Mass spectrometry profiling of non-enzymatic deamidation of articular cartilage components suggests slower protein turnover in deep regions and in hips compared with knees

s / Osteoarthritis and Cartilage 24 (2016) S8eS62 S17 16 MASS SPECTROMETRY PROFILING OF NON-ENZYMATIC DEAMIDATION OF ARTICULAR CARTILAGE COMPONENTS SUGGESTS SLOWER PROTEIN TURNOVER IN DEEP REGIONS AND IN HIPS COMPARED WITH KNEES M.-F. Hsueh yz, A. Khabut x, S. Kjellstr€ om k, R.D. Zura ¶, M.P. Bolognesi ¶, P. Onnerfjord x, V. Kraus yz. yDuke Univ., Durham, NC, USA; zDuke Molecular Physiology Inst., Durham, NC, USA; xDept. of Clinical Sci., Lund Univ., Lund, Sweden; kDept. of Biochemistry and Structural Biology, Lund Univ., Lund, Sweden; Duke Univ. Med. Ctr, Durham, NC, USA Purpose: Articular cartilage has an abundant extracellular matrix (ECM) consisting of many proteins with long half-lives. Osteoarthritis (OA) can arise due to increased catabolic activity with insufficient protein synthesis to maintain tissue homeostasis. Although it is assumed that certain matrix proteins are maintained in low-turnover states, the turnover rates in different cartilage locations are not uniform but are believed to vary with distance from the cartilage surface. We hypothesized that the quantification of post-translational modifications could serve as an index of in vivo protein turnover. We focused on deamidation, a non-enzymatic covalent modification produced by the hydrolysis of the amide group on the side chains of Asn or Gln to form Asp and Glu. There are no repair systems in the cartilage ECM for deamidation and, as a consequence, deamidated proteins can accumulate in a time-dependent manner. In the present study, we identified deamidated protein in extracts of chondrocytes and matched ECM from the knee joint. Further, we profiled the deamidated proteins within ECM from different joint types (knee and hip), disease states (healthy and OA), and the depth of cartilage (superficial, middle and deep zones). Methods: We collected hip and knee cartilage (both healthy and OA) as surgical waste from surgical repair of trauma or joint replacement. Serial transverse cryosections of 12mm thickness were generated at different depths. Chondrocyte proteins were separated from ECM by one rapid freeze/thaw cycle in hypotonic solution and cartilage ECM proteins were extracted by quanidine-HCl. All extracts were processed for proteomic analysis performed with an EasyLC nanoflow HPLC connected to a LTQ-Orbitrap Velos Pro mass spectrometer (Thermo Fisher Scientific) equipped with a nanoEasy spray ion source. Protein classification analysis was performed by the PANTHER classification system (www.pantherdb.org). Multivariable analyses were performed to evaluate for differences in abundance of deamidated epitopes due to joint site, disease state and depth to identify the independent association of each of these factors with deamidation. Results: To obtain a profile of the deamidated proteins within cartilage, we performed non-targeted discovery proteomic analysis to identify the deamidated peptides within chondrocytes and ECM from knee cartilage. In the chondrocyte extracts, representing primarily intracellular proteins, we identified 79 deamidated proteins (out of 341 identified proteins) with only 12% of the Asn and Gln being deamidated on average. In the cartilage ECM extracts we identified 108 deamidated proteins (out of 330 identified proteins) with 34% of the Asn and Gln deamidated on average. Expanding our analyses to cartilage from different joint types (knee and hip), disease states (healthy and OA), and the depths (superficial, middle and deep), we identified 165 deamidated proteins (out of 454 identified proteins). Protein classification analysis confirmed that up to 70% of the proteins with deamidated epitopes were ECM associated. We quantified the deamidation ratio (deamidated NþQ/ total NþQ) and performed a multivariable regression analysis with all three factors simultaneously on proteins, including fibronectin, cartilage oligomeric matrix protein (COMP) and aggrecan. Controlling for joint site and disease state, there were more deamidated fibronectin and COMP epitopes identified in the deep zone (Figure. 1). Controlling for the other two factors, there were also more deamidated COMP epitopes in hip cartilage and in OA cartilage compared with knee and healthy cartilage, respectively. In addition, more deamidated aggrecan G1 was identified in hip compared with knee cartilage. Conclusions: We have evaluated the presence of deamidated epitopes in chondrocyte intracellular proteins and ECM proteins. The number of identified deamidated proteins, deamidated epitopes and the protein classification results suggest that proteins associated with ECM are more susceptible to deamidation. The multivariable analyses suggests that deamidation can vary by joint type, disease states, and the depth of cartilage. Analyses of three canonical cartilage proteins, COMP, fibronectin and aggrecan demonstrated a general accumulation of deamidated eptiopes in hip cartilage and in deep cartilage of both hip and knee. These data are consistent with slower turnover of the ECM of hip cartilage and slower protein turnover in the deep zone of cartilage. Figure 1. Deamidated residue in cartilage matrix. 17 SITE SPECIFIC PROTEOGLYCAN CONTENT IS BETTER MAINTAINED IN THE PERICELLULAR THAN EXTRACELLULAR MATRIX IN EARLY POSTTRAUMATIC OSTEOARTHRITIS S.P. Ojanen yz, M.A. Finnil€a yz, A.E. Reunamo y, A.P. Ronkainen y, W. Herzog x, S. Saarakkala zk, R.K. Korhonen y. yUniv. of Eastern Finland, Kuopio, Finland; zUniv. of Oulu, Oulu, Finland; xUniv. of Calgary, Calgary, AB, Canada; kOulu Univ. Hosp., Oulu, Finland Purpose: One of the characteristics of early osteoarthritis (OA) is the loss of proteoglycans (PGs) in the superficial zone of articular cartilage. Previous studies have mainly focused on investigating alterations of the PG content in the extracellular matrix (ECM) during the disease, while possible local changes in the pericellular matrix (PCM) are not fully understood. It is still debated whether PG content is better maintained in the PCM than ECM for instance due to increasing PG production at the early stage of OA. We hypothesized that PG content is reduced less in the PCM compared to the ECM in an animal model of early posttraumatic OA. Methods: Unilateral anterior cruciate ligament transection (ACLT) was performed to eight 14-month old rabbits and they were sacrificed four weeks after the surgery. The non-operated knees were used as contralateral controls (C-L). Furthermore, an intact control group (CNTRL) consisted of six knees from three age-matched rabbits. Lateral and medial femoral condyles, femoral grooves, tibial plateaus and patella were extracted from the joints. After sample processing, 3 mm thick sections were cut from the primary load bearing areas and stained with Safranin-O. PG content was quantified from images collected with light microscope and CCD cooled camera under monochromatic illumination. Images were calibrated against neutral density filters to obtain digital densitometry (DD) images (Fig. 1A). Analyses were conducted using a custom-made MATLAB (R2012a, Mathworks Inc., Natick, Massachusetts, United States) script. From each section an average of five cells were chosen from each cartilage zone (superficial, middle and upper-deep). Zones were defined by cell shapes (flat, round, ellipsoidal) and the upper-deep zone was up to ~50% of tissue thickness. For each cell vertical (axial) and horizontal (transversal) optical density (OD) profiles were calculated (Fig. 1B). Differences of the PG content between the groups were analyzed by comparing the absolute and normalized OD values of the PCM and ECM. Profiles were normalized to highlight the changes in the PG content in the PCM relative to that in the ECM (Fig. 1C). One-way ANOVA statistical analyses were performed using SPSS (Ver. 21, IBM Corp., Armonk, NY). Results: PG content in the PCM (absolute OD values) was reduced (p < 0.05) in ACLT compared to CNTRL knees at all analyzed depths in femoral condyle, tibial plateau, and patellar cartilage (Table 1). Comparing C-L with CNTRL group, significant differences in the PG content (p < 0.05) occurred only in tibial plateau and patellar cartilage particularly in the superficial and middle zones. Significantly (p < 0.05) higher normalized PG content in the PCM in both ACLT and C-L groups compared to the CNTRL group samples was found in the superficial zone of cartilage in femoral condyles, tibial plateaus and patella. In the middle zone, significant differences (p < 0.05) in the normalized PG content of the PCM between ACLT and CNTRL groups were found at all sites, and between C-L and CNTRL groups in tibial plateau and patellar cartilages. Increased normalized PG content of the PCM in the deep zone of cartilage (p < 0.05, ACLT vs. CNTRL) was observed only in femoral condyles and patella. Conclusions: ACLT caused less PG loss in the PCM than in the ECM, especially in the superficial and middle zones of cartilage, at all sites