Ankylosing spondylitis (AS) is a common, highly heritable, inflammatory arthritis primarily affecting the spine and pelvis. Disease progression in AS is characterised by inflammation and excessive bone formation that gradually bridges the gap between joints, leading to eventual fusion of affected joints in severe cases. Genetic factors are the main determinants of disease susceptibility, with an estimated heritability above 90%. Disease prevalence estimates vary between ethnic populations, and these range between 1 in a 1,000 to 1 in a 100; this variability is partly explained by the population frequency of the main genetic risk factor, the HLA-B*27 allele.
In the past six years, genome-wide association studies have robustly identified 12 loci outside of the MHC associated to disease susceptibility in individuals of European descent and two loci identified in individuals of Asian descent. In aggregate, these finding do not account for the observed heritability of the disease estimated with epidemiological studies, suggesting that further genomic loci are associated to disease susceptibility. Furthermore, it has been demonstrated that genetic factors are important determinants of disease severity and in the osteoproliferative process observed at the sites of inflammation. Much has been learned about the biological mechanisms involved in disease aetiology following previous genetic studies in disease susceptibility. However, the mechanisms underlying the transition between the inflammatory state to that of bone formation, characteristic of AS pathology, are poorly understood.
This thesis describes three studies aimed at better characterising the genetic factors that affect disease susceptibility and radiographic severity in AS. In the first study, the hypothesis was tested that variants in genes involved in anabolic and catabolic bone pathways are associated with the osteoproliferative process observed at sites of inflammation in AS patients. A quantitative measure of bone formation was derived from radiographs of patients scored with the modified Stoke Ankylosing Spondylitis Spinal Score, a clinical measurement used to assess radiographic severity in patients. In this study 498 SNPs were selected and genotyped in 688 samples with lumbar and cervical radiographs and a further 15 SNPs were genotyped in a replication cohort of 830 cases. One variant in the gene RANK was found to be associated (experiment-wise significance; p<10-4) with radiographic severity, and a second variant in the gene PTGS1 showed suggestive association in the discovery and replication phase analyses. RANK encodes a TNF-superfamily receptor which interacts with RANK-ligand to regulate osteoclast activation. These findings suggest that factors linking inflammation and bone resorption are important in the pathogenesis of osteoproliferation in AS, with obvious potential therapeutic implications. This study is described in Chapter Two.
The second study describes a case-control association study using the Illumina Immunochip microarray which aimed to identify further loci associated with AS. The Immunochip is a custom-design genotyping platform with close to 200,000 SNPs designed for immunogenetic studies. As part of this study nine strategies for deriving genotypes from Illumina’s BeadChips raw intensity files were compared, as this has not previously been described for custom-design microarrays with high content of low frequency variants. This comparison is described in Chapter Three.
Chapter Four describes the results from the case-control association study involving 10,619 individuals with AS and 15,145 controls using the Immunochip. This study identified 13 new risk loci and 12 additional AS-associated haplotypes at 11 loci. These findings have increased the number of associated loci to 31 with a total to 43 independent genetic signals.
In Chapter Five, an investigation of the major histocompatibility complex associations of AS using the Immunochip genotypes is described. Using imputation, association of SNPs and HLA class I and II classical alleles and their amino acid determinants was tested. This study identified three HLA-B risk and two protective alleles, and found evidence of association with the MHC class I HLA-A*02:01 allele and the class II locus HLA-DPB1. The interaction between the ERAP1 risk allele and the HLA-B*27 allele was confirmed and further evidence of interaction with the HLA-B*40:01 risk allele was observed. Amino acid residue analysis suggested that the associations observed at the three loci, HLA-A, -B and -DPBP, are highly correlated with polymorphic amino acid positions in the peptide-binding groove of these proteins.
In summary, through gene-mapping experiments several genetic loci associated with disease susceptibility that implicates several biological mechanisms in AS aetiology were identified. Amongst these findings, four independent signals were found at two loci encoding four aminopeptidases that are involved in peptide processing before MHC class I presentation. Furthermore, it was observed that the association of the ERAP1 risk allele was restricted to carriers of HLA-B*27 or HLA-B*40:01 alleles, supporting the hypothesis that the mechanisms by which HLA-B*27 confers risk to disease is through peptide presentation. IL-23R signalling and IL-17/IL-22 mediated immunity were first implicated in AS aetiology with the discovery of IL23R risk variants, and this study further identified susceptibility variants at other members of this pathway, including IL6R, TYK2, IL27, IL1R2 and IL1R1. In combination, the mounting evidence of the importance of peptide presentation and IL-23R signalling in AS pathogenesis supports the development of new therapeutics targeting these pathways.