5-9. Many proteins involved in antigen processing and presentation are encoded by genes within the major histocompatibility complex

The major histocompatibility complex is located on chromosome 6 in humans and chromosome 17 in the mouse and extends over some 4 centimorgans of DNA, about 4 × 10⁶ base pairs. In humans it contains more than 200 genes. As work continues to define the genes within and around the MHC, both its extent and the number of genes are likely to grow; in fact, recent studies suggest that the MHC may span at least 7 × 10⁶ base pairs. The genes encoding the α chains of MHC class I molecules and the α and β chains of MHC class II molecules are linked within the complex; the genes for β₂-microglobulin and the invariant chain are on different chromosomes (chromosomes 15 and 5, respectively, in humans and chromosomes 2 and 18 in the mouse). Figure 5.10 shows the general organization of the MHC class I and II genes in human and mouse. In humans these genes are called Human Leukocyte Antigen or HLA genes, as they were first discovered through antigenic differences between white blood cells from different individuals; in the mouse they are known as the H-2 genes.

There are three class I α-chain genes in humans, called HLA-A, -B, and -C. There are also three pairs of MHC class II α- and β-chain genes, called HLA-DR, -DP, and -DQ. However, in many cases the HLA-DR cluster contains an extra β-chain gene whose product can pair with the DRα chain. This means that the three sets of genes can give rise to four types of MHC class II molecule. All the MHC class I and class II molecules can present peptides to T cells, but each protein binds a different range of peptides (see Sections 3-16 and 3-17). Thus, the presence of several different genes of each MHC class means that any one individual is equipped to present a much broader range of peptides than if only one MHC molecule of each class were expressed at the cell surface.

The two TAP genes lie in the MHC class II region, in close association with the LMP genes that encode components of the proteasome, whereas the gene for tapasin, which binds to both TAP and empty MHC class I molecules, lies at the edge of the MHC nearest the centromere (see Fig. 5.10). The genetic linkage of the MHC class I genes, whose products deliver cytosolic peptides to the cell surface, with the TAP, tapasin, and proteasome genes, which encode the molecules that generate peptides in the cytosol and transport them into the endoplasmic reticulum, suggests that the entire MHC has been selected during evolution for antigen processing and presentation.

When cells are treated with the interferons IFN-α, -β, or -γ, there is a marked increase in transcription of MHC class I α-chain and β₂-microglobulin genes, and of the proteasome, tapasin, and TAP genes. Interferons are produced early in viral infections as part of the innate immune response, as described in Chapter 2, and so this effect increases the ability of cells to process viral proteins and present the resulting peptides at the cell surface. This helps to activate the appropriate T cells and initiate the adaptive immune response in response to the virus. The coordinated regulation of the genes encoding these components may be facilitated by the linkage of many of them in the MHC.

The HLA-DM genes, which encode the DM molecule whose function is to catalyze peptide binding to MHC class II molecules (see Section 5-7), are clearly related to the MHC class II genes. The DNα and DOβ genes, which encode the DO molecule, a negative regulator of DM, are also clearly related to the MHC class II genes. The classical MHC class II genes, along with the invariant-chain gene and the genes for DMα, β, and DNα, but not DOβ, are coordinately regulated. This distinct regulation of MHC class II genes by IFN-γ, which is made by activated T cells of T_H1 type as well as by activated CD8 and NK cells, allows T cells responding to bacterial infections to upregulate those molecules concerned in the processing and presentation of intravesicular antigens. Expression of all of these molecules is induced by IFN-γ (but not by IFN-α or -β), via the production of a transcriptional activator known as MHC class II transactivator (CIITA). An absence of CIITA causes severe immunodeficiency due to nonproduction of MHC class II molecules. (MHC Class II Deficiency, in Case Studies in Immunology, see Preface for details)

5-10. A variety of genes with specialized functions in immunity are also encoded in the MHC

Although the most important known function of the gene products of the MHC is the processing and presentation of antigens to T cells, many other genes map within this region; some of these are known to have other roles in the immune system, but many have yet to be characterized functionally. Figure 5.11 shows the detailed organization of the human MHC.

Figure 5.11

Detailed map of the human MHC. The organization of the class I, class II, and class III regions of the human MHC are shown, with approximate genetic distances given in thousands of base pairs (kb). Most of the genes in the class I and class II regions (more...)

In addition to the highly polymorphic ‘classical’ MHC class I and class II genes, there are many genes encoding MHC class I-type molecules that show little polymorphism; most of these have yet to be assigned a function. They are linked to the class I region of the MHC and their exact number varies greatly between species and even between members of the same species. These genes have been termed MHC class IB genes; like MHC class I genes, they encode β₂-microglobulin-associated cell-surface molecules. Their expression on cells is variable, both in the amount expressed at the cell surface and in the tissue distribution.

One of the mouse MHC class IB molecules, H2-M3, can present peptides with N-formylated amino termini, which is of interest because all bacteria initiate protein synthesis with N-formylmethionine. Cells infected with cytosolic bacteria can be killed by CD8 T cells that recognize N-formylated bacterial peptides bound to H2-M3. Whether an equivalent MHC class IB molecule exists in humans is not known. The large number of MHC class IB genes (50 or more in the mouse) means that many different MHC class IB molecules can exist in a single animal. They may, like the protein that presents N-formylmethionyl peptides, have specialized roles in antigen presentation. Some of them are known to be recognized by NK cell receptors, as we discuss further below.

Yet other MHC class IB genes have functions unrelated to the immune system. The HFe gene lies some 3 × 10⁶ base pairs from HLA-A. Its product is expressed on cells in the intestinal tract, and has a function in iron metabolism, regulating the uptake of iron into the body, most likely through interactions with the transferrin receptor. Individuals defective for this gene have an iron-storage disease, hemochromatosis, in which an abnormally high level of iron is retained in the liver and other organs. Mice lacking β₂-microglobulin, and hence defective in the expression of all class I molecules, show a similar iron overload. Exactly how this gene product regulates the levels of iron within the body is not known, but it is unlikely to involve an immunological mechanism.

The other genes that map within the MHC include some that encode complement components (for example, C2, C4, and factor B) and some that encode cytokines—for example, tumor necrosis factor-α (TNF-α) and lymphotoxin (TNF-β)—all of which have important functions in immunity. These have been termed MHC class III genes, and are shown in Fig. 5.11. The functions of these genes are discussed in Chapters 2 and 9.

Many studies have established associations between susceptibility to certain diseases and particular alleles of MHC genes, and we now have considerable insight into how polymorphism in the classical MHC class I and class II genes can affect resistance or susceptibility. But although most of these MHCinfluenced diseases are known or suspected to have an immune etiology, this is not true of all of them, and it is important to remember that there are many genes lying within the MHC that have no known or suspected immunological function. One of these is the enzyme 21-hydroxylase which, when deficient, causes congenital adrenal hyperplasia and, in severe cases, salt-wasting syndrome. Even where a disease-related gene is clearly homologous to immune system genes, as is the case with HFe, the disease mechanism may not be immune related. Disease associations mapping to the MHC must therefore be interpreted with caution, in the light of a detailed understanding of its genetic structure and the functions of its individual genes. Much remains to be learned about the latter and about the significance of all the genetic variation localized within the MHC. For instance, the C4 genes are highly polymorphic, but the adaptive significance of this genetic variability is not well understood.

5-11. Specialized MHC class I molecules act as ligands for activation and inhibition of NK cells

Some class IB genes, for example the members of the MIC gene family, are under a different regulatory control from the classical MHC class I genes and are induced in response to cellular stress (such as heat shock). There are five MIC genes, but only two—MICA and MICB—are expressed and produce protein products. They are expressed in fibroblasts and epithelial cells, particularly in intestinal epithelial cells, and may play a part in innate immunity or in the induction of immune responses in circumstances where interferons are not produced. The MICA and MICB molecules are recognized by a receptor that is present on NK cells, γ:δ T cells and some CD8 T cells and is capable of activating these cells to kill MIC-expressing targets. The MIC receptor is composed of two chains. One is NKG2D, an ‘activating’ member of the NKG2 family of NK-cell receptors whose cytoplasmic domain lacks an inhibitory sequence motif found in other members of this family that act as inhibitory receptors (see Section 2-28); the other is a protein called DAP10, which transmits the signal into the interior of the cell by interacting with and activating intracellular protein tyrosine kinases.

Other MHC class IB molecules may inhibit cell killing by NK cells, as described in Chapter 2. Such a role has been suggested for the MHC class IB molecule HLA-G, which is expressed on fetus-derived placental cells that migrate into the uterine wall. These cells express no classical MHC class I molecules and cannot be recognized by CD8 T cells but, unlike other cells lacking classical MHC class I molecules, they are not killed by NK cells. This appears to be because HLA-G is recognized by an inhibitory receptor, ILT-2, on the NK cell, which prevents the NK cell killing the placental cell.

Another MHC class IB molecule, HLA-E, also has a specialized role in cell recognition by NK cells. HLA-E binds a very restricted subset of peptides, derived from the leader peptides of other HLA class I molecules. These peptide:HLA-E complexes can bind to the receptor NKG2A, which is present on NK cells in a complex with the cell-surface molecule CD94. NKG2A is an inhibitory member of the NKG2 family, and on stimulation inhibits the cytotoxic activity of the NK cell. Thus a cell that expresses either HLA-E or HLA-G is not killed by NK cells.

5-12. The protein products of MHC class I and class II genes are highly polymorphic

Because of the polygeny of the MHC, every person will express at least three different antigen-presenting MHC class I molecules and three (or sometimes four) MHC class II molecules on his or her cells (see Section 5-9). In fact, the number of different MHC molecules expressed on the cells of most people is greater because of the extreme polymorphism of the MHC and the codominant expression of MHC gene products.

The term polymorphism comes from the Greek poly, meaning many, and morphe, meaning shape or structure. As used here, it means within-species variation at a gene locus, and thus in its protein product; the variant genes that can occupy the locus are termed alleles. There are more than 200 alleles of some human MHC class I and class II genes (Fig. 5.12), each allele being present at a relatively high frequency in the population. So there is only a small chance that the corresponding MHC locus on both the homologous chromosomes of an individual will have the same allele; most individuals will be heterozygous at MHC loci. The particular combination of MHC alleles found on a single chromosome is known as an MHC haplotype. Expression of MHC alleles is codominant, with the protein products of both the alleles at a locus being expressed in the cell, and both gene products being able to present antigens to T cells (Fig. 5.13). The extensive polymorphism at each locus thus has the potential to double the number of different MHC molecules expressed in an individual and thereby increases the diversity already available through polygeny (Fig. 5.14).

Figure 5.12

Human MHC genes are highly polymorphic. With the notable exception of the DRα locus, which is functionally monomorphic, each locus has many alleles. The number of different alleles is shown in this figure by the height of the bars. The figures (more...)

Figure 5.13

Expression of MHC alleles is codominant. The MHC is so polymorphic that most individuals are likely to be heterozygous at each locus. Alleles are expressed from both MHC haplotypes in any one individual, and the products of all alleles are found on all (more...)

Figure 5.14

Polymorphism and polygeny both contribute to the diversity of MHC molecules expressed by an individual. The high polymorphism of the classical MHC loci ensures a diversity in MHC gene expression in the population as a whole. However, no matter how polymorphic (more...)

Thus, with three MHC class I genes and a possible four sets of MHC class II genes on each chromosome 6, a human typically expresses six different MHC class I molecules and eight different MHC class II molecules on his or her cells. For the MHC class II genes, the number of different MHC molecules may be increased still further by the combination of α and β chains encoded by different chromosomes (so that two α chains and two β chains can give rise to four different proteins, for example). In mice it has been shown that not all combinations of α and β chains can form stable dimers and so, in practice, the exact number of different MHC class II molecules expressed depends on which alleles are present on each chromosome.

All MHC products are polymorphic to a greater or lesser extent, with the exception of the DRα chain and its mouse homologue Eα. These chains do not vary in sequence between different individuals and are said to be monomorphic. This might indicate a functional constraint that prevents variation in the DRα and Eα proteins, but no such special function has been found. Many mice, both domestic and wild, have a mutation in the Eα gene that prevents synthesis of the Eα protein. They thus lack cell-surface H-2E molecules, so if H2-E molecules do have a special function it is unlikely to be an essential one. All other MHC class I and class II genes are polymorphic.

5-13. MHC polymorphism affects antigen recognition by T cells by influencing both peptide binding and the contacts between T-cell receptor and MHC molecule

The products of individual MHC alleles can differ from one another by up to 20 amino acids, making each variant protein quite distinct. Most of the differences are localized to exposed surfaces of the outer domain of the molecule, and to the peptide-binding groove in particular (Fig. 5.15). The polymorphic residues that line the peptide-binding groove determine the peptide-binding properties of the different MHC molecules.

Figure 5.15

Allelic variation occurs at specific sites within MHC molecules. Variability plots of the amino acid sequences of MHC molecules show that the variation arising from genetic polymorphism is restricted to the amino-terminal domains (α₁ and α (more...)

We have seen that peptides bind to MHC class I molecules through specific anchor residues (see Section 3-16), and that the amino acid side chains of these residues anchor the peptide by binding in pockets that line the peptide-binding groove. Polymorphism in MHC class I molecules affects which amino acids line these pockets and thus their binding specificity. In consequence, the anchor residues of peptides that bind to each allelic variant are different. The set of anchor residues that allows binding to a given MHC class I molecule is called a sequence motif. These sequence motifs make it possible to identify peptides within a protein that can potentially bind the appropriate MHC molecule, which may be very important in designing peptide vaccines. Different allelic variants of MHC class II molecules also bind different peptides, but the more open structure of the MHC class II peptide-binding groove and the greater length of the peptides bound in it allow greater flexibility in peptide binding (see Section 3-17). It is therefore more difficult to predict which peptides will bind to MHC class II molecules.

In rare cases, processing of a protein will not generate any peptides with a suitable motif for binding to any of the MHC molecules expressed by an individual. When this happens, the individual fails to respond to the antigen. Such failures in responsiveness to simple antigens were first reported in inbred animals, where they were called immune response (Ir) gene defects. These defects were identified and mapped to genes within the MHC long before the function of MHC molecules was understood. Indeed, they were the first clue to the antigen-presenting function of MHC molecules, although it was only much later that the ‘Ir genes’ were shown to encode MHC class II molecules. Ir gene defects are common in inbred strains of mice because the mice are homozygous at all their MHC loci and thus express only one type of MHC molecule from each gene locus. This limits the range of peptides they can present to T cells. Ordinarily, MHC polymorphism guarantees a sufficient number of different MHC molecules in a single individual to make this type of nonresponsiveness unlikely, even to relatively simple antigens such as small toxins. This has obvious importance for host defense.

Initially, the only evidence linking Ir gene defects to the MHC was genetic—mice of one MHC genotype could make antibody in response to a particular antigen, whereas mice of a different MHC genotype, but otherwise genetically identical, could not. The MHC genotype was somehow controlling the ability of the immune system to detect or respond to specific antigens, but it was not clear at the time that direct recognition of MHC molecules was involved.

Later experiments showed that the antigen specificity of T-cell recognition was controlled by MHC molecules. The immune responses affected by the Ir genes were known to be dependent on T cells, and this led to a series of experiments in mice to ascertain how MHC polymorphism might control T-cell responses. The earliest of these experiments showed that T cells could only be activated by macrophages or B cells that shared MHC alleles with the mouse in which the T cells originated. This was the first evidence that antigen recognition by T cells depends on the presence of specific MHC molecules in the antigen-presenting cell. The clearest example of this feature of T-cell recognition came, however, from studies of virus-specific cytotoxic T cells, for which Peter Doherty and Rolf Zinkernagel were awarded the Nobel Prize in 1996.

When mice are infected with a virus, they generate cytotoxic T cells that kill self cells infected with the virus, while sparing uninfected cells or cells infected with unrelated viruses. The cytotoxic T cells are thus virus-specific. A particularly striking outcome of these experiments was that the specificity of the cytotoxic T cells was also affected by the polymorphism of MHC molecules. Cytotoxic T cells induced by viral infection in mice of MHC genotype a (MHC^a) would kill any MHC^a cell infected with that virus but would not kill cells of MHC genotype b, or c, and so on, even if they were infected with the same virus. Because the MHC genotype restricts the antigen specificity of T cells, this effect is called MHC restriction. Together with the earlier studies on both B cells and macrophages, this work showed that MHC restriction is a critical feature of antigen recognition by all functional classes of T cells.

Because different MHC molecules bind different peptides, MHC restriction in immune responses to viruses and other complex antigens might be explained solely on this indirect basis. However, it can be seen from Fig. 5.15 that some of the polymorphic amino acids in MHC molecules are located in the α helices flanking the peptide-binding cleft in such a way that they are exposed on the outer surface of the peptide:MHC complex and can be directly contacted by the T-cell receptor (see Fig. 3.27). It is therefore not surprising that when T cells are tested for their ability to recognize the same peptide bound to different MHC molecules, they readily distinguish the peptide bound to MHC^a from the same peptide bound to MHC^b. Thus, the specificity of a T-cell receptor is defined both by the peptide it recognizes and by the MHC molecule bound to it (Fig. 5.16). This restricted recognition may sometimes be caused by differences in the conformation of the bound peptide imposed by the different MHC molecules rather than by direct recognition of polymorphic amino acids in the MHC molecule itself. MHC restriction in antigen recognition therefore reflects the combined effect of differences in peptide binding and of direct contact between the MHC molecule and the T-cell receptor.

Figure 5.16

T-cell recognition of antigens is MHC restricted. The antigen-specific T-cell receptor (TCR) recognizes a complex of antigenic peptide and MHC. One consequence of this is that a T cell specific for peptide x and a particular MHC allele, MHC^a (left panel), (more...)

5-14. Nonself MHC molecules are recognized by 1–10% of T cells

The discovery of MHC restriction, by revealing the physiological function of the MHC molecules, also helped explain the otherwise puzzling phenomenon of recognition of nonself MHC in the rejection of organs and tissues transplanted between members of the same species. Transplanted organs from donors bearing MHC molecules that differ from those of the recipient—even by as little as one amino acid—are invariably rejected. The rapid and very potent cell-mediated immune response to the transplanted tissue results from the presence in any individual of large numbers of T cells that are specifically reactive to nonself, or allogeneic, MHC molecules. Early studies on T-cell responses to allogeneic MHC molecules used the mixed lymphocyte reaction. In this reaction T cells from one individual are mixed with lymphocytes from a second individual. If the T cells of one individual recognize the other individual's MHC molecules as ‘foreign,’ the T cells will divide and proliferate. (The lymphocytes from the second individual are usually prevented from dividing by irradiation or treatment with the cytostatic drug mitomycin C.) Such studies have shown that roughly 1–10% of all T cells in an individual will respond to stimulation by cells from another, unrelated, member of the same species. This type of T-cell response is called alloreactivity because it represents the recognition of allelic polymorphism in allogeneic MHC molecules.

Before the role of the MHC molecules in antigen presentation was understood, it was a mystery why so many T cells should recognize nonself MHC molecules, as there is no reason the immune system should have evolved a defense against tissue transplants. However, once it was appreciated that T-cell receptors have evolved to recognize foreign peptides in combination with polymorphic MHC molecules, alloreactivity became easier to explain. From experiments in which T cells from animals lacking MHC class I and class II molecules have been artificially driven to mature, it has been shown that the ability to recognize MHC molecules is inherent in the genes that encode the T-cell receptor, rather than being dependent on selection for MHC recognition during T-cell development. The high frequency of alloreactive T cells clearly reflects the commitment of the T-cell receptor to the recognition of MHC molecules in general.

Mature T cells have, however, survived a stringent selection process for the ability to respond to foreign, but not self, peptides bound to self MHC molecules. It is therefore thought that the alloreactivity of mature T cells reflects the cross-reactivity of T-cell receptors normally specific for a variety of foreign peptides bound by self MHC molecules. Given a T-cell receptor that normally binds a foreign peptide displayed by a self MHC molecule (Fig. 5.17, left panel), there are two ways in which it may bind to nonself MHC molecules. In some cases, the peptide bound by the nonself MHC molecule interacts strongly with the T-cell receptor, and the T cells bearing this receptor are stimulated to respond. This type of cross-reactive recognition arises because the spectrum of peptides bound by nonself MHC molecules on the transplanted tissues differs from those bound by the host's own MHC, and it is known as peptide-dominant binding (Fig. 5.17, center panel). In a second type of cross-reactive recognition, known as MHC-dominant binding, allo-reactive T cells respond because of direct binding of the T-cell receptor to distinctive features of the nonself MHC molecule (Fig. 5.17, right panel). In these cases the recognition is less dependent on the particular peptide bound; T-cell receptor binding to unique features of the nonself MHC molecule generates a strong signal because of the high concentration of the nonself MHC molecule on the surface of the presenting cell. Both these mechanisms may contribute to the high frequency of T cells that can respond to nonself MHC molecules on transplanted tissue.

Figure 5.17

Two modes of crossreactive recognition that may explain alloreactivity. A T cell that is specific for one peptide:MHC combination (left panel) may cross-react with peptides presented by other, nonself (allogeneic), MHC molecules. This may come about in (more...)

5-15. Many T cells respond to superantigens

Superantigens are a distinct class of antigens that stimulate a primary T-cell response similar in magnitude to a response to allogeneic MHC. Such responses were first observed in mixed lymphocyte reactions using lymphocytes from strains of mice which were MHC identical but otherwise genetically distinct. The antigens provoking this reaction were originally designated minor lymphocyte stimulating (Mls) antigens, and it seemed reasonable to suppose that they might be functionally similar to the MHC molecules themselves. We now know that this is not the case, however. The Mls antigens found in these mice strains are encoded by retroviruses which have become stably integrated at various sites into the mouse chromosomes. They act as superantigens because they have a distinctive mode of binding to both MHC and T-cell receptor molecules that enables them to stimulate very large numbers of T cells. Superantigens are produced by many different pathogens, including bacteria, mycoplasmas, and viruses, and the responses they provoke are helpful to the pathogen rather than the host.

Superantigens are unlike other protein antigens, in that they are recognized by T cells without being processed into peptides that are captured by MHC molecules. Indeed, fragmentation of a superantigen destroys its biological activity, which depends on binding as an intact protein to the outside surface of an MHC class II molecule which has already bound peptide. In addition to binding MHC class II molecules, superantigens are able to bind the V_β region of many T-cell receptors (Fig. 5.18). Bacterial superantigens bind mainly to the V_β CDR2 loop, and to a smaller extent to the V_β CDR1 loop and an additional loop called the hypervariable 4 or HV4 loop. The HV4 loop is the predominant binding site for viral superantigens, at least for the Mls antigens encoded by the endogenous mouse mammary tumor viruses. Thus, the α-chain V region and the CDR3 of the β chain of the T-cell receptor have little effect on superantigen recognition, which is determined largely by the germline-encoded V sequences of the expressed β chain. Each superantigen is specific for one or a few of the different V_β gene segments, of which there are 20–50 in mice and humans; a superantigen can thus stimulate 2–20% of all T cells.

Figure 5.18

Superantigens bind directly to T-cell receptors and to MHC molecules. Superantigens can bind independently to MHC class II molecules and to T-cell receptors, binding to the V_β domain of the T-cell receptor (TCR), away from the complementarity-determining (more...)

This mode of stimulation does not prime an adaptive immune response specific for the pathogen. Instead, it causes a massive production of cytokines by CD4 T cells, the predominant responding population of T cells. These cytokines have two effects on the host: systemic toxicity and suppression of the adaptive immune response. Both these effects contribute to microbial pathogenicity. Among the bacterial superantigens are the staphylococcal enterotoxins (SEs), which cause food poisoning, and the toxic shock syndrome toxin-1 (TSST-1), the etiologic principle in toxic shock syndrome. (Toxic Shock Syndrome, in Case Studies in Immunology, see Preface for details)

The role of viral superantigens in human disease is less clear. The T-cell responses to rabies virus and the Epstein-Barr virus indicate the existence of superantigens in these human pathogens but the genes encoding them have not yet been identified. The best characterized viral superantigens remain the mouse mammary tumor virus superantigens which are common as endogenous antigens in mice. We will see in Chapter 7 how these have made it possible to observe the deletion of self-reactive T cells as they develop in the thymus, and in Chapter 11 how the virus uses the response to its superantigen to promote its own transmission.

5-16. MHC polymorphism extends the range of antigens to which the immune system can respond

Most polymorphic genes encode proteins that vary by only one or a few amino acids, whereas the different allelic variants of MHC proteins differ by up to 20 amino acids. The extensive polymorphism of the MHC proteins has almost certainly evolved to outflank the evasive strategies of pathogens. Pathogens can avoid an immune response either by evading detection or by suppressing the ensuing response. The requirement that pathogen antigens must be presented by an MHC molecule provides two possible ways of evading detection. One is through mutations that eliminate from its proteins all peptides able to bind MHC molecules. The Epstein-Barr virus provides an example of this strategy. In regions of south-east China and in Papua New Guinea there are small isolated populations in which about 60% of individuals carry the HLA-All allele. Many isolates of the Epstein-Barr virus obtained from these populations have mutations in a dominant peptide epitope normally presented by HLA-All; the mutant peptides no longer bind to HLA-All and cannot be recognized by HLA-All-restricted T cells. This strategy is plainly much more difficult to follow if there are many different MHC molecules, and the presence of different loci encoding functionally related proteins may have been an evolutionary adaptation by hosts to this strategy by pathogens.

In large outbred populations, polymorphism at each locus can potentially double the number of different MHC molecules expressed by an individual, as most individuals will be heterozygotes. Polymorphism has the additional advantage that individuals in a population will differ in the combinations of MHC molecules they express and will therefore present different sets of peptides from each pathogen. This makes it unlikely that all individuals in a population will be equally susceptible to a given pathogen and its spread will therefore be limited. That exposure to pathogens over an evolutionary timescale can select for expression of particular MHC alleles is indicated by the strong association of the HLA-B53 allele with recovery from a potentially lethal form of malaria; this allele is very common in people from West Africa, where malaria is endemic, and rare elsewhere, where lethal malaria is uncommon.

Similar arguments apply to a second strategy for evading recognition. If pathogens can develop mechanisms to block the presentation of their peptides by MHC molecules, they can avoid the adaptive immune response. Adenoviruses encode a protein that binds to MHC class I molecules in the endoplasmic reticulum and prevents their transport to the cell surface, thus preventing the recognition of viral peptides by CD8 cytotoxic T cells. This MHC-binding protein must interact with a polymorphic region of the MHC class I molecule, as some allelic variants are retained in the endoplasmic reticulum by the adenoviral protein whereas others are not. Increasing the variety of MHC molecules expressed therefore reduces the likelihood that a pathogen will be able to block presentation by all of them and completely evade an immune response.

These arguments raise a question: if having three MHC class I loci is better than having one, why are there not far more MHC loci? The probable explanation is that each time a distinct MHC molecule is added to the MHC repertoire, all T cells that can recognize self peptides bound to that molecule must be removed in order to maintain self tolerance. It seems that the number of MHC loci present in humans and mice is about optimal to balance out the advantages of presenting an increased range of foreign peptides and the disadvantages of an increased loss of T cells from the repertoire.

5-17. Multiple genetic processes generate MHC polymorphism

MHC polymorphism appears to have been strongly selected by evolutionary pressures. However, for selection to work efficiently in organisms that reproduce slowly, such as humans, there must also be powerful mechanisms for generating the variability on which selection can act. The generation of polymorphism in MHC molecules is an evolutionary problem not easily analyzed in the laboratory; however, it is clear that several genetic mechanisms contribute to the generation of new alleles. Some new alleles are the result of point mutations, but many arise from the combination of sequences from different alleles either by genetic recombination or by gene conversion, a process in which one sequence is replaced, in part, by another from a different gene (Fig. 5.19).

Figure 5.19

Gene conversion can create new alleles by copying sequences from one MHC gene to another. Sequences can be transferred from one gene to a similar but different gene by a process known as gene conversion. For this to happen, the two genes must become apposed (more...)

Evidence for gene conversion comes from studies of the sequences of different MHC alleles. These reveal that some changes involve clusters of several amino acids in the MHC molecule and require multiple nucleotide changes in a contiguous stretch of the gene. Even more significantly, the sequences that have been changed frequently derive from other MHC genes on the same chromosome, which is a typical signature of gene conversion. Genetic recombination between different alleles at the same locus may, however, have been more important than gene conversion in generating MHC polymorphism. A comparison of sequences of MHC alleles shows that many different alleles could represent recombination events between a relatively small set of hypothetical ancestral alleles (Fig. 5.20).

Figure 5.20

Genetic recombination can create new MHC alleles by DNA exchange between different alleles of the same gene. Conventional meiotic recombination differs from gene conversion in that DNA segments are exchanged between alleles on the two homologous chromosomes (more...)

The effects of selective pressure in favor of polymorphism can be seen clearly in the pattern of point mutations in the MHC genes. Point mutations can be classified as replacement substitutions, which change an amino acid, or silent substitutions, which simply change the codon but leave the amino acid the same. Replacement substitutions occur within the MHC at a higher frequency relative to silent substitutions than would be expected, providing evidence that polymorphism has been actively selected for in the evolution of the MHC.

5-18. Some peptides and lipids generated in the endocytic pathway can be bound by MHC class I-like molecules that are encoded outside the MHC

Some MHC class I-like genes map outside the MHC region. One family, called CD1, expressed on dendritic cells and monocytes as well as some thymocytes, functions in antigen presentation to T cells, but the molecules it encodes have two features that distinguish them from classical MHC class I molecules. The first is that the CD1 molecule, although similar to MHC class I molecules in its subunit organization and association with β₂-microglobulin, behaves like an MHC class II molecule. It is not retained within the endoplasmic reticulum by association with the TAP complex but is targeted to vesicles, where it binds its peptide ligand. The peptide antigens bound by CD1 are therefore derived from the breakdown of extracellular proteins within acidified endosomal compartments and, like the peptides that bind to MHC class II molecules, tend to be longer than the peptides that bind to classical MHC class I molecules.

The second unusual feature of CD1 molecules is that they are able to bind and present glycolipids, in particular the mycobacterial membrane components mycolic acid, glucose monomycolate, phosphoinositol mannosides, and lipoarabinomannan. These are derived either from internalized mycobacteria or from the uptake of lipoarabinomannans by the mannose receptor that is expressed by many phagocytic cells (see Section 2-15). These ligands will thus be delivered into the endocytic pathway, where they can be bound by CD1 molecules. The relationship between the peptide-binding and lipid-binding capacities of CD1 molecules is not clear. Structural studies show that the CD1 molecule has a deep binding groove into which the glycolipid antigens bind. Whether the peptide antigens also bind in this deep groove is not yet known; although CD1-binding peptides are predominantly hydrophobic in character, it is thought unlikely that they bind to the same site as the lipids. It appears that the CD1 genes have evolved as a separate lineage of antigen-presenting molecules able to present microbial lipids and glycolipids, as well as a subset of peptides, to T cells.

Summary

The major histocompatibility complex (MHC) of genes consists of a linked set of genetic loci encoding many of the proteins involved in antigen presentation to T cells, most notably the MHC class I and class II glycoproteins (the MHC molecules) that present peptides to the T-cell receptor. The outstanding feature of the MHC molecules is their extensive polymorphism. This polymorphism is of critical importance in antigen recognition by T cells. A T cell recognizes antigen as a peptide bound by a particular allelic variant of an MHC molecule, and will not recognize the same peptide bound to other MHC molecules. This behavior of T cells is called MHC restriction. Most MHC alleles differ from one another by multiple amino acid substitutions, and these differences are focused on the peptide-binding site and adjacent regions that make direct contact with the T-cell receptor. At least three properties of MHC molecules are affected by MHC polymorphism: the range of peptides bound; the conformation of the bound peptide; and the direct interaction of the MHC molecule with the T-cell receptor. Thus the highly polymorphic nature of the MHC has functional consequences, and the evolutionary selection for this polymorphism suggests that it is critical to the role of the MHC molecules in the immune response. Powerful genetic mechanisms generate the variation that is seen among MHC alleles, and a compelling argument can be made that selective pressure to maintain a wide variety of MHC molecules in the population comes from infectious agents.