Exploring the early molecular pathogenesis of osteoarthritis using differential network analysis of human synovial fluid

This MS data set was used as validation data for an exploratory proteomics study of biomarkers in osteoarthritis (OA) using SOMAscan assay data. We analyzed human synovial fluid from three groups with the two methods: 1) deceased donors with mildly degenerated cartilage and/or meniscus, which is a well-known characteristic of OA, 2) deceased donors with no degeneration as controls and 3) total knee replacement patients with end-stage knee osteoarthritis. We conducted gaussian graphical network analyses of proteins from group 1 and 2 using SOMAscan assay data, and included the 800 aptamers/proteins with largest absolute log2 fold changes in the network construction. We found 102 differentially connected proteins, which of 31 were also present in the MS data set. After filtering out non-proteotypic peptides, proteins with only one peptide used for identification across all samples and proteins with more than 50% missing values, we ended up with 21 proteins. We analyzed these proteins by 1) estimating correlation between the methods on log2 standardized data sets and 2) compared log2 fold changes and 95%CIs by linear mixed effect model with quantification on the protein level as the outcome variable, adjusted for age and sex. Most proteins shown a positive correlation for the two methods. The comparison of log2 fold changes and 95%CIs between the methods show similar results for mild degeneration group when compared to healthy controls, whereas the results for the comparison between healthy controls and late stage OA group show dissimilarities for a few proteins.

Shared first authors: Martin Rydén, Amanda Sjögren
Authors: Patrik Önnerfjord, Aleksandra Turkiewicz, Jon Tjörnstrand*
Last authors: Neserin Ali ([e-mail]), Martin Englund ([e-mail])
Department of Clinical Sciences Lund, Lund University
*Department of Orthopedics, Skåne University Hospital, Sweden