Identification of genetic risk factors within the human leukocyte antigen region for ulcerative colitis

The human leukocyte antigen (HLA) region is a highly variable gene cassette located on chromosome 6p21 that is involved in diverse functions of the host-immune system. The HLA is the strongest but also the most complex susceptibility locus for inflammatory bowel disease (IBD) and its subforms Crohn’s disease (CD) and ulcerative colitis (UC). Tight structures of linkage disequilibrium within this highly complex gene locus complicate the identification of causal variants to an association signal. In previous publications, the DRB1*01:03 allele was identified as a determining factor for IBD in the Caucasian population1. However, for most of the association signal such a delineation was not possible. The aim of this thesis was therefore to further characterise the HLA signal and to analyse whether tightly linked signals between the HLA-DQ and -DR loci could be resolved using a trans-ethnic approach. The focus of this study was UC. To this end, we assembled a unique dataset of individuals from 8 different ethnicities in collaboration with partners of the International IBD Genetics Consortium. The individuals were typed on lllumina’s ImmunoChip. With NGS-based typing of HLA alleles being expensive for a larger number of samples, we opted to inferring HLA alleles using an imputation approach. First, we compiled an HLA imputation reference panel specially tailored to our cohorts of interests. Next, we imputed HLA alleles and performed HLA fine mapping across the different populations. For the first time, we identified genetic factors shared across different ethnicities both in effect direction and magnitude. We found whole allele groups that are associated with UC, highlighting alleles of the HLA-DRB1*15 group. Using haplotype phasing of HLA alleles, we were able to explain the tightly linked signals of the HLA-DQ and -DR loci and showed that there are entire HLA-DQ-DR haplotypes shared across different ethnicities. In addition, we identified the previously reported DRB1*01:03 as a population-specific signal that is mostly present in individuals of Western European descent and hardly present in non-Caucasian individuals. Furthermore, by analysis of physico-chemical properties of the respective HLA proteins and analysis of peptides that are preferentially bound by them, we were able to further reduce the complexity of the signal, showing that risk and protective proteins, respectively, share common features both in regard to their chemical properties as well as to the peptides they preferentially bind. Our findings should be of broad interest to the research community, both to genetics and peptidomics experts also in the context of the analysis of other immune-related diseases, for instance multiple sclerosis or leprosy, which are also associated with the HLA-DRB1*15 group.