrete both T h1 and T h2 cytokines. Some CD4 T cells may remain as T h0 cells without further differentiation.

Figure 18. Dendritic cell regulation of CD4 T cell differentiation. DCs are important regulators of both activation and differentiation of CD4 T cells. Different cytokines, and perhaps cell surface molecules such as ICOS expressed by DCs induce distinct patterns of gene expression in responding naïve T cells; some key examples are shown. These stimuli initiate signalling pathways that activate transcription factors that are signatures of a particular differentiation pathway; these are shown in the nuclei of the T cells. Contributions from other cell types are also likely to be very important. For example, IFN‐γ contributes to T h1 activation and may originate from NK cells; IL‐4 is crucial for T h2 differentiation and may come from mast cells, basophils or iNKT cells; and TGF‐β, important for both T reg and T h17 cells, can be produced by many cell types, including DCs.

Figure 19. Effector functions of T h1 CD4 T cells. Two important cytokines secreted by T h1 cells are IFN‐γ and IL‐2. These act on several different cell types. IFN‐γ acts on B cells to promote secretion of opsonizing IgG antibodies and on macrophages to make them potent antimicrobial cells. IFN‐r acts, with other cytokines, to increase NK cell cytotoxicity and on CD8 T cells with IL‐2 to promote their activation, survival and proliferation. IgG2a is a major opsonizing antibody in mice.

Figure 21. CD4 T cell polarization in Leishmania infection in mice. Leishmania major is a human parasite that also infects mice. C57BL/6 mice infected intradermally develop a local lesion that heals spontaneously and leaves the mouse resistant to re‐infection. This is due to a polarized T h1 response. BALB/c mice, in contrast, develop a systemic infection and die. This is because their response is biased to T h2. If, however, C57BL/6 mice are treated with a blocking antibody to IFN‐γ, they make a T h2 response and develop a fatal systemic infection. In contrast, if infected BALB/c mice are treated with a blocking antibody to IL‐4, they make a T h1 response, recover and are resistant to re‐infection. These experiments show the crucial roles of IFN‐γ and IL‐4 in polarizing CD4 T cell responses to T h1 and T h2, respectively.

Figure 20. Effector functions of T h2 CD4 T cells. Four important cytokines that T h2 cells secrete are IL‐4, IL‐5, IL‐9 and IL‐13. IL‐4 promotes activated B cells to make barrier antibodies, including IgE. IL‐4 and IL‐13 act on macrophages to promote alternative activation – these cells have crucial functions in repair and healing. IL‐5 or IL‐9 (in concert with other cytokines and chemokines) act to promote esosinophil and basophil recruitment to, and survival in, inflamed tissues.

Figure 22. Effector functions of T h17 CD4 T cells. The major cytokine secreted by T h17 T cells is IL‐17. IL‐17 acts on stromal cells and endothelial cells, inducing them to secrete growth factors, and cytokines and chemokines, which stimulate neutrophil production and recruitment to inflamed tissues.

Figure 23. Activation of CD8 T cells. In many cases, CD8 T cells need help from CD4 T cells to become activated. Evidence suggests that in some cases DCs first activate CD4 T cells, which then in turn signal back and super‐activate the DCs. This may involve CD40–CD40 ligand interactions leading to increased expression of additional costimulatory molecules by the DCs (such as those shown). These additional signals are required for activation of CD8 T cells, which at the same time are recognizing peptide–MHC complexes on the DCs. IL‐2 secreted by activated CD4 T cells may also lead to proliferation or expansion of CTLs, perhaps particularly in peripheral tissues.

Figure 24. Effector functions of CD8 T cells. Activated CD8 T cells can possess cytoplasmic granules that contain the cytotoxic machinery. Following recognition of peptide–MHC class I complex on a target cell, the T cell cytoskeleton is activated and the granules move to the immunological synapse. The granules discharge perforin and granzymes into the cleft. These molecules may be endocytosed by the target cell. Perforin may then polymerize to form a pore in the endocytic membrane. This allows entry of granzymes to the cytosol, which induce activation of the apoptosome, leading to caspase activation and apoptosis. Activated CD8 T cells can also express Fas ligand, which on binding to Fas on a target cell, may also induce apoptosis. Activated CD8 T cells can also secrete cytokines such as TNF‐α. Binding of this cytokine to the respective cytokine receptor can in some cases also trigger apoptosis.

Figure 25. Major histocompatibility complex tetramers. MHC tetramers (and pentamers) are used to quantitate the numbers of T cells that can recognize a particular peptide–MHC complex. The affinity of a TCR for its peptide–MHC complex is too low to permit stable binding in vitro. MHC tetramers are composed of four copies of a particular peptide–MHC complex and the avidity of the tetramer is sufficient for stable binding. Tetramers are made by refolding recombinant MHC class I (α and β 2‐microglobulin) molecules in the presence of a specific peptide. Biotin is then bound enzymatically to the MHC. Fluorescent streptavidin, which has four binding sites for biotin, is added to complete the tetramer. The FACS contour plot shows an example in which a lymphocyte population has been labelled and around 1% of the cells have bound the tetramer.

Figure 26. Memory T cells. Following an adaptive immune response many effector cells apoptose but memory T cells persist; these are responsible for quicker, stronger responses following secondary exposure to the antigen and which can transfer immunity to a naïve animal. Some, effector memory T cells, are similar to effector T cells. For example, they express chemokine receptors such as CXCR5 that enable recruitment to inflamed tissues and they do not recirculate. Other central memory T cells resemble naïve T cells. For example, they express L‐selectin and CCR7 and can recirculate. Both forms of memory T cell can be reactivated to become active effector cells.

Figure 27. T cell receptor generation. Complete TCR genes are only present in T cells. In germline DNA, separate gene segments encode parts of the V region or the C region. The β chain region contains numerous V β segments, a limited number of D β segments and multiple J β segments. In developing T cells, one segment from each group is selected (largely randomly) and the selected segments are brought together to form the complete V region. The V β region is paired to a C β region segment. For α chains the principle is similar, and there are numerous V α and J α segments, but no D region segments. Additional diversity is generated by modification of the bases at the junctions between V, (D) and J segments.

Figure 28. Junctional diversity in antigen receptor gene formation. When antigen receptors are being formed in T or B cells, RAG molecules bind to recombination signal sequences (RSSs) that flank the V, D and J segments. Spacer segments of 12 or 23 base pairs in the RSSs ensure the correct joining of the different segments. The RAG molecules interact to cleave the flanking sequences of the gene segments, forming a hairpin between the coding sequences. This hairpin is cleaved by the enzyme Artemis, and subsequently repaired to form palindromic ends and P‐regions. Additionally random nucleotides can be added to the free DNA ends by the enzyme TdT to form N‐regions. The result is that additional bases can be added to the junctions between the segments, thus greatly increasing receptor diversity.

Figure 29. Thymic selection of αβ T cells. The αβ TCRs generated in thymocytes can potentially recognize any MHC molecules expressed in a species. In the cortex, T cells that cannot bind with sufficient affinity to self MHC expressed on cortical epithelial cells undergo apoptosis while those that do recognize self MHC are rescued from apoptosis. This is called positive selection (1). Surviving thymocytes enter the medulla and interact with DCs and medullary epithelial cells that express AIRE. Those that now recognize self peptide–MHC complexes with medium or high affinity undergo apoptosis. This is called negative selection (2). After passing through both processes, T cells exit the thymus into the periphery. These cells are self MHC‐restricted, very weakly or non‐autoreactive, but can potentially recognize with high affinity foreign peptides bound to self MHC (3).

Figure 30. Determination of αβ or γδ T cell receptor expression. According to one model which type of TCR is generated in a thymocyte depends on which gene(s) first rearrange successfully. If a β chain gene rearranges first, the β chain pairs with a surrogate α chain and γδ rearrangement is suppressed. If, however, both γ and δ genes rearrange successfully before the β chain, αβ rearrangement is suppressed and the γδ TCR is expressed. Other models to explain αβ versus γδ T cell development have also been proposed.

Figure 31. Effector functions of regulatory T cells. T reg can be natural (formed in the thymus) or induced during adaptive immune responses. The mechanisms by which T reg suppress responses may include contact‐dependent actions on DCs, which are then either killed or become unable to fully activate naïve T cells. Alternatively, or in addition, T reg may secrete IL‐10 and/or TGF‐β, which may act on naïve T cells during their activation to suppress full activation. Receptor–ligand interactions, such as Notch and its ligands, may also be important for some functions of T reg (not shown).