Human MHC architecture and evolution: implications for disease association studies

Abstract

Major histocompatibility complex (MHC) variation is a key determinant of susceptibility and resistance to a large number of infectious, autoimmune and other diseases. Identification of the MHC variants conferring susceptibility to disease is problematic, due to high levels of variation and linkage disequilibrium. Recent cataloguing and analysis of variation over the complete MHC has facilitated localization of susceptibility loci for autoimmune diseases, and provided insight into the MHC's evolution. This review considers how the unusual genetic characteristics of the MHC impact on strategies to identify variants causing, or contributing to, disease phenotypes. It also considers the MHC in relation to novel mechanisms influencing gene function and regulation, such as epistasis, epigenetics and microRNAs. These developments, along with recent technological advances, shed light on genetic association in complex disease.

Figure 1

Annotation of MHC reference haplotypes in VEGA. Representation of the PGF, COX and QBL haplotypes and variation in the Vega Genome Browser ‘MultiContigView’. MultiContigView allows simultaneous display and navigation of genome annotation between the human MHC regions in the different reference haplotypes. Manually curated gene structures within the HLA-DR region are shown (see Fig. 3 and accompanying text).

Figure 3

Simplified gene configurations in the RCCX and HLA-DRB regions on MHC reference haplotypes. The COX and QBL haplotype are monomodular for RCCX, whereas the PGF, SSTO and DBB haplotypes are bimodular. Each C4 gene may encode a C4A or a C4B protein. C4L and C4S are long and short C4 genes, respectively. 21A and 21B are CYP21A ( pseudogene) and CYP21B, respectively. DRB2, DRB6, DRB7, DRB8 and DRB9 are pseudogenes.

Figure 2

A reduced map of the MHC illustrating clustering of immune system genes. Two kinds of gene clustering are apparent. First, sequence-related duplicates that have allowed diversification of duplicates, e.g. HLA or C4 complement genes. Second, sequence unrelated genes with related immune function, e.g. PSMB8/9 — proteosome components, TAP1/2 — peptide transporters, TAPBP — peptide chaperone, and HLA — peptide presentation.

Figure 4

A shared ancestral block between two DR3 -DQ2 MHC haplotypes. The COX and QBL haplotypes share an ~158 kb ancestral block, which contains the HLA-DRB2, HLA-DRB1, HLA-DQA1, HLA-DQB1 and MTCO3P1 loci. Within this block there are only 14 single nucleotide polymorphisms (SNPs) distinguishing the two DR3-DQ2 haplotypes, whereas there are 3808 SNPs that differ between the DR3-DQ2 haplotypes and the DR15-DQ6 haplotype (PGF). This difference can be calculated to correlate with ~3400 generations that elapsed since the two DR3-DQ2 haplotypes separated vs. greater than 20 million years that separate the DR15-DQ6 haplotype from the DR3-DQ2 haplotypes. Vertical lines represent SNPs. HLA types for each haplotype are shown.

Figure 5

Frequencies of the DR3-DQ2 haplotype within 18 populations. The figure represents the ethnic diversity of DR3-DQ2 haplotype distribution (NCBI dbMHC).

Figure 6

Synergy of HLA-KIR (human leucocyte antigen-killer immunoglobulin receptor) compound genotypes and duality between autoimmunity and infection. The general paradigm for HLA-KIR epistasis based on current disease association studies. Different HLA-KIR combinations provide different levels of activation and inhibition of NK (natural killer) and T cells resulting in differing susceptibility and protection against infection and autoimmunity.

Figure 7

A summary of potential factors involved in an MHC association with a complex disease. CNV, copy number variation; HLA, human leucocyte antigen; KIR, killer immunoglobulin receptor; SNP, single-nucleotide polymorphism.