Proteomic Clustering Reveals the Kinetics of Disease Biomarkers in Bovine and Human Models of Post-Traumatic Osteoarthritis.

Objectives: In this study, we apply a clustering method to proteomic data sets from bovine and human models of post-traumatic osteoarthritis (PTOA) to distinguish clusters of proteins based on their kinetics of release from cartilage and examined these groups for PTOA biomarker candidates. We then quantified the effects of dexamethasone (Dex) on the kinetics of release of the cartilage media proteome. Design: Mass spectrometry was performed on sample medium collected from two separate experiments using juvenile bovine and human cartilage explants (3 samples/treatment condition) during 20- or 21-day treatment with inflammatory cytokines (TNF-α, IL-6, sIL-6R) with or without a single compressive mechanical injury. All samples were incubated with or without 100 nM Dex. Clustering was performed on the correlation between normalized averaged release vectors for each protein. Results: Our proteomic method identified the presence of distinct clusters of proteins based on the kinetics of their release over three weeks of culture. Clusters of proteins with peak release after one to two weeks had biomarker candidates with increased release compared to control. Dex rescued some of the changes in protein release kinetics the level of control, and in all conditions except control, there was late release of immune-related proteins. Conclusions: We demonstrate a clustering method applied to proteomic data sets to identify and validate biomarkers of early PTOA progression and explore the relationships between the release of spatially related matrix components. Dex restored the kinetics of release to many matrix components, but not all factors that contribute to cartilage homeostasis.