Cindy Boer

Cartilage Thickness and Hip Osteoarthritis | 83 2.1 not PIK3R1 but rather a long-non-coding RNA (lncRNA), lnc-PIK3R1-4:1 , is causal, since a variant in LD with the lead SNP is located in the predicted transcription start site of this lncRNA potentially affecting its expression. Although conserved in mice and zebraf- ish, thus far no function has been ascribed to this lncRNA [34]. We confirmed a locus previously associated with cartilage thickness, the DOT1L locus. Our identified SNP, rs11880992 is in high LD with the previously reported SNP rs129827744, and both are associated with cartilage thickness and hip OA[1]. Despite the previously presented suggestive evidence for involvement of DOT1L in chondrogen- ic differentiation, DOT1L did not receive a high score in our systematic prioritization study; the gene GADD45B, located in the region 500Kb downstream of the lead SNP, received a similar score. GADD45B is a transcriptional co-factor for C/EBP-β , a master regulator of chondrocyte differentiation[35]. Thus, it remains unclear which gene or genes in this locus contribute to the cartilage phenotype. Further research is needed to determine whether DOT1L is the true causal gene in this locus. Our analyses suggest that the majority of prioritized genes in hip OA associated loci are involved in cartilage and bone developmental pathways; including TGFA , RUNX2 , FGFR3 , PTHLH , COL12A1 and others that seem to affect bone and/or cartilage develop- ment such as PIK3R1 and KMT2A We hypothesize that the mJSW and OA associated variants influence gene expression regulation. The dysregulation of these genes and mechanisms during development may, later in life, result in an increased risk for OA. The identified mJSW SNPs are associated with hip OA, but not with knee OA. We have analysed the identified SNPs also for association with knee OA in the TREAT-OA meta-analysis dataset[36], but found no association. This observation fits in the overall finding that many of the identified genetic loci for OA seem to be site-specific [37], and support the hypothesis that the aetiology of OA is different in each joint. Nevertheless, this observation can still be a result of low power in the GWAS studies that have been done for OA till now[38], and final conclusions on this aspect cannot be drawn at this point. This is the first report linking TGFA to human OA most likely by affecting mJSW. It may serve as a new target for future therapies. We have identified multiple mJSW asso- ciated loci which have previously been associated with other bone and cartilage related phenotypes such as bone mineral density and height, displaying a possible pleiotropic effect for the analysed traits. It will be important to understand how mJSW and OA as- sociated variants can affect the developmental processes that regulate morphometry of the hip joint, including the formation of articular cartilage. Therefore further expres- sion and functional studies are warranted of genes identified to be associated with hip OA phenotypes.