Faculty Bibliography

Tissue-engineering techniques have been successful in developing cartilage-like tissues in vitro using cells from animal sources. The successful translation of these strategies to the clinic will likely require cell expansion to achieve sufficient cell numbers. Using a two-dimensional (2D) cell migration assay to first identify the passage at which chondrocytes exhibited their greatest chondrogenic potential, the objective of this study was to determine a more optimal culture medium for developing three-dimensional (3D) cartilage-like tissues using human cells. We evaluated combinations of commonly used growth factors that have been shown to promote chondrogenic growth and development. Human articular chondrocytes (AC) from osteoarthritic (OA) joints were cultured in 3D environments, either in pellets or encapsulated in agarose. The effect of growth factor supplementation was dependent on the environment, such that matrix deposition differed between the two culture systems. ACs in pellet culture were more responsive to bone morphogenetic protein (BMP2) alone or combinations containing BMP2 (i.e. BMP2 with PDGF or FGF). However, engineered cartilage development within agarose was better for constructs cultured with TGFbeta3. These results with agarose and pellet culture studies set the stage for the development of conditions appropriate for culturing 3D functional engineered cartilage for eventual use in human therapies

INTRODUCTION: The synovial membrane encapsulates diarthrodial joints to create a fluid-filled cavity that provides a lubricating environment for the articulating cartilage surfaces within. Fibroblast-like synoviocytes (FLS) regulate the composition of this synovial fluid, maintaining its viscous and lubricating properties. Fluid shear arising during joint articulation is recognized as a relevant stimulus for FLS in both healthy and pathologic conditions1. Gentle oscillatory shear has been shown to influence the response of FLS to inflammatory cytokines overexpressed in osteoarthritis (OA)2. Cartilage wear particles (CWP) are also recognized as a key factor in OA3,4, but their influence on FLS mechanosensing sensing is unknown. The specific mechanisms of FLS shear sensing and its downstream consequences, as well as the influence of the OA environment on this sensing are still unknown. This work seeks to characterize the real-time intracellular calcium ([Ca2+]i) response of FLS to physiologic fluid shear and determine how this response is modulated by both chemical and physical OA factors, Interleukin-1alpha (IL1) and CWP. Intracellular calcium is an important cell signaling molecule in cell mechanotransduction. The aim of the current study is to test the hypothesis that these OA factors heighten FLS shear sensitivity as determined using an in vitro model system. METHODS: Cell Culture. FLS were isolated from juvenile bovine synovium, expanded in alphaMEM containing 10% fetal bovine serum (FBS), 100 U/mL penicillin, 100 mg/mL streptomycin, and 5 ng/mL bovine FGF for two passages to obtain a pure population, and plated into silicone wells on 5 mug/cm2 collagen type 1 coated glass slides at high density to obtain a confluent monolayer after 24hr attachment. To mitigate artifacts due to cellular starvation the media was switched to 0.5% FBS for 12hr prior to imaging, with 10 ng/mL IL1 or CWP at 50 particles/cell for preconditioned groups. CWP were generated by mechanical abrasion of cartilage explants and sorting via Coulter counter to obtain a size of 0.9 mum5. Calcium Signaling. FLS were loaded with 5muM Fura Red to track relative [Ca2+]i in real time, where increasing [Ca2+]i results in decreasing fluorescence. A parallel plate flow chamber was employed to expose cells to 1 dyne/cm2 shear stress. Hank's Buffered Salt Solution with 0.1% FBS was flowed, with 10 ng/mL IL1 or 1 mM octanol for indicated groups. Fluorescence intensity over time was determined for individual cells and normalized as relative [Ca2+]i with respect to average intensity during equilibrium period prior to flow. Cells were considered responders if relative [Ca2+]i rose 20% above equilibrium level, which was found to be sufficient to exclude any false responses due to natural variation in equilibrium levels. Percentile response, peak magnitude, area, and latency were determined for 100 cells/slide across 3 slides/group. Dye Transfer Assay. 'Parachute' FLS were loaded with Calcein and DiI and plated at low density onto a non-dyed FLS monolayer, where Calcein transfer to neighboring cells was quantified after 4hr as a measure of gap junction communication, for 40 parachute cells (identified by non-transferable DiI) across triplicate slides per group. Preconditioned monolayers received 10 ng/mL IL1, or 50 particle/cell CWP for 24hr prior, while 1 mM octanol was applied during imaging for the negative control. Statistical Analysis. Parametric statistics were performed via ANOVA with Tukey HSD post-hoc testing (alpha<0.05), and nonparametric statistics via Fischer Exact Test with Holms-Sidak correction (alpha<0.05). RESULTS: FLS display a single Ca2+ peak in response to fluid shear (Figure 1a) coordinated among a percentage of the population, dependent upon a minimum level of serum. IL1 showed differing effects on shear response depending on timing of exposure (Figure 1b). Exposure during flow alone (CTL/IL1) significantly decreased percentile response, while preconditioning had no significant effect alone (IL1/CTL) but abolished the decreasing effect of IL exposure during flow when combined (IL1/IL1). A separate vehicle-controlled experiment showed that FLS do not respond to IL exposure itself. CWP attachment to FLS (Figure 2b,c) also significantly increased percentile response (Figure 2a) as well as peak magnitude, area, and latency. The proposed mechanism for both IL1 and CWP preconditioning enhancement of shear sensing is increase gap junction formation and communication. Gap junction blocking via octanol significantly reduced percentile response by 22%, indicating that cell-cell passing of Ca2+ signals makes up a percentage of responding cells in normal conditions, while dye transfer experiments showed that IL1 and CWP both increased gap junction communication compared to control (Figure 3). DISCUSSION: FLS demonstrate a robust Ca2+ signaling response to fluid shear which is modulated by chemical and physical factors of the OA environment, namely by increasing percentile response and thus confirming the hypothesis of heightened sensitivity. While mechanisms for the suppressing influence of IL1 exposure alone are still being explored, the influence of IL1 and CWP preconditioning may be explained by an increase in gap junction communication leading to more cell-cell signaling and higher percentile response, which supports previous implications of increased gap junction communication between FLS in the pathology of OA6. Future work will focus on the mechanisms for individual cell shear sensing, and how they are influenced by OA factors, thus contributing to the phenomenon observed in the present study. The primary cilium is of particular interest as it is both involved in mechanosensing in other cell types7, and implicated in OA pathology in its modulation by the inflammatory environment8. (Figure Presented)

Background/Purpose: We and others have previously shown that periostin (Postn) expression is dramatically elevated in cartilage and sub-chondral bone in OA patients and surgical models of OA (medial meniscectomy and anterior crucial ligament resection or PMX, partial meniscectomy) in rodents. In vitro Postn promotes collagen and proteoglycan degradation in human chondrocytes by upregulating MMP-13 and ADAMTS4 expression. Postn controls gene expression in bone cells by interacting with avb3 integrin. However, the nature of periostin receptor(s) in chondrocytes is unknown. DDR1, a collagen-binding receptor tyrosine kinase highly expressed in chondrocytes, controls MMP-13 expression during chondrogenesis. Therefore, we hypothesized that the effect of Postn on chondrocytes is mediated by DDR1 and Postn-deficient mice (Postn^-/-) are protected from surgically-induced post-traumatic OA. Methods: (Postn^-/-) mice were purchased from Jackson Laboratory (B6;129-Postntm1Jmol/J Stock No: 009067). We subjected 3- months old littermates (Postn+/+, Postn+/- and Postn^-/-) to partial medial meniscectomy (PMX) or sham surgery, and harvested the knee joints 8 week post-surgery for histological assessment of OA progression. Human OA chondrocytes cultures were incubated in the presence or absence of the DDR1 inhibitor DDR1-IN-1 dihydrochloridein (100-500 nM) for 2 h before addition of Postn (1 mug/ml) or control vehicle to the culture medium. MMP-13 levels were determined by ELISA 24 h post stimulation. Results: We observed abundant expression of DDR1 mRNA in human chondrocytes and we found comparable levels of DDR1 in OA and normal cartilage. However, Postn expression was 3-4 times as high in OA than in normal cartilage. Pre-incubation of human cartilage explants or cultured chondrocytes with DDR1-IN-1 dihydrochloridein inhibited both constitutive and Postn-induced MMP-13 expression in a dose-dependent manner. In contrast, neutralizing antibody to alphavbeta3 integrin had no effect on Postn induction of MMP-13 expression. Co-immunoprecipitation experiments showed that Postn physically interacts with DDR1 in human chondrocytes. Furthermore, Postn^-/- mice showed reduced PMX-induced cartilage degeneration and osteophyte formation, and both Postn+/- and Postn^-/- mice had reduced subchondral bone thickening, relative to Postn+/+ mice. Conclusion: Postn^-/- mice are protected from surgically-induced post-traumatic OA, showing that Postn promotes cartilage degeneration. DDR1 mediates the stimulatory effect of Postn on MMP-13 expression. Further studies are in progress to investigate the potential of periostin as a druggable target for the treatment of OA

Background/Purpose: Obesity is a common risk factor for knee osteoarthritis (KOA). While it is intuitive that bariatric weight loss improves knee pain, it is not clear how much is due to decreased mechanical load vs metabolic changes. Methods: Patients were screened for knee pain prior to sleeve gastrectomy, gastric bypass, or laparoscopic gastric banding. We required pain for >15 days/month and VAS pain > 30, excluding lupus, inflammatory arthritis, crystal disease, psoriasis, and bilateral knee replacement. Enrolled patients took standing knee xrays for Kellgren-Lawrence (KL) grading. We measured BMI and used the Knee Injury and Osteoarthritis Outcome Score (KOOS) questionnaire at baseline and 1, 3, 6 and 12 months, calculating % excess weight loss (%EWL) and deltaKOOS. We collected blood at baseline and followup to study biomarkers for predicting KOOS scores. Results: Of 536 patients considering bariatric surgery, we found 308 with knee pain and enrolled 176 (91.5% female; BMI 43.6 kg/m2+/-7, 32-61; age 42 +/-11, 18-73) well distributed in xray severity (KL0-4). For the 150 patients who had surgery, knee improvement paralleled weight loss at the followups. At 1 year, %EWL correlated well with deltaKOOS pain (R = .262, n = 114, p = 0.005), similar to other intervals and to other KOOS measures. The sleeve and bypass (n=72 and 27) vs banding (n=15) resulted in higher deltaKOOS pain at 1 year: 32.9 +/-21.3 and 30.7 +/-22.6 vs 10.2 +/-21.4, p=0.001. Sleeve and bypass patients also achieved a higher % of their potential deltaKOOS pain improvement than did banding (65.2% and 60.1% vs 16.8% of remaining KOOS points to 100), and a higher % of patients improved to any degree (93.1% and 88.9% vs 66.7%). Radiographic severity did not predict deltaKOOS at 1 year, nor did the presence of key comorbidities. Patients lost weight in a near-linear fashion through 1 year (Fig. 1), but their KOOS improvements plateaued at 1 month. This held true in sleeve and bypass subgroups (with altered anatomy), while banding showed less consistent deltaKOOS despite a similar trend in %EWL. Baseline leptin levels in obese KOA were higher than non-obese KOA and non-obese/non-OA controls from other cohorts (100.2 +/-61.9 vs 26.2 +/-16.7 and 15.4+/-13.8, p<0.001). Similarly, IL-1Ra, a potential marker of OA progression, was much higher than non-obese KOA or controls (1123+/-940 vs 324.0+/-145.6 and 272+/-130.0, p<0.001). Within obese KOA, higher leptin levels predicted worse xrays (KL0/1 vs KL2/3/4, p = 0.037). After 1 year, mean leptin and IL1-Ra from obese KOA patients had decreased (p<0.001). Conclusion: Bariatric surgery improves knee OA symptoms proportionally to %EWL. Most relief occurs during the 1st month before much weight loss, suggesting a metabolic impact beyond mechanical load reduction on joints - at least with the sleeve and bypass that alter digestive anatomy. Leptin and IL-1Ra serum levels are elevated in obese KOA vs non-obese KOA and controls - and fall after bariatric surgery which could contribute to knee pain relief

Background/Purpose: We aimed to understand the mechanism by which membrane-type 1 matrix metalloproteinase (MT1-MMP, MMP-14) controls bone and cartilage homeostasis. MT1-MMP, a cell-membrane-bound proteinase with an extracellular catalytic site and a 20-amino acid cytoplasmic tail, plays a key role in postnatal bone formation. The genetic deficiency of MT1-MMP in the mouse causes dwarfism, osteopenia and severe arthritis. Deletion of MT1-MMP in bone marrow-derived mesenchymal progenitor cells (BM-MSC) recapitulates this phenotype, showing that MT1-MMP controls osteogenic differentiation in MSC. The phenotype of MT1-MMP-/- mice has been proposed to result from lack of MT1-MMP proteolytic activity. However, mounting evidence shows a variety of proteolysis-independent signaling functions of MT1-MMP. The unique tyrosine (Y573) in the MT1-MMP cytoplasmic tail is fundamental for the control of intracellular signaling. Methods: We generated a mouse with the Y573D mutation in MT1-MMP (MT1-MMP Y573D) and characterized its skeletal phenotype by histological and microCT analyses. Isolated BM-MSC were induced to differentiate into osteoblasts, chondrocytes and adipocytes, using qRT-PCR to analyze gene expression. Mouse C3H10T1/2 MSC were transfected with MT1-MMP cDNA and analyzed for Wnt signaling by luciferase reporter assays. Results: MT1-MMP Y573D mice had increased trabecular bone relative to wt littermates, marked thinning of articular cartilage with disorganized tissue architecture, clustering and cloning of chondrocytes, and pronounced decrease in bone marrow-associated and total body fat. We induced BM-MSC from wt and MT1-MMP Y573D littermates to differentiate into osteoblast and chondrocytes, and myeloid precursors into osteoclasts. The Y573D mutation dramatically increased MSC expression of osteoblast markers and strongly downregulated chondrocyte and osteoclast markers. These findings indicated that Wnt signaling is upregulated in MT1-MMP Y573D-expressing MSC. Therefore, we analyzed Wnt signaling. We transiently transfected C3H10T1/2 MSC cells in osteoblast medium with the cDNAs for wt MT1-MMP and MT1-MMP Y573D. As controls the cells were transfected with the empty vector (pcDNA) or with MT1-MMP E240A, a mutant devoid of proteolytic activity. MT1-MMP Y573D dramatically upregulated Wnt signaling relative to wt MT1-MMP and MT1-MMP E240A. Conclusion: MT1-MMP controls Wnt signaling by a mechanism independent of extracellular proteolysis and mediated by its cytoplasmic tail. MT1-MMP is a bifunctional protein, with an extracellular proteolytic activity that promotes bone formation through ECM remodeling and a cytoplasmic tail that controls osteogenesis by interacting with a key pro-osteogenic signaling pathway

Background/Purpose: Osteoarthritis (OA) etiopathogenesis includes an inflammatory component. Published reports indicate that synovial fluid urate levels, even in patients without gout, associate with OA prevalence/severity. Whether serum urate (sUA), the precursor for gout and a biomarker for cardiovascular and kidney disease, may serve as a biomarker to convey or predict OA risk is not known. We investigated whether sUA levels associate with knee OA radiographic severity and contrast MRI-measured quantitative synovial volume (SV), and whether sUA levels predict radiographic progression, in a gout-free knee OA cohort. Methods: We assessed sUA in 88 gout-free subjects who completed a 24-month prospective, natural history knee OA study. Subjects had symptomatic medial knee OA, met ACR knee OA criteria and had BMI <33 at study entry. sUA was measured (enzyme-colorimetry) in serum frozen and banked at baseline. At baseline and 24 months, patients underwent standardized weight-bearing fixed-flexion posteroanterior knee radiographs (SynaFlexerTM). Twenty-seven subjects additionally had a dynamic gadolinium-enhanced 3.0T knee MRI that was read for quantitative synovial volume (SV). A musculoskeletal radiologist, blinded to subject data, determined joint space width (JSW) and Kellgren-Lawrence (KL) grades at each time point. Joint space narrowing (JSN) was determined as JSW change from baseline to 24 months. Pearson's correlations, student's t-tests, one-way ANOVA with post hoc Tukey-Kramer tests, ROC and AUC curves were used in statistical analyses, as appropriate. Results: sUA correlated with JSN in both univariate (r=0.40, p<0.01) and multivariate analyses (adjusting for age, gender and BMI, r=0.28, p=0.010). There was a significant difference in mean JSN after dichotomization of sUA at 6.8mg/dL, the solubility point for serum urate, even after adjustment for age, gender and BMI (JSN [+/-SEM] of 0.90mm+/-0.20mm for sUA>6.8; JSN [+/-SEM] of 0.31mm+/-0.09mm for sUA<6.8, p<0.01). Baseline sUA distinguished progressors (JSN>0.2mm), and fast progressors (JSN>0.5mm), from non-progressors (JSN<0.0mm) in multivariate analyses (area under the receiver operating characteristic curve [AUC] 0.626, p=0.027; AUC 0.620, p=0.045, respectively). sUA also correlated with SV (r=0.44, p=0.0040), a possible marker of JSN, though this correlation did not persist after controlling for age, gender and BMI (r=0.13, p=0.562). Conclusion: In non-gout patients with knee OA, sUA levels predict JSN and may serve as a biomarker for OA progression. (Figure presented)

Background/Purpose: Growing numbers of studies show increased expression in Osteoarthritis (OA) of inflammatory cytokines, such as IL-1beta and TNFalpha, in joint tissues and peripheral blood mononuclear (PBM) cells. The IL1 receptor antagonist (IL1RN) gene cluster region has been associated with susceptibility to knee OA, thereby further implicating inflammation in OA pathogenesis. In these studies, we examined the association of IL-1RN haplotype with the radiographic severity of symptomatic knee OA (SKOA). Methods: Genomic DNA from SKOA patients from three cohorts (NYU I, NYU-II and O

Background/Purpose: Osteoarthritis (OA) is a disease characterized by pain and tissue destruction, in some cases concomitant with inflammation. The link between pain and tissue destruction is yet unknown, and there is a lack objective quantifiable parameters. Collagens are the main structural proteins of the joint extracellular matrix. The degradation of especially type I (connective tissue), II (cartilage), III (synovium) and IV (basement membrane) collagens have been shown to be elevated in OA. So we investigated whether biomarkers reflecting collagen degradation were associated with knee OA representing with different pain and inflammatory phenotypes. Methods: 111 knee OA patients, 62% women, from NYUHJD progression cohort study with varying degree of OA were included: mean (SD) age, 62 (10); mean(SD) BMI, 27(4); NSAID users, 23%; radiographic OA (KL>2) 68%; and bilateral knee OA; 87%. Pain was assessed by VASpain and WOMAC at baseline (BL) at a 2- year follow-up (FU) visit. Median (IQR) were 39 (13-69) and 37 (13-52) for BL VASpain and WOMACpain. 4 BL serum biomarkers of type I, II, III and IV collagen degradation (C1M, C2M, C3M, C4M), and the 2 inflammatory biomarkers CRPM and hsCRP, were assessed. Data were log2 transformed. Associations between BL biomarkers, BL pain and change (CHG) pain scores were assessed by multivariate linear model including gender, age, BMI, KLsignalknee, bilateral knee OA and NSAID use. Patients with cont. mild/moderate pain had a BL VASpain<54 and FU VASpain<30, cont. moderate/severe pain had VASpain>30 at baseline and FU, and transitional severe pain had either VASpain-BL<30 and VASpain-FU>54 or VASpain-BL>54 and VASpain-FU<30 (ref). Patients with; low biochemical disease activity index (bDAI) low in CRPM (<12nM) moderate bDAI were high in CRPM but low in hsCRP (<5), and high bDAI (flare) were high in CRPM and hsCRP. Results: BL association between pain and biomarkers C2M (beta -17.9, p<0.0001) and KLsignalknee (beta -5.4, p=0.0031) were significantly associated with WOMAC pain. C2M (beta -12.4, p=0.0033), C3M (beta -19.9, p=0.059), age (beta -0.84, p<0.0018), KLsignalknee (beta 8.9, p=0.0021) and bilateral knee OA (beta -12.2, p=0.087) were associated with VASpain. Association between BL biomarkers and CHG pain C2M (beta 13.3, p=0.0016), age (beta 0.5, p=0.029) and bilateral OA (beta -12.0, p=0.043) were significantly associated with delta WOMACpain. Only age, BMI and NSAID use was associated with CHG VASpain. Association between pain phenotypes and BL biomarkers Patients with cont. mild/moderate pain had significantly higher C2M compared patients with transitional severe pain (p=0.0014) and cont. moderate/severe pain (p=0.04). Biomarker, BL pain and CHG pain in patients w. inflammatory OA Patient with low bDAI had lower WOMACpain (p<0.05) and VASpain(p<0.1). C1M was higher (p<0.05) in the flare group compared to the low and moderate bDAI groups. C3M was higher (p<0.05) in the moderate bDAI group than the low DAI group. Conclusion: Different collagen degradation products are linked differentially to different phenotypes. Cartilage degradation (C2M) was consistently linked to CHG pain phenotypes, whereas it was not associated with an inflammatory phenotype. In contrast, C1M and C3M were linked to inflammatory and flared OA

Background/Purpose: Persistent inflammation is a characteristic of several joint diseases, including OA. It is nowadays appreciated that this could be a result of a failure to (optimally) activate inflammation resolution pathways. Therefore, we investigated the presence of specialized pro-resolving lipid mediators (SPM) and their precursors as pathway markers of the resolution process in the joint of OA patients and controls. Methods: SF was obtained from knee OA (2 populations) and rheumatoid arthritis (RA) patients fulfilling the ACR criteria for OA and RA, respectively, and healthy controls. Lipid mediators (LMs) were determined by targeted lipidomics using liquid-chromatography mass spectrometry. Sixty different lipids including pro-inflammatory (e.g. prostaglandins, leukotrienes) and anti-inflammatory/pro-resolving LM (e.g. SPM), as well as their precursors can be detected with our technique. Results: SF from 24 OA and 12 RA patients were first studied. Thirty-seven lipids were detected in the soluble fraction of SF, including polyunsaturated fatty acids (PUFA) and their lipoxygenase (LOX) and cyclooxygenase (COX) pathway markers in both OA and RA patients. Among these, pro-inflammatory LM such as PGE2 and thromboxane B2, as well as the pathway markers of resolution and precursors of SPM, 17-HDHA and 18-HEPE, were detected. Except for the LOX products of arachidonic acid: 15-HETE, 6-trans-LTB4 and 20-OH- LTB4, which were lower in OA than in RA SF, all other lipid mediators and PUFA were comparable between OA and RA samples. Ratios of metabolites to their precursors indicated that both pro- (e.g. LTB4) and anti-inflammatory LOX products (e.g. 17- HDHA) are more efficiently generated in RA than in OA patients, while no differences were observed in COX products. Interestingly, the SPM resolvin D2 (RvD2) could also be detected, but only in the insoluble fraction (cells and undigested matrix), indicating that the resolution pathways are activated in OA. This expands our previous publication showing activation of resolution in RA patients. To assess the efficiency of activation of resolution in OA, we have performed targeted lipidomics on total SF in an additional study with 32 OA patients and 10 healthy controls. Confirming earlier data, most LMs were also detected in this study, including the pro-inflammatory PGE2, the SPM precursors 17-HDHA and 18-HEPE, and the SPM RvD2. Additionally, we detected 18S-resolvin E3 (18S-RvE3). Remarkably, both the absolute concentrations of the SPM RvD2 and 18S-RvE3, and the ratio to their precursors, 17-HDHA and 18-HEPE, were lower in OA compared to healthy SF, indicating less efficient generation of SPM in OA compared to healthy joints. In contrast, the proinflammatory lipid PGE2was higher in OA than in healthy SF, indicating that the lower activation of resolution is paired by a higher inflammatory load in OA compared with healthy individuals. Conclusion: By using a state-of-the-art technique, we show for the first time that resolution pathways are activated in OA patients. Importantly, resolution seems to be less efficiently activated than in healthy individuals, which could account for the persistent inflammation observed in OA and RA patients

Background: Osteoarthritis (OA) is heterogeneous disease characterized by pain and tissue destruction which is in some case concomitant with inflammation. However, the link between pain and tissue destruction is yet unknown, and there is a lack objective quantifiable parameters. Collagens are the main structural proteins of the joint extracellular matrix. The degradation type I (connective tissue), II (cartilage), III (synovium) and IV (basement membrane) collagens have been shown to be elevated in joint degenerative diseases. Objectives: To investigate whether biomarkers reflecting collagen degradation were associated with symptomatic knee OA representing with different pain and inflammatory phenotypes. Methods: 111 knee OA patients, 62% women, from NYUHJD progression cohort study with varying degree of OA were included: mean (SD) age, 32 (10); mean (SD) BMI, 27 (4); NSAID users, 23%; radiographic OA (KL>2) 68%; and bilateral knee OA; 87%. Pain was assessed by VASpain and WOMAC at baseline (BL) at a 2-year follow-up (FU) visit. Median (IQR) were 39 (13-69) and 37 (13-52) for BL VASpain and WOMACpain. 4 BL serum biomarkers of type I, II, III and IV collagen degradation (C1M, C2M, C3M, C4M), and the 2 inflammatory biomarkers CRPM and hsCRP, were assessed. Data were log2 transformed. Associations between BL biomarkers, BL pain and change (CHG) pain scores were assessed by multivariate linear model including gender, age, BMI, KLsignal knee, bilateral knee OA and NSAID use. Patients with cont. mild/moderate pain had a BL VASpain<54 and FU VASpain<30, cont. moderate/severe pain had VASpain>30 at baseline and FU, and transitional severe pain had either VASpain-BL<30 and VASpain-FU>54 or VASpain-BL>54 and VASpain-FU<30 (ref). Patients with; low biochemical disease activity index (bDAI) low in CRPM (<12nM) moderate bDAI were high in CRPM but low in hsCRP (<5), and high bDAI (flare) were high in CRPM and hsCRP. Results: BL association between pain and biomarkers. C2M (beta -17.9, p<0.0001) and KLsignal knee (beta -5.4, p=0.0031) were significantly associated with WOMAC pain. C2M (beta -12.4, p=0.0033), C3M (beta -19.9, p=0.059), age (beta -0.84, p<0.0018), KLsignal knee (beta 8.9, p=0.0021) and bilateral knee OA (beta -12.2, p=0.087) were associated with VASpain. Association between BL biomarkers and CHG pain. C2M (beta 13.3, p=0.0016), age (beta 0.5, p=0.029) and bilateral OA (beta -12.0, p=0.043) were significantly associated with delta WOMACpain. Only age, BMI and NSAID use was associated with CHG VASpain. Association between pain phenotypes and BL biomarkers. Patients with cont. mild/moderate pain had significantly higher C2M compared patients with transitional severe pain (p=0.0014) and cont. moderate/severe pain (p=0.04). Biomarker, BL pain and CHG pain in patients w. inflammatory OA. Patient with low bDAI had lower WOMACpain (p<0.05) and VASpain (p<0.1). C1M was higher (p<0.05) in the flare group compared to the low and moderate bDAI groups. C3M was higher (p<0.05) in the moderate bDAI group than the low DAI group. Conclusions: Different collagen degradation products are linked differentially to different phenotypes. Cartilage degradation (C2M) was consistently linked to pain CHG and phenotypes, whereas it was not associated with an inflammatory phenotype. In contrast, C1M and C3M were linked to inflammatory OA and flared OA