NEW YORK (GenomeWeb) – A Wellcome Trust Sanger Institute-led team has identified nine novel loci associated with osteoarthritis risk.
Osteoarthritis, the most common form of arthritis, can afflict any joint — commonly the hands, hips, and knees — as the cartilage between bones breaks down. According to the Centers for Disease Control and Prevention, more than 30 million adults in the US are affected by osteoarthritis, particularly people over the age of 70.
To search for genetic loci associated with the condition, the team conducted a genome-wide association study using samples from the UK Biobank and other cohorts. As they reported in Nature Genetics, the researchers found nine new osteoarthritis loci, and traced the likely causal variants to a number of non-coding regions, including in genes that are differentially expressed in affected cartilage.
"These results are an important step towards understanding the genetic causes of osteoarthritis and take us closer to uncovering the mechanism behind the disease," Eleni Zengini, joint first author from the University of Sheffield, said in a statement. "Once we know that, it opens the door to developing new therapies for this debilitating disease."
Zengini and her colleagues first turned to samples from the UK Biobank, both from people clinically diagnosed with osteoarthritis and from people who self-reported having osteoarthritis, for the discovery phase of their GWAS. While they noted relying on self-reports of osteoarthritis increased phenotype uncertainty, they said that the increased sample size and power they garnered by folding in those self-report samples overcame that drawback.
With this discovery cohort, they conducted five GWAS based on whether the individual's osteoarthritis was self-reported or clinically diagnosed and whether the osteoarthritis affected their hips or their knees, or was at any joint.
The researchers then took 173 variants they uncovered in the discovery phase forward, for replication testing in an Icelandic cohort from DeCode Genetics with up to 18,069 cases and 246,293 controls, and for combined meta-analyses in both cohorts. After the meta-analyses, they found six loci with genome-wide significant associations and three loci that fell just below significance for association with osteoarthritis.
These nine loci were all located within non-coding sequences, the researchers noted. For instance, rs2521349, which was linked to hip osteoarthritis, fell within an intron of MAP2K6 — which encodes a protein involved in cellular processes influencing bone, muscle, and fat-tissue homeostasis and differentiation — while rs375575359, which was linked to osteoarthritis at any site, resides within an intron in the zinc finger-encoding gene ZNF345 and the intergenic variant rs2820436 fell in between a long-noncoding RNA gene and a pseudogene and was associated with osteoarthritis at any joint.
The other osteoarthritis-linked loci could be traced to introns of and intergenic regions near genes involved in cytoskeletal structures, in integrating cellular signaling and function, and in musculoskeletal development in mice.
Through a combination of quantitative proteomics and RNA sequencing, Zengini and her colleagues found that the expression of a number of coding genes near these nine loci differed between healthy and diseased chondrocytes obtained from patients undergoing joint replacement surgery.
For example, PCYOX1, near the rs3771501 variant, was differentially expressed at both the RNA and protein levels in degraded cartilage, showing increased RNA expression, but decreased protein expression. The protein it encodes, prenylcysteine oxidase 1, catalyzes the degradation of prenylated proteins. Additionally, FAM136A, also near rs3771501 had lower transcriptional levels in degraded cartilage. Four other genes likewise exhibited differential expression. The researchers noted that these genes could represent targets for future therapeutics.
Zengini and her colleagues also noted a strong genome-wide correlation between knee osteoarthritis and BMI.