Two independent proteomic approaches provide a comprehensive analysis of the synovial fluid proteome response to Autologous Chondrocyte Implantation

Abstract

BackgroundAutologous chondrocyte implantation (ACI) has a failure rate of approximately 20%, but it is yet to be fully understood why. Biomarkers are needed that can pre-operatively predict in which patients it is likely to fail, so that alternative or individualised therapies can be offered. We previously used label-free quantitation (LF) with a dynamic range compression proteomic approach to assess the synovial fluid (SF) of ACI responders and non-responders. However, we were able to identify only a few differentially abundant proteins at baseline. In the present study, we built upon these previous findings by assessing higher-abundance proteins within this SF, providing a more global proteomic analysis on the basis of which more of the biology underlying ACI success or failure can be understood.MethodsIsobaric tagging for relative and absolute quantitation (iTRAQ) proteomic analysis was used to assess SF from ACI responders (mean Lysholm improvement of 33; n = 14) and non-responders (mean Lysholm decrease of 14; n = 13) at the two stages of surgery (cartilage harvest and chondrocyte implantation). Differentially abundant proteins in iTRAQ and combined iTRAQ and LF datasets were investigated using pathway and network analyses.ResultsiTRAQ proteomic analysis confirmed our previous finding that there is a marked proteomic shift in response to cartilage harvest (70 and 54 proteins demonstrating ≥ 2.0-fold change and p < 0.05 between stages I and II in responders and non-responders, respectively). Further, it highlighted 28 proteins that were differentially abundant between responders and non-responders to ACI, which were not found in the LF study, 16 of which were altered at baseline. The differential expression of two proteins (complement C1s subcomponent and matrix metalloproteinase 3) was confirmed biochemically. Combination of the iTRAQ and LF proteomic datasets generated in-depth SF proteome information that was used to generate interactome networks representing ACI success or failure. Functional pathways that are dysregulated in ACI non-responders were identified, including acute-phase response signalling.ConclusionsSeveral candidate biomarkers for baseline prediction of ACI outcome were identified. A holistic overview of the SF proteome in responders and non-responders to ACI has been profiled, providing a better understanding of the biological pathways underlying clinical outcome, particularly the differential response to cartilage harvest in non-responders.Electronic supplementary materialThe online version of this article (10.1186/s13075-018-1573-4) contains supplementary material, which is available to authorized users.

Identification of Candidate Synovial Fluid Biomarkers for the Prediction of Patient Outcome After Microfracture or Osteotomy

“…The second, targeted approach was to assess a panel of SF and plasma proteins that we routinely test and that have been previously associated with OA 32,45 or that we have identified as changed in response to ACI. 11,12 This method meant that any candidate proteins identified from the proteomic analysis could be assessed in a larger cohort to determine whether they added any predictive value over and above other more commonly assessed OA proteins, baseline OA severity (as deemed by Kellgren-Lawrence score), and patient demographic characteristics, together contributing to the development of predictive models for patient outcome to microfracture or osteotomy surgery. Further, particularly for the microfracture cohort, there was a wide range of patient ages, which in itself could account for differential abundance of some of the measured proteins (eg, serum HA 19 ); therefore, this statistical approach could account for collinearity between parameters, such as between age and serum HA.…”

“…45 Concentrations of MMP-1, MMP-3, and S100A13 were assayed using duo-sets (MMP-1 and S100A13; Biotechne) or Quantikine ELISAs (MMP-3; Biotechne). 11,12 All assays were optimized to determine the appropriate sample dilution factor (over and above dilution due to lavage) as follows: ADAMTS-4, undiluted; sCD14, 1:200; COMP, 1:1000; HA, 1:3000; MMP-1, 1:3; MMP-3, 1:100; and S100A13, 1:20. Plasma samples were diluted 1:50 for COMP and assayed neat for HA.…”

“…As alternatives to assessing OA-related proteins, unbiased discovery-based approaches have been used to identify within-treatment proteome shifts that are unique to ACI nonresponders. 11,12 In the study described here, we used both these targeted and nontargeted approaches. We hypothesized that these approaches can be used to identify protein biomarkers that may predict an individual’s probability of a successful outcome in response to microfracture or osteotomy.…”

“…13 However, far fewer studies have focused on the identification of biomarkers to predict clinical outcome in response to a surgical intervention to prevent or delay the onset of OA. To our knowledge, the only published studies that have used human samples with this aim have been related to ACI 11,12,45 or anterior cruciate ligament (ACL) reconstruction. 18 Wright et al 45 demonstrated that assessment of proteins of known biological relevance to OA within the synovial fluid (SF) can identify protein markers that, when combined with known demographic or surgical risk factors, can be used to predict the success of surgical treatment with ACI.…”

Biomarkers of Joint Damage in Osteoarthritis: Current Status and Future Directions

“…In this method, different types of fluorescent dyes are used for the labeling of proteins, and these signals can also be used to identify proteins at the same spots in the gel [110]. There are different proteomics techniques used for the identification of proteins including matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF), liquid chromatography with tandem mass spectroscope (LC-MS/MS), surface-enhanced laser desorption/ionization-time-of-flight (SELDI-TOF), selected reaction monitoring (SRM), and multiple reaction monitoring (MRM) [111,112].…”