The HLA-DRB9 gene and the origin of HLA-DR haplotypes

Construction and benchmarking of a multi-ethnic reference panel for the imputation of HLA class I and II alleles

Genotype imputation of the human leukocyte antigen (HLA) region is a cost-effective means to infer classical HLA alleles from inexpensive and dense SNP array data. In the research setting, imputation helps avoid costs for wet lab-based HLA typing and thus renders association analyses of the HLA in large cohorts feasible. Yet, most HLA imputation reference panels target Caucasian ethnicities and multi-ethnic panels are scarce. We compiled a high-quality multi-ethnic reference panel based on genotypes measured with Illumina’s Immunochip genotyping array and HLA types established using a high-resolution next generation sequencing approach. Our reference panel includes more than 1,300 samples from Germany, Malta, China, India, Iran, Japan and Korea and samples of African American ancestry for all classical HLA class I and II alleles including HLA-DRB3/4/5. Applying extensive cross-validation, we benchmarked the imputation using the HLA imputation tool HIBAG, our multi-ethnic reference and an independent, previously published data set compiled of subpopulations of the 1000 Genomes project. We achieved average imputation accuracies higher than 0.924 for the commonly studied HLA-A, -B, -C, -DQB1 and -DRB1 genes across all ethnicities. We investigated allele-specific imputation challenges in regard to geographic origin of the samples using sensitivity and specificity measurements as well as allele frequencies and identified HLA alleles that are challenging to impute for each of the populations separately. In conclusion, our new multi-ethnic reference data set allows for high resolution HLA imputation of genotypes at all classical HLA class I and II genes including the HLA-DRB3/4/5 loci based on diverse ancestry populations.

A trans-specific polymorphism in ZC3HAV1 is maintained by long-standing balancing selection and may confer susceptibility to multiple sclerosis

The human ZC3HAV1 gene encodes an antiviral protein. The longest splicing isoform of ZC3HAV1 contains a C-terminal PARP-like domain, which has evolved under positive selection in primates. We analyzed the evolutionary history of this same domain in humans and in Pan troglodytes. We identified two variants that segregate in both humans and chimpanzees; one of them (rs3735007) does not occur at a hypermutable site and accounts for a nonsynonymous substitution (Thr851Ile). The probability that the two trans-specific polymorphisms have occurred independently in the two lineages was estimated to be low (P = 0.0054), suggesting that at least one of them has arisen before speciation and has been maintained by selection. Population genetic analyses in humans indicated that the region surrounding the shared variants displays strong evidences of long-standing balancing selection. Selection signatures were also observed in a chimpanzee population sample. Inspection of 1000 Genomes data confirmed these findings but indicated that search for selection signatures using low-coverage whole-genome data may need masking of repetitive sequences. A case–control study of more than 1,000 individuals from mainland Italy indicated that the Thr851Ile SNP is significantly associated with susceptibility to multiple sclerosis (MS) (odds ratio [OR] = 1.47, 95% confidence intervals [CI]: 1.08–1.99, P = 0.011). This finding was confirmed in a larger sample of 4,416 Sardinians cases/controls (OR = 1.18, 95% CI: 1.037–1.344, P = 0.011), but not in a population from Belgium. We provide one of the first instances of human/chimpanzee trans-specific coding variant located outside the major histocompatibility complex region. The selective pressure is likely to be virus driven; in modern populations, this variant associates with susceptibility to MS, possibly via the interaction with environmental factors.

Balancing selection is common in the extended MHC region but most alleles with opposite risk profile for autoimmune diseases are neutrally evolving

Background

Several susceptibility genetic variants for autoimmune diseases have been identified. A subset of these polymorphisms displays an opposite risk profile in different autoimmune conditions. This observation open interesting questions on the evolutionary forces shaping the frequency of these alleles in human populations.

We aimed at testing the hypothesis whereby balancing selection has shaped the frequency of opposite risk alleles.

Results

Since balancing selection signatures are expected to extend over short genomic portions, we focused our analyses on 11 regions carrying putative functional polymorphisms that may represent the disease variants (and the selection targets). No exceptional nucleotide diversity was observed for ZSCAN23, HLA-DMB, VARS2, PTPN22, BAT3, C6orf47, and IL10; summary statistics were consistent with evolutionary neutrality for these gene regions. Conversely, CDSN/PSORS1C1, TRIM10/TRIM40, BTNL2, and TAP2 showed extremely high nucleotide diversity and most tests rejected neutrality, suggesting the action of balancing selection. For TAP2 and BTNL2 these signatures are not secondary to linkage disequilibrium with HLA class II genes. Nonetheless, with the exception of variants in TRIM40 and CDSN, our data suggest that opposite risk SNPs are not selection targets but rather have accumulated as neutral variants.

Conclusion

Data herein indicate that balancing selection is common within the extended MHC region and involves several non-HLA loci. Yet, the evolutionary history of most SNPs with an opposite effect for autoimmune diseases is consistent with evolutionary neutrality. We suggest that variants with an opposite effect on autoimmune diseases should not be considered a distinct class of disease alleles from the evolutionary perspective and, in a few cases, the opposite effect on distinct diseases may derive from complex haplotype structures in regions with high genetic diversity.

Reassessing the role of HLA-DRB3 T-cell responses: evidence for significant expression and complementary antigen presentation

In humans, several HLA-DRB loci (DRB1/3/4/5) encode diverse beta-chains that pair with alpha-chains to form DR molecules on the surface of APC. While DRB1 and DRB5 have been extensively studied, the role of DRB3/4 products of DR52/DR53 haplotypes has been largely neglected. To clarify the relative expression of DRB3, we quantified DRB3 mRNA levels in comparison with DRB1 mRNA from the same haplotype in both B cells and monocytes, observing quantitatively significant DRB3 synthesis. In CD19+ cells, DRB1*03/11/13 was 3.5-fold more abundant than DRB3, but in CD14+ this difference was only two-fold. Monocytes also had lower overall levels of DR mRNA compared with B cells, which was confirmed by cell surface staining of DRB1 and DRB3. To evaluate the functional role of DRB3, tetramer-guided epitope mapping was used to detect T cells against tetanus toxin and several influenza antigens presented by DRB3*0101/0202 or DRB1*03/11/13. None of the epitopes discovered were shared among any of the DR molecules. Quantitative assessment of DRB3-tetanus toxin specific T cells revealed that they are present at similar frequencies as those observed for DRB1. These results suggest that DRB3 plays a significant role in antigen presentation with different epitopic preferences to DRB1. Therefore, DRB3, like DRB5, serves to extend and complement the peptide repertoire of DRB1 in antigen presentation.

Comparative genetics of MHC polymorphisms in different primate species: duplications and deletions

Gene products of the major histocompatibility complex (MHC) play a crucial role in the activation of adaptive (antigen-dependent) immune responses. In this paper similarities and dissimilarities among the MHCs of different primate species and their functional implications are reviewed. The human HLA system represents the most thoroughly investigated MHC of any contemporary living primate species, and so it will serve as a reference.

Genomic organization of the mammalian MHC

The Human Genome Project transformed the quest of more than 50 years to understand the major histocompatibility complex (Mhc). The sequence of the Mhc from human and mouse, together with a large amount of sequence and mapping information from several other species, allows us to draw general conclusions about the organization and origin of this crucial part of the immune system. The Mhc is a mosaic of stretches formed by conserved and nonconserved genes. Surprisingly, of the approximately 3.6-Mb Mhc, the stretches that encode the class I and class II genes, which epitomize the Mhc, are the least conserved part, whereas the approximately 1.7-Mb stretches that encode at least 115 other genes are highly conserved. We summarize the available data to answer the questions (a) What is the Mhc? and (b) How can we define it in a general, not species-specific, way? Knowing what is essential and what is incidental helps us understand the fundamentals of the Mhc, and defining the species differences makes the model organisms more useful.

Mhc-DQ-DRB haplotype analysis in the rhesus macaque: evidence for a number of different haplotypes displaying a low allelic polymorphism

In the HLA-DRB subregion of man, five major groups of haplotypes, often displaying a remarkable polymorphism, are distinguishable. The polymorphism is thought to be generated by point mutation, microgene conversion and gene rearrangement by recombination. In order to gain insight into the organization of the rhesus macaque major histocompatibility complex (MHC) class II region, DRB genes from monkeys of different origins previously typed for their DQ genes were analyzed. At first DRB haplotypes were deduced from DQ-homozygous monkeys. The stability of these haplotypes was then examined in DQ-heterozygous monkeys by sequence-based typing for the presence of members of the DRB1*03 and DRB1*04 lineage, and for seven single alleles detected on the haplotypes. Six DRB haplotypes linked to the five most frequent and three haplotypes linked to less frequent DQ haplotypes were identified. Six novel DRB alleles were detected. The number of DRB genes per haplotype varied between two and four. The results altogether suggest that in rhesus macaques, in comparison to man, the DQ haplotypes are linked to only a small number of DRB haplotypes, the number and diversity of DRB haplotypes is larger, and the allelic polymorphism of a given haplotype is smaller. The diversity of the DRB haplotypes was partly due to the varying number and identity of genes linked to DRB1*03 and DRB1*04. Furthermore, the number of DRB1 genes themselves varied from zero to two.

Major histocompatibility complex class II polymorphisms in primates

In the past decade, the major histocompatibility complex (MHC) class II region of several primate species has been investigated extensively. Here we will discuss the similarities and differences found in the MHC class II repertoires of primate species including humans, chimpanzees, rhesus macaques, cotton-top tamarins and common marmosets. Such types of comparisons shed light on the evolutionary stability of MHC class II alleles, lineages and loci as well as on the evolutionary origin and biological significance of haplotype configurations.

Phylogenetic history of hominoid DRB loci and alleles inferred from intron sequences

The evolutionary relationships among the MHC class II DRB4, DRB5 and DRB6 loci as well as the allelic lineages and alleles of the DRB1 locus were studied based on intron 1 and intron 2 sequences from humans, chimpanzee (Pan troglodytes), bonobo (Pan paniscus) and gorilla (Gorilla gorilla). The phylogenetic trees for these sequences indicate that most of the DRB1 allelic lineages predate the separation of the hominoid species studied, consistent with previous analysis of the coding sequences of these lineages. However, the intron sequence variation among alleles within DRB1 allelic lineages is very limited, consistent with the notion that the majority of the contemporary alleles have been generated within the last 250,000 years. The clustering of the DRB1 allelic lineages *08 and *12 with *03 supports a common ancestry for the DR8 and DR52 haplotypes. Similarly, the clustering of DRB1 allelic lineages *15 and *01 with the DRB3 locus is consistent with a common ancestry for the DR1 and DR51 haplotypes. Two cases of recombination around the second exon were observed: 1) the HLA-DRB6 locus appears to have been generated through a recombination between a DRB5 allele and an ancestral DRB6 allele, and 2) the gorilla sequence Gogo-DRB1 *03 appears to have been generated through a recombination between the DRB3 locus and an allele from the DRB1 *03 allelic lineage. The nucleotide substitution rate of DRB introns was estimated to 0.85-1.63 x 10(-9) per site per year, based on comparisons between the most closely related sequences from different hominoid species. This estimate is similar to the substitution rate for other intronic regions of the primate genome.

Retroelements in the human MHC class II region

Molecular genetic studies of the human major histocompatibility complex (MHC) have led to the identification of more than 200 genes. Besides the large number of genes in the MHC, densely clustered areas of retroelements have been identified. These include short and long interspersed elements (SINEs and LINEs), and human endogenous retroviruses (HERVs). The presence of retroelements in the MHC provides a clear example of how these elements affect the genome plasticity of the host. Comparative analyses of these retroelements have proven highly useful in evolutionary studies of the MHC. Recently, HERV-encoded superantigens have been implicated as candidate autoimmune genes in type I diabetes and multiple sclerosis. In addition, genetic analyses have revealed that autoimmune diseases show strong associations with MHC class II genes. The intriguing correlations between retroviral encoded antigens, MHC class II genes and the development of autoimmune disease merit intense future investigations of retroelements, in particular those endogenous retroviruses located in the MHC class II region proper.

Presence of retroelements reveal the evolutionary history of the human DR haplotypes

Comparison of intron sequences has been a successful tool for drawing major conclusions about the evolutionary relationship of DRB genes. This complex family of genes is discussed in this review as well as a proposed model for the evolution of HLA-DR haplotypes. The model is based both on phylogenetic analysis of intron sequences as well as presence of ERV9 LTR elements located at identical position in intron 5 of a number of DRB genes. According to this model, two main evolutionary branches of DR haplotypes exist. The DR53 haplotype represents one branch, and the second branch contains the DR51, DR52, DR1, and DR8 haplotypes. After the divergence of the DR53 haplotype, an ERV9 LTR element was inserted in a primordial gene. Consequently, all DRB1 genes as well as the DRB3 gene within haplotypes of the second branch, contain this LTR element. In addition, conserved regulatory sequence motifs are found present within these LTR elements that might regulate DRB gene expression. Novel haplotypes are generated by recombinations and the maintenance of the DR haplotype variation as well as the frequent genetic rearrangements observed might be evolutionary advantageous.

Neutral behavior of shared polymorphism

Several cases have been described in the literature where genetic polymorphism appears to be shared between a pair of species. Here we examine the distribution of times to random loss of shared polymorphism in the context of the neutral Wright-Fisher model. Order statistics are used to obtain the distribution of times to loss of a shared polymorphism based on Kimura's solution to the diffusion approximation of the Wright-Fisher model. In a single species, the expected absorption time for a neutral allele having an initial allele frequency of 1/2 is 2.77 N generations. If two species initially share a polymorphism, that shared polymorphism is lost as soon as either of two species undergoes fixation. The loss of a shared polymorphism thus occurs sooner than loss of polymorphism in a single species and has an expected time of 1.7 N generations. Molecular sequences of genes with shared polymorphism may be characterized by the count of the number of sites that segregate in both species for the same nucleotides (or amino acids). The distribution of the expected numbers of these shared polymorphic sites also is obtained. Shared polymorphism appears to be more likely at genetic loci that have an unusually large number of segregating alleles, and the neutral coalescent proves to be very useful in determining the probability of shared allelic lineages expected by chance. These results are related to examples of shared polymorphism in the literature.