Genetic and functional studies of osteoarthritis susceptibility at COL11A1 and GDF5

Abstract

Osteoarthritis (OA) is the most common musculoskeletal disease and is characterised by joint pain and dysfunction resulting from the progressive focal loss of the articular cartilage of the joint. It is a multifactorial disease arising from the interplay between genetic and environmental risk factors, with a continuous distribution between the two extremes of predominantly genetic and predominantly environmental. Whilst several environmental risk factors for the disease are known, including body mass index (BMI), injury and occupation, known genetic risk factors are less well understood. Two major strategies have been employed in order to identify genetic factors conferring OA susceptibility: the investigation of candidate genes, which are principally chosen on the basis that the proteins that they code for are known to have a role in joint formation and maintenance, and genome wide association scans (GWASs), which search agnostically for disease associated polymorphic DNA variants. Within this thesis I report research on two genes highlighted using both strategies, although each is a compelling candidate. I report on the investigation of the functional effects of common polymorphism within COL11A1, which harbours a single nucleotide polymorphism (SNP) that showed evidence of association to OA via the arcOGEN GWAS. I also report on the search for and analysis of rare variants of GDF5, a gene previously discovered as harbouring OA susceptibility by candidate gene studies. I tested for an allelic expression imbalance (AEI) of COL11A1 using RNA from patient cartilage but I did not detect a correlation between genotype at the GWAS OA associated SNP rs2615977 and the expression of the gene. I did however discover that genotype at a second polymorphism within COL11A1, rs1676486, did correlate with COL11A1 AEI. rs1676486 has previously been reported to be associated with lumbar disc herniation. However, my analysis of the arcOGEN dataset revealed that this SNP is not associated with OA. To identify rare variants in GDF5 that could impact upon OA susceptibility I sequenced the protein coding region of the gene, its untranslated regions, exon-intron boundaries and its proximal promoter in 962 OA cases and controls. Six novel and very rare variants with minor allele frequencies (MAFs) of ≤0.0006 were discovered and I confirmed the existence of known variants with common MAFs. The absence of variants with intermediate MAFs implies that the gene may have been subjected to a genetic bottleneck. One rare variant, within the proximal promoter of GDF5, was carried forward for functional analysis using luciferase reporter assays. I discovered that the novel A-allele of this variant increased the expression of the reporter plasmid relative to its common C-allele. I also demonstrated that this A-allele is able to counteract the reduced gene expression mediated by the T allele of the previously reported OA associated GDF5 SNP rs143383. I subsequently performed electrophoretic mobility shift assays (EMSAs) and identified the transcriptional activator/repressor YY1 as the trans-acting factor binding differentially to the alleles of the variant. Overall, this thesis demonstrates that the OA association to COL11A1 identified by GWAS is not caused by AEI of that gene within the articular cartilage of OA patients, and that the deep sequencing of current OA susceptibility loci to identify novel risk alleles is also a means to identify mechanisms to counteract the effects of susceptibility alleles that are already known.