The Major Histocompatibility Complex And Disease

Overview Of The Human Major Histocompatibility Complex Dna Sequence

Completion of the MHC deoxyribonucleic acid (DNA) sequence preceded completion of the total human genome sequence by almost 4 years ( ). The rush to complete the entire MHC sequence followed the need to understand the biology and genetics of the mouse MHC (H-2) and HLA regions, which were initially provided by murine inbred experiments and serologic typing ( ). Although the MHC was discovered more than 50 years ago, its nature has been resolved only in the last 2 decades, with the advent of DNA cloning and discovery of the structure of class I and class II molecules ( ; ).

The MHC located on the short arm of chromosome 6 spans approximately 4 megabases (Mb) and contains more than 250 identified loci, by far the most gene-dense region of the human genome sequenced to date ( ). It encodes the most polymorphic human proteins—the HLA class I and class II molecules, with over 24,000 allelic variants identified so far ( ). More than 40% of the total expressed loci in the MHC are associated with immune-related functions. This clustering of immune-related genes in the MHC region may not be coincidental and may be the result of evolutionary forces joining genes with similar functions.

Traditionally, the MHC region has been classified into three regions from telomere to centromere: HLA class I, non-HLA class (MHC class III), and HLA class II. HLA class I and class II gene functions were reviewed in detail in Chapter 50 . We will concentrate on the MHC class III region and will later return to HLA class I and class II genes to discuss disease associations.

Genes In The Central Or Class Iii Region

The human MHC class III, or central MHC region, located between the HLA class I and class II regions, is the most gene-dense region of the human genome. Specifically, the MHC class III region contains 61 protein-coding genes in ∼900-kb sequence, with an average gene size of ∼8.5 kb ( ), compared with the entire human genome, which contains, on average, fewer than 11 genes/Mb and has an average gene size of 27 kb ( ). Although HLA class I– and class II–encoded proteins are distinguished by structural and functional similarities shared by the members of each class, MHC class III molecules can be defined only as non-I and non-II because their genes and their products have no common features and are not recognized by T cells. Figure 51.1 shows a gene map of the MHC class III region.

Despite convincing genetic association studies showing the importance of the MHC class III region in health and disease, neither the genes responsible for disease nor the underlying mechanism of pathology is known for many of the associated conditions. One reason for this is the fact that the functions of nearly half the genes in the MHC class III region are not completely known. A high-throughput yeast two-hybrid system was used to look into the function of several intracellular proteins encoded within the MHC class III region ( ). This study revealed that approximately one third of the analyzed proteins encoded within the MHC class III region may have a role in mRNA processing, which again suggests clustering of functionally related genes within this region of the human genome. Specifically, three proteins (STK19, PBX2, and NELFE) of those in the MHC class III region interact with proteins implicated in transcriptional regulation. Four proteins (LSM2, DDX39B, DXO, and SKIV2L) are orthologs of yeast proteins involved in mRNA processing, and another five proteins (PRRC2A, STK19, CLIC1, PBX2, and ABHD16A) interact with proteins previously implicated in RNA processing. Proteins involved in RNA processing are usually categorized as “housekeeping” proteins, but they are often critical for innate immunity against viral RNA. Metabolism of intracellular nucleic acids is also critical for limiting autoimmunity against DNA and RNA. SKIV2L (superkiller viralicidic activity 2-like)–deficient patients have an elevated interferon response ( ) and present with trichohepatoenteric syndrome, an inflammatory bowel disease–like intractable diarrhea ( ).

The region closer to HLA class II genes contains several genes of known function, including TNXB ( ), which encodes tenascin X, a large extracellular matrix protein expressed in connective tissues, and NEU1 , which encodes neuraminidase 1, a lysosomal sialidase, and SLC44A4 , which encodes a choline transporter. Mutations of TNXB are a cause of Ehlers-Danlos syndrome ( ), primary myopathy ( ), and vesicoureteral reflux ( ; ). Mutations of the NEU1 gene are responsible for congenital sialidosis ( ; ). Mutations in SLC44A4 cause autosomal dominant nonsyndromic deafness ( ).

The region toward the class I genes also contains several genes of known function, including VARS , the cytoplasmic valyl-tRNA synthetase, MSH5 , a homolog of Escherichia coli MutS recombinase, and CSNK2B , casein kinase type II. Mutations in VARS lead to neurodevelopmental encephalopathy and microcephaly ( ), mutations in MSH5 lead to primary ovarian insufficiency ( ), and pathogenic variants in CSNK2B cause intellectual disability and global developmental delay with or without epilepsy ( ).

Almost one half of genes within the MHC class III region play roles in the innate immune system, including members of the complement fixation cascade (C4, C2, CFB) and the tumor necrosis factor family (TNF, LTα, LTβ), which will be discussed in greater detail later in this chapter. Other members of the immunoglobulin superfamily are also located within the MHC class III regions: NCR3 (natural cytotoxicity triggering receptor 3) and the less characterized LY6G genes ( ). The NCR3 gene encodes the NKp30 protein responsible for triggering natural killer (NK) cells ( ). NK cells play a variety of roles in the immune system, including defense against infectious organisms as well as tumors, and have also been implicated in autoimmunity. NKp30 is involved in the killing of cells infected with Plasmodium falciparum ( ) and cytomegalovirus ( ), fungi ( ), hematologic and solid tumor cells ( ) and contributes to the pathogenesis of primary Sjögren syndrome ( ). Variants in NCR3 gene contribute to genetic risk of malaria attacks ( ). Increased expression of NCR3 and the adjacent LST1 (leukocyte-specific transcript 1) gene has been found in the blood of rheumatoid arthritis patients ( ); and haplotype-specific variants in these genes alter susceptibility to arthritis in animal models ( ). LST1 encodes a palmitoylated transmembrane signaling adaptor and membrane scaffolding protein in myeloid cells ( ). Both genes are also significantly upregulated in response to lipopolysaccharide, interferon-γ, and bacterial infection, suggesting a role for these gene products in autoimmune-related inflammation and dendritic cell/NK cell–associated functions. The Ly6 (lymphocyte antigen 6 complex) family of proteins were originally described using antisera to lymphocytes and were used as markers for lymphocyte subset differentiation, but now we know that they belong to a family of 35 genes in humans, 5 of which are located in the MHC class III region. The functions of many Ly6 proteins are still not well understood, but it ranges from neutrophil migration, cell-cell interaction, complement regulation, macrophage activation, to modulation of viral infections ( ).

The NFKBIL1 (NF-κB-inhibitor–like 1) and BTNL2 (Butyrophilin-like protein 2) genes are also located in the MHC class III region ( ; ). Much interest has been focused on the NFKBIL1 ( ) and BTNL2 ( ) proteins after single mutations in their promoter regions were linked to rheumatoid arthritis ( ) and sarcoidosis ( ), respectively. BTNL2 may also contribute to the risk of type 1 diabetes ( ). AIF1 (Allograft inflammatory factor-1) encodes a calcium-binding protein highly expressed in monocytes and macrophages and involved in phagocytosis in these cells and in antigen uptake in intestinal M cells, and have also been implicated in rheumatoid arthritis and other inflammatory processes ( ; ). Finally, the genes for the microsomal enzyme steroid P450 21-hydroxylase and three members of the major heat-shock protein 70 family are also contained in the MHC class III region and will be discussed later.

Extended Haplotypes

Based on the study of the distribution of complotypes in relation to HLA-B and HLA-DR specificities on normal Caucasian haplotypes determined in family studies, striking linkage disequilibrium involving the whole region became evident. One could easily recognize HLA-B, complotype, DR allele sets that showed statistically significant three-point linkage disequilibrium ( Fig. 51.2 ) and that define what are regarded as extended haplotypes ( ). A shorthand nomenclature for extended haplotypes has been designed by enclosing HLA-B, complotype, and HLA-DR variants in brackets. Thus the most common extended haplotype in Caucasians is [HLA-B8, SC01, DR3] . Extended haplotypes are highly characteristic of an ethnic subgroup and have lower frequencies or do not occur at all in other ethnic groups ( ; ).

Initial studies defined extended haplotypes as the genomic interval between HLA-B and DR . It then became evident that other uncharacterized MHC alleles in the interval were likely to be included. Despite the fact that HLA-A alleles have shown limited variation for extended haplotypes, only half such haplotypes show unique and significant HLA-A allele associations ( ). The HLA-Cw locus was by far the last to be characterized. Based on genetic distance and previously incomplete typing data of HLA-B / Cw pairs, it was evident that conserved haplotypes would also include the HLA-Cw genetic region. Thus, different HLA-Cw alleles have been recently associated with different extended haplotypes ( Table 51.2 ) ( ). Furthermore, several TNF and HSP allele systems and microsatellites have been studied in relation to extended haplotypes ( Table 51.3 ). More recently, models have been created to describe the variable sizes of stretches of conserved DNA in the MHC, using the known frequencies of four different types of small blocks (<0.2 Mb) of relatively conserved DNA sequence: HLA-Cw / B; TNF; complotype; and HLA-DR / DQ ( ). Using HLA allele identification and TNF microsatellites, Yunis and colleagues have shown that some extended haplotypes extend to the HLA-A and HLA-DPB1 loci, which form fixed genetic units of at least 3.2 Mb of DNA. Intermediate fragments of extended haplotypes also exist, which are, nevertheless, larger than any of the four small blocks. This complexity of genetic fixity at various levels should be taken into account in studies of genetic disease association, immune response control, and human diversity.

Extended haplotypes, which account for at least 30% of normal Caucasian haplotypes, have relatively fixed gross structure and DNA sequence and carry very similar (if not identical) alleles even when they are found in apparently unrelated individuals. The frequency of extended haplotypes is more likely to be underestimated than overestimated because haplotypes must be determined from family studies involving thousands of unrelated index cases. The original MHC reference sequence ( ) has been used as the starting point for numerous studies undertaken to ascertain multiple common and rarer extended haplotypes ( ; ; ; ), and then eight haplotypes have been completely sequenced, and the haplotype [HLA-A3, B7, DR15] was designated as the new MHC reference sequence incorporated into the reference human genome assembly ( ). The advent of next generation (or massively parallel) sequencing technologies enabled the sequencing of individual genomes. However, the current technologies in clinical use struggle with complex portions of the genome, like the MHC. These challenges are now being overcome, and there are novel technologies that can sequence the entire MHC, including some that are capable of phasing the entire diploid MHC, not just the individual HLA genes ( ; ; ; ). These studies will facilitate the identification of precise disease loci and may help us to better understand events such as recombination and polymorphism.

Mhc Disease Associations

The MHC is associated with more diseases than any other region on the human genome. Identification of the key genetic association within the MHC has proven difficult for at least four reasons: (1) MHC has the highest gene density in the genome, (2) and these large numbers of genes are often in strong linkage disequilibrium forming complex haplotype structures, (3) extended haplotypes contain multiple equally plausible candidate loci, and (4) the associated diseases often show non-Mendelian inheritance.

We have reviewed concepts about complotypes and extended haplotypes to enhance our understanding of the relevance of a rather extensive literature on MHC disease associations. Most of the MHC disease associations reported to date are contained within extended haplotypes. It is probably because of this simple fact that so many MHC-allele–associated diseases have been reported, because it is not a single base pair but more than 3 million base pairs of conserved DNA that constitute the marker. For example, complement alleles such as CFB∗F1 and C4B∗3 have been linked to type 1 diabetes, because the extended haplotypes that carry them, [HLA-B18 , F1C30 , DR3] and [HLA-B62 , SC01 , DR3] , are more common in these patients than in the general population. This association means that haplotypes comprising 3 million bases or more in length of conserved genomic DNA carry susceptibility alleles for type 1 diabetes. These 3 million base-long extended haplotypes often contain multiple candidate loci, best illustrated by psoriasis, a chronic inflammatory skin disorder. Psoriasis was associated with HLA-Cw6 over 30 years ago ( ), and many studies attempted since to identify the causal gene or genes. Although HLA-C itself is the strongest candidate gene, multiple additional HLA class I and class II, and MICA/MICB variants appear to be independently associated with psoriasis ( ; ). Moreover, there are at least four genes (PSORS1C1-3 and CDSN) located next to HLA-C that encode skin-expressed genes providing additional candidate genes for this complex skin disease ( ; ). It is also critical to know the ethnic distribution of an associated MHC allele to evaluate whether the allele in patients is increased compared with that in ethnically matched control populations or is in fact truly a marker for the ethnic distribution of the disease ( ). Furthermore, it is more difficult to map the genes responsible for disease in Caucasian Americans than in African Americans or Asian Americans because the level of genetic diversity is lower in Caucasians.

Most diseases associated with MHC are autoimmune and do not show clear-cut Mendelian inheritance. Many problems confound attempts to understand the inheritance mechanisms by which these diseases occur, and analysis of MHC markers in patients and their families has clarified the picture only marginally. A major problem with MHC gene associations is their incomplete penetrance, making ordinary formal segregation and linkage studies very difficult (if not impossible) to carry out. This is seen most clearly in monozygotic twins, who presumably have identical genes. If one such twin has one of these diseases (e.g., type 1 diabetes), the other twin does not necessarily have the same disease. For type 1 diabetes, the concordance rate appears to be no higher than 50% ( ; ). This suggests that penetrance of a disease in a completely susceptible host is incomplete. Although there is excellent reason to consider genes in the MHC as determinants of type 1 diabetes susceptibility, only 15% of MHC-identical siblings of diabetic patients have insulin-dependent diabetes. This difference between 50% and 15% provides evidence for the influence of genes at a second, non–MHC-linked locus (or loci), and suggests the influence of other factors, such as environmental factors, in determining susceptibility.

Incomplete penetrance makes assignment of a specific mode of inheritance difficult and makes the likelihood of finding families with more than one affected member within one or more generations low. Families are frequently used to determine modes of inheritance. Another complicating factor is the inability to determine whether we are studying a group of patients that are homogeneous in terms of genetic determination. About 5% to 6% of random families with a type 1 diabetic proband will have a second affected child. Of these sibling pairs, approximately 60% will be MHC identical, 35% will be haploidentical, and a few percent will share no MHC haplotypes. This pattern suggests recessive inheritance of an MHC-linked susceptibility gene for the disease. At the very least, however, family studies provide highly useful haplotype data and usually allow the assignment of homozygosity for an HLA marker in probands. They have also established that MHC association in most diseases of interest is based on linkage between a susceptibility gene and the MHC.

Yet another unknown is the number of different susceptibility alleles for a disease in any specific population. In this regard, extended haplotypes can be helpful because, if increased in patients, they probably represent a single susceptibility allele that could be anywhere in the region of fixity. However, some patients have only portions of these haplotypes, and this provides little information about the location of specific allele involvement. Furthermore, it appears likely that many MHC-associated diseases are polygenic. Because of this, we will first discuss diseases associated with mutations of a single MHC gene; this will be followed by a discussion of polygenic diseases.