History

What has become known as the major histocompatibility complex (MHC) was initially identified in the early 1900s, but it was not until the late 1930s that studies began to focus on graft acceptance (histocompatibility) and antigen response phenotypes (H-2) in different strains of mice.^1,2 In the 1950s, Dausset detected the first histocompatibility antigens in humans, the MHC class I antigens, with antibodies from multiply transfused patients.^3,4 These antibodies revealed in the human population differing patterns of binding to white blood cells (leukocytes) and each pattern of binding came to define a human leukocyte antigen (HLA) specificity.^5,6 These HLA specificities were later determined to be encoded by three distinct polymorphic loci, HLA-A, HLA-B and HLA-C. The human MHC class II antigens were initially described via their ability to stimulate the proliferation of T-cells from one individual when mixed with lymphocytes from a second individual.⁷ Each pattern of T-cell reactivity (allorecognition) to a panel of homozygous typing cells (HTC) was assigned an HLA-D phenotype.⁸ It is now known that the HLA-D phenotypes are due to T-cell allorecognition of the products of the MHC class II loci, primarily HLA-DR. HLA-DQ and HLA-DP make minor contributions to these phenotypes. The genes specifying both class I and class II antigens are tightly clustered in a single chromosomal region, the MHC.

Genomic organization

The human MHC is a genetic region located on the short arm of chromosome 6 (6p21.3) extending approximately 4 megabases (Mb) (Fig. 39.1). The MHC encodes over 250 genes and pseudogenes of which at least 150 are expressed as proteins.^9,10 This genetic complex is divided into three regions: (centromeric) class II, class III and class I (telomeric). Although the proteins encoded within the MHC participate in a variety of functions, approximately 40% are devoted to immune system functions.

Fig. 39.1 Map of the human major histocompatibility complex (MHC) located on the short arm of chromosome 6. The map (drawn closely to scale) is divided into three regions (class II, class III and class I) and shows the relative positions of some of the MHC encoded genes and pseudogenes (ψ). The class II region encodes the expressed classical HLA class II molecules (yellow). The number of HLA-DR B (DRB) genes and pseudogenes differs for different haplotypes (DR52 haplotype shown; Fig. 39.2). Genes encoding protein products intimately involved in MHC class I and class II molecule assembly and in peptide processing and binding (green) also are shown. The class III region contains several immune system relevant genes and varies in length among individual chromosomes in the area encoding C4B (asterisk). The class I region encodes the classical HLA class I heavy chain proteins (yellow). This region also encodes the expressed nonclassical HLA class I genes (red) and MHC class I chain-related genes (blue).

(The figure was generated from data contained in references.^9,13)

The class II region (~1.2 Mb) contains at least 34 expressed genes and 16 pseudogenes and spans from SynGAP (Ras-GTPase-activating protein, centromeric) to TSBP (testis-specific basic protein, telomeric).¹¹ This region includes the genes that encode for the classical class II molecules (HLA-DR, -DQ and -DP). In addition, gene products involved in MHC class I antigen processing (LMP2 and LMP7, the large multifunctional proteosome genes), peptide transport (TAP1 and TAP2, the transporter associated with antigen processing genes) and complex assembly (tapasin) and gene products involved in MHC class II complex assembly (HLA-DM and HLA-DO) are encoded in this region.

The class III region (~1 Mb) extends from NOTCH4 (transmembrane receptor involved in cell differentiation and development, centromeric) to MCCD1 (mitochondrial coiled-coil domain protein 1, telomeric).^12,13 On average, this region contains 1 gene every 10 kilobases (kb) and is the most gene dense region in the human genome with approximately 72% of it being transcribed. The class III region encodes at least 62 expressed genes including complement components (e.g., C2 and C4B), heat shock proteins (e.g., Hsp70-1 and Hsp70-2) and cytokines of the tumor necrosis factor family (e.g., TNF and LTA). Some individuals have duplications in the area encoding complement component C4B; thus, this area can vary in length.

The class I region (~1.85 Mb) from MICB (centromeric) to HLA-F (telomeric) encodes at least 118 genes; 57 expressed genes and 61 pseudogenes.¹⁴ In some instances, an additional ~0.95 Mb to TRIM27 (tripartite motif 27, a transcription factor, telomeric) is included and called the extended class I region.⁹ This region includes the classical class I genes (HLA-A, -B and -C) and the nonclassical class I genes (HLA-E, -F, and -G). This region also encodes the MHC class I chain-related (MIC) genes (MICA and MICB). The focus of this chapter is on the MHC encoded gene products involved in histocompatibility, primarily the classical human leukocyte antigens. Other MHC and nonMHC encoded genes that participate in histocompatibility are also covered.

Genomic organization

The heavy chains of the classical MHC class I (class Ia) molecules are encoded in the class I region of the MHC and associate with beta 2 microglobulin (encoded on chromosome 15) to form the mature class I molecule. The gene order within the MHC is shown in Fig. 39.1. Each of the classical MHC class I heavy chains is encoded by a single gene that is divided into 8 exons.¹⁵ Exon 1 encodes the 5′ untranslated region (UTR) and the hydrophobic signal sequence. The signal sequence directs insertion of the protein into the membrane at the cell surface and is cleaved from the mature protein. The extracellular portion of the class I heavy chain is encoded by exons 2–4. Exon 5 encodes the transmembrane region and exons 6 and 7 encode the intracellular cytoplasmic tail. The 3′ UTR and the polyadenylation (poly(A)) site are encoded by exon 8. The mRNA, which is translated into protein, includes all eight exons after removal by splicing of the intervening sequences (introns).

The class II region of the MHC contains three subregions (centromeric) HLA-DP, -DQ and -DR (telomeric). Each encodes at least one cell surface class II molecule (Fig. 39.1). The class II molecules are noncovalently associated heterodimers that consist of an α chain and a β chain.^15,16 Each chain is encoded by a separate gene, an A gene for the α chain and a B gene for the β chain. The expressed HLA-DP heterodimer is encoded by the DPA1 and DPB1 genes. The HLA-DP subregion contains two DP pseudogenes, DPA2 and DPB2. The HLA-DQ subregion contains two A (DQA1 and DQA2) and three B (DQB1, DQB2 and DQB3) genes. The expressed HLA-DQ heterodimer is encoded by the DQA1 and DQB1 genes, while the remaining DQ genes are pseudogenes. Each individual has two copies of chromosome 6 and, thus, two copies of each of the expressed HLA-DP and HLA-DQ genes. These genes are polymorphic and, consequently, an individual can have two different expressed A genes and two different expressed B genes for HLA-DP and for HLA-DQ. While not all combinations form,¹⁷ the products of some of these genes can associate in several αβ combinations, regardless of chromosomal origin. Therefore, an individual could express up to four different HLA-DP and up to four different HLA-DQ molecules.

The HLA-DR subregion is more complex.^18,19 A HLA-DR molecule composed of a conserved α chain encoded by the DRA gene and a polymorphic β chain encoded by the DRB1 gene is almost always present. This is the major class II molecule expressed on the cell surface. An additional eight DRB genes and pseudogenes have been identified in the HLA-DR subregion. The number of DRB genes present and the number of expressed DRB gene products is characteristic of each chromosome (haplotype) a person inherits (Fig. 39.2). For example, the DR1 haplotype carries two DRB genes, the expressed DRB1 gene and a DRB6 pseudogene. The DR8 haplotype carries only one DRB gene, the expressed DRB1 gene. Other DR subregion haplotypes can encode a second expressed HLA-DR molecule composed of the DRA gene product associated with a DRB3 gene product (DR52 molecule), a DRB4 gene product (DR53 molecule) or a DRB5 gene product (DR51 molecule) and can contain one to three DRB pseudogenes. The designations DR51, DR52 and DR53 are antibody (serologically) defined.

Fig. 39.2 Relative gene organization of the DR subregion of different DR haplotype groups. The protein products of the DRA and DRB1 genes (yellow) combine to form the primary expressed HLA-DR molecule. The protein products of the DRB5 (green), DRB3 (blue) and DRB4 (red) genes encoded by the DR51, DR52 and DR53 haplotype groups, respectively, also form expressed functional DR molecules in combination with the DRA polypeptide. The DRB pseudogenes (ψ) are shown in black. Table 39.4 lists the alleles associated with each haplotype.

The A genes, which encode the α chains of class II molecules, contain five exons.^15,16 The 5′ UTR and hydrophobic signal sequence are encoded by exon 1, like the class I genes. Exons 2 and 3 encode the extracellular domains. Exon 4 encodes the connecting peptide, the transmembrane region, the intracellular cytoplasmic tail and a portion of the 3′ UTR. The remainder of the 3′ UTR and the poly(A) signal are encoded by exon 5. Each class II β chain is encoded by a B gene divided into six exons. Exons 1–3 are similar to that of the A genes. Exon 4 encodes the connecting peptide and transmembrane region, while exon 5 encodes the cytoplasmic tail. The 3′ UTR and poly(A) signal are encoded by exon 6. All exons and introns are transcribed into RNA for the class II A and B genes. Again, introns are removed by splicing to form the mRNA that is translated into protein.

Diversity

The nucleotide sequences of many of the HLA genes differ among individuals. These sequence variants are termed alleles; genes with many alleles are termed polymorphic. Alleles of a locus may differ by a single nucleotide to many nucleotides potentially resulting in changes in the amino acid sequence of the protein specified by that gene. The classical HLA class I and class II loci are the most polymorphic loci in humans. The HLA-B and HLA-DRB1 loci have over 1100 and 650 known alleles, respectively (Tables 39.1, 39.2).^20,21 In contrast to non-HLA genes, the nucleotide differences found among HLA alleles are usually nonsynonymous (alter the amino acid sequence) and are focused in the exon(s) encoding the most functionally important region of the HLA molecule, the antigen binding site.¹⁸ It is thought that this diversity has been maintained to provide the human population with the capacity to recognize a diverse repertoire of pathogenic peptides.^22–24 Unfortunately, the allelic differences in HLA molecules expressed on the cells of different individuals can be recognized as foreign when tissue is grafted from one individual to another.

Table 39.2 HLA class II allelesa,b

^a Listing of class II alleles assigned by July 2009.^20,21 The number of HLA class II alleles continues to increase as more individuals are studied. Alleles are defined by DNA sequencing. An expanded and updated table can be found at http://hla.alleles.org. A description of the nomenclature in this table can be found in Table 39.9.

^b Each column is independent.

^c Alleles of the DRA locus. The differences among these DR alpha chain alleles are not considered important for transplantation matching.

^d Alleles of the DRB1 locus. Most haplotypes contain a DRB1 locus.

^e Alleles of the DRB3 locus, the second expressed DR molecule in haplotypes carrying DRB1*03, *11, *12, *13, *14 alleles.

^f Alleles of the DRB4 locus, the second expressed DR molecule in haplotypes carrying DRB1*04, *07, *09 alleles.

^g Alleles of the DRB5 locus, the second expressed DR molecule in haplotypes carrying DRB1*15, *16 alleles.

^h Alleles of the DQA1 locus. DQA1 allelic products pair with DQB1 allelic products to form the DQ molecule.

ⁱ Alleles of the DQB1 locus.

^j Alleles of the DPA1 locus. DPA1 allelic products pair with DPB1 allelic products to form the DP molecule.

^k Alleles of the DPB1 locus. The approach to naming DPB1 alleles was slightly different than that used for other loci because of the lack of serologic information.²⁰

Several mechanisms are hypothesized to have generated this HLA diversity over evolutionary time. The majority of the polymorphism is hypothesized to have arisen by the non-reciprocal exchange of short polymorphic regions or cassettes among alleles, a process referred to as gene conversion. As a result, the HLA alleles are patchworks of polymorphic cassettes, each cassette shared by some of the other alleles at the locus, embedded in a conserved framework (Fig. 39.3). Exchange of cassettes among alleles at different loci, reciprocal recombination involving the exchange of entire exons, and mutation also have contributed to HLA diversity.^18,25

Fig. 39.3 Representation of a polymorphic exon of several HLA alleles. Each allele is a combination of conserved (solid line, nucleotides not shown) and polymorphic (colored blocks) nucleotide sequences. The polymorphic nucleotides form ‘cassettes’ which are shared among HLA alleles. The combination of cassettes within an HLA gene characterizes a specific allele. To identify the exact HLA alleles present in an individual, DNA-based typing must be able to identify these multiple cassettes and to link the cassettes to one another.

Each individual inherits two copies of the chromosome carrying the MHC, one from each parent, and thus has two copies of each gene in the MHC. Individuals who carry two identical alleles at a locus are homozygous; individuals with two different alleles at a locus are heterozygous. Because the HLA genes are located within a small genetic distance, they are usually inherited as a block unless separated by recombination. The block, a specific set of alleles at the multiple HLA loci inherited together from a parent, is termed a HLA haplotype. Fig. 39.4 illustrates the inheritance of HLA-A, -B, and -DRB1 alleles within a family. By convention, the paternal haplotypes are generally designated a and b and the maternal haplotypes, c and d. Thus, there are four possible MHC genotypes in the offspring: ac, ad, bc, and bd. Because the chances of inheriting a given genotype are random, the probability of occurrence of any one of the four genotypes is one in four in a mating. In a family with five children, at least two of the children will be MHC identical unless recombination has occurred.

Fig. 39.4 The inheritance of HLA alleles and haplotypes within a family. Paternal haplotypes are labeled a,b; maternal haplotypes, c,d. Sibling 5 inherits a recombinant chromosome from the mother. Not all HLA genes are shown. The HLA nomenclature used in the figure is described in Table 39.9.

The alleles in the MHC complex can be reshuffled by crossing over between homologous chromosomes during the generation of sperm or eggs. The frequency of recombination across the MHC from HLA-A to HLA-DPB1 can range from 0.7 to 4.3%.²⁶ Studies in humans suggest that there are several sites at which recombination preferentially occurs within the MHC, particularly between HLA-B and HLA-DRB1 and between HLA-DQB1 and HLA-DPB1.^26,27 Studies comparing MHC-identical sib pairs versus haplotype mismatched sib pairs and unrelated individuals show that recombination rates can vary up to sixfold between individuals with different MHC haplotypes and suggest a genetic influence within the MHC on recombination rates.²⁶ On average, the frequency of recombination between HLA-A and HLA-B is 0.7%, between HLA-B and HLA-DRB1 is 1.0%, and between DQB1 and DPB1 is 0.8%. Recombinations between DQA1 and DRB1 loci and between B and C loci are very rare.

The HLA alleles and haplotypes found in individuals depend on their racial and ethnic backgrounds.^28,29 For example, the allele DRB1*0302 is found in African Americans, but is only rarely observed in individuals of northern European or Asian descent. Likewise, the frequency of a combination of alleles on a single copy of chromosome 6 can differ among population groups. Table 39.3 lists the ten most common haplotypes identified in the African American population compared to the ranking of these haplotypes in several other US populations.²⁹ For example, the most common haplotype in African Americans is A*3001, Cw*1701, B*4201, DRB1*0302 found at a frequency of 1.5%. This haplotype was ranked 37th in Hispanic Americans but was not observed in this sampling of European Americans and Asian Americans.

There are, however, some alleles and haplotypes that are found in many populations. For example, the allele A*0201 is a common allele in many US populations, being found in 29.6% of European Americans, 12.5% of African Americans, 9.5% of Asian Americans and 19.4% of Hispanic Americans.²⁹ The most common haplotype in European Americans, A*0101, Cw*0701, B*0801, DRB1*0301, is the second most frequent haplotype in African Americans and Hispanic Americans but is the 191st most frequent haplotype in Asian Americans.

When large databases of HLA typed individuals are analyzed, only a small percent of potential HLA phenotypes are found. Using serologic assignments from an unrelated donor registry, of the predicted 19 536 660 HLA-A,-B,-DR phenotypes, only 1.6% were observed.³⁰ This suggests that not all HLA allele combinations will be found. Indeed, some HLA haplotypes appear more frequently than expected. Linkage disequilibrium measures the degree of non-random association between alleles of separate loci. Apparently high disequilibrium across the DR-DQ subregion coupled with a lack of recombination have resulted in specific associations between DQA1 and DQB1 alleles and between DRB1 and DQ alleles although a single allele such as DQB1 may be associated with one of several partner alleles. For example, DQB1*0602 is found on the same chromosome as the DQA1 alleles DQA1*0102, or *0103 or *0104, but has not been observed with DQA1*0201, *0301, *0401, or *0501.^31,32 These associations may differ among individuals of different racial/ethnic backgrounds. For example, DRB1*0901 is associated with DQB1*0201 in African Americans but with DQB1*0303 in individuals of northern European descent. Within the DR subregion, specific allele combinations at the several DRB loci are associated with families of DR haplotypes (Fig. 39.2). For example, the DRB3 locus is found in haplotypes carrying specific DRB1 alleles including DRB1*0301, *1101, *1201, *1301, and *1401 alleles (Table 39.4). In the class I region, associations between HLA-B and -C alleles have also been noted.²⁹

Table 39.4 DR haplotypes^a

^a It should be noted that there are exceptions to these associations. For examples, DRB1*15 haplotypes have been observed that lack a DRB5 locus and some DRB5 positive haplotypes lack a DRB1 locus. In addition, some DRB1*01 haplotypes carry a DRB5 locus.

Extension of linkage disequilibrium across longer regions of the MHC has resulted in associations between specific class I and class II alleles. The associations of multiple alleles result in extended haplotypes.^33,34 The most well known extended haplotype is A*0101, Cw*0701, B*0801, DRB1*0301 which is common in northern Europe, appearing at a frequency of approximately 5–15%.²⁸ It has been hypothesized that these associations may have been maintained in the population by selection, that is, associations between DR and DQ as well as associations within an extended haplotype might represent optimal combinations of immune response molecules. It is also likely that features of the genome structure limiting recombination or changes in the structure of the population, such as through admixture of different ethnic groups, have caused the linkage disequilibrium. Because alleles at various HLA loci are non-randomly associated, these associations enhance the frequency with which individuals share alleles across multiple HLA loci facilitating the selection of HLA identical individuals as tissue donors.

Expression

Classical MHC class I proteins are expressed by most nucleated cells, but the level of expression on the cell surface varies for different cell types. Cis-acting sequence blocks (enhancer A, interferon-stimulated response element (ISRE), W/S box, X box (previously known as site α) and Y box (previously known as enhancer B)) in the regulatory (promoter) region upstream of each class I gene control gene expression (Fig. 39.5A).³⁵ Each promoter sequence block binds numerous proteins (transcription factors) that regulate the level of transcription of the gene and ultimately the amount of protein at the cell surface. For example, enhancer A binds stimulating protein 1 (Sp1) and various members of the NF-κB/rel family of transcription factors. Collectively, the complex is termed NF-κB. Normal class I gene expression requires the coordinated action of each of these regulatory elements; however, disruption of any one sequence block reduces, but does not appear to ablate, MHC class I expression.

Fig. 39.5 (A–C) The promoter regions of the classical MHC class I and class II genes. (A) MHC class I gene expression is regulated by sequence blocks: enhancer A, E box, interferon-stimulated response element (ISRE) and the SXY module (W/S, X and Y boxes). Each block binds specific transcription factor complexes such as interferon response factor 1 and 3 (IRF-1/-3), regulatory factor X (RFX) and nuclear factor Y (NF-Y) that regulate expression. (B) Nucleotide sequences of each regulatory block of several MHC class I alleles compared to that of HLA-A*0201 with nucleotide variations underlined. Nucleotide sequence variations dictate the set of transcription factors that comprise each complex and the binding affinity of the complex to the regulatory block. These differences cause variations in the level of transcription of each class I gene and ultimately to differences in the level of specific MHC class I molecules expressed at the cell surface. (C) MHC class II gene expression is regulated by the SXY module. Again, each block of this module binds specific transcription factor complexes. Unlike class I gene expression, class II gene expression is entirely dependent on the class II transactivator (CIITA) (yellow). Similar to class I gene expression, differences in the nucleotide sequence of class II allele regulatory sequence blocks (examples not shown) ultimately affect the level of specific MHC class II molecules expressed at the cell surface. Transcription start site in (A) and (C) denoted with an arrow.

The amounts of HLA-A, -B and -C molecules expressed at the surface of a cell are not equal.³⁶ HLA-A and -B are abundant with HLA-A expressed at somewhat higher levels than HLA-B, in many instances. HLA-C is expressed at very low levels in comparison accounting for about 10% of cell surface class I molecules. This is due to sequence variations in the regulatory blocks of each class I locus (Fig. 39.5B) which alter the type and affinity of transcription factor binding. For example, only NF-κB/rel family members bind to enhancer A in the HLA-A promoter, while enhancer A in the HLA-B promoter binds SP1 in addition to NF-κB/rel family members. The inclusion of SP1 binding may lead to less efficient expression of the HLA-B gene. There are also allele specific differences in the nucleotide sequence of these regulatory elements such that, for example, different HLA-B alleles are expressed at different levels. Additionally, some HLA-B allele promoter regions encode an E box regulatory element that binds the transcription factors upstream stimulatory factor 1 (USF-1) and USF-2. The presence of the E box correlates with reduced basal expression levels of these HLA-B alleles.³⁵

MHC class I gene expression can be up-regulated by various cytokines.^35,36 Interferon-γ (IFN-γ) performs a fundamental role in enhancing MHC class I expression by inducing increased gene transcription via transactivation of the ISRE. Again, locus and allele specific sequence differences in the ISRE result in different levels of transcriptional enhancement for each of the class I genes. Additionally, IFN-γ up-regulates expression of the class II transactivator (CIITA) further enhancing MHC class I gene expression. Other cytokines, such as tumor necrosis factor (TNF), can enhance the stimulatory effect of INF-γ on MHC class I gene expression via up-regulation of NF-κB that acts through enhancer A.

MHC class II protein expression is more limited than that of MHC class I. Cell surface HLA-DR, -DQ and -DP molecules are found primarily on professional antigen presenting cells (APC) and on other immune system cells such as T-lymphocytes.^36,37 Professional APC are bone marrow derived cells dedicated to the task of peptide presentation by MHC molecules and include B-lymphocytes, macrophages, dendritic cells, thymic epithelial cells and Kupffer cells. IFN-γ can induce class II expression in other cell types. Like the class I genes, the promoter regions of the class II genes contain the cis-acting sequence blocks W/S box, X box and Y box (termed the SXY module) that regulate gene expression (Fig. 39.5C). This is a conserved regulatory module present in the promoters of a wide range of genes involved in antigen presentation which functions as a single regulatory block. The X and Y boxes bind several ubiquitous transcription factors, such as regulatory factor X (RFX) and nuclear factor Y (NF-Y), respectively, in a complex termed the enhanceosome.^38,39

Occupancy of the class II gene regulatory elements by the enhanceosome is absolutely required, but not adequate for expression and IFN-γ induction of the A and B genes of HLA-DR, -DQ and -DP. CIITA also is required for class II gene expression.^35,38 Cell type specific and IFN-γ induction of class II expression is the direct result of CIITA expression patterns. Constitutive expression of CIITA is confined to professional APC and other immune system cells and can be induced by IFN-γ in other cell types, paralleling expression of MHC class II. CIITA is recruited to the enhanceosome of the class II promoter by the S box. This is not the case for class I gene expression for which the S box apparently plays no role. All of the other transcription factors that regulate class II gene transcription are ubiquitously expressed and constitutively occupy the regulatory sequence blocks in the MHC class II gene promoters. Of interest, patients with bare lymphocyte syndrome (MHC class II deficiency) can have a defect in any one of a number of the transcription factors that bind to the SXY module or in CIITA. These patients still express MHC class I albeit at reduced levels.

HLA-DR, -DQ and -DP are not expressed at the same levels on cell surfaces, similar to expression of the different MHC class I molecules. HLA-DR is the most abundant MHC class II molecule expressed by cells. DRB1 is expressed at a higher level compared to DRB3 and DRB4.^40,41 HLA-DQ is expressed at reduced levels and HLA-DP is the least abundant cell surface class II molecule. Like the regulatory elements in the promoters of class I genes, there are both locus and allele specific sequence differences in the regulatory elements of the class II genes.³⁶ These sequence differences account for the dissimilar levels of class II molecule expression in two ways: 1) alter the binding affinity of the transcription factors and 2) allow binding of proteins that repress gene transcription of specific class II loci. For example, the X box in the HLA-DPA1 gene promoter region specifically binds the X box repressor protein which diminishes transcription of the HLA-DPA1 gene and reduces the overall level of HLA-DP on the cell surface.

Some pathogenic micoorganisms and many malignant cells down-regulate HLA gene expression to avoid recognition by the immune system.^42–44 For example, human cytomegalovirus (CMV) interferes with IFN-γ induction of MHC class I and class II gene expression.⁴⁵ To avoid detection by T-cells, many carcinomas and lymphomas lack cell surface HLA-A and HLA-B molecules due to defects in the expression or the binding of specific transcription factors to the promoter regulatory blocks of these genes. In many instances, expression of HLA-C in these malignant cells is unaffected, allowing the cell to avoid recognition by natural killer (NK) cells. In fact, alterations in MHC class I expression is very common (reported to be 70% or greater) in a variety of tumor types such as cervical, breast and colorectal cancers.⁴²

Structure of class I and class II molecules

The classical class I molecules (HLA-A, -B and -C) are expressed on cell surfaces as a trimolecular complex. This complex is composed of the HLA class I polypeptide (heavy chain), beta 2 microglobulin (β₂m) and a peptide. The MHC encoded class I heavy chains are glycosylated transmembrane proteins of approximately 340 amino acids (~44 kilodaltons (kD)) that belong to the immunoglobulin (Ig) superfamily of proteins.^15,46 The extracellular portion of the class I heavy chain is composed of the amino-terminal 275 amino acids. The following 40 amino acids make up the hydrophobic transmembrane region and the carboxy-terminal 25 amino acids comprise the intracellular cytoplasmic tail. As an aside, soluble isoforms of the classical HLA molecules are produced and may have immunoregulatory roles.⁴⁷

The extracellular portion of the class I heavy chain is divided into three domains, termed α1, α2 and α3 (Fig. 39.6A). Each domain is encoded by a separate exon (exons 2–4, respectively) and is approximately 90 amino acids long. The 3D structures of the extracellular portion of several class I molecules were resolved by X-ray crystallography.^48,49 The α1 and α2 domains fold together to form a groove (termed the antigen binding groove) distal to the cell membrane that consists of a floor of eight antiparallel beta strands topped by two alpha helices fashioned into the walls. The membrane proximal α3 domain folds into a structure which is similar to that of the constant region domains of immunoglobulins (antibodies). This domain is composed of two antiparallel beta sheets, one with four strands and one with three strands. The sheets are linked by a disulfide bond. β₂m (~12 kD) is non-covalently associated with the α3 domain of the class I heavy chain and is required for cell surface expression. Its 3D structure is identical to that of the α3 domain of the heavy chain.

Fig. 39.6 (A–D) Classical MHC class I and class II structures. (A) Ribbon diagram of the extracellular portion of a representative classical MHC class I molecule encoded by HLA-A*0201. The antigen binding groove is formed by the α1 (green) and α2 (blue) domains of the class I heavy chain. An eight-stranded β pleated sheet (broad arrows) forms the floor of the groove which is overlaid on two sides by α helical walls (twists). The backbone of a peptide (yellow) is seen within the groove. The α3 domain (brown) of the class I heavy chain non-covalently associates with beta 2 microglobulin (β2m; red). The transmembrane region and cytoplasmic tail (not visualized) would extend toward the bottom of the figure. The relative position of the HLA-Bw4/Bw6 and NK KIR C1/C2 epitopes on the α1 helix is circled. (B) Ribbon diagram of the extracellular portion of a representative MHC class II molecule, HLA-DR1; DR(α,β1*0101). The antigen-binding groove is formed by first domains of the α chain (α1; green) and β chain (β1; blue), similar to that of MHC class I. The backbone of a peptide (yellow) is shown in the groove. The class II α chain α2 domain (red) and class II β chain β2 domain (brown) form structures similar to β2m and the class I heavy chain α3 domain, respectively. The transmembrane region and cytoplasmic tail (not visualized) would extend toward the bottom of the figure. Top view of the electrostatic surface of the antigen binding grooves of the (C) class I molecule and (D) class II molecule shown in A and B, respectively. Negatively and positively charged surfaces are denoted by red and blue, respectively. The respective peptides (yellow) are depicted as stick models showing the carbon backbone and side chains. Pockets binding amino acid side-chains of each peptide are clearly visible in each figure. Ribbon diagrams were generated with the program Molescript and electrostatic surface models were generated with the program Grasp on a Silicon Graphics workstation from Protein Data Bank accession codes 1b0g (MHC class I) and 1aqd (MHC class II).

The peptide, the third component of the trimolecular complex, is generally 8–10 amino acids in length and non-covalently bound to the class I heavy chain.^48,49 It lies in the groove formed by the α1 and α2 domains (Fig. 39.6C). The peptide is anchored at its amino- and carboxy-terminal ends by non-covalent bonds to amino acid residues in the class I heavy chain. There are pockets along the groove which accommodate amino acid side chains at various positions along the peptide. The pockets are unique for each class I molecule because polymorphic residues from the α1 and α2 domains participate in their formation. Each pocket has specific physical and chemical characteristics that are determined by the conserved and polymorphic class I residues that form the pocket. These characteristics, in turn, dictate which amino acid side chains are accommodated at the corresponding peptide position. This defines the peptide binding motif of each class I molecule and defines the overall character of the set of peptides bound (Table 39.5).^50,51 For example, the protein encoded by A*1101 will accommodate a variety of ‘small’ amino acids at peptide postions 2, 3 and 6 and prefers basic amino acids at peptide position 9. However, each pocket does not make an equal contribution to peptide binding. Certain pockets, specific to each class I molecule, play a more predominant role and the corresponding peptide position is termed an anchor position. The preferred amino acids at an anchor position are termed anchor residues. Using the protein encoded by A*1101 again as an example, although peptide positions 2, 3 and 6 contribute to peptide binding, it is peptide position 9 that is the anchor position. Lysine and arginine are the anchor residues at this position with lysine preferred over arginine. It is of note that not all peptides that bind to an HLA molecule fully adhere to the defined peptide binding motif and that amino acids other than anchor residues can be found at anchor positions in these peptides. In the end, each class I molecule does bind a unique, large set of peptides and the peptide set shares particular sequence characteristics which are dictated by the amino acid residues that make up the groove of the HLA class I heavy chain.

There are benefits to determining the peptide binding motif for HLA molecules. For example, these motifs can be used to identify antigenic peptides from pathogen proteins as candidates for use in peptide based vaccines. As another example, expression of specific HLA allelic products is associated with an increased risk of developing many autoimmune diseases.^52,53 In most cases, this is thought to be the result of the differential binding capacity of HLA molecules for particular peptides. Thus, knowing the binding motif for a HLA molecule aids in the identification of the culprit peptide and allows for the design of synthetic peptides that mimic disease associated peptides for use in blocking autoimmune responses. Because the number of HLA allelic products is so large, algorithms have been developed for prediction of peptide binding to HLA molecules.^54–57

The class II molecules are expressed on cell surfaces as a trimolecular complex, structurally analogous to the class I molecules, and consist of the class II α chain, the class II β chain and a peptide. Both the class II α (34 kD) and β (28 kD) chains are transmembrane glycoproteins and are Ig superfamily members, comparable to the class I heavy chain.^15,16 The extracellular portion of the α and β chains are divided into two domains, the membrane distal α1 and β1 domains and the membrane proximal α2 and β2 domains. Similar to the class I heavy chains, each domain is encoded by a separate exon (exons 2 and 3, respectively) and is about 90 amino acids in length. Three additional regions complete each chain of the class II molecule; a connecting peptide of 12 amino acids which is highly hydrophilic and links the membrane proximal domain to the transmembrane region, a 23 amino acid hydrophobic transmembrane region and an intracellular cytoplasmic tail that consists of the carboxy terminal 8–15 amino acids.

The 3D structures of the extracellular portion of several class II molecules have been determined by X-ray crystallography.^48,49,58 These structures are strikingly similar to that of class I molecules. The α1 and β1 domains fold together to form a peptide binding groove like the groove formed by the class I heavy chain α1 and α2 domains (Fig. 39.6B). The α2 and β2 domains of the class II chains fold to form Ig constant region domain like structures similar to that of the class I α3 domain and β₂m.

The peptides that bind to class II molecules are anchored to the class II antigen binding groove by non-covalent bonds to the peptide backbone and by binding of peptide amino acid side chains into pockets along the groove (Fig. 39.6D), similar to the class I molecules.^49,58,59 Because of the polymorphic nature of the MHC class II proteins, each class II molecule, like each class I molecule, also binds a large set of peptides which share a peptide binding motif specific to that class II molecule (Table 39.5). The peptides that bind to class II molecules are heterogenous in length and generally 13–25 amino acids long. The low and open ends of the class II groove allow peptides of varying lengths to bind in an extended conformation with the ends of the peptide overhanging the ends of the groove. This is in contrast to the class I molecules which bind peptides of 8–10 amino acids. The ends of the class I groove are high and closed; thus, MHC class I molecules optimally accommodate shorter peptides whose ends are tucked into the groove (compare Fig. 39.6C and 39.6D).

Function

Peptide processing and binding

Mature cell surface class I molecules are formed in the endoplasmic reticulum (ER) with the aid of several resident ER proteins including tapasin, calnexin and calreticulin.^60,61 Initially, the class I heavy chain and β₂m fold and associate facilitated by calnexin and calreticulin (Fig. 39.7). This complex then transiently associates with the transporter associated with antigen processing (TAP) where a peptide is loaded into the groove of the class I heavy chain. Finally, the trimolecular complex is dispatched to the cell surface.

Fig. 39.7 MHC class I and class II peptide processing and binding. MHC class I peptide processing and binding is depicted in the left portion of the figure. Class I heavy chain (blue) and beta 2 microglobulin (β₂m) (lavender) fold and associate in the endoplasmic reticulum (ER) aided by resident ER proteins. Cytosolic derived proteins are converted into peptides (orange) by the proteosome, shuttled into the ER by the transporter associated with antigen processing (TAP) and trimmed by the ER aminopeptidase I (ERAPI) where they bind to the class I antigen binding groove. Stable class I + β₂m + peptide complexes are transported from the ER to the Golgi where the class I polypeptide is glycosylated. Mature MHC class I molecules then are shuttled to the cell surface. MHC class II peptide processing and binding is depicted in the right portion of the figure. Like MHC class I and aided by resident ER proteins, the class II α (green) and β (yellow) chains associate in the ER along with the invariant chain (Ii) (light gray). The class II αβ + Ii complex is glycosylated as it is transported through the Golgi to the MHC class II compartment (MIIC). Here, proteases cleave Ii leaving only the Ii derived peptide CLIP bound to the class II binding groove (class II + CLIP). HLA-DM (dark gray) catalyzes the exchange of CLIP for peptides (orange) derived by proteolytic degradation from internalized microorganisms and proteins. Stable class II + peptide complexes then traffic to the cell surface and represent mature MHC class II molecules.

Peptides, derived from both self (normal cellular) proteins (potential autoantigens) or foreign proteins (antigens), are generated in the cytosol by the proteosome.^60,61 The proteosome is a macromolecular structure that proteolytically cleaves proteins into peptides (a process termed antigen processing) and consists of members of the large multifunctional proteosome (LMP) family and other protein subunits. Two LMP family members (LMP2 and LMP7) are encoded in the class II region of the MHC. The proteosome is tightly associated with the TAP molecule which shuttles the peptides into the lumen of the ER.^60,61 TAP is formed by the association of the products of the TAP1 and TAP2 genes also encoded in the MHC class II region. Another MHC encoded gene product, tapasin, stabilizes the TAP heterodimers, links the class I heavy chain to TAP for peptide loading and facilitates loading of peptides onto the class I molecule. An endoplasmic reticulum aminopeptidase plays a dual function in class I antigen presentation. It trims longer peptides that are more readily shuttled by TAP to the optimal length (8–9 amino acids) enhancing binding to class I molecules, but also destroys many peptides limiting the pool available for binding to class I.^62,63

The class II α and β chains fold and associate in the ER with the assistance of resident ER chaperones such as calnexin (Fig. 39.7),⁶⁴ similar to class I molecules. Unlike class I, full maturation of the class II molecule does not take place in the ER. Instead, class II heterodimers are directed primarily to specialized endosomal compartments (MHC class II compartments, MIIC), or, in some instance, first traffic to the cell surface and then to the MIIC.⁶⁵ Once in the MIIC, peptides are loaded into the antigen binding groove and mature class II molecules are dispatched to the cell surface.

MHC class II molecules bind peptides derived from endocytosed microorganisms and from self and foreign proteins degraded by proteases in the endocytic pathway.⁶⁶ This is in contrast, yet complementary, to the peptides bound by MHC class I molecules. In general, class II molecules bind peptides from cell surface and extracellular sources, while class I molecules bind peptides from intracellular sources. Thus, proteins from the whole environment of a cell can be surveyed by the immune system. Invariant chain (Ii), a nonMHC encoded glycoprotein, plays a key role in facilitating this division of function.

Ii performs several functions in assuring proper antigen presentation by class II molecules. Ii chain serves as a chaperone in the folding and assembly of class II αβ heterodimers and protects the class II peptide binding groove from binding peptide in the ER via a 25 residue internal peptide segment termed CLIP (class II associated invariant chain peptide).^65,66 Ii also provides the intracellular targeting signals that direct the complex to the MIIC. Under the acidic conditions of the MIIC, Ii is proteolytically cleaved and dissociates from the class II molecule, while CLIP remains bound to the class II antigen binding groove. CLIP is exchanged for antigenic peptides in a reaction catalyzed by HLA-DM, a resident MIIC protein that tightly associates with the class II heterodimer.^67–69 HLA-DM is a critical component in shaping the repertoire of peptides bound to class II molecules as it retains the class II molecules in the MIIC until a stable, high affinity complex between the class II molecule and a peptide is formed. In certain APC types, B-cells and thymic epithelial cells, another resident MIIC protein, HLA-DO, negatively regulates the actions of HLA-DM.^67,68

Analogous to the MHC class II molecules, both HLA-DM and HLA-DO are class II related Ig superfamily members expressed as heterodimers that consist of an α chain and a β chain.^67–69 The HLA-DM and HLA-DO α and β chains are encoded by A and B genes, respectively, in the MHC class II region and regulation of these genes is similar to that of the MHC class II genes.⁶⁹ The 3D structure of HLA-DM resembles that of the MHC class II molecules except that its peptide binding groove is almost entirely obscured.⁴⁸

Components of the peptide processing and binding pathways of MHC class I and class II molecules are a favorite target of disruption by many pathogens and malignant cells to avoid detection by the immune system.^42–44 For example, two proteins (US3 and US6) encoded by human CMV block cell surface expression of MHC class I molecules and thus, detection of the infected cell by cytotoxic T-cells. US3 binds to MHC class I molecules and retains them in the ER and US6 inhibits peptide transport into the ER by TAP. Lack of cell surface MHC class I, however, renders the CMV infected cell susceptible to lysis by NK cells. To circumvent NK cell recognition, CMV encodes a class I like decoy termed UL18 which is recognized by NK cells and inhibits their function.⁴⁵ Other pathogens and malignant cells employ a variety of unique strategies to block MHC expression.

T lymphocyte recognition of HLA molecules

The predominant function of classical MHC class I and class II molecules is to bind peptides from the cell’s environment and present these peptides on the surface of the cell for inspection by the immune system. The archetypical receptor involved in the inspection process is the T-cell receptor (TCR) on T lymphocytes (Fig. 39.8). There are two types of TCR (αβ and γδ) both of which are multipolypeptide complexes that consist of a TCR α chain covalently paired with a TCR β chain or of a TCR γ chain covalently paired with a TCR δ chain tightly, but noncovalently, associated with several chains of the CD3 family (α, δ, ε, ζ, η).^70,71 The TCR chains are involved in the direct recognition of the HLA molecule and of the peptide bound to the HLA molecule.^71,72 CD3 is involved in the signaling process which activates the T-cell after recognition of the ligand by the TCR chains.⁷⁰

Fig. 39.8 CD4⁺ versus CD8⁺ T-cells. T-cells are classically divided into two groups based on their expression of either CD4 (top, dark gray) or CD8 (bottom, light gray). Both types of T-cell express the TCR (blue) and CD3 complex (CD3) (lavender). CD4⁺ T-cells are primarily of the helper phenotype (Th) and CD8⁺ T-cells are generally of the cytotoxic phenotype (Tcyt). The TCR expressed by CD4⁺ Th-cells is usually restricted to MHC class II molecules (green), while the TCR expressed by CD8⁺ Tcyt cells is generally restricted to MHC class I molecules (yellow). Peptide bound to each MHC molecule is depicted in orange.

T lymphocytes are classically divided into two groups based on expression of co-receptors (CD4 and CD8) which are intimately involved in recognition of the HLA molecule.^71,73 CD4 expressing (CD4⁺) T-cells are usually MHC class II restricted; that is, their TCR recognizes either an HLA-DR, -DQ or -DP molecule which all have a CD4 binding site (Fig. 39.8). CD8 expressing (CD8⁺) T-cells are generally MHC class I restricted recognizing either an HLA-A, -B or -C molecule which all have a CD8 binding site. CD4 and CD8 enhance the interaction between the TCR and HLA molecule and provide signals, like the CD3 chains, to activate the T-cell.

The CD4/CD8 division in MHC restriction, for the most part, also extends to the general phenotypic function of the T-cell. CD4⁺ T-cells are mostly of the helper phenotype (helper T-lymphocyte; Th-cell). Th-cells are dedicated to the initiation and generation of immune responses to specific antigens, including antibody production by B-lymphocytes and cytotoxic cellular responses. CD8⁺ T-cells are usually of the cytotoxic phenotype (cytotoxic T-lymphocyte; Tcyt-cell) and are the principal component of the cytotoxic cellular response to specific antigens. Tcyt-cell recognition of a foreign peptide bound to a MHC class I molecule leads to killing of the abnormal or infected cell.

Like immunoglobulins, the TCR chains are encoded by novel genes formed by the combining of gene segments termed variable, diversity, joining and constant.^74,75 Multiple gene segment alternatives and variable joining lead to an enormous potential repertoire of TCR structures. The outcome of this diversity is the generation of a large pool of T-cells expressing a wide range of TCR that can recognize a wide variety of antigenic peptides. Simplistically, each T-cell, through its TCR, recognizes a specific MHC molecule (MHC restriction) in combination with a distinct peptide (antigen specific recognition) (Fig. 39.9). In actuality, a single TCR can recognize with different affinities many (estimated to be approximately 10⁶) different MHC/peptide complexes that share common structural features.⁷²

Fig. 39.9 T-cell recognition is antigen specific. The TCR (blue) expressed by different individual T-cells differ and, thus, each recognizes a specific peptide bound to a specific MHC molecule. In the example shown, the four CD4⁺ (dark gray) Th-cells have differing specificities for particular peptides (orange) due to variations in the sequence of their TCR. However, each recognizes their respective peptide bound to the same MHC class II molecule (green). The CD3 complex (CD3) is depicted in lavender.

MHC class I and class II molecules are essential to the creation of the T-cell pool available for immune surveillance. T-cell maturation and selection occurs in the thymus in a process called thymic education.^76,77 In the thymus, T-cells are in contact with MHC class I and class II molecules that are complexed with a variety of self peptides. The TCR must be able to interact with a self MHC molecule complexed with a self peptide. Many T-cells have a TCR that is unable to bind to self MHC plus self peptides complexes and these die in the thymus. T-cells whose TCR has a low affinity interaction with and are partially activated by self MHC plus self peptide complexes (positive selection) are self tolerant and leave the thymus for peripheral tissues. These T-cells maintain some degree of partial autoreactivity and constitute the pool available for immune responses to foreign peptides. T-cells whose TCR has a high affinity for and are reactive to (autoreactive) these complexes are deleted (negative selection) from the T-cell pool via the generation of a variety of intracellular signals. It has become apparent in recent years that a subset of these autoreactive cells are not deleted from the thymus, but instead develop into regulatory T-cells (Treg). Treg-cells are not themselves reactive to self antigens, but instead suppress the activity of other potentially autoreactive T-cells.^76,78

The diversity of MHC molecule + self peptide complexes is central not only to the selection of a diverse repertoire of T-cells in the thymus, but also to the maintenance of the multiplicity of available T-cells in peripheral tissues.⁷⁹ Disruptions in this process can result in autoreactivity and autoimmune diseases. Other histocompatibility antigens (nonclassical MHC class I molecules, MHC class I related chains and CD1 molecules) are expressed in the thymus and by other tissues and likely help to select and maintain the available T-cell pool. In a human tissue transplant, amino acid sequence differences (mismatches) between any of the HLA molecules expressed by the cells of the donor and of the recipient, theoretically, can cause T-cell reactivity to the foreign HLA molecule. This reactivity can manifest itself as either graft rejection or GVHD and results because the T-cells were not educated for tolerance to the disparate HLA molecule.

The TCR is not the only receptor that binds MHC molecules as its ligand. Other receptors are discussed in the section on ‘Other immune system receptors’ and their interactions with MHC molecules regulate a variety of functions by immune system cells.

Allorecognition and transplantation

Allorecognition results when T-cells of one individual react to foreign (allogeneic) MHC molecules of another individual. There are three distinct pathways of allorecognition termed direct, indirect, and semi-direct^80,81 (Fig. 39.10). Direct allorecognition occurs when recipient T-cells recognize as foreign an intact allogeneic MHC molecule expressed by a donor cell.⁸¹ Indirect allorecognition occurs when recipient T-cells recognize a self MHC molecule that has bound a peptide derived from the foreign MHC molecule which contains the amino acid difference(s). This type of allorecognition involves the uptake of grafted cellular material, from which the foreign MHC molecules are processed into peptides and bound to self MHC molecules via the normal MHC class I and class II peptide processing and binding pathways. The semi-direct pathway involves the uptake and surface expression of intact foreign MHC+peptide complexes by recipient APC. Since these APC can also process and present foreign MHC molecules as peptides, this enables both direct and indirect recognition of foreign MHC on the surface of the same APC by recipient T-cells. When immune cells are present in the graft, donor T-cells may recognize recipient MHC by these same pathways. These pathways may contribute differentially to the alloresponse over time.

Fig. 39.10 Allorecognition is the recognition of and reaction to determinants of foreign MHC molecules by T-cells. Two of the types of allorecognition, direct (top) and indirect (bottom), are shown. In direct allorecognition (top), the TCR (blue) expressed by a CD4⁺ (dark gray) Th-cell is directly recognizing the differences of the foreign class II molecule (red) complexed with self peptides (yellow) which may or may not play a role in recognition of the complex. In indirect allorecognition (bottom), the TCR expressed by a CD4⁺ Th is recognizing a peptide (red) derived from the foreign class II molecule in complex with a self class II molecule (green). This peptide contains one or more amino acids that differ from the comparable peptide (green) derived from the self class II molecule.

Allorecognition is purely a transplantation phenomenon. Successful transplantation of tissues, especially HSC grafts, relies heavily on the HLA identity of the donor and the recipient.⁸² Unfortunately, HLA mismatches between a donor and recipient are often unavoidable. Any HLA disparity has the potential to cause allorecognition which increases the risk of graft rejection or GVHD. However, not all HLA disparity leads to destructive alloreactive T-cell responses.

The frequency of T-cells of any individual that respond to allogeneic MHC is substantially greater than the frequency of T-cells that respond to a specific foreign antigen.⁸³ The high frequency of potentially alloreactive T-cells is most likely due to the fact that the available T-cell pool consists of T-cells that are inherently reactive to MHC molecules complexed with self peptides.^81,84 Furthermore, because the TCR itself is flexible, a single TCR is able to recognize multiple MHC+peptide complexes and to accommodate differences in structure.

Non-classical MHC class I molecules

The non-classical MHC class I (class Ib) molecules, HLA-E, HLA-F and HLA-G, are encoded in the class I region of the MHC (Fig. 39.1) and the proteins expressed by these genes are structurally similar to the classical class I molecules (class Ia). However, the class Ib molecules, in general, are noticeably less polymorphic and expressed at lower levels on cell surfaces than the class Ia molecules. These glycoproteins also display unique tissue distribution patterns and participate in specialized immune system functions. The impact of the class Ib molecules on immune responses directed toward foreign tissue is not yet known.

MHC class I chain-related molecules (MIC)

The genes encoding the MHC class I chain-related molecules (MIC) are located in the class I region of the MHC centromeric to HLA-B (Fig. 39.1).¹⁰¹ MICA and MICB are relatively polymorphic.^21,102 To date, 67 alleles of MICA that encode 56 unique proteins and 30 alleles of MICB that encode 19 unique proteins have been described.

Several features are unique to MICA and MICB. Expression of MICA and MICB is up-regulated in response to stress because of a heat shock response element in the promoter region of their genes.¹⁰¹ In fact, the promoter regions of the MICA and MICB genes do not contain any of the classical MHC class I regulatory elements (Fig. 39.5). Instead, their promoters share homologies with the heat shock protein 70 promoter. MICA and MICB protein expression is limited to endothelial cells, fibroblasts and epithelial cells primarily in the gastrointestinal tract and thymus and does not require association with β2m, peptide or TAP.¹⁰³ The 3D structure of the extracellular domains of MICA revealed that the folding resembles that of the class Ia molecules, but that the overall orientation of these domains is dissimilar.¹⁰⁴ The MICA α1 and α2 domains are tilted to expose a portion of the underside of the groove, as well as the groove itself to interactions with receptors. The peptide binding groove is mostly obscured and unlikely to bind peptide antigens.

MICA and MICB serve as ligands for NKG2D, an activating NK cell receptor.^101,103 The interaction of NKG2D with the MIC molecules leads to NK cell mediated lysis of the epithelial cell. In fact, MIC expression is up-regulated in a wide variety of epithelial tumors such as melanoma, colon and kidney and it has been shown that the response to these cells is mediated through the interaction of NKG2D with MICA and MICB.¹⁰⁵ MICA and MICB also may play a role in solid organ transplantation. Both molecules have been found to be expressed on epithelial and endothelial cells of transplanted kidney, heart and pancreas that show evidence of rejection and/or cellular injury.¹⁰⁶ Further, the presence of MIC-specific antibodies in solid organ transplant recipients shows an association with graft failure.¹⁰⁷ Thus, MIC molecules appear to be transplantation antigens that can mediate an alloresponse by NK cells and B-cells that leads to graft rejection.

CD1 molecules

CD1 is a family of proteins whose nonpolymorphic genes are encoded on an MHC paralogous region on chromosome 1.¹⁰⁸ CD1 molecules are structurally close to MHC class I molecules at the gene and protein levels, while they are functionally close to MHC class II molecules as CD1 molecules associate with invariant chain and utilize the class II antigen presentation pathway.^65,109 Like both MHC class I and class II molecules, CD1 molecules present antigens to T-lymphocytes for immune recognition.

Five CD1 molecules have been described which are classified into two groups based on sequence homology and tissue distribution.¹⁰⁸ Members of group I (CD1a, -b, -c, -e) are expressed primarily on professional APC, such as dendritic cells and B-cells, and on cortical thymocytes; while members of group II (CD1d) are expressed on most cells of hematopoietic origin and on hepatocytes, keratinocytes and intestinal epithelium. The crystal structures of CD1a, CD1b and CD1d reveal a general conformation similar to MHC class Ia molecules except that the antigen-binding grooves are narrower, deeper and very hydrophobic.^108,110 There are two deep pockets at either end of the groove that accommodate the lipid tails of lipid antigens. The depth of these pockets varies (CD1d has the deepest and CD1a has the shallowest), conferring different lipid antigen binding characteristics on each of the CD1 molecules. Several lipid antigens have been identified from bacterial, synthetic and self sources. For example, CD1b binds mycolates derived from mycolic acid of Mycobacterium species, while CD1a binds lipopeptides also from mycobacteria. Lipid antigens also have been identified from Borrelia burgdorferi (Lyme disease) and Sphingomonas.

Group I CD1 molecules present lipid antigens to T-cells that express either γδ TCR or αβ TCR.^108,110 CD1d (group II) is the ligand primarily for invariant NK T-cells (iNKT). iNKT cells are usually double negative (CD4⁻/CD8⁻) or CD4⁺ T-cells that express a specific αβ TCR (Vα24-Jα18 with Vβ11) together with receptors found on NK cells, CD161 (NK-RP1A) in particular. iNKT cells develop in the thymus from double positive (CD4⁺/CD8⁺) T-cells like other T-cells, but require interaction with CD1d in the thymus for their differentiation and maturation. iNKT cells play a large role in enhancing the immune response to certain microbial pathogens by stimulating APCs, T-cells and B-cells. These cells can also be directly cytotoxic. Because the CD1 molecules are not polymorphic, they are unlikely to have an impact on donor selection for tissue transplantation.

Minor histocompatibility antigens

Clinically significant GVHD is observed in HLA identical sibling transplants of HSC implicating histocompatibility antigens specified by genes other than the classical HLA genes in the allorecognition of foreign tissue.¹¹¹ These so-called minor histocompatibility antigens (mHag) were identified coincident with the major histocompatibility antigens. mHag are peptides derived from self proteins, other than HLA, whose sequence differs among individuals in the population. Theoretically, any polymorphic protein that differs between the tissue donor and recipient has the potential to provide a peptide which functions as a mHag. Many mHag loci have been described in humans utilizing mHag specific T-cell clones in cytotoxicity assays. However, only a handful of peptides derived from these loci have been identified (Table 39.6).^112,113 The frequency of the mHags alleles varies in different population groups and some can be quite common.¹¹⁴ The peptides containing the variant amino acid(s) are processed through the normal MHC class I and class II antigen presentation pathways, are bound to an HLA molecule and are recognized as foreign by T-cells.^111,115

The HLA alleles expressed by a donor and recipient determine whether a mHag might contribute to allorecognition in transplantation. Furthermore, the contribution of a mHag on transplant outcome might be contingent on the sex of the donor and recipient and on the type of tissue transplanted (Table 39.6). HLA alleles are important because each HLA allelic product binds peptides that fit a specific binding motif and, thus, each mHag is only bound by certain HLA molecules. Therefore, the mHag(s) of significance differ widely depending on which HLA alleles are carried by the tissue donor and recipient (transplant pair). For example, mHag HA-1 could have an impact if the donor or recipient expresses HLA-A*0201 and there is HA-1 disparity between the transplant pair.^113,115 Recognition of the H allelic form of the HA-1 peptide bound to the A*0201 encoded molecule stimulates allorecognition. HA-1 would be of no consequence to the graft when, for instance, both donor and recipient are homozygous for HLA-A*0101, as HA-1 is not able to bind to this HLA-A molecule and consequently is never presented to T-cells. The HA-1 R peptide does not elicit an immune response. HA-1 also can be used to demonstrate the importance of the type of tissue involved in the transplant. HA-1 is expressed only on cells of hematopoietic origin. Therefore, HA-1 most likely would not be pertinent in the case of a kidney transplant, even if the transplant pair expresses HLA-A*0201, since HA-1 would not be expressed by the transplanted kidney.

It is now apparent that mHags not only have an impact on the development of GVHD, but also can be beneficial in generating a graft versus tumor (GVT) response^111–113115 as cancer cells of both hematopoietic and epithelial origin have been found to aberrantly express mHag. Once characterized, mHags can be included in matching protocols when they are pertinent to transplantation of tissue or to the elimination of residual malignant disease.¹¹²

Killer cell immunoglobulin-like receptors

The killer cell immunoglobulin-like receptors (KIR) are, at least, a 14 member family of proteins expressed by NK cells and a subset of T-cells.¹¹⁶ Like the HLA molecules and TCR, KIR belong to the immunoglobulin (Ig) protein superfamily and possess either 2 (KIR2D) or 3 (KIR3D) glycosylated extracellular Ig like domains (Fig. 39.11). The type of signal (activating or inhibiting) generated by a KIR is dictated by its cytoplasmic tail. Members that generate inhibitory signals, preventing cytotoxicity and cytokine release, have a long cytoplasmic tail (e.g., KIR3DL1 and KIR2DL2) which contains immunoreceptor tyrosine based inhibitory motifs (ITIM). Binding of the inhibitory KIR to its ligand leads to phosphorylation of the tyrosine residue in each ITIM and ultimately to inhibition of NK and T-cell functions. Activating KIR family members have a short cytoplasmic tail without ITIM (e.g., KIR2DS1 and KIR3DS1). Instead, these members associate via a conserved positively charged residue in their transmembrane regions with an adaptor protein, often DAP12, that contains immunoreceptor tyrosine-based activation motifs (ITAM) in its cytoplasmic tail. Interaction of an activating KIR with its ligand causes phosphorylation of the ITAM in the adaptor which may lead to activation of NK and T-cell functions. Since NK and T-cells express multiple activating and inhibiting receptors including KIR, cellular function is finely regulated and represents an integration of signals.¹¹⁷

Fig. 39.11 Examples of the three receptor families that recognize HLA molecules as their ligands. The killer cell immunoglobulin-like receptors (KIR, green, top) have either two or three extracellular Ig like domains. Inhibitory KIR have long cytoplasmic tails and signal via immunoreceptor tyrosine based inhibitory motifs (ITIM, gray box). Activating KIR with short cytoplasmic tails associate with a signaling protein called DAP12 (yellow) that has an immunoreceptor tyrosine based activation motif (ITAM) (black box) in its cytoplasmic tail. KIR2DL4 has an ITIM but is activating by associating with Fc epsilon receptor gamma chains (FcεRγ, magenta) that contain ITAM in their cytoplasmic tails. The ligands for many of the KIR are groups of class I HLA molecules (heavy chain, light green; β₂m, blue; peptide, orange). For example, KIR2DL1 recognizes group 2 HLA-C class I molecules; KIR2DL3 recognizes group 1 HLA-C class I molecules; KIR3DL3 recognizes HLA-B class I molecules with a HLA-Bw4 public epitope (Fig. 39.6, Table 39.8); and KIR2DL4 recognizes soluble HLA-G. The ligands for some KIR are unknown (denoted by question mark). NKG2A, -B, -C, -E and -H (orange, center) are expressed as heterodimers with CD94. NKG2D is expressed as a homodimer. Like the KIR, inhibitory members contain ITIM in their cytoplasmic tails, while activating members associate with DAP12. The exception is NKG2D, an activating receptor, which associates with another signaling protein called DAP10 (lavender) that possesses a phosphatidylinositol-3 (PI-3) kinase binding site (white box) instead of an ITAM. The ligand for all CD94/NKG2 heterodimers is the non-classical MHC class I molecule, HLA-E. NKG2D recognizes the MHC class I chain-related A and B (MICA/B) molecules (light blue). The leukocyte immunoglobulin-like receptors (LILR) have either two or four extracellular Ig like domains (cyan, bottom). Again, inhibitory members possess ITIM in their cytoplasmic tails. Activating members associate with FcεRγ (magenta) like KIR2DL4. LILRB1 and LILRB2 recognize a variety of HLA molecules. HLA-B27 allelic products are the ligand for LILRA1 and LILRA2 recognizes soluble HLA molecules. The ligand(s) for other family members is not known.

Classical MHC class I molecules serve as the ligand for several of the KIR family members (Fig. 39.11).^116,118 Some KIR with two extracellular Ig like domains interact with groups of HLA-C molecules.^116,119 Almost all HLA-C molecules can be divided into one of two groups (C1 and C2) based on the presence of a shared (public) epitope determined by amino acid residue 80 in the α1 domain of the HLA-C molecule.¹²⁰ The C1 group (e.g. Cw1 encoded by Cw*010201) has an asparagine at residue 80 and the C2 group (e.g. Cw2 encoded by Cw*020201) has a lysine at residue 80. C1 group molecules are ligands for KIR2DL2 and KIR2DL3, while C2 group molecules are ligands for KIR2DL1. A KIR with three extracellular Ig like domains, KIR3DL1, interacts with a group of HLA-B molecules carrying the Bw4 public epitope. HLA-G is recognized by KIR2DL4.¹⁰⁰ Ligands for some KIR have not been identified.

The genes encoding KIR are located in the leukocyte receptor complex on chromosome 19.¹²¹ Haplotypes vary in the number of KIR genes they encode.¹²² One common haplotype includes only seven of the 14 KIR genes: KIR3DL3, KIR2DL3, KIR2DL1, KIR2DL4, KIR3DL1, KIR2DS4, and KIR3DL2. Other haplotypes include more of the stimulatory KIR genes. In addition to gene content variability, the KIR genes are polymorphic.¹²³ For example, there are over 50 KIR3DL1 alleles. KIR3DL1 allelic variation alters its level on the cell surface and impacts the affinity of interaction with its HLA ligand.¹²⁴

At the level of the individual NK cell, only a subset of the KIR genes carried by an individual are expressed.¹²⁵ During development, individual NK cells acquire specific KIR, apparently at random so that an individual who carries KIR2DL1 and KIR2DL3, for example, will have NK cells that express only KIR2DL1, only KIR2DL3, and KIR2DL1+KIR2DL3. This variegated expression increases the sensitivity of the NK cells to the loss of individual HLA molecules during tumorgenesis or viral infection. Because KIR and HLA genes are found on different chromosomes, individuals may carry a KIR gene yet lack its cognate HLA ligand. For example, an individual may express KIR3DL1 but not express an HLA-B molecule carrying a Bw4 epitope. To avoid autoreactivity by absence of a ligand, it is thought that individual NK cells are educated by interaction with MHC class I molecules and that this process controls the functional responsiveness of the NK cells.¹²⁶ In KIR3DL1 positive, HLA-Bw4 negative individuals, NK cells expressing KIR3DL1 as their only inhibitory receptor will be hyporesponsive, unable to react to the absence of the HLA-Bw4 ligand.¹²⁷

Specific KIR genes have been associated with susceptibility or resistance to a variety of infectious and autoimmune diseases.¹¹⁸ The impact of NK cells, and inhibitory KIR in specific, in eradicating residual malignant disease was first noted in T-cell depleted haploidentical HSC transplants.¹²⁸ Subsequent studies have underscored the difficulty in predicting NK alloreactivity in the context of differing transplantation protocols and recent studies have continued to investigate the utility of considering KIR when selecting a HSC donor.¹²⁹ In general, algorithms to predict NK alloreactivity have focused on KIR and HLA genotypes of the donor and the recipient.¹³⁰ The goal is to select a donor expressing an inhibitory KIR ligand for a patient who does not express the ligand for that receptor, for example, to select a KIR2DL1 positive donor for a patient who is negative for the HLA-C2 group. It is then expected that donor NK cells, released from their inhibition, will attack residual malignant cells expressing NK cell activating ligands. Observations of NK reactivity in HSC transplantation have lead to clinical trials using adoptive immunotherapy with NK cells for patients with myeloid leukemia and with solid tumors.¹³¹

CD94 and NKG2

CD94 and the NKG2 family (NKG2A-NKG2F and NKG2H) are primarily transmembrane glycoproteins and their genes are encoded in the NK complex on chromosome 12.¹³² NKG2A, NKG2C, NKG2E and NKG2F share greater than 94% protein sequence homology in their extracellular domains. NKG2A, 2C and 2E are expressed as covalently linked heterodimers with CD94. NKG2B and NKG2H are alternative mRNA splice products of the NKG2A and NKG2E genes, respectively, and also are expressed as heterodimers with CD94. NKG2F does not form heterodimers with CD94 and is expressed in intracellular compartments. NKG2D is discussed separately below.

Like the KIR, CD94/NKG2 heterodimers and NKG2F are expressed by NK cells and subsets of CD8⁺ T-cells and can either inhibit or activate cellular functions.¹³² The action of the CD94/NKG2 heterodimeric receptors is determined by the NKG2 family member. CD94/NKG2A and CD94/NKG2B are inhibitory with two ITIM in their cytoplasmic tails. NKG2C, NKG2E and NKG2H lack ITIM in their cytoplasmic tails and, like activating KIR, are associated noncovalently with DAP12 and are activating receptors. NKG2F, although intracellular, also associates with DAP12 and is an activating receptor.

All of the CD94/NKG2 heterodimers recognize the non-classical class I molecule HLA-E as their ligand; NKG2A/B and NKG2E/H with equal affinity and NKG2C with tenfold less affinity.¹³² Polymorphism has not been demonstrated in these receptors, nor is it likely since their ligand, HLA-E, is highly conserved in the population. On the other hand, it has been demonstrated that the sequence of the bound peptide can affect CD94/NKG2 recognition of HLA-E.^132,133 This could have consequences on the regulation of NK and T-cell recognition of tissues in transplantation, particularly stimulating a GVT response, especially since there is polymorphism in the HLA class I signal sequences bound by HLA-E.¹³²

NKG2D shares only ~20% protein identity with other NKG2 family members and is expressed as a homodimer on the cell surface of NK, γδ T-cells and subsets of CD4⁺ and CD8⁺ T-cells.^134,135 NKG2D is an activating receptor and signaling is achieved by association with two dimers of DAP10 carrying phosphatidylinositol-3 kinase binding sites in their cytoplasmic tails. The ligands for NKG2D are MICA and MICB. There is now evidence that the interaction of NKG2D with its ligands may play a role in the transplantation of tissues, particularly with respect to graft rejection of solid organ transplants.¹³⁴

Leukocyte immunoglobulin-like receptors

The leukocyte immunoglobulin-like receptors (LILR) or immunoglobulin-like transcripts (ILT) are a multigene family located on the long arm of chromosome 19 closely linked to the KIR genes.¹³⁶ Eleven expressed LILR members have been identified, ten of which are glycosylated transmembrane proteins and one being a glycosylated soluble molecule.^136,137 The LILR, like the KIR, belong to the Ig superfamily and have either two or four Ig-like extracellular domains (see Fig. 39.11). There are inhibitory (LILRB1-5) and activating (LILRA1-6) receptors, similar to the KIR and CD94/NKG2 receptor families. Inhibitory LILR have long cytoplasmic tails that contain two to four ITIM and activating LILR have short cytoplasmic tails devoid of known signaling motifs. Activating LILR associate with the Fc epsilon receptor γ chain which contains an ITAM in its cytoplasmic tail.

LILR expression varies widely on immune system cells, but they are found primarily on APC of the myeloid lineage such as macrophages and dendritic cells^136–138 as well as on B-cells and on subsets of NK and T-cells. Two of the inhibitory LILR, LILRB1 and LILRB2, interact with HLA-A, -B, -C, -F, and -G molecules, but their specificities for HLA allelic products and affinities differ. Two of the activating LILR also bind class I molecules. LILRA1 binds some allelic products of HLA-B*27 and LILRA2 binds soluble class I molecules. There is a fair degree of polymorphism in the LILR genes which clusters in several discrete areas of the extracellular domains¹³⁶ which could be the reason for the differences in allelic specificity and affinity for class I molecules.

Evidence is beginning to mount for a possible role of LILRs in transplantation. The inhibitory LILR mediate potent tolerogenic effects on immune responses.^137,138 Thus, engagement of these receptors on APCs presenting alloantigens can result in down-regulation of the alloresponse. For instance, LILR recognition of HLA-G in kidney transplant patients was shown to be associated with a reduction in T-cell alloproliferation and induction of a regulatory/suppressive T-cell population. The full extent of LILR involvement in transplantation requires further study.

Serologic detection of class I and class II allelic products

Lymphocyte microcytotoxicity testing has been used for HLA typing since the 1960s; however, typing for HSC transplantation is performed mainly by DNA-based testing. The HLA serologic phenotype is determined by testing unseparated lymphocyte preparations or T-lymphocytes (for HLA-A, -B, -C) or enriched B-lymphocytes (for HLA-DR, -DQ) for the presence of specific HLA molecules as detected by a panel of well-characterized HLA antibody preparations (alloantisera or monoclonal antibodies).¹³⁹ Lymphocytes are incubated with a panel of antibody reagents. If the lymphocytes carry a cell surface HLA molecule recognized by a complement-fixing antibody, the antibody binds to the cells and the cells are subsequently lysed by the addition of complement in excess. Following termination of the reaction and staining to distinguish live from dead cells, the percent lysis of the cells is determined for each antibody reagent using a microscope and a numerical grade is assigned (e.g., 1 = 0–10% lysis, 6 = 51–80%, 8 = 81–100%). Scores of 6–8 indicate that the specific HLA molecule detected by the antibody reagent is present.

The antibodies used to detect specific HLA molecules (or ‘specificities’) are derived primarily from the sera of alloimmunized individuals including multiparous women, transplant recipients, and multi-transfused patients. Because of the complex reactivity patterns of alloantisera, several alloantisera are used to define each specificity. To address this complexity, monoclonal antibodies detecting HLA specificities have been generated; however, the availability of these reagents is limited.

Cellular detection of HLA disparity

Cellular assays were historically used to detect HLA differences between individuals, but are not commonly used today. The response of one cell (responder) in tissue culture to the foreign histocompatibility molecules on the surface of a second cell (stimulator) is called the mixed leukocyte culture (MLC) or mixed lymphocyte reaction (MLR).¹⁴⁰ The response is made unidirectional by preventing cells from one of the two individuals from replicating by treating those cells with radiation or an alkylating agent prior to addition to the culture. The MLC represents a summation response of a responder cell to differences in the HLA class II molecules (HLA-DR, -DQ, and -DP) encoded by the irradiated stimulator cell haplotypes. The response to DR molecules predominates. The use of cellular assays to determine the similarity between two individuals has declined significantly because DNA based testing procedures can now be used to identify class II disparity between individuals and because the MLC can be influenced by a variety of factors including the health of the individuals contributing the cells, the patient’s disease, and the history of prior transfusion.¹⁴¹

Other cellular assays measure the frequency of cytotoxic or helper T-lymphocytes in the potential HSC donor. Limiting dilution is used to isolate and measure the frequency of T-lymphocyte precursors which respond to recipient cells. The differences stimulating these responses are thought to be HLA class I differences or minor histocompatibility antigens for cytotoxic precursors and HLA class II differences for helper precursors. The correlation between the precursor frequency and transplantation outcome is controversial;^142,143 thus, DNA-based testing is more commonly used to identify HLA disparity between potential donor and recipient.

DNA-based identification of class I and class II alleles

DNA-based methods target differences in the nucleotide sequences of alleles for HLA typing. Commercial kits employing all of the methodologies described below are available although some laboratories continue to design their own reagents. Any cell with a nucleus can be used as a source of DNA; however, DNA is usually prepared from a small quantity (0.2–1 ml) of whole blood. Alternative sample types such as epithelial cells from a buccal swab might provide a more reliable source of genomic DNA in patients with hematologic malignancies,¹⁴⁴ in patients treated with immune suppressive agents or in patients who have been transfused.¹⁴⁵ The sensitivity of detection of HLA alleles is enhanced greatly by the amplification of DNA encoding HLA genes using the polymerase chain reaction (PCR) (Fig. 39.12). Many typing reactions utilize PCR primer sequences that are broadly specific and are shared by all alleles at an HLA locus. Other typing procedures utilize primer sets that are narrowly specific and are shared by only a subset of alleles at a locus. In this way, the laboratory can isolate large quantities of specific HLA alleles for identification as described below.

Fig. 39.12 One procedure used for DNA-based typing is sequence specific oligonucleotide probe hybridization. An HLA gene is amplified by the polymerase chain reaction using primers (P1, P2) that flank the polymorphic exon (in this example, exon 2 of DRB1). The DNA is labeled (indicated by a star) during amplification. The labeled DNA is hybridized to a series of oligonucleotide probes that are fixed to a solid support (membrane, microspheres or plate). Each probe detects a polymorphic region of exon 2 (indicated by colored blocks). The denatured DNA binds to those probes that contain sequences complementary to HLA allele sequences contained within the amplified DNA. Binding of the labeled DNA is detected by a variety of methods. The positive and negative probe hybridization results are interpreted by comparison of the nucleotide sequences of the probes with the sequences of known HLA alleles to determine those present in the sample.

Serologic specificities

Historically, serological and cellular reagents were used to define HLA types. These reagents defined specificities (or determinants) localized to one or several HLA molecules; these specificities were used as surrogates of alleles. The specificities identified through these methods were defined and standardized during international workshops in which typing reagents and cells were exchanged among participating laboratories and a consensus reached on the definition of each specificity⁶ (Table 39.7). These specificities were named by the HLA molecule and by a numerical designation based on the order of discovery. As additional reagents were identified, serologically defined specificities were often subdivided (i.e., ‘split’). For example, the serologic specificity HLA-B5 defined in the 1967 international workshop was subdivided into HLA-B51 and -B52 as a result of the reagents tested in the 1975 workshop. Table 39.7 lists the broader, shared (supertypic) specificities for the current ‘split’ (subtypic) specificities.

The numbering system used to define HLA ‘types’ is based on the date of assignment so that the relationships among various types is not immediately clear. Thus, for example, subdivisions of B5 were assigned as B51 and B52 since they were the 51st and 52nd serologic specificities to be assigned. Their relationship to B5 is not immediately clear unless their more complete nomenclature, B51(5) and B52(5), is provided. The assignments associated with the HLA-A and HLA-B serologic specificities are further complicated because it was not initially appreciated that these were specified by two different loci. Thus, the naming system is based on a single numerical series^5,6 such that numerical designations of HLA-A include HLA-A1, -A2, -A3, -A9, -A10, -A11, etc. while the HLA-B designations include the remaining numbers in the series, HLA-Bw4, -B5, -Bw6, -B7, -B8, -B12, etc.

The HLA-Bw4 and -Bw6 serologic specificities are broad, supertypic determinants which reside on the same molecule as the B locus subtypic specificities. HLA-B locus molecules (and some HLA-A locus molecules) characteristically carry either Bw4 or Bw6 specificities. In the case of Bw4 and Bw6, the use of the ‘w’ in the designation indicates the shared nature of the determinant. The region of the HLA-B molecule carrying the Bw4/Bw6 specificities has been identified to lie in the α1 alpha helix between amino acid residues 77 and 83 (Fig. 39.6A).^157,158 (A similarly located epitope identified by NK cell reactivity divides the HLA-C molecules in to group 1 and group 2^116,119,120.)

Antibody reactivity patterns have been used to group HLA-A and HLA-B molecules into cross-reactive epitope groups (CREG).^159–161 Examples of CREGs are listed in Table 39.8. An antigenic determinant (epitope) shared among members of a CREG is called a public specificity. Class I molecules within a CREG share one or more determinants that are not shared by molecules in another CREG. These groupings have been used as the basis for matching schemes for selection of solid organ donors^162,163 and to predict the reactivity of HLA-specific antibodies in sensitized patients.¹⁶⁰ In contrast, in HSC transplantation, mismatching schemes based on CREGs or other shared epitopes have been shown not to impact survival or the occurrence of GVHD.^164,165

Table 39.8 HLA class I cross-reactive epitope groups^a

^{a 159,160} The inclusion of HLA specificities within a CREG is not standard (i.e., CREG nomenclature has not been assigned by consensus of the HLA community) although there is considerable overlap between the various groupings reported in the literature.

^b A listing of antigens carrying Bw4 and Bw6 specificities can be found at http://hla.alleles.org. Not all alleles encoding a specific antigen may carry the usual Bw4 or Bw6 epitope. For example, B*0736, an allele whose product appears to carry the B7 serologic specificity,¹⁶⁸ carries a Bw4 epitope rather than the Bw6 epitope carried by most of the B*07 allelic products. See also B*2715 in Table 39.11.

Limitations in serologic testing reagents have resulted in inconsistent assignments of some HLA types.¹⁶⁶ This can complicate the selection of HLA matched donors from HSC donor registries and umbilical cord blood banks. With the increased use of DNA based HLA typing and its increased resolution and consistency, these difficulties are expected to disappear as the number of DNA typed donors and cord blood units increases.

DNA-based allele designations

Nucleotide sequencing has been used to identify the many alleles encoding the HLA molecules. Each HLA allele is designated by the name of the gene followed by an asterisk and a multiple digit number indicating the allele.²⁰ Additional and optional letters added to the end of the numerical designation relate to the expression of the allele at the protein level. The present nomenclature system is summarized in Tables 39.1, 39.2, and 39.9 and is described below.

Currently, allele names may contain four two-digit fields. The first two-digit field was originally based on the serologic type of the resultant molecule. Thus, the first HLA-A allele characterized from a cell bearing the HLA-A1 serologic specificity was assigned as A*0101 and the first HLA-B allele characterized from a cell bearing the HLA-B8 serologic specificity was assigned as B*0801. Identification of new alleles from other A1 or B8 positive cells led to the assignment of alleles A*0102 and B*0802. The second two-digit field of the allele name indicates the order of discovery. Alleles that share digits in the first field of their name share extensive sequence homology. This sequence similarity is also a criterion for nomenclature assignment and may take precedence over the serologic association in assigning an allele name.

Since most allelic differences lie within the antigen-binding groove of the HLA molecule, it is theoretically possible that these differences will stimulate allorecognition. There are two exceptions in which different allele names may not indicate different HLA molecules: 1) alleles which differ in their DNA sequence but do not differ in their encoded polypeptides are designated by sharing the digits in the first two fields of their allele names and are distinguished by the two digits in the third field of their name (e.g., B*070201, B*070202, B*070203). The differences among alleles in these clusters are not important to consider in transplantation matching since these alleles encode the same HLA proteins. 2) Alleles which have identical nucleotide sequences in their coding regions but differ in the nucleotide sequences outside of the coding region (i.e., in 5′ regulatory regions or introns) are designated by a sharing of the digits in the first three fields of their allele names and are distinguished by the two digits in the fourth field of their name (e.g., A*01010101 and A*01010102N). The differences among alleles in these clusters may or may not be important to consider in transplantation matching depending on the effect of the sequence difference. Allelic differences that result in loss of protein expression are important considerations in the selection of a donor while a difference which has no impact on the expression of the allele is not an important consideration.

Alleles that are not expressed as proteins at the cell surface may have a single letter, N, appended to their names (e.g., A*01010102N) to indicate ‘null’. Null alleles are important to consider in transplantation since individuals who carry a null allele are functional homozygotes. For example, an individual who carries A*01010102N, A*0207 alleles would be serologically typed as A2 only compared to an individual who carries A*01010101, A*0207 who would be serologically typed as A1,A2. Other designations (e.g., L, S, Q) signify alterations in HLA expression and are described on the HLA nomenclature website (http://hla.alleles.org).²¹

While this chapter was being written, the HLA nomenclature underwent a revision to allow an extension of the number of digits designating each field in an allele name and to use colons to separate the fields making up the designation. Examples of the new nomenclature are shown in Table 39.9.

National Marrow Donor Program® (NMDP) DNA nomenclature

The ability to distinguish among alleles (the level of resolution) by DNA-based typing methods is controlled by the choice and number of reagents used and the typing technique. Large scale volunteer donor typing for a HSC registry or umbilical cord blood bank is usually carried out at low to intermediate resolution in which the alleles at each HLA locus are narrowed down to a few possibilities. For example, the HLA-DRB1 test results of a volunteer might narrow the possible alleles at this locus to DRB1*1301 or 1320 and to DRB1*0403 or *0406 or *0407 or *0411 or *0417. In order to include these possibilities within their database, the NMDP has developed a letter coding system for these multiple allele combinations. Thus, the combination DRB1*1301 or 1320 is designated DRB1*13VS where VS means 01 or 20 and DRB1*0403 or *0406 or *0407 or *0411 or *0417 is designated DRB1*04DZ (DZ means 03 or 06 or 07 or 11 or 17). The letter codes are purely descriptive, with codes being created only for allele combinations which have been observed during HLA testing. A listing of codes can be found on the NMDP website (http://bioinformatics.nmdp.org); the NMDP codes are used internationally in the exchange of HLA information among registries and cord blood banks.¹⁶⁷

Limitations in the assignment of HLA alleles

Since identification of HLA alleles may often be based on partial nucleotide sequence information obtained through the use of a small set of primers and probes, it is possible that the interpretation of the assay results will miss alleles which were unknown at the time of the typing.¹⁶⁸ For example, the identification of the DNA sequence GGTAAGTATAAG encompassing codons 11–14 in the DRB1 gene was once interpreted to mean that the individual carried the allele DRB1*0701 since only DRB1*0701 was known to carry that polymorphic sequence cassette. In 1998, two new alleles were described which also carry this sequence, DRB1*0703 and DRB1*0704. Thus, individuals tested prior to 1998 and characterized as carrying DRB1*0701 might alternatively carry DRB1*0703 or DRB1*0704. The continued discovery of new HLA alleles has significant implications for the HLA assignments since HLA types of individuals are stored and used for many years in HSC registries or cord blood banks. Since a patient may be HLA typed during the initial evaluation of treatment options but donor selection and transplant may not be the first line of therapy, the clinical laboratory should update the HLA assignments for that patient when the decision to seek an unrelated donor is made. Patients with newly described or rare alleles may require additional search strategy assistance.

Because of the similarity among HLA alleles, a DNA-based assay may assign several alternative genotypes; for example, a patient may carry the genotype A*01010101 + A*02010101 or the genotype A*0114 + A*9204 or the genotype A*0236 + A*3604. While clinical laboratories are required to report all alternative genotypes, the laboratory may indicate that one of the genotypes is more likely based on the expected frequencies of HLA alleles in the population group to which the patient belongs. The decision to perform additional HLA testing to determine which genotype is indeed carried by an individual is a decision made by the transplant center and its clinical laboratory and should be based on a sound knowledge of the HLA profile of the relevant human population.

Correlation between serologic specificities and DNA-based allele assignments

Today, patients requiring HSC transplants are often typed by DNA-based technologies while registries and banks of potential donors include many serologically typed volunteers or cord blood units. Thus, a HSC registry’s matching algorithm must be designed to compare a variety of HLA assignments and to determine the potential for an allele match. Table 39.10 provides some examples of the association between HLA types. In many cases, digits in the first field of the allele name are based on the serologic assignment of the HLA molecule so, for example, alleles such as A*0101 and A*6801 correspond to A1 and A68, respectively. In addition, several alleles may encode the same serologic type, such as B*2701, *2702 and *2703, which all encode a HLA-B molecule carrying a B27 serologic specificity (Table 39.11).

Table 39.11 Association between some of the HLA-B*27 alleles and serologic specificities^a

^{a 169}

^b Based on serologic typing of specific B*27 allelic products. Data in [ ] describes unusual serologic patterns or the prediction of serologic typing based on an artificial neural network.

^c Undefined indicates that the allelic product does not have a known serologic specificity.

It is important to note, however, the many examples where the associations between an allele and the serologic type of the HLA molecule encoded by that allele are not known or are apparently different. For example, even though B*2708 encoded a molecule which bore a B7 serologic determinant, it was found to have a closer structural relationship to alleles in the B*27 group which took precedence in the assignment of the allele name. In this case, the name of the allele cannot be used to predict the serologic type of its allelic product. Information on the serologic assignments associated with each HLA allele is routinely updated and reported.¹⁶⁹ This ‘dictionary’ can be found in a searchable format at http://www.ebi.ac.uk/imgt/hla/dictionary.html.

Autoimmunity

Autoimmune diseases are characterized by an abnormal immune response to self antigens. Many of these diseases are associated with the presence of specific HLA types.^52,53 The association of HLA-DQ2 and DQ8 with celiac disease is one example. This disease is a chronic inflammatory disease developing in response to ingested wheat gluten or related proteins.¹⁷⁰ The majority of patients carry DQ2 and a minority DQ8. Yet, since only a fraction of individuals with DQ2 or DQ8 develop celiac disease, HLA-DQ is most likely just one out of several genetic and/or environment factors that combine to cause susceptibility to the disease.

A variety of mechanisms involving the HLA molecule have been evoked to explain the observed HLA-disease associations. One hypothesis suggests that a peptide from a pathogen which mimics a self peptide binds to the HLA molecule associated with the disease to initially stimulate autoreactive T-cells. A second hypothesis suggests that disease occurs as the result of altered thymic selection in which self reactive T-cells escape selection in the thymus as a result of poor self peptide binding properties of the disease-associated HLA molecules. Still another hypothesis suggests that it is not the HLA molecule itself, but the product of another as yet unknown gene within the MHC which affects susceptibility. It is likely that the immune mechanisms resulting in a loss of tolerance to self antigens will vary among autoimmune diseases; thus, elucidation of the role that HLA plays will await the unraveling of each complex disease.

Drug hypersensitivity

Hypersensitivity to several drugs has been associated with specific HLA alleles: 1) the HIV reverse transcriptase inhibitor abacavir and B*5701; 2) the gout prophylactic allopurinol and B*5801; and 3) the antiepileptic carbamazepine and B*1502. Hypersensitivity to abacavir has been linked to HLA antigen presentation to T-lymphocytes although it is not yet known how abacavir or its by-products create the novel antigen triggering the response.¹⁷¹

Cell, tissue and organ transplantation

Allogeneic grafts of HSCs, organs and other tissues have been everyday life-saving procedures since the mid 20th century.⁴ The necessity for HLA matching of donor and recipient differs for each type of graft. For HSC transplantation, close matching for HLA is required for successful outcome. For kidney transplantation, close matching increases the likelihood of long-term organ survival,¹⁷² but survival of grafts in mismatched donor–recipient pairs is also good. For logistical reasons, HLA matching plays a smaller role in transplantation of organs other than HSC and kidneys and outcome relies on management of graft rejection using immune suppression. Grafts of non-viable tissues rarely take HLA matching into consideration, as these non-viable tissues are generally used either temporarily (skin) or as ‘scaffolding’ for replacement with autologous tissue.