Osteoarthritis subtypes show distinguished transcriptomic landscapes

Abstract

Purpose: Osteoarthritis affects over 10% of elderly, yet, no disease modifying drugs are available. In part, development of such drugs may be hindered due to existing disease heterogeneity which thus far has not been taken into account. Subtypes likely depend on different molecular mechanisms underlying the divers clinical features and would, as such, likely benefit from different therapeutic modes of action. To identify potential subtype-specific druggable targets, differences in the transcriptomic landscapes of articular cartilage from atrophic (osteophytes absent) and normotrophic (osteophytes present) osteoarthritis subtypes were determined. Methods: Articular cartilage samples were collected from the ongoing RAAK study. Whole genome expression profiles of macroscopically unaffected (preserved) cartilage of atrophic and normotrophic patients were established with microarray (Illumina HumanHT-12 v3; N=5 atrophic versus N=20 normotrophic) and RNA-sequencing (Illumina HiSeq 2000; N=9 atrophic versus N=8 normotrophic). A meta-analysis was performed for 6,257 genes overlapping between microarray analyses and RNA-sequencing using the Log2 Fold-Difference in combination with the Standard Error of the Mean. Enrichment of genes involved in specific pathways was explored with online available tools (STRING and DAVID; PFDR ≤5.0x10-5). Results: Periostin (POSTN), a gene functioning in osteogenesis, was identified to be higher expressed in atrophic cartilage with transcriptome wide significance (PFDR = 2.7x10-2; fold difference: 2.2). Significant enrichment was observed for genes involved in skeletal system development (PFDR = 1.7x10-5) and in glycosaminoglycan binding (PFDR = 3.4x10-4) for the 366 unique genes identified with nominal significant difference in atrophic compared to normotrophic cartilage (P ranging from 3.5x10-5 to 5.0x10-2). Conclusions: To our knowledge, we are the first to explore the transcriptomic landscapes of cartilage from osteoarthritis subtypes, and we have identified POSTN to be differentially expressed between atrophic and normotrophic patients. Currently, RT-qPCR is being performed with additional samples to further follow-up on the results. Given that POSTN plays a role in osteogenesis and atrophic patients do not have osteophytes, possibly the gene is involved in mineralization of the cartilage. This needs to be further investigated in functional studies.