High-Density Genotyping of Immune-Related Genes Implicates Host Response to Microbes in Behcets Disease Risk To specifically target immune-related genes, we genotyped 2014 Turkish Behcets disease cases and 1826 controls (i.e., the Turkish Behcets disease GWAS discovery and Turkish replication samples from our previous disease association studies) using an Illumina iSelect Custom array, the Immunochip. The directly genotyped markers revealed three novel loci (IL1A-IL1B, IRF8, and CEBPB-PTPN1) with variant associations that exceeded the threshold for genome-wide significance. Imputation and fine mapping of suggestive loci with P < 5X10-5 in these Turkish samples identified significantly associated variants within EGR2. Associations of markers within EGR2 and IRF8 were replicated in 608 Behets disease cases and 737 controls from Japan and a meta-analysis of data from both populations identified genome-wide significant associations in LACC1 and RIPK2. Thus, we identified six novel loci, IL1A-IL1B, IRF8, CEBPB-PTPN1, EGR2, LACC1, and RIPK2. In addition, our Immunochip study led us to clarify reported associations of FUT2 and IL12A. The minor allele of rs4402765 located between IL1A and IL1B was associated with Behcets disease in the Turkish cohort with P = 2.22X10-9. This study is the first to report a disease-associated common variant in the IL1A-IL1B locus for any human illness, an important finding since rare mutations in NLRP3 (a component of the IL-1 activating inflammasome) and the IL-1 receptor antagonist cause autoinflammatory disease, and IL-1 inhibitors have been shown to be effective treatment for the ocular inflammation of Behcets disease. We found strong evidence that the most significant disease-associated IL1A genetic variant is also the most significant IL1A expression-associated variant in a large lymphoblastoid cell eQTL database. Somewhat surprisingly, the Behcets disease-associated allele is associated with reduced IL1A gene expression, and we also confirmed this finding in our own gene expression studies of freshly isolated healthy donor monocytes. IL-1alpha is highly expressed in the epidermis and plays an important role in skin barrier functions against pathogens, and thus a reduced capacity to express IL1A could lead to a reduced host response to invading or colonizing microorganisms. Interestingly, this concept of disease risk is similar to that which we previously proposed for the chemokine receptor, CCR1, for which the Behcets disease-related allele is associated with reduced gene expression and reduced chemotaxis, suggesting that the disease-associated allele reduces the ability to recruit infection controlling cells to sites of microbial colonization/infection. Thus, these variants could lead to Behcets disease susceptibility by promoting a more protracted or more extensive inflammatory response to common microbes. The SNP rs913678, located between CEBPB and PTPN1, associated with genome-wide significance in the Turkish cases and controls, P = 1.10 x 10-9, is significantly associated with decreased CEBPB expression in whole blood, but not with expression of PTPN1. CEBPB encodes C/EBPbeta, a member of the C/EBP family of basic region-leucine zipper proteins. C/EBPbeta is known to play a fundamental role in the antibacterial activity of macrophages; in fact, C/EBPbeta-/- mice are characterized by enhanced susceptibility to Candida albicans, Listeria monocytogenes, and Salmonella typhimurium infection. This is another example in which the Behcets disease-related allele is associated with a deficient host response to microbes. Our Immunochip study clarified the reported association of variants in the FUT2 locus, encoding alpha (1, 2) fucosyltransferase, which synthesizes secreted H antigen, the precursor of the ABO histo-blood group antigens in body fluids and the intestinal mucosa. We found that homozygosity for two common FUT2 non-secretor alleles of the ancestry specific SNPs, rs601338 (Turkish) and rs1047781 (Japanese), showed strong disease association (P=1.95X10-11). Non-secretor status has also been associated with Crohns disease and with the microbiome composition of the gut. Taken together, the reduced expression or function of Behcets disease-associated alleles of IL1A, CCR1, CEBPB, and FUT2 suggest that impaired host response to microbes and an altered microbiome composition contribute to Behcets disease. A manuscript describing these findings has been submitted for publication. Analysis of Killer Immunoglobulin-like Receptor (KIR) Genes in Behcets Disease The Behcets disease associated type, HLA-B*51, is a ligand for a pair of allelic KIR receptors present on cytotoxic cells, KIR3DL1, which inhibits their cytotoxicity, and KIR3DS1, which activates their cytotoxic activity. To determine whether KIR3DL1/KIR3DS1 genotypes are associated with Behcets disease risk, we genotyped the Turkish cases and controls from the Immunochip study with two sequence-specific PCR assays. KIR genotypes were obtained in 1799 of the cases and 1710 of the controls. Classical HLA types were imputed using the Immunochip HLA region genotypes and a reference of 5225 European individuals with SNP genotype and HLA type data. Presence of the cytotoxic activating KIR3DS1 or the inhibitory KIR3DL1 allele did not differ significantly between cases and controls. Furthermore, neither allele appeared to interact with HLA-B alleles (HLA-B*51 or HLA-B types bearing the Bw4 motif). Due to the complexity of this locus (i.e., sequence and copy number variation), these negative association results do not exclude a role of KIRs in the pathogenesis of Behcets disease. We are now exploring methods to more comprehensively evaluate the KIR genomic region. A manuscript describing our findings to date is currently in preparation. Analysis of ER-associated aminopeptidase 1 (ERAP1) protein structural variants and their association with Behcets disease The ERAP1 protein is responsible for trimming intracellular proteosome-derived peptides for efficient loading and presentation by HLA class I proteins. A coding ERAP1 variant (p.Arg725Gln) is recessively associated with Behcets disease and similar to that found in two other HLA class I-associated diseases, ankylosing spondylitis and psoriasis. The ERAP1 association is limited to individuals who carry the disease-associated HLA type. The ERAP1 protein is highly polymorphic with 10 missense amino acid variants with frequency greater than 1% found in the 1000 Genomes Project EUR super-population. A single disease-associated variant amino acid does not act in isolation, but acts instead in the context of all the variants present in the complete protein structure. We therefore used haplotype analysis to identify the common protein allotypes of ERAP1 and to evaluate their contributions to risk of Behcets disease. We identified 10 haplotypes with frequency greater than 1% in the EUR population. Eight of the 10 haplotypes were found at greater than 1% frequency in the Turkish population and only one, Hap10 (bearing 5 non-ancestral alleles, M349V, K528R, D575N, R725Q, and Q730E) was recessively associated with disease in 1900 cases and 1779 controls (P = 3.13 x 10-6). In the Turkish population, individuals who carry HLA-B*51 and are homozygous for the Hap10 haplotype have 10.96 fold increased disease risk compared with those with neither risk factor. This ERAP1 allotype has poor peptide-trimming activity; thus it is likely to contribute to Behcets disease risk by influencing the nature of peptide pool available for binding HLA-B*51.
Showing the most recent 10 out of 14 publications
